0% found this document useful (0 votes)

291 views

Final Report Compressed

This document is a final year project report for the design and construction of an autonomous hovercraft for landmine detection. A group of six mechanical engineering students designed and built a prototype hovercraft with the goal of providing a proof of concept for using hovercraft in mine detection. The hovercraft was designed with independent lift and thrust systems, metal detector sensors for mine detection, and the capability for autonomous navigation and sensor integration. Testing confirmed the hovercraft met design specifications but full autonomy was not achieved within the project timeline. The report provides details on the conceptual and detailed design of the hovercraft systems, manufacturing, and test results.

Uploaded by

Enas Al-khawaldeh

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

291 views

Final Report Compressed

Uploaded by

Enas Al-khawaldeh

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 309

School of Mechanical Engineering

Final Year Project 2009

DESIGN AND BUILD OF AN AUTONOMOUS HOVERCRAFT FOR LANDMINE DETECTION

Authors: Reuben Brown Beau Krieg Rahim Kurji Paul Hocking Adrian McLay Jarrad Norton
October 30, 2009

Supervisor: Dr. Maziar Arjomandi

Executive Summary
This report outlines the development of a autonomous hovercraft platform for the express purposes of land mine detection. Design, construction and testing was undertaken by a group of six nal year engineering students from the School of Mechanical Engineering at the University of Adelaide, during 2009. The aim of this project was to develop a prototype vehicle that would provide a proof of concept for the application of hovercraft to mine detection. The scope of this project included an extensive manufacturing eort ranging from woodworking to composite manufacture. This manufacturing was partaken by the group, and supported by the Mechanical Engineering Workshop. The application of a hovercraft to mine detection oers several advantages over established mine detection methodologies. The low pressure footprint of of a hovercraft minimises the probability of detonation. The application of autonomous features to the platform, including way point navigation and automated mine detection and marking, removes the operator from the eld and hence from the hazard. This unique solution however, raises a series of challenges, including the minimisation of metallic components, designing a highly controllable yet manoeuvrable craft and successfully integration of sensitive sensor congurations. To meet these challenges, a hovercraft conguration with a axial fan lift system feeding a segmented skirt and a dierential and vectored thrust system mounted on a composite base structure, was designed. Autonomy was incorporated by adapting a commercially available autopilot for use as a two-dimensional planar control system which was integrated with the hovercrafts actuation systems via a microcontroller. O the self metal detector coils integrated into the hull provided detection capabilities for the platform. The output from each coil was extracted and processed to provide a signal that could be used to actuate a physical marking system, specically designed to paint the ground with a high visibility mark within a one metre halo radius of the target. This sensor conguration had a tested accuracy of 68% in detection of a representative target. Performance testing of the hovercraft conrmed that design specications were met. Unfortunately, due to time constraints, full autonomy was not realised, however the detailed design and integration provides promising groundwork for future work, with active yaw control implemented. Testing of all other systems conrmed their suitability for the application, however the complete prototype was not tested under the entire operational scenario at the completion of this project and is reserved for future work.

iii

Acknowledgements
The authors of this report would like to formally recognise the valued contribution of numerous parties, without whom this project would not have reached its current state. In particular, Project Hover would like to formally acknowledge the Defense Science and Technology Organisation, for its nancial and technical support that made this project possible. In particular, we thank Canicious Abeynayake and the sta at Object Detection Technologies Weapons Systems Division for their support and expertise. In addition, the authors would like to thank AMCOR and Soil Testing Services for their nancial support of this project. The authors would also like to express their appreciation to this projects supervisor, Dr Maziar Arjomandi, not only for his expertise, but also for the genuine interest he has taken in our education. Finally, the authors would like to tank the Mechancial Engineering Academic and Workshop sta for their interest and advice, as well as the administrative sta who made our life that much easier.

Disclaimer
We, the authors, declare that the material contained with this report is entirely our own, unless otherwise specied.

Reuben Brown a1132198 ........................

Beau Krieg a1132286 ........................

Rahim Kurji a1126547 ........................

Paul Hocking a1132751 ........................

Adrian McLay a1118636 ........................

Jarrad Norton a1147683 ........................

Contents
Executive Summary Acknowledgements Disclaimer Table of Contents List of Figures List of Tables Nomenclature Acronyms . . . . List of Symbols . List of Units . . List of Subscripts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii iii iv viii xiii xvi xvii . xvii . xvii . xviii . xix xix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 2 2 3 5 5 6 7 9 9 11 11 13 15 16 19 19 19 19 20 21 22 22 23 23 24 26 28 28 28 28

Coordinate System 1 Introduction 1.1 Aims . . . . . . . . . . . . . . . . . . 1.2 Project Objectives . . . . . . . . . . 1.2.1 Primary Project Objectives . 1.2.2 Extended Project Objectives 1.3 Project Denition . . . . . . . . . . . 1.3.1 General Requirements . . . . 1.3.2 Operational Environment . . 1.3.3 Operational Scenario . . . . . 1.3.4 Performance Specications . 1.4 Scope . . . . . . . . . . . . . . . . . 2 Feasibility Study 2.1 Hovercraft Benchmarking 2.2 Landmines . . . . . . . . . 2.2.1 Detection Methods 2.3 Autonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3 Conceptual Design 3.1 Preliminary Sizing . . . . . . . . . . . . . . 3.1.1 Statistical Analysis . . . . . . . . . . 3.1.2 Beam Ratio . . . . . . . . . . . . . . 3.1.3 Planform Sizing . . . . . . . . . . . . 3.1.4 Weight Estimation . . . . . . . . . . 3.2 Prototype Developement . . . . . . . . . . . 3.2.1 Integrated Lift and Thrust Systems 3.2.2 Thrust Dominant System . . . . . . 3.2.3 Independant Thrust Lift Systems . . 3.2.4 Prototype Generation . . . . . . . . 3.2.5 Prototype Selection . . . . . . . . . 3.2.6 Prototype Specic Requirements . . 3.3 Power Requirements . . . . . . . . . . . . . 3.3.1 Lift Power . . . . . . . . . . . . . . . 3.3.2 Propulsion Power . . . . . . . . . . . v

vi 3.3.3 Preliminary Thrust and Power Estimates . . . Planform Design . . . . . . . . . . . . . . . . . . . . . Payload Area . . . . . . . . . . . . . . . . . . . . . . . Landmine Detection Sensor . . . . . . . . . . . . . . . 3.6.1 Specications . . . . . . . . . . . . . . . . . . . 3.6.2 Sensor Selection . . . . . . . . . . . . . . . . . Static Stability . . . . . . . . . . . . . . . . . . . . . . 3.7.1 Application of Jupe Skirt to Enhance Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 31 32 33 33 34 37 40 43 43 45 47 53 53 55 55 55 56 58 61 62 63 64 64 64 66 67 70 71 73 80 80 81 82 82 84 90 92 98 99 99 104 107 113 113 116 120 125 125 126 130 135 136 136 136

3.4 3.5 3.6

3.7

4 Detailed Design 4.1 Payload Area . . . . . . . . . . . . . 4.1.1 Sensor Selection . . . . . . . 4.1.2 Sensor Conguration . . . . . 4.1.3 Signal Analysis . . . . . . . . 4.1.4 Physical Marking System . . 4.2 Lift . . . . . . . . . . . . . . . . . . . 4.2.1 Design Specications . . . . . 4.2.2 Design Calculations . . . . . 4.2.3 Fan Design and Selection . . 4.2.4 Engine Selection . . . . . . . 4.2.5 Inlet Bell Design . . . . . . . 4.2.6 Mount Design . . . . . . . . . 4.2.7 Shaft Extension Design . . . 4.2.8 Design Conclusion . . . . . . 4.3 Propulsion . . . . . . . . . . . . . . . 4.3.1 Design Specications . . . . . 4.3.2 Conguration Selection . . . 4.3.3 Propeller Selection . . . . . . 4.3.4 Motor Selection . . . . . . . . 4.3.5 ESC and Battery Selection . 4.3.6 Mount Design . . . . . . . . . 4.4 Structural . . . . . . . . . . . . . . . 4.4.1 Primary Design Criteria . . . 4.4.2 Design Specications . . . . . 4.4.3 Specic Loading Conditions . 4.4.4 The Sandwich Principle . . . 4.4.5 Skin Selection . . . . . . . . . 4.4.6 Core Selcetion . . . . . . . . 4.4.7 Design Calculations . . . . . 4.5 Skirt . . . . . . . . . . . . . . . . . . 4.5.1 Design Specication . . . . . 4.5.2 Conguration Selection . . . 4.5.3 Material Selection . . . . . . 4.5.4 Template Generation . . . . . 4.6 Control System . . . . . . . . . . . . 4.6.1 Components . . . . . . . . . 4.6.2 Sensors and Actuators . . . . 4.6.3 Yaw Control . . . . . . . . . 4.7 Navigation System . . . . . . . . . . 4.7.1 Selection Criteria . . . . . . . 4.7.2 Selection . . . . . . . . . . . 4.7.3 Autopilot Selection . . . . . . 4.7.4 Integration . . . . . . . . . . 4.8 Platform Analysis . . . . . . . . . . 4.8.1 Platform Weight . . . . . . . 4.8.2 Trim and COG Calculations .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

CONTENTS 5 Manufacturing 5.1 Manufacturing Methods . . . . . 5.2 Lift System Construction . . . . 5.3 Inlet Bell Construction . . . . . . 5.4 Propulsion System Construction 5.5 Hull Construction . . . . . . . . 5.6 Skirt Construction . . . . . . . . 5.7 Fairing Construction . . . . . . . 5.8 Final Assembly . . . . . . . . . . 5.8.1 Lift System Installation . 5.8.2 Propulsion System . . . . 5.8.3 Payload . . . . . . . . . . 5.8.4 Skirt . . . . . . . . . . . . 5.8.5 Fairing . . . . . . . . . . . 5.9 Quality Control . . . . . . . . . .

vii 139 139 140 140 141 143 144 145 145 146 147 147 148 148 148 151 151 151 153 154 156 156 158 159 162 163 163 165 166 166 166 167 169 169 169 171 171 172 172 172 172 174 174 174 175

. . . . . . . . . . . . . .

6 Testing 6.1 Component Testing . . . . . . . . . . . . . . . . . . . . . . 6.1.1 Metal Detector Range Testing . . . . . . . . . . . 6.1.2 Metal Detector Signal Testing . . . . . . . . . . . . 6.1.3 Landmine Detection Operational Scenario Testing 6.2 System Testing . . . . . . . . . . . . . . . . . . . . . . . . 6.2.1 Lift System Airow Testing . . . . . . . . . . . . . 6.2.2 Propulsion System Testing . . . . . . . . . . . . . 6.2.3 Static Hover Testing . . . . . . . . . . . . . . . . . 6.3 Longnitudinal and Transverse Behaviour Testing . . . . . 6.4 Performance Testing . . . . . . . . . . . . . . . . . . . . . 6.4.1 Acceleration Testing . . . . . . . . . . . . . . . . . 6.4.2 Deceleration . . . . . . . . . . . . . . . . . . . . . . 6.4.3 Additional Acceleration Performance Testing . . . 6.4.4 Climb Gradient Tesing . . . . . . . . . . . . . . . . 6.4.5 Additional Functionality Testing . . . . . . . . . . 6.4.6 Turning circle . . . . . . . . . . . . . . . . . . . . . 7 Project Management and Finance 7.1 Practical Project Management Methodology 7.2 Resource Management . . . . . . . . . . . . 7.2.1 Organisational Hierarchy . . . . . . 7.2.2 Work Distribution . . . . . . . . . . 7.3 Review Process . . . . . . . . . . . . . . . . 7.4 Time Management . . . . . . . . . . . . . . 7.4.1 Deadline Specication . . . . . . . . 7.5 Risk Management . . . . . . . . . . . . . . 7.6 Finance . . . . . . . . . . . . . . . . . . . . 7.6.1 Sponsorship . . . . . . . . . . . . . . 7.6.2 Financial Position . . . . . . . . . . 7.6.3 Labour Costs . . . . . . . . . . . . .

. . . . . . . . . . . . . . . .

. . . . . . . . . . . .

8 Conclusions 177 8.1 Primary Project Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177 8.1.1 Extended Project Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179 8.2 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180 References Appendix A Autopilot Conguration Files 180 187

viii Appendix B Segmented Skirt Stability Equations Appendix C Simulink Appendix D External Milestones Appendix E Internal Time Line Appendix F Lift Calculations Appendix G Lift Mount Strength Calculations Appendix H Hull Appendices Appendix I Bill Of Materials

CONTENTS 197 199 201 203 205 207 209 217 221 261 269 271 273 275 281 283

Appendix J Manufacturing Drawings Appendix K Risk Treatment Appendix L Centre of Gravity Calculations Appendix M Landmine Detection Code Appendix N Landmine Statistical Analysis Appendix O Landmine Detection Sensors Appendix P Breakdown of Hours and Costs Appendix Q Sensitivities of Lift and Thrust

List of Figures
1 1.1 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 Coordinate System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mission Prole . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hovertechnics HoverJet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AirCommander AC4 Hovercraft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Amphibious Marine Vanguard 14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pacic Hovercraft Slider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AT Landmine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AP Landmine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Using a probe for landmine detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . Vehicle mounted landmine sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Autonomous Vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi 7 11 12 12 13 14 15 15 16 17 17 20 21 23 25 25 29 30 30 31 32 33 34 39 39 40 40 44

2.10 Automated Turf Management Vehicle . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 Beam Ratio Histogram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Linear Regression Model for Dry Weight Estimation . . . . . . . . . . . . . . . . . . . Splitter Conguration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prototype 1 and Prototype 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prototype 3 and Prototype 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensitivity of Lift Power to Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sensitivity of Thrust to Weight and Terrain Slope . . . . . . . . . . . . . . . . . . . .

FBD of Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Skirt drag over dierent terrains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.10 Planform Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.11 Hovercraft hull showing front and rear payload bays . . . . . . . . . . . . . . . . . . . 3.12 Simulation landmine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.13 Change in area due to cushion deformation in pitch motion . . . . . . . . . . . . . . . 3.14 Reaction arm of segmented skirt in pitch motion . . . . . . . . . . . . . . . . . . . . . 3.15 Reaction arm of segmented skirt in roll motion . . . . . . . . . . . . . . . . . . . . . . 3.16 Weight vector from picth or roll motion . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1 Dimensioned Front and Rear Payload Areas . . . . . . . . . . . . . . . . . . . . . . . . ix

x 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9

LIST OF FIGURES Required Detection Width Based on Hovercraft Dimensions . . . . . . . . . . . . . . . DSTO Metal Detector Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bounty Hunter Tracker IV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dual Metal Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Composite Metal Detector Mount Design . . . . . . . . . . . . . . . . . . . . . . . . . Switching Impulse Interference Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . Metal Detector Switching Dead Period . . . . . . . . . . . . . . . . . . . . . . . . . . . Scanning Pattern With Periodic Switching . . . . . . . . . . . . . . . . . . . . . . . . . 45 46 46 48 50 51 51 52 52 53 54 55 58 60 61 61 62 63 64 65 67 67 69 69 70 71 72 72 74 75 75 77 77 79

4.10 Relay Circuit Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.11 Implemeted Threshold Levels for Landmine Detection . . . . . . . . . . . . . . . . . . 4.12 Marking System Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.13 Marking System Actuation Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.14 Fan Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.15 Engine Power Curve . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.16 Inlet Bell Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.17 Diagram of Inlet Bell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.18 Duct Diameter vs r/d . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.19 Fan and Engine Mount . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.20 Engine Shaft Extension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.21 Assembly of Mount with Inlet Bell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.22 Air Jet Vectoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.23 Air-Screw Vectoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.24 CT estimatation curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.25 Selected Propeller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.26 CP estimatation curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.27 Typhoon 600-43 HET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.28 Dual DualSky 90A brushless controller . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.29 3800 mAh 8 cell LiPo battery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.30 General Propulsion Mount Conguration . . . . . . . . . . . . . . . . . . . . . . . . .

4.31 ANSYS Analysis of Propulsion Deection . . . . . . . . . . . . . . . . . . . . . . . . . 4.32 SKF Composite Bushing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.33 Ocean Controls Stepper Motor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.34 Ocean Controls Stepper Motor Controller . . . . . . . . . . . . . . . . . . . . . . . . . 4.35 Propulsion System Parts List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

LIST OF FIGURES 4.36 Area Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.37 Specic Loading Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.38 Weight, Strength and Stifness in Sandwitch Panel . . . . . . . . . . . . . . . . . . . . . 4.39 Force Distribution in a Sandwich Panel . . . . . . . . . . . . . . . . . . . . . . . . . . 4.40 Density Vs Tensile Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.41 Density Vs Tensile Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.42 Laminate Impact Strength . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.43 Tensile Strength and Tensile Modulus . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.44 Flexural Ridigity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.45 Weight Per Ply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.46 Comparitve Weight for Bending . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.47 Density and Compressive Modulud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.48 Density and Shear Modulud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.49 Density and Compressive Modulud . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.50 Loading 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.51 Weight and Deection of R80 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.52 Rib-Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

xi 80 83 83 84 85 86 86 87 89 89 90 91 91 94 94 96 97

4.53 Bag Skirt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.54 Finger Skirt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 4.55 Bag and Finger Skirt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 4.56 Pericell Skirt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 4.57 Skirt Attachment Bracket . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 4.58 Bi-conical Skirt Segment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 4.59 Skirt attachment bracket . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 4.60 Skirt attachment bracket . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 4.61 Skirt Test Rig . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 4.62 Skirt conguration arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 4.63 Propulsion moment during operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 4.64 Jupe Template Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113 4.65 Control System Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114 4.66 MiniDragon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 4.67 Remote and Receiver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.68 Servo Pulse Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4.69 Voltage Divider . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117

xii

LIST OF FIGURES 4.70 Stepper Control Circuit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 4.71 Stepper Driver Signal Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 4.72 Yaw Control Prototype . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121

4.73 Yaw Prototype Closed Loop Response . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 4.74 Gyro Step Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 4.75 Closed Loop Yaw Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 4.76 Tiny Autopilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 4.77 Kestrel Autopilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 4.78 Microbot APS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

4.79 Ardupilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 4.80 Carrot Placement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 4.81 Torrens Parade Ground Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133 4.82 Yaw Angle Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 4.83 3 View Drawing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 5.1 5.2 5.3 5.4 5.5 5.6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 Bell Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Hull Mould . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144 Hull Glassing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145 Hull Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146 Top Lip, Top Plate and Rib System . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 Fully Assembled Platform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149 Test apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Metal detector vertical range testing results . . . . . . . . . . . . . . . . . . . . . . . . 152 Metal detector range horizontal testing results . . . . . . . . . . . . . . . . . . . . . . . 152 Signal testing apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153 Periodic switching interference and positive signal . . . . . . . . . . . . . . . . . . . . 154

Positive detection and Object Interference Levels . . . . . . . . . . . . . . . . . . . . . 154 Operational landmine detection scenario . . . . . . . . . . . . . . . . . . . . . . . . . . 155 Detection accuracy over hovercraft width . . . . . . . . . . . . . . . . . . . . . . . . . 155 Landmine physical marking results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156

6.10 Volumetric ow rate as a function of RPM and eciency . . . . . . . . . . . . . . . . . 157 6.11 Thrust Testing Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 6.12 Thrust Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159 6.13 Local Coordinate System for Static Hover Test . . . . . . . . . . . . . . . . . . . . . . 160

LIST OF FIGURES

xiii

6.14 Static Hover Test Arrangement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 6.15 Variation is drift due to lift engine speed . . . . . . . . . . . . . . . . . . . . . . . . . . 162 6.16 Longintudinal and Transverse Response . . . . . . . . . . . . . . . . . . . . . . . . . . 163 6.17 Acceleration over 5 m intervals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 7.1 7.2 7.3 Organisational Chart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171 Expenditure and Monthly Breakdown . . . . . . . . . . . . . . . . . . . . . . . . . . . 175 Monthly Breakdown of Hours Spent . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

M.1 Landmine detection code, main method . . . . . . . . . . . . . . . . . . . . . . . . . . 271 M.2 Landmine detection code, timer interrupt . . . . . . . . . . . . . . . . . . . . . . . . . 272 O.1 Basic GPR operation diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276 O.2 GPR pulse detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 O.3 Metal detector coil operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 278 O.4 Detection range for 9 inch coil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Q.1 Sensitivity of Lift Power to Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284 Q.2 Sensitivity of Thrust to Weight and Terrain Slope . . . . . . . . . . . . . . . . . . . . 286

Q.3 Sensitivity of Thrust to Head Wind Speed . . . . . . . . . . . . . . . . . . . . . . . . . 287 Q.4 Eect of Mass of Propulsion Power Requirements . . . . . . . . . . . . . . . . . . . . . 287

This Page Intentionally Left Blank

List of Tables
1 1.1 3.1 3.2 3.3 4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8 4.9 Coordinate Specications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Summary of Performance Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . Pearson Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Beam Ratio Descriptive Statistics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Landmine detection sensor decision matrix . . . . . . . . . . . . . . . . . . . . . . . . . Front payload bay specication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bounty Hunter Tracker IV specications . . . . . . . . . . . . . . . . . . . . . . . . . . Metal Detector Specications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fan Types and Characteristics (Bleier, 1997) . . . . . . . . . . . . . . . . . . . . . . . Fan Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HGA-100L Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . KT100J Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Engine Specications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxi 9 19 20 37 43 47 47 56 58 59 59 60

Skirt Conguration Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

4.10 Skirt Conguration Comparison, Adapted from (Fitzgerald and Wilson, 1995, pg. 12) & (Amyot, 1989) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 4.11 Elastomer Material Comparison, Adapted from (Amyot, 1989) . . . . . . . . . . . . . 105 4.12 Fabric Material Comparison, Adapted from (Amyot, 1989) . . . . . . . . . . . . . . . . 105 4.13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 4.14 Summary of Inputs and Outputs for Microcontroller . . . . . . . . . . . . . . . . . . . 115 4.15 Selection Criteria for Autopilot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 4.16 Tiny v2 Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 4.17 Kestrel Autopilot Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 4.18 Microbot APS Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 4.19 Ardupilot Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 6.1 6.2 6.3 6.4 Propeller thrust output at 75% throttle . . . . . . . . . . . . . . . . . . . . . . . . . . 158 Drift Results for Trim Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Drift response to lift engine speed during static hover . . . . . . . . . . . . . . . . . . 162

Acceleration Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 xv

xvi 6.5 6.6 7.1

LIST OF TABLES Theoretical Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164 Deceleration performance of the hovercraft . . . . . . . . . . . . . . . . . . . . . . . . . 165 Financial Position . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 174

D.1 External Milestones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 E.1 Internal Time line . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203 O.1 List of current landmine senor types . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275 O.2 Various magnetometer types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277 P.1 Hours and Costs Breakdown . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

Nomenclature
Acronyms
AP Anti Personnel. 14, 38 AT Anti Tank. 14, 38 ATD Analog to Digital. 121 BOM Bill of Materials. 225 COG Centre of Gravity. 78, 145 CPT Cure Ply Thickness. 94 CPU Central Processing Unit. 135 DGPS Dierential GPS. 17, 18, 133, 137 ESC Electronic Speed Controller. 76

LiPo Lithium Ion Polymer. 76 MCU Microcontroller. 56, 59, 76, 121, 133, 144 MOI Moment of Inertia. 81 OS Operating System. 139 PC Personal Computer. 133 PID Proportional Integral Derivative. 126, 138 PWM Pulse Width Modulation. 134137 RC Remote Control. 72, 74, 126, 189 RPM Rotations Per Minute. 72 SOP Safe Operating Procedure. 184, 185 SPI Serial Peripheral Interface. 135, 144 UAV Unmanned Aerial Vehicle. 133, 135 UN United Nations. 14, 48 VLF Very Low Frequency. 49, 50

FBD Free Body Diagram. 34 FVF Fibre Volume Fraction. 93, 94 FWF Fibre Weight Fraction. 94 J Advance Ratio. GPL GNU Public License. 134 GPR Ground Penetrating Radar. 39, 245, 246 GPS Global Positioning System. 4, 17, 18, 120, 133, 135137, 139, 141, 176, 178, 191 GRP Glass Reinforced Plastics. 87 GUI Graphical User Interface. 4 I2C Inter-Integrated Circuit. 135, 144 ICT Input Capture Timer. 121 B Beam Length. C Coecient. S Characteristic Coecient. D Loss Coecient. Density. Angle. A Area.

List of Symbols

IMU Inertial Measurement Unit. 17, 18, 120, W Design Weight. d 133, 135, 136 T Thrust. KF Kalman Filter. 17, 18, 133 D Diameter. xvii

xviii DA Aerodynamic Drag. DC Cushion Momentum Drag. DS Skirt Interaction Drag. DT Terrain Slope Drag. e(t) Error Term. Q Volumetric Flow Rate. W Force of Engine on Each Support. L Lift Force. g Acceleration due to gravity. h Hover Height. V Velocity. Ixp Moment of Inertia (Propulsion). Deection. C Circumference. H Height. a Characteristic Length. b Characteristic Length. d Characteristic Length. h Characteristic Length. s Characteristic Length. mAh Milli-amp Hours. y Characteristic Length. m metre. L Length. min minute. t Thickness. ms millisecond. M Mass. Pa Pascals. M moment. I Moment of Inertia. N Number of Fan Blades. Pd Pitch to diametre ratio. rpm Revolutions per Minute. s second. V Volt. Hz Hertz. HP Horse Power. h hour. k kilo. Pitch Angle. Pitch Angle. XMP Pitch Moment Arm. P Power. P Pressure. R Radius. Roll Angle. n Rotation Speed. Skirt Contact Angle. E Elastic Modulus.

List of Units

degrees (rotation).

List of Subscripts W Watt.

xix Lift. Maximum. Net.

List of Subscripts
a

max

N ET

Air.
b

Per Blade. pivot. Planform. Power. Propeller. Ratio of air to total pressure requirement. Skirt. Theoretical. Thrust. T Section.

Beam.
p

cg c

Centre of Gravity. Cushion. Drag. Fan Ring. Footprint. Generated. Hover. Hovercraft.

p P p

R f

gen

t T ts

H hc

This Page Intentionally Left Blank

Coordinate System
The coordinate system and reference frames described in the gure and table below will be the conventions used throughout this report.

Figure 1: Coordinate System Table 1: Coordinate Specications Force or Linear or Position or Moment Angular Velocity Euler Angle X u x Y v y Z w z K p M q N r

Motion Surge Sway Heave Roll Pitch Yaw

xxi

This Page Intentionally Left Blank

Chapter 1 Introduction

1.1

Aims

There are over 100 million buried and active land mines across the world today. Hundreds of millions more are stored in military stockpiles, and over 10 million more are produced annually. Land mines and explosive remnants of war aect at least 78 countries and injure or kill up to 20,000 people annually. Land mines, cluster bombs and unexploded ordinance provide a real and enduring threat for local populations, wildlife and ecosystems. There is a need to develop a method of detection which is sustainable, has low environmental impact and is of low risk to human life. There are currently several techniques of mine and unexploded ordinance detection which require time consuming, manual surveying of potential mineelds. These methods include human screening, placing an operator at extreme risk, or eective vehicle mounted/destructive methods causing extreme environmental damage through the detonation of mines. This project proposes the use of an unmanned hovercraft as a mounting platform for mine detection technologies. A hovercraft is less likely to initiate pressure sensitive munitions as it has a low ground pressure footprint. Furthermore, autonomous features will remove the operator from the mineeld and the threat. Environmental damage through the travel of the craft will be signicantly reduced from traditional vehicle mounted methods. There are several design issues which will need to be targeted specically in the design of this platform. Maneuverability and controllability of the craft are important and detection and detonation cannot be performed by the same platform a degree of accuracy will be required in order to mark the positions of detected bodies. This in itself raises several issues which need to be addressed through design. Any horizontal drift of the craft will need to be kept to a minimum to ensure a desired orientation can be maintained. Accurate turning and motion capabilities will also be required. Furthermore, the eects of cambered or sloped terrains on the crafts direction of travel may need to be compensated for. A marking system will need to be devised which identies the location of any detections and alerts the operators to these nds. Current mine detecting sensor arrays are quite heavy and expensive, and thus the hovercraft should be capable of lifting a payload of sensors which must be isolated from potential damage. Similarly the hovercraft is required to hover at a height which optimises the performance of 1

2 these sensors.

CHAPTER 1. INTRODUCTION

The design considerations which come about through the application of a hovercraft to this function lead to a set of primary objectives for the platforms development.

1.2

Project Objectives

The following section outlines the objectives to which this project aims. These objectives are classied as either primary or extended, dependent on their diculty and scope.

1.2.1

Primary Project Objectives

Primary project objectives benchmark the minimum outcomes that should be achieved for the project. 1. To design and manufacture a remotely controlled hovercraft Outline: The project group will design and manufacture a remote controlled, unmanned hovercraft to the specications outlined in Section 1.3.4.

Measure of Completion: Testing will be composed of a variety of test ights designed to assess the performance of the craft across the performance parameters outlined in Section 1.3.4. A linear run will be used to assess the acceleration, operational speed and deceleration performance of the craft. Turning radius will be assessed using a u-turn maneuver, and extended ight tests will be used to conrm the endurance of the craft. Successful completion of this goal will be achieved when all design parameters (Table 1.1) are physically tested and conrmed.

2. To provide adequate control and stability to the hovercraft such that it can hover while stationary. Outline: Control and stability mechanisms, will be designed such to limit the horizontal drift of the hovercraft to within a halo radius of two metres.

Measure of Completion: The hovercraft will be tested under the static hover scenario to ensure that it will remain within

1.2. PROJECT OBJECTIVES

a 2 metre halo radius of its starting position for at least 30 seconds. The hovercraft will be tested indoors, away from any signicant external airow and on a at, level surface.

3. To manufacture a hovercraft to accommodate an o-the-shelf metal detector array Outline: Control and stability mechanisms must be adequate enough to accommodate the limitations of the sensor. Furthermore, the platform must have sucient means to function with a payload of 20kgs.

Measure of Completion: The sensor will be installed and tested by placing a target within the sensors line of sight as it traverses a linear path. Should the hovercraft be able to perform all of the functionalities outlined in objectives 1 and 2 while under payload, this objective will be declared complete.

4. Hovercraft must be able to identify and physically mark the location of mines. Outline: The completed hovercraft, with installed payload, will be able to identify a target representing a land mine. When the hovercraft passes over the mine it will leave a physical mark within a 1 metre halo radius of the target.

Measure of Completion: The completed hovercraft will be set on a linear path whereby it will pass over a target representing a land mine. By completion of its run, it will be required to leave a physical mark within a 1 metre halo radius of the target.

1.2.2

Extended Project Objectives

Extended project objectives represent an increased project scope that will compliment and enhance, but not dene the project.

1. The hovercraft will autonomously follow a predened path over at terrain with no obstacles Outline: Provision will be made as a part of the control system, for the hovercraft to follow a predened

CHAPTER 1. INTRODUCTION path, based on Global Positioning System (GPS) feedback. The path will be over at terrain, with no obstacles.

Measure of Completion: A basic path though a eld composed of straight lines and u-turns, will be predened and uploaded to the hovercrafts host. Upon activation of the hovercrafts automatic control systems, the GPS location of the hovercraft will be continuously recorded and compared to the desired path to access the precision between the two. 2. GPS tracking of the hovercraft combined with graphical user interface Outline: An installed GPS on the hovercraft will be utilised such that the location of the hovercraft may be tracked remotely, in real time, via a Graphical User Interface (GUI).

Measure of Completion: The hovercraft will be manually navigated around a large course, across a predened track. During the test, the GPS data will be streamed in real time to produce a comprehensive account of the hovercrafts position for the duration of the test. The same track will then be traversed, by foot, with a hand held GPS to produce a series of way points along the path. Both points will then be compared for precision. Successful completion requires that the GPS measurements be within 2 times the accuracy of the least accurate system.

3. Electronically mark the position of mines Outline: Positive responses from the payload sensors will be recorded electronically, in real time, using GPS co-ordinates.

Measure of Completion: Targets will be manually placed around a eld, and their GPS co-ordinates recorded using a hand held GPS unit. The hovercraft will then be remotely controlled such that the sensors pass over the target mines. The co-ordinates recorded by the on board GPS will be compared to the initially recorded points to assess the precision between the two. Successful completion requires that the GPS measurements be within 2 times the accuracy of the least accurate system.

4. Operation in non-ideal environments

1.3. PROJECT DEFINITION Outline:

The hovercraft will be tested in more challenging environments including cambered, sloped and uneven terrains, and in more challenging environmental conditions including light rain and wind (weather dependent).

Measure of Completion: Additional test sites will be located and surveyed in terms of their environmental parameters. The basic tests used to assess the competition of Goal 1 will be repeated within these varied environments to conrm required operating performance for additional environments and conditions.

1.3
1.3.1

Project Denition
General Requirements

This following considerations must be taken into account when designing the platform.
Sizing Restrictions

Given that this is a prototype platform, a large amount of testing will be required. In order simplify transport, it desirable that the platform should t entirely into the tray of small trailer, and hence should be no larger than a 1800 by 1200 mm rectangle. Designs that maximise the width of the sensor sweep of the craft are to be given preference.
Weight Restriction

To make transportation easier, the platform should not require additional equipment to load and unload it from the transportation vehicle, and as such the platform, including payload, is limited to 100 kg. Lighter designs are to be preferred to minimise power requirements and to ease transport.
Endurance

As this is a proof-of-concept design, the prototype platform requires minimal endurance for testing purposes. As such, a minimum endurance of 10 minutes is required for testing purposes.
Control

The hovercraft is required to be unmanned, and should be controlled by remote and operate autonomously for some functions. Remote control should be used for testing purposes, as well as overriding autonomy in the event that these features fail.

6
Hover Height

CHAPTER 1. INTRODUCTION

The hover height of the platform determines the height of the obstacle that the platform can traverse, hence higher hover heights are advantageous to the application as it expands the range of operating environments that the hovercraft can be applied to. Despite this, the hover height is also restrained by the operating range of the sensors.
Payload

The platform should be capable of housing and operating mine detection hardware. To allow for a range of sensors, and processing equipment, a payload capacity of 20 kg has been dened.

1.3.2

Operational Environment

Test Environment The test environment represents the ideal operating environment for the hovercraft, that is a static environment away from inputs that will degrade the performance of the craft in operations. This environment expresses the simplest and minimum requirements upon which the hovercraft is required to meet the specied performance levels. Such an environment will provide minimal drag, and as such will be inside, on a at, smooth concrete surface, away from wind and other weather eects. The features of this environment include:
Inside Negligible Slope (<1.5 ) Negligible Wind (<3 m/s) Concrete oor Room temperature

Field (Design) Environment Land mines are located worldwide, and hence inhabit a variety of environments, some which are more suitable than others for the application of a hovercraft. The eld environment expresses the upper limit of the scope of this project, and relates to extended goal four (Section 1.2.2). This environment will be outside, with some slope, low vegetation (below hover height), light breeze (up to 3 m/s). This platform will not be required to operate in either marine or aquatic environments.

1.3. PROJECT DEFINITION

1.3.3

Operational Scenario

Mission Prole The mission prole for the platform will follow the process as outlined in Figure 1.1 This mission prole with be achieved using a series of modes of operations for the hovercraft.

Figure 1.1: Mission Prole

Modes of Operation
Home

Upon entering the eld the initial mode should be to move to the home position

8
Mine Sweep

CHAPTER 1. INTRODUCTION

The Mine Sweep mode of the hovercraft will consist of autonomous navigation by traversing between preprogrammed way points.
Manual Navigation

Manual, remote control of the platform will be invoked for advanced manoeuvres, such as obstacle avoidance and course corrections, in the event that they are required.
Post-Detection

When a positive signal is received from the mine detection sensor payload, the craft will perform the following action sequence: 1. Log GPS co-ordinates 2. Perform back-up loop to conrm the reading 3. Positive response from back-up loop initiate physical marking system 4. Negative response from back-up loop, the platform should store the GPS co-ordinates in a list of possible positives 5. Return to Mine Sweep mode
Return Home

Upon commencement of sweep, or due to depletion of power supply, the platform should proceed along a preprogrammed path back to the home position.

Manoeuvrability Requirements
Lift o

The hovercraft will be required to lift o the ground to the appropriate hover height.
Stationary Hover

The platform will be capable maintaining a stationary position while hovering.

Forwards

The platform will be required to accelerate to its operational speed and continue forwards while maintaining a straight course. The operating speed is dened as walking speed, that is 4-6 km/hr as this is the operating speed of current mine detection methodologies. The platform should accelerate to this speed within 10 seconds.

1.4. SCOPE
Reverse

The reverse manoeuvre is required for the hovercraft to be able to rescan a potential sensor target, as to conrm its presence. The reverse manoeuvre is similar to the forwards manoeuvre in that is must accelerate to operational speed (5 km/hr) and maintain a straight course. Acceleration to this speed should also be within 10 seconds.
Left and Right Turns

The platform must be capable of pivoting in yaw such that it may successfully transverse the mineeld. Tight turns are required, and hence a maximum turning radius of 2 times the craft length is required.
Braking

The braking manoeuvre is required to bring the platform to rest from operational speed, and into the Stationary Hover mode. Deceleration to rest should be within 10 seconds.

1.3.4

Performance Specications

Table 1.1 summarises the performance specications that have arisen from the various requirements which have been discussed in this section. Table 1.1: Summary of Performance Requirements Parameter Value Length <1800 mm Width <1200 mm Total Weight <100 kg Nominal Operational Speed 5 km/hr Turning Radius <2x length Acceleration >0.14 m/s Deceleration >0.14 m/s

1.4

Scope

This project is restricted to a proof of concept design by which we seek to conrm that mine detection hardware can be successfully mounted and operated on an unmanned hovercraft with autonomous features. This report outlines the initial stages of design from literature review, through feasibility study, conceptual and destailed design, manufacture and testing of the protoype platform.

This Page Intentionally Left Blank

Chapter 2 Feasibility Study

2.1

Hovercraft Benchmarking

In order to determine the hovercraft technology available a market study of hovercraft was undertaken to identify dierent congurations implemented to date. The majority of the hovercraft looked at are hovercraft sold for leisure or commercial purposes. Four smaller hovercraft in particular were identied and are summarised below based on there variances in conguration. The only unmanned hovercraft identied were small remote control hovercraft models.

Hovertechnics HoverJet

Figure 2.1: Hovertechnics HoverJet (Hovertechnics, 2009)

Hoverjet is a leisure hovercraft sold by hovertechnics and is able to seat 2 people. The hoverjet implements a segmented skirt conguration and has integrated lift and propulsion systems. This means that the hovercraft may only hover when moving forward. The fan is run from a rotax engine and is of axial conguration. The hull is made of a Kevlar composite structure and the craft is able to obtain maximum speeds of 64 k/h. 11

CHAPTER 2. FEASIBILITY STUDY

Figure 2.2: AirCommander AC4 Hovercraft (Air-Commander, 2009)

AirCommander AC4 Hovercraft The AirCommander AC4 is used as a reghting hovercraft and is principally designed to traverse water and swampy regions. The lift and propulsion system of this hovercraft are independant meaning the hovercraft can hover while stationary, and run o independant motors. As can be seen in Figure 2.2 the lift fan is installed into the front of the hovercraft and propulsion at the rear. The AC4 again implements a segmented type skirt conguration and uses a composite hull. Tis hovercraft is able to seat 4 persons and lift a 1 tonne payload.

Amphibious Marine Vanguard 14

Figure 2.3: Amphibious Marine Vanguard 14 (Amphibious, 2009) The Vanguard 14 by Amphibious Marine is a 3-4 persons hovercraft capable of lifting a 350 kg payload.

2.2. LANDMINES

THe craft structure is primarily a composites, aluminium hybrid, with aluminium being used to mount engines. The thrust and lift are again independant systems, and a ducted fan is used as propulsion vectored using rudder control surfaces. The Vanguard 14 implements a segmented bag type skirt which is made up of 4 main compartments (fore, aft, port and starboard). The front bag is able to be dropped to cause a drag and eective braking of the craft.

Pacic Hovercraft Slider

Figure 2.4: Pacic Hovercraft Slider (Pacic-Hovercraft, 2009) Finally the slider by Pacic Hovercrafts is a New Zealand made vehicle and implements a bag and nger type skirt. It is the only small hovercraft benchmarked under 4 m length to do so. Although there is only one motor (does not have mechanical isolation), the lift and thrust systems run o independant fans. Two centrifugal fans are used for lift and an axial fan for thrust. This hovercraft is able to obtain max speeds of 80 k/h and has been utilised in military applications. Its average payload capacity is 275 kg with a length to beam ratio of 1.8.

2.2

Landmines

It is estimated there are currently more than 100 million active landmines around the world, with 100 million more in storage (Bruschini and Gros, 1998). Of these there are more than 350 known land mine designs, varying in weight, size, material and power (Unicef, 2000). Landmine statistics have indicated over 65 eected countries around the world. Landmines are responsible for approximately 26,000 deaths worldwide, per year; half of which are caused from severe injuries, before the victim is found or taken to hospital (Buse, 1999). The cost to deploy a landmine can vary between $3 and $30, whist is can cost up to $1000 to remove a single landmine (Products, 2007). It is estimated it would cost a minimum of 33 billion dollars to

CHAPTER 2. FEASIBILITY STUDY

remove all currently active landmines and at the current removal rate it would take over 1000 years.

The United Nations (UN) have set detection goal of 99.6% for the detection and removal accuracy of landmines and a false alarm rate of one every 1.25 square meters. Currently no detection methods are capeable of acheiveing these standards. This is due to a number of factors such as varying landmine types, environmental conditions and technology limitations (Oce, 2001).

Operational landmines are normally buried in the rst 300 mm of soil and are found in all open terrain types (Stull, 1997). Buried landmines are most often pressure activated and often have an alternate ring mechanism . Landmines types can typically be classied into two categories, Anti Tank (AT) mines and Anti Personnel (AP) mines, also referred to as blast mines and fragmentation mines, respectively.

AT mines, as their name suggests, are target towards tanks and vehicles and are commonly found in open areas such as roads and elds. They are typically the larger of the two mines and contain signicant explosive content, having an average weight and diameter of approximately 10 kg and 280mm, respectively (Figure 2.5). AT mines are most commonly activated using a pressure trigger, requiring an activation pressure in the order of 100s of kg to detonate (Das et al., 2002).

Figure 2.5: Physical composition of an AT landmine (HowStuWorks, 2009) and (Gudmundson, 2008)

AP mines, targeted towards humans, are far smaller than AT mines, often found in dense environmental locations such as scrub and walking trails. AP mines not only contain explosive content but metallic fragments, used as high speed projectiles (Figure 2.6). They may be triggered from a number of high sensitivity activation mechanisms such as a tripwire, sound levels and pressure.

The landmines detailed above are considered traditional landmines, laid post World War 2 and are made mainly from metallic components. More recent technology has lead to the invention of composite landmines, which contain little to no metallic components making them incredibly dicult to detect.

2.2. LANDMINES

Figure 2.6: Physical composition of an AP landmine (HowStuWorks, 2009) and (News, 2009)

2.2.1

Detection Methods

Direct Detection
Direct detection traditionally involves a human operator physically searching through the mineeld (Figure 2.7). This is one of the slowest and most expensive methods of landmine detection, costing approximatley $1000 and taking up to four hours to remove a single landmine. Further it is the most dangerous form of detection, resulting in one death for every 1800 mines that are removed (Bruschini and Gros, 1998).

Figure 2.7: Ocer using a probe for landmine detection (Wikipedia, 2009) Subsequent forms of direct detection involve an operator with a metal detector or another form of landmine detection sensor, such detection methods are quicker than a probe, but are less accurate and still place the operator in great danger. More abstract direct detection methods use animals, such as dogs, ferrets, rats and bees to detect the explosive content in landmines. Animals are highly accurate for specic landmine types and environmental conditions, but are extremely expensive and

16 time consuming to train.

CHAPTER 2. FEASIBILITY STUDY

Mechanically Assisted Detection

Vehicle mounted sensors are the primary form of direct detection (Figure 2.8). Vehicle mounted sensors are far larger and heavier than hand held sensors and provide a considerably increased ground coverage (Gooneratne et al., 2004). These methods produce a great risk of premature detonation, placing the equipment and more importantly the operator at risk. Such methods are also very expensive.

Figure 2.8: A vehicle mounted landmine detection sensor (qsine, 2005) More severe mechanical methods aim to detonate landmines, rather than detect them. The most common method of vehicle landmine detonation is a front mounted ail. A ail operates from a number of large metal balls mounted on the end of a chain and subsequently to a rotating axle, that are used to pound the ground and trigger any landmines. This method is one of the most accurate and fastest forms of landmine deactivation, however results in signicant environmental damage.

2.3

Autonomy

Kelly et al. (2006) designed and built an autonomous vehicle to operate in a challenging environment and also to map an area following a predened path, consisting of multiple way points (Figure 2.9). The sensors used were odometry sensors, Dierential GPS (DGPS), ground speed radar and an Inertial Measurement Unit (IMU). These sensors are input into a Kalman Filter (KF) to provide accurate state estimates. In order to incorporate obstacle avoidance, the vehicle also used 2D scanning ladars, stereo cameras, infrared cameras and a pressure sensitive bumper. To process all the input data the vehicle uses ve commercial desktop computers, one for sensor interfacing, one for creating depth maps from

2.3. AUTONOMY

the stereo cameras, one to log the video which can be analysed at a later date, one for auxiliary interfaces and one for all the actuators and low-level sensors.

Figure 2.9: Autonomous Vehicle (Kelly et al., 2006) Roth and Batavia (2002) developed a formal method for comparing and evaluating the performance of path tracking algorithms for their automated turf management vehicle, shown in Figure 2.10. The design goals of the vehicle were to nd small objects with no false positives, detect all true obstacles to keep the vehicle safe, operate within cm level precision and must be able to cover area in an ecient and eective manner. The sensors used were a GPS, IMU, stereo based homography and a laser sensor. The combined GPS and IMU are used to determine accurate position combined with a Zhang path tracker (Zhang et al., 1997) to move the autonomous vehicle in complex manoeuvres with little deviation from a dened path. To avoid obstacles the stereo based homography is used to detect objects which lie above the ground plane with the laser scanner used as a backup detection system for instances where vision-based methods fail.

Figure 2.10: Automated Turf Management Vehicle

CHAPTER 2. FEASIBILITY STUDY

Mandapat (2001) developed an autonomous system for use with their navigation test vehicle and then performed a comparison between three dierent congurations of IMU, GPS and KF. The rst conguration used an expensive high accuracy DGPS and a high data rate IMU, which showed positioning accuracy of under 10 cm at a data rate of 10 Hz. The second conguration was a GPS with ability to measure position and orientation, which had accuracy within 20 cm at a data rate of 5 Hz. The third conguration uses a KF on an IMU with a low cost DGPS, and then a second KF on the combined IMU and GPS data. This conguration achieved an accuracy of 20 cm at the fastest data rate of 12.5 Hz. The IMU/GPS compares favorably in that its adequate performance and low cost makes it ideal for autonomous ground vehicle navigation (Mandapat, 2001).

Chapter 3 Conceptual Design

3.1
3.1.1

Preliminary Sizing
Statistical Analysis

To aid in the preliminary sizing of the hovercraft, a market survey of 26 light hovercraft (less than one tonne) was performed and their dimensional data, namely empty weight, payload weight, length and width, was collated for statistical analysis. The Pearson correleation coecents where calculated for the empty, payload and total weights of the craft, as well as their length, width and beam ratios. These results are summarised in Table 3.1. Table 3.1: Pearson Table WP L WT OT AL Length (L) 0.89275 0.96568 0.68036 < 0.001 < 0.001 < 0.001 1 0.97913 0.80203 0 < 0.001 0.97913 1 0.76922 0 < 0.001 0.80203 0.76922 1 < 0.001 < 0.001 0.57483 0.51398 0.45559 0.002 0.007 0.002 0.36739 0.37581 0.68 0.065 0.058 < 0.001

WE WP L WT OT AL Length (L) Width (W) L/W

Pearson Sig. Pearson Sig. Pearson Sig. Pearson Sig. Pearson Sig. Pearson Sig.

WE 1 0.89275 < 0.001 0.96568 < 0.001 0.68036 < 0.001 0.40484 0.040 0.36366 0.068

Width (W) 0.40484 0.040 0.57483 0.002 0.51398 0.007 0.45559 0.019 1 -0.3335 0.096

L/W 0.36366 0.068 0.36739 0.065 0.37581 0.058 0.68 < 0.001 -0.3335 0.096 1 -

3.1.2

Beam Ratio

As part of the market survey, the beam ratio of 26 hovercraft were sampled. Given that there was no independant, statistically signicant correlations between the beam ratio and the other deign constraints (Table 3.1), the beam ratio can be considered an independant quantity, and hence the use of any commonly used beam ratio is warrented. Figure 3.1 displays a frequency histogram of the beam ratio for the sample, while the desciptive statistics are summarised in Table 3.2. The beam 19

CHAPTER 3. CONCEPTUAL DESIGN

F re q u e n c y
3 2 1 0

1 .0

1 .2

1 .4

1 .6

1 .8

2 .0

2 .2

2 .4

B e a m

R a tio ( L /W )

Figure 3.1: Beam Ratio Histogram Table 3.2: Beam Ratio Descriptive Statistics N 26 Mean 1.788 Standard Deviation 0.31164 Lower 95% CI of Mean 1.66182 Upper 95% CI of Mean 1.91357 Minimum 1 1st Quartile 1.63 Median 1.8 3rd Quartile 2.03 Maximum 2.25 ratios for the sample encompass a wide range of possible values, with possible values ranging from 1 to 2.25. Despite this, the histogram shows that the far majority of the data is grouped between 1.5 and 2.3, which is conrmed by the borders of the interquatile range. Hence, while it is possible to design hovercrafts with a beam ratio of unity, the majority of craft have ratios higher than 1.5. This minimum beam ratio was selected as per the requirement to maximise the possible sensor sweep of the craft for a given craft area.

3.1.3

Planform Sizing

As per the general requirments for the platform, the entire hovercraft is required to t in a 1800 x 1200 mm rectangle. Allowing 100 mm for the protrustion of additional equipmemnt such as the skirt and skirt attachments, reduces the maximum planform sizing to 1700 x 1100 mm. Given the

3.1. PRELIMINARY SIZING

5 0 0

4 0 0

3 0 0

E m p ty (k g )

2 0 0

1 0 0

0 0 1 0 0 2 0 0 3 0 0 4 0 0 5 0 0

P a y lo a d ( k g )

Figure 3.2: Linear Regression Model for Dry Weight Estimation

precribed beam ratio of 1.5, width becomes the limiting factor, and a planform sizing of 1100x1650 is appropriate. The length of the hovercaft was reduced to 1600 for simplicity, given a nal planform sizing of 1600 x 1100 mm.

3.1.4

Weight Estimation

Statistical relations based on the aforementioned market survey were used as the basis for the initial weight extimation of the platform. To determine the best available predictor variable of empty wight; payload weight, length and width were assessed for their correlation with dry weight, for a variety of hovercraft. The corresponding Pearson correlations are summarised in Table 3.1. Both length and payload weight display high correlations with empty weight, with Pearsons coecients of 0.89275 and 0.68036 respectively. Both correlations are statistically signicant at the 99% condence level(p<0.001) but it is clear that payload weight is the best linear predictor of the empty weight of the hovercraft. The linear regession model for this relation was calculated (see Figure 3.2) , and the result is Equation 3.1. The model was a good t, with a coecient of determination of 0.788.

WP L = 1.432WE 31.29

(3.1)

Based on to dened payload weight of 20 kg, an initial, dry weight estimation of 55.8 kg. Appling a safety factor of 1.5 gives a design dry weight of approximately 85 kg.

CHAPTER 3. CONCEPTUAL DESIGN

3.2

Prototype Developement

Four prototype designs were generated using a combination of possible systems and congurations. These prototypes are described and assessed in this section. Development of the overall conguration was heavily inuenced by existing congurations and benchmarking, but the unique requirements of this platform also needed to be taken into consideration. Three major congurations were considered: 1. Integrated Lift and Thrust Systems
Lift Dominant Systems Thrust Dominant Systems

2. Independent Thrust and Lift system

3.2.1

Integrated Lift and Thrust Systems

Lift Dominant Systems

With a lift dominant system, the craft structure is centered primarily around lift system. The primary advantage of such a conguration is the reliability of the lift system and limited eect on pitch and roll stability of the craft given the low moment created by an integrated thrust system. One of the initial designs conceived was a platform build around a centrifugal fan with the integration of pu port around the craft which could be actuated by a series of buttery values A high daylight clearance is more probable given the nature of the lift system and its overcall inuence on the vehicle. One of the major conceived disadvantage would be high levels of precession excepted with a dominant centralised lift system, however with the addition of pu ports the control of the hovercraft could be highly rened mitigating such a risk. However, exceptional levels of maneuverability would come at the price of low levels of thrust. Limitation in operational terrain would apply and most importantly the inability to traverse ground with variable slopes would be lost. An additional disadvantage is that pu porting could potentially cause variations in the pressure distributions of the air cushion. This is undesireable for the application of mine detection as if the pressure lowers enough to cause contact with the ground there is high risk of setting o mines.

3.2. PROTOTYPE DEVELOPEMENT

Figure 3.3: Splitter Conguration

3.2.2

Thrust Dominant System

A vehicle designed around its thrust system can be considered to be thrust dominant. The primary advantages of such a system are the high levels of thrust and consequently high operational velocities. Air from the thrust system is drawn o, in most cases by installing a splitter behind the axial thrust fans, and directed at high velocity into the plenum chamber. This principle is shown in Figure 3.3. The primary disadvantage of this system is that the platform is unable to statically hover as thrust is required to generate lift. Due to the operational scenarios of the hovercraft, in particular the desired ability to stop and conrm positive detection signals, this characteristic is undesireable.

3.2.3

Independant Thrust Lift Systems

A combination of the two systems can be seen in the independant thrust lift system. In such a conguration, no one system is dominant, allowing for better control of design variables. Although a weight penalty must be endured for incorporating a independent systems, greater levels of reliability and high levels of maneuverability are achieved. An independent thrust and lift system allows for an inherent fail safe mechanism in the platform. In the event the thrust system fails, the craft is still in hover. This is ideal for the operational environment of the platform.

CHAPTER 3. CONCEPTUAL DESIGN

3.2.4

Prototype Generation

Some characteristics of the platform were able to be selected prior to the generation of dierent congurations. It was decided initially that a skirted craft was required to maintain an acceptable level of eciency and controllability to the platform. This left the main prototype decisions to be between independant and integrated lift and thrust, and the method of manoeuverability control.

Prototype 1

Prototype 1 represents a typical small hovercraft design in that a single, large, engine and fan is used to provide both lift and propulsion. The ow from the single ducted fan is split into two streams, generally using a splitter plate, with approximately one third of the ow being redirected in the plenum chamber to generate lift. The remaining ow is used for propulsion. Control of the craft is provided by rudders to adjust the direction of the thrust vector and cause a torque in yaw. The design is skirted to increase stability and eciency of the craft.

Prototype 2

Prototype 2 incorporates three fans and engines, one for lift and the other two for propulsion. The lift fan is the standard horizontally orientated, axial fan and petrol engine combination. The two propulsion sources are electric motor controlled propellers free to rotate around the vertical axis to provide vectored thrust. These two fans are positioned at the aft edges of the craft, and each system has two degrees of freedom, angular position and throttle. This conguration boasts high manoeuvrability and the ability to move in reverse and brake.

Prototype 3

Prototype 3 was used to demonstrate the possibility of dierential thrust as a control mechanism. The dual hull conguration has unecessary complexity in that two skirts and two air sources (or splitting of one source) is required to achieve lift. The main disadvantages of this design is stability analysis of the platform becomes complex and as the cushion does not cover the entire area of the platform, pressures may exceed the limit for preventing the detonation of landmines.

3.2. PROTOTYPE DEVELOPEMENT

Figure 3.4: Prototype 1 and Prototype 2

Figure 3.5: Prototype 3 and Prototype 4

26 Prototype 4

CHAPTER 3. CONCEPTUAL DESIGN

Prototype 4 implements integrated lift and thrust systems by using a single, horizontally orientated axial fan and engine combination. The outlet air from this system is split, with part of the ow being directed into the air cushion as lift. The remaining ow is fed into a duct network with several xed outlets positioned around the perimeter of the craft. Each of these outlets have variable valves attached to them to vary the output thrust from each. Selective operation of these valves relative to each other allows the hovercrafts net thrust vector to be manipulated for propulsion and control of the craft. The main disadvantage of this system is that dicult control strategies are required to achieve control with so many variables, and an accumulator may be required to maintain sucient pressures in the piping for thrust.

3.2.5

Prototype Selection

The dominating factor for the selection of lift and thrust system integration was the ability to statically hover. Static hover is dened as a necessary component of performance to achieve the operational requirements of the platform. It was therefore selected to use independant lift and thrust systems and account for extra weight requirements.

Manoeuvrability Control Mechanism

Three primary methods of manoeurability control were recognised

Rudder & Elevon Congurations

Control surface maneuvrability, as implemented on conventional aircraft, is inecient at low speeds (Amyot, 1989). Implementation on a hovercraft therefore requires installation within the slip stream of a propeller or air jet. There are two main control surfaces implemented on hovercraft, rudders and elevons. Rudders are used for controlling the hovercraft in yaw and allow the generation of a turning moment through the redirection of the thrust. Rotation of the rudder surface eects this force. Elevons work on the same principal as rudders however are orientated horizontally to create a moment in the direction of pitch (Bliault and Yun, 2000). They are primarily used to dynamically trim the craft and ensure pitch stability during operation. The limitation of such control surfaces is that they are only eective for a range of less than 180 degrees.

3.2. PROTOTYPE DEVELOPEMENT Dierential Thrust

Dierential thrust may only be implemented in multiple propeller congurations and involves the creation of a yawing moment through a dierential in output between two sources. Dierential thrust is eective at driving yaw, however has the disadvantage that decreasing the thrust on one side decreases the overall speed of the craft (Amyot, 1989). Additionally there is no mechanism to produce a sidewards force, meaning this form of propulsion has inherent low maneuverability relative to other thrust vectoring methods. (Amyot, 1989) Dierential thrust is relatively simple to implement as no complex mechanisms are required.

Direct Thrust Vectoring Direct vecotoring of a thrust stream is the most eective method in terms of maneuverability as it allows 360 degree rotation of the control vector. Direct thrust vectoring is complex to implement relative to other methods as a complex mechanism is required to rotate the propeller assembly itself. It maintains the high eciency of air-screw propulsion systems, however, safety must be considered due to the exposure of propeller blades. Additionaly, rotation of the propeller assembly demands expensive spacial requirements. This method is eective as yawing moments can be created without slowing the speed of the craft. Propellers may be completely reversed to provide backward thrusting or braking of the platform. The ability to statically hover was considered critical to the operational scenarios of the platform, and therefore lift dominant methods were eliminated. Traditionally integrated propulsion methods do not allow the hovercraft to remain stationary while in hover. This is because the propulsion fan must be on to produce lift, and results in an accompanying thrust force. This incurred the elimination of thrust dominant systems. Due to the application of the platform in mine detection it was necessary to ensure the platform does not lower in operation. In order to achieve this, it was necessary to ensure that pressure distribution in the air cushion remained constant. As such, air cushion integrated methods such as pu ports were also eliminated. Control surface based control systems were considered however were unable to provide the 360 degree of rotation required to provide the backward movement and braking capability necessary to achieve the platforms operational behaviour. Another braking mechanism was identied to integrate with control surface systems. The ability to drop or expand the skirt at the front of the craft induces a drag at the front resulting in very eective braking. These methods were disregarded due to high drag increasing the risk of mine detonation. Based on these evaluations, Prototype 2 was selected. This conguration include

CHAPTER 3. CONCEPTUAL DESIGN

3.2.6

Prototype Specic Requirements

Prototype selection brought about a range of specic criteria used as limitations in detail design. These requirements are summarised below:
The craft should be skirted to optimise eciency A plenum chamber conguration of the hull should be adopted. The propulsion system should be vectored thrust and be capable of 360 degree vectoring, two

fans were selected so that smaller propellers could be used.

The lift system should idependently achieve the ow rates required to lift 100kg.

3.3

Power Requirements

Weight minimisation is a key feature of hovercraft design. While ideally the overall weight of a hovercraft should be as low as is possible, in reality, there are only so many resources that should be invested to this end. As such, sensitivity analysis is a valuable design tool that aids the designer when comparing the selection of certain congurations and items. The following section considered two such analyses, being the sensitivity of propulsion and lift requirements to the overall weight of platform.

3.3.1

Lift Power

The lift system must produce a steady state cushion pressure that will provide enough force to counteract the weight of the craft, shown by equation 3.2

Pc = f (M, g )

(3.2)

The sensity of lift requirements to weight were derived by dierentiating the governing equations (calculations detail in Appendix Q).This resultant graph is approximately linear, giving a sensitivity of 29 W/kg for the lift system.

3.3.2

Propulsion Power

The power requirements for propulsion are dictated by the equation of drag, and thrust required to overcome the inertia of the craft. The gross thrust is described by equation Q.5. Substituting in

3.3. POWER REQUIREMENTS

2.2

2.0

Power Lift System (kW)

1.8

1.6

1.4

1.2

1.0

100

Figure 3.6: Sensitivity of Lift Power to Mass the equations for drag, equation Q.7 respresents the total thrust requirement. By dierentiating this equation we obtain the sensitivity of propulsion to weight.

TGross = TN ET + DA + DC + DT + Df

(3.3)

1 QV TGross = (ma) + ( V 2 ) + + (mgsinT ) 2 Figure Q.2 shows a calculated sensitivty of 7.8 W/N.

(3.4)

3.3.3

Preliminary Thrust and Power Estimates

As discussed is section 3.3.2, the gross thrust requirements can be calculated as a function of net forces required to overcome drag. A Free Body Diagram (FBD) of the craft in Figure 3.8 shows the force loading on the craft. The drag forces associated with aerodynamics, the slope of the ground, and momentum of the cushion air can be calculated using the standard equations in section 3.3.2. The preliminary calculations are done respective to the design weight of 100kg. DA can be estimated using equation 3.5, giving a value of 3.5N. This calculation is based on a light breeze of 2m/s acting against a wedge shape cross sectional area. For this geometry Cd can be estimated at 0.5. (Munson et al., 2002)

CHAPTER 3. CONCEPTUAL DESIGN

1 0 0 9 0 8 0 7 0

N o S 1 d e 2 d e 3 d e 4 d e 5 d e g g g g g

lo p e

T h ru s t (N )

6 0 5 0 4 0 3 0 2 0 1 0 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0

W e ig h t ( k g )

Figure 3.7: Sensitivity of Thrust to Weight and Terrain Slope

Figure 3.8: FBD of Forces

1 D A = C d a V 2 2

(3.5)

DT is a drag force based on the slope of the terrain and may be estimated using equation Q.7. DT is calculated to be 25.7N. This calculation was done for a slope of 1.5 degree, giving the hovercraft some ability to manage uneven ground that may be required for testing. DC is a drag force associated with a change in momentum of the air underneath the cushion. The drag can be estimated from equation Q.7 and is calculated to be 2.9N. The TN ET is dened as the raw thrust required to to accelerate the craft from rest to its operational speed, disregarding all other forces. This has been calculated using newtons second law and is

3.4. PLANFORM DESIGN estimated at 11.9N based on the 85kg preliminary weight.

No empirical formula is available to calculate DS and thus it must be estimated using statistical methods highlighted in Bliault and Yun (2000). Figure 3.9 shows the estimated drag over dierent terrains. The y-axis represents the hoverheight of the craft in m, and the x-axis the skirt drag as a percentage of weight. Although the design limitation on clearance of the craft is 15mm, a height of 5mm has been chosen to get an upper limit for skirt interaction drag. Testing of the platform will only occur on short grass and concrete and so the short grass has been used to estimate a skirt drag of 5N.

Figure 3.9: Skirt drag over dierent terrains (Bliault and Yun, 2000) This concludes a total thrust requirement of 49N to operate the craft at a 1.5 degree slope, on short grass.

3.4

Planform Design

The planform of the hovercraft has two main proposes/eects:

Provides a surface upon which all equipment is mounted/housed Provides lift surface

Hence, the planforms upper surface must be sized and shaped to accomodate all component systems, while it lower surface provides the area upon which the lifting force is applie, as well as the point upon which this force is centred (centre of pressure), that is the centroid of the area. Given that the platforms propulsion system is located entirely in the aft section of the craft, this region should be at a maximum. In addition, the location of the propulsion equipment here means that the hovercraft will

CHAPTER 3. CONCEPTUAL DESIGN

Figure 3.10: Planform Design be back heavy, and hence the centre of gravity of the craft shall shift aft of the rectangular centroid. This condiditon will destabilise the hovercraft was the centre of gravity will no longer co-inside with he centre of pressure of the cushion. As such, planform area in the fore section of the craft was removed to reduce the decit between the centre of gravity and centre of pressure locations. Figure 3.10 shows the nal planform of the craft. Curves were incorporated into the design for asthetic purposes and to remove sharpe corners that would complicate skirt design. The central area is resevered for the lift system, the aft for propulsion and the fore section for payload.

3.5

Payload Area

The hovercraft hull is designed with two designated payload areas, a front and rear payload bay (Figure 3.11). The front payload bay is required to accommodate a number of interchangeable landmine detection sensors and and all associated components, such as power and data processing equipment. The payload bay is situated in the front third of the hovercraft and includes both the internal hull and top plate area. The rear payload bay is situated at the rear of the hovercraft, between the two thrust modules and is

3.6. LANDMINE DETECTION SENSOR

Figure 3.11: Hovercraft hull showing front and rear payload bays

to house the physical marking system and on board electronics. The overall size of the rear payload bay is signicantly smaller than the front payload bay, whilst still including the internal hull and top plate area. The physical marking system is to be situated along the central axis of the hovercraft, providing the highest accuracy mark.

3.6

Landmine Detection Sensor

The landmine detection sensors tted to the hovercraft are required to detect the presence of an object resembling a landmine. Multiple sensor types and congurations have been considered for integration into the hovercraft, with the aim to develop a number of interchangeable sensor congurations.

3.6.1

Specications

The landmine detection sensors are to be mounted in the allocated front payload bay, located in the bow of the hovercraft (Figure 3.11). The front payload bay is limited to a maximum width, 1100 mm, depth, 600 mm and weight, 20 kg; from the platform design specications. It is desired to maximize the detection width of the landmine detection sensors, increasing search eciency and accuracy.

CHAPTER 3. CONCEPTUAL DESIGN

Simulation Landmine

The hovercrafts landmine detection sensors are required to detect the presence of landmines, however in tests scenarios it is not possible to use real landmines due to practicality and availability. To represent a landmine in test scenarios a simulation landmine has been developed. The simulation landmine was based on an AT landmine rather than a AP landmine. AT landmines were chosen for a number of reasons. Firstly their environmental location is far more suited to the hovercrafts physical size and maneuverability. Secondly they are signicantly larger than AP landmines and contain more metallic content, making them easier to detect and more suited to our proof of concept platform. The development of a simulation landmine was done through a statistical analysis of over 100 AT mines (see Appendix N). The simulation landmine developed is shown in gure 3.12. It is made from 1.5 mm thick sheet steel, determined from the average case weight distributed over the landmine surface area.

Figure 3.12: Simulation landmine

3.6.2

Sensor Selection

Various landmine detection sensor types have been evaluated based on their detection ability and practical integration as a module into the hovercraft. Further their ability to detect composite landmines has been considered, as an extended goal.

3.6. LANDMINE DETECTION SENSOR

Ground Penetrating Radar

Ground Penetrating Radar (GPR) have been used extensively in landmine detection for their detailed ground imagery and detection capabilities of both traditional and composite landmines. Data provides a detailed subsurface density spectrum image and indicates the presence of landmines by identifying edges between density mediums. Landmine detection is most suited to high frequency GPR, providing the detail and depth penetration required, which low frequency GPR can not (for literature review of GPR see Appendix O). The practical operation of GPR is greatly limited by their operational requirements, mainly their low ground clearance required to be maintained within approximately 30mm of the surface. This is particularly problematic for integration into the hovercraft as a substantial ground clearance is of 150mm, equal to that of the hover height, is required for operational conditions. A subsequent limitation for the implementation of a GPR its physical detection width. From expert consultancy (Alpha GeoScience, 2009) it has been determined that a GPR with an operational frequency in the order of 2 GHz would be required for our application. For the required frequency standard sizes indicate a detection width of approximately 160mm (Geoscience-AB, 2009), resulting in an array of 6 GPR units to achieve a reasonable, 1 metre, detection range. Such a large number of units is not desirable due to the severely increased data processing requirements from both multiple signals, potentially removing the functionality for real time data analysis.

Magnetometer Magnetometers have limited standalone use in landmine detection, more often used in conjunction with a number of other sensors. Their limited individual use is most accredited to their high number of false positives, resulting from their high sensitivity. However magnetometers also benet from such a high degree of sensitivity, able to easily detect the small ring pin in composite landmines. They are further beneted from a fast sampling rate, allowing high speed operation (for literature review of magnetometers see Appendix O). The sensitivity and substantial detection range of magnetometers raise a number of issues with nearby metallic components on board the hovercraft. With an average detection radius of up to 1.5 meters, all components on board the hovercraft are potentially in the magnetometers range. Stationary metallic components, such as mounts and bolts on board the hovercraft, produce a constant interference level which may be to an extent ltered through electronics and post analysis. However the extent of ltering is limited by the level of interference and its regularity, where vibrations can cause the interference

CHAPTER 3. CONCEPTUAL DESIGN

level to uctuate preventing eective ltering. Variable uctuations in moving metallic components, such as a petrol engine, produce a interference level which most often cannot be ltered. To prevent interference other forms of isolation may be used such as a Faraday cage or an extended mounting arm. A Faraday cage is able to shield the magnetometer from external interference, however they are made from a large number of metal windings and are required to encase the sensor with no gaps, making them extremely heavy and large, which is not suitable for integration into the hovercraft due payload size and weight restrictions. An extended mounting arm is a viable solution based on ground clearance and sensor damage issues alone. However an arm in excess of 1.5 meters would be required to ensure no interference from the lift engine, which is undesirable greatly increasing the hovercrafts overall length and reducing its maneuverability and stability.

Metal Detector

Metal detectors are the most common form of hand held landmine detection and are also widely used in vehicle mounted congurations. They operate in a similar fashion to magnetometers, with signicantly reduced sensitivity and range. A reduced sensitivity signicantly reduces the number of false positives ignoring small metallic components, however at the same time limits the ability to detect composite landmines. A smaller detection range reduces the interference eects from nearby metallic components, in both number and severity (for literature review of metal detectors see Appendix O). Whilst the reduced detection range limits mounting congurations, it results in the overall detection range not exceeding the payload bay totally eliminating interference eects from the majority of metallic components, such as the petrol engine and thrust system. With the detection range conned to the payload bay, interference prevention may be used as an alternative to shielding. Whereby all components capable of interference are removed from the payload bay, namely all metallic components, resulting in a metallically isolated payload bay.

Decision Matrix
The decision matrix weighs the relative benets and limitations of each sensor against their application into the hovercraft. It is clear from the decision matrix that a metal detector is the most appropriate sensor, based on the criteria considered. The criteria chosen consider strongly the sensor integration into the hovercraft and their limitations and to a lesser extent the overall performance and properties of the sensor.

3.7. STATIC STABILITY

Table 3.3: Landmine detection sensor decision matrix Category Weight GPR Magnetometer Metal Detector Ground Clearance 5 15 50 30 Interference 5 45 10 35 Sensitivity 4 32 40 20 Size 4 8 32 32 Weight 4 16 32 40 Data Processing 3 12 9 24 Detection width 3 30 9 21 Depth Penetration 3 21 27 12 Maneuverability Constraints 3 12 12 30 Cost 2 10 4 20 Physical Mount Complexity 2 8 8 16 Power Requirements 1 5 4 9 Total 214 237 289

3.7

Static Stability

The selected prototype has no specic control surfaces that directly control either pitch or roll motion. Instead, this congeration relies entirely on the inherent response of the air cushion to produce retorative moments in both pitch and roll to ensure the stability of the platform. This retorative response arises from the movement of the air cushions center of pressure relative to the centre of gravity of the craft.

To ensure the static stability of the craft in both pitch and roll, the projection of the centre of pressure on the ground must always be outside of the projection of the centre of gravity for the same input. Given that the hovercraft has a limited range of possible pitch and roll inputs as the hull makes contact with the surporting surface, static stability need only be assured over this range.

Based on the geometric properties of the craft, the maximum pitch and roll can be approximated by modelling the hovercraft as a at plat, pivoting around its geometrical centre at its hover height above a rigid surface. The angle at which it makes contact with the ground is the maximum pitch angle, and hence is determined by Equation 3.6. Similarly, the maximum roll angle can be approximated by Equation 3.7.

CHAPTER 3. CONCEPTUAL DESIGN

h = sin1 ( Lc )
2

(3.6) (3.7) (3.8)

h = sin1 ( B c )
2

Subsituting in the geometrical parameters in for the concept design, the maximum pitch and roll angles are approximately 3 and 4.5 respectively. The mechanism by which the air cushion shifts the centre of pressure depends on the conguration of the skirt. For the segmented skirt, as the craft heels, the skirt deforms as it contacts the ground, changing the shape of the air cushion on the ground and hence shifting the centre of pressure in response. Graham (1992) descripes a mathematical method by which this pressure shift can be simulated. Figure 3.13 shows the change in area that accompanies skirt deformation in pitch. The corresponding areas and moment arms can be calculated using the equations listed in Appendix B and assume the following:
Partial contact of skirt with ground Neglegibile momentum ux out of cushion Rectangular planform area Centre of gravity and centre of pressure located at the planform centroid for no pitch or roll

input
Assume height of centre of gravity to be 40 mm above hull top plate No heave motion accompanying pitch or roll
5

Mp = p c
i=1

Afi XM Pi

(3.9)

Afi XM Pi xc p =
i=1 5

(3.10) Afi

i=1

In order to ensure static stability, the moment arm (xc p) of the skirt must be negative accross the

3.7. STATIC STABILITY

entire range of possible motion for pitch. Solving these equation for [3, 3] produces Figures 3.14. Figure 3.15 was produced by exhanging the length and beam quantities, as well as the pitch and roll angles accross the possible roll range ( [4.5, 4.5]) . As can be seen by these graphs, the segmented skirt is only capable of producing a positive corrective moment for small angular deviations less (less than 1 ) for both pitch and roll inputs. While in practice, a greater range of stability is anticipated due to eects of reaction moments transferred though the skirt and arising from skirt stiness, as well as unbalanced momentum ux (Graham, 1992), this analysis makes it clear that without additional actuation systems such as horizontal rudders or active trim control, the segmented skirt alone will be insucient to provide stability in either pitch or roll. As such, other means of ensuring stability must be considered.

Figure 3.13: Change in area due to cushion deformation in pitch motion

Figure 3.14: Reaction arm of segmented skirt in pitch motion

CHAPTER 3. CONCEPTUAL DESIGN

Figure 3.15: Reaction arm of segmented skirt in roll motion

3.7.1

Application of Jupe Skirt to Enhance Stability

The jupe skirt is a more stable skirt variation. Like the segmented design, the jupe skirt induces reaction moments by altering the position of the centre of pressure of the overall air cusion, however they dont rely on the movement of the centroid of the cushion area to achieve this aim. Rather, as the jupe skirt come in contact with the ground, it essentially cuts o ow out of the jupe, allowing the pressure within the jupe to approach the pressure of the plenumn chamber, and hence exerting producing a localised area of higher pressure, producing a reaction moment. The arm of this reaction moment is determined by the geometric position of the jupe skirt in relation to the center of gravity. Hence, neglecting any other eects such as irregularities in air momentum ux out of the skirt, and assuming the total seal condition of the jupe in pitch or roll motion, the reaction moment of the hovercraft will always be positive provided that the projection of the platforms weight vector doesnt fall outside of the geometrical centre of the jupes.

Figure 3.16: Free body diagram of movement of weight vector due to pitch or roll motion (Sullivan et al., 1978)

3.7. STATIC STABILITY

Figure 3.16 shows graphically the movement of the projection of the weight vector along the plane of the hull. To ensure a positive restorative moment, the position of the jupe along the hull section (xR ) must be greater than the project of the weight vector (xCG ), as expressed in Equation 3.11, accross the entire range of pitch and roll motion. From this free body diagram, the movement of the centre of gravity projection is related by Equation 3.12.

xR < xCG

(3.11)

cos =

ZCG sin xCG

(3.12) (3.13) (3.14)

x CG = ZCG tan

Solving Equation 3.12 for all [3, 3], [4.5, 4.5] creates an ellipical area of potential centre of gravity movement. This semi-major and semi-minor axes of the ellipse are dened by the maximum and minimum values of the pitch and roll angles. The equation of the ellipse is summarised in Equation 3.15. Given the stability criterion (Equation 3.11), the jupe will be capable of producing a positive restorative moment. Provided that a minimum of four jupes are installed to allow for angular displacement, in both directions along the longnitudinal and transverse axes, static stability can be ensured.

y2 x2 + tan = 1 ZCG tan ZC G

(3.15)

This Page Intentionally Left Blank

Chapter 4 Detailed Design

4.1

Payload Area

Payload Bay
The hovercrafts payload is split into two main areas, namely a front and rear payload bay (as dened in Section 3.5). The front payload bay is required to accommodate the selected landmine detection sensors and the rear payload bay, the physical marking system and on board electronics. The front payload bay specications have been determined from the hovercrafts physical dimensions (Table 4.1, Figure 4.1). The payload bay includes the top plate area, but has been limited to 100mm above the top plate, level with that of the bell, ensuring a neat fairing design. Table 4.1: Front payload Specication Maximum Width Minimum Width Maximum Length Minimum Length Top Plate Area Hull Volume Total Volume Load Capacity bay specication Quantity 1100 mm 600 mm 530 mm 230 mm 0.51 m2 0.047 m3 0.098 m3 20 kg

The rear payload bay extends over the entire back third over the hovercraft, restricted by the placement and movement of the propulsion modules (Figure 4.1). The payload area includes both the internal and top plate area, with a height restriction of 70 mm above the top plate to ensure adequate clearance for the propulsion modules.

Sensor
From the specied hover height of 150 mm and average landmine depth of 100 mm, an overall minimum sensor detection depth of 250 mm is required. Further a specied detection width of 80% of the 43

CHAPTER 4. DETAILED DESIGN

Figure 4.1: Dimensioned Front and Rear Payload Areas

hovercrafts overall width, or 808 mm, was chosen. An 80% detection width allows for considerable overlap in each search pass, whilst maintining eective and ecent search algoritms. A sensor operating frequency of 24.4 Hz was determined from, a maximum operational speed of 5 km/h and a scan resolution of 57 mm, proportional to 5 scans over the width of the simulation landmine (Equation 4.3).

scan width = 20% 285 = 57.00 [mm] operational speed = 5 [km/h] = 1388.89 [mm/s] frequency = operational speed/scan width = 1388.89/57 = 24.37 [Hz]

(4.1) (4.2) (4.3)

A target detection level of 50% accuracy has been set for operational landmine detection. This value is extremely low when compared to the UN target accuracy level of 99.6%, however it is adequate to determine a proof of concept operation of the overall system. The selected sensors must be able to operate continuously for at lest 15 minutes, the hovercrafts maximum operational time.

4.1. PAYLOAD AREA

Figure 4.2: Required Detection Width Based on Hovercraft Dimensions

4.1.1

Sensor Selection

A metal detector was the chosen sensor type for the hovercraft. It was selected based on its low interference levels and variable mounting options; further it is beneted from low power requirements, light weight and fast data processing. There are a number of metal detector types, analog, digital and metal detector arrays; operating using a variety of dierent methods, pulse induction, beat frequency oscillation and Very Low Frequency (VLF) (see Section O). Beat frequency oscillation is the oldest form of metal detector and can be very erratic due to sensitive interference levels. Such a high level of interference is not suitable for landmine detection, requiring high accuracy and a low percentage of false positives. A metal detector array incorporates a number of pulse induction coils in an overlapping conguration. The size of a metal detector array can vary from 500mm to 4000mm, based on the number of overlapping coils. The technology used in a metal detector array to prevent interference from similar frequency coils and remove the need for periodic sweeping, is underdeveloped and very expensive, to the value of $500,000. However the technology is extremely accurate and ideal for landmine detection, able to detect metallic components while stationary, which conventional metal detectors cannot. The array presented for use on the hovercraft is a prototype unit from DSTO and Minelab (Figure 4.3), is approximately 3520mm long and weighs over 50 kg. Such a large unit far exceeds the hovercrafts payload specications and is not suitable for integration. Should a scale prototype become available

CHAPTER 4. DETAILED DESIGN

it would be ideal, and should be consider for future integration into the hovercraft.

Figure 4.3: DSTO Metal Detector Array Both analog and digital metal detectors operate on similar principles. The major dierence being digital metal detectors contain a large amount of inbuilt circuitry, used for signal analysis, whilst analog metal detectors only produce the raw detection signal. Digital metal detectors are often capable of indicating such features as object depth, size and position relative to the coil. Although highly benecial to the user, the complex circuitry makes it extremely dicult for integration into the hovercraft, built as a self contained unit with no simple output for a signal interface. An analog metal detector provides a high degree of customization and a simplistic electrical interface, allowing for complete integration into the hovercraft. The interface allows for direct coupling to the analog detection signal, produced as a direct correlation to magnetic induction. Further the unit may be easily altered to t within the limited payload mounting area. The nal metal detector chosen was a Bounty Hunter Tracker IV (Figure 4.4). This particular model incorporates a VLF coil, with a direct coupled analog signal and audio tone. The standard 8 inch coil and processing unit operates at 6.7 kHz, well above the required operating frequency. The complete metal detector specications are shown below;

Figure 4.4: Bounty Hunter Tracker IV

4.1. PAYLOAD AREA Specication Length Weight Power Modes Search Coil Audio Feedback Operating Frequency Quantity 1270mm 1.9kg 2 x 9V Motion all-metal Progressive Discrimination 8 Inch Open Waterproof Coil 10 Inch Open Waterproof Coil Two-Tone 6.7 kHz

Table 4.2: Bounty Hunter Tracker IV specications Table 4.3: Metal Detector Specications Single Coil Detection Parameters Array Detection Parameters Width(mm) Depth(mm) Depth Limit Range(mm) Dead Zone(mm) 305 305 82.02% 305 503 305 305 82.02% 610 198 305 305 82.02% 914 -106 381 381 65.62% 381 427 381 381 65.62% 762 46 381 381 65.62% 1143 -335

Dia.(in.) 8 8 8 10 10 10

No. Coils 1 2 3 1 2 3

4.1.2

Sensor Conguration

The main inuencing factors on sensor conguration are the detection range and depth penetration, directly inuencing the overall performance and eectiveness of the landmine detection sensors. There is no set method to determine the exact detection range of a coil, with specications most often determined from experimentation; however the detection range may be approximated from proportionality statements. Basic VLF coil proportionality rules state a detection depth and width proportional to 1.5 times the coil diameter. Based on the basic proportionality rules, the following table details a number of sensor congurations and their relative performance, in detection depth and width (Figure 4.3). The nal sensor conguration must achieve a detection depth of over 250mm and cover width of at least 808 mm, to meet the specied performance parameters. For an 8 inch coil the required detection depth is in the upper 20% of the detection range. Although adequate to meet the desired 250 mm range, the outer limits of the detection range are less accurate and a produce a larger number of false positives. To achieve the desired detection range, of 808 mm, using 8 inch coils, it would require an array of 3 coils, which is undesirable due to resticted payload mounting area and complex signal processing. If two coils were to be used a dead zone of 198m would be present, which equates to approximately 80% of the sinulation landmine width. Such a large dead

CHAPTER 4. DETAILED DESIGN

zone is detremental to accuracy and eciency, greatly decreasing the chances of detection. A 10 inch coil is capable of producing a more substantial detection depth than the 8 inch coil, within 65% of its overall range, providing a far high degree of accuracy. However their detection range is still not adequate to achieve the desired 808 mm detection width with 2 coils, having a dead zone of 46 mm, or 16% of sinulation landmine width. However this dead zone is tiny in comparison to the 8 inch dead zone and is small enough to safely assume that only a small percentage of landmines will pass through undetected.

Physical Conguration
From the performance characteristics a dual 10 inch coil conguration was selected. The sensors are to be mounted within the hull, achieving a maximum depth penetration of 230 mm below the surface, well over the 100 mm average landmine depth. A dead zone separation gap equal to 25% the simulation landmine width, or 70 mm, was selected; providing a small dead zone in which few to no landmines may pass through undetected and providing an overall detection range of 832 mm (Figure 4.5).

Figure 4.5: Dual Metal Detector Coil Arrangement and Theoretical Detection Range

3 8 1 mm

7 0 mm 8 3 2 mm

Metallic Isolation
Metal detectors are able to detect changing magnetic elds, produced from the motion of metallic components relative to the coil (see Appendix O). The strength of the signal is proportional to the size of the object and distance from the coil, with large, close object producing a stronger signal

4.1. PAYLOAD AREA

than small, distant objects. As such any metallic objects within the detection range may cause an interference signal, with movement caused from vibrations relative to the detection coil.

To reduce interference with the metal detector the payload bay was designed as an metalically isolated area. The isolation area was limited to the payload bay as the sensors range do not extend beyond its limits. To remove any further metallic components the front payload bay has been designed with the restriction to only mount the landmine detection sensors themselves.

The payload bay was constructed of composite materials and wood, joined with resin and micro balloons, ensuring no metallic components were used.

Although designed to be completely metallically isolated, various metallic components were required around the edge of the payload bay. Firstly the aluminum skirt mount, positioned around the edge of the top plate and secondly small steel bolts used to join the removable top plate to the hull. Although undesirable these components have a limited eect on the metal detectors, being on the edge of the metal detectors range and the metal detector being able to discriminate/ignore various metal types, such as aluminum and small metallic objects.

In order to mount the coils within the hull a composite mount was designed, ensuring no metallic components were used.

Initial composite mount designs considered the implementation of a plastic arm, utilizing the existing pin mounting feature on the coil. However a number of issues were raised in both the design and implementation of such a bracket. Firstly the intricate plastics manufacturing, possibly requiring a rapid prototyping machine and/or extensive CNC, both of which are expensive and time consuming. Secondly the implementation of the bracket presents potential complications, with a single mounting point and no gorund contact, inducing a number of vibrations into the coil, resulting in a large amount of interference.

The chosen mount design incorporates a plywood, CNC, locating center piece and two outer curves to restrict is planar movement along with a top brace restricting its vertical movement (Figure 4.6). The CNC pieces are micro ballooned to the hull as a t on assembly piece. Followed by the top plywood braced being doweled in place. Dowels are used to ensure the brace and sensor are removable for interchangeable sensors and maintenance. This mount achieves multiple large surface contact points in all planes, greatly restricting induced vibrations in the sensor.

CHAPTER 4. DETAILED DESIGN

Figure 4.6: Composite Metal Detector Mount Design

Coil Interference
The dual coil sensor conguration raises signal interference issues between the two coils. Two metal detector coils operating at similar frequencies are known to cancel each other out when within close proximity of each other. The cancellation phenomenon exists up to 20 meters between coils, making it a major issue. Coil interference may be overcome through a number of methods, shielding, frequency shifting and operational switching. Shielding, done through the implementation of a Faraday cage, is not possible for integration into the payload bay, due to considerable size and weight restrictions. Frequency switching is a fast and eective solution for overcoming coil interference. Whereby one coil is able to shift operating frequencies away from the other coil, completely removing all interference. However this feature is only available on high end digital metal detectors and not on our selected metal detector. Operational switching has been implemented as the chosen system for coil interference removal. Operation switching involves switching between the two coils periodically, whereby only one coil is on at any given time, completely removing all interference. Operational switching is highly eective, able to completely remove all interference eects; however there is an inherent undersired switching impulse. The switching impulse is caused from electrical interference as the coil is charged, whereby it detects

4.1. PAYLOAD AREA

its own change in magnetic ux from the induced coil charge. The interference produced is in the form of a short impulse, with a magnitude of approximatley 4 V, 50% of the overall resolution and a total period of 250 ms (Figure 4.7). This impulse is highly undesirable and greatly restricts the performance of the overall system. The greatest eect is the resultant dead period, caused from the metal detector powering up, inwhich nothin can be detected. The dead period lasts for a period of 50 ms, following each switching impulse (Figure 4.8).

Figure 4.7: Switching Impulse Interference Levels

Figure 4.8: Metal Detector Switching Dead Period The coils are switched at a frequency of 2 Hz, through an external relay circuit. The switching frequency is determined from the overall impulse period, allowing a 500 ms detection period, double the total period. The switching frequency is quite low, well below the desired 24.4 Hz specied previously. This is highly undesirable, but unavoidable to remove interference eects. A faster switching frequency of up to 10 Hz may be used, however faster switching speeds do not allow the coil to fully disharge between switching and inturn create higher and longer switching impulses, further reducing accuracy and complicating signal processing.

CHAPTER 4. DETAILED DESIGN

The set switching frequency of 2 Hz, correlates to a scanning distance of 695 mm at operational speed (Figure 4.9). This scanning gap produces a 70 mm eld overlap between subsequent on periods of the same coil. This ensures no area passes under the hovercraft without scanning.

6 9 5mm

7 0mm

Figure 4.9: Scanning Pattern With Periodic Switching

The external relay circuit (Figure 4.10) incorporates a relay, in a dual output, single input, isolated circuit. The relay circuit is powered from a isolated 12 volt battery. The 2 Hz signal is produced from the Microcontroller (MCU) on a interrupt timer (see Section 4.6.2). The on/o switch of one coil is then connected to the normally open gate of the relay and the other to the normally closed gate.

Figure 4.10: Relay Circuit Design

4.1. PAYLOAD AREA

4.1.3

Signal Analysis

Detection signals are produced from each metal detector processing unit as individual signals, proportional to the induced magnetic ux and are sent to an A2D port of the MCU via a voltage reduction circuit (see Section 4.6.2). Positive detection signals are then ltered from interference and switching signals, using a threshold value (Figure 4.11). Input signals are limited with a maximum 10 bit resolution, over a 5V range. Switching signals have a peak input of 4 V, low level interference below 3 V and a positive detection signal of approximately 8.2 V, providing signicant resolution for a positive detection above the threshold value of 7 V.

8 . 2V o l t s

4V o l t s

0V o l t s

Figure 4.11: Implemeted Threshold Levels for Landmine Detection The threshold is implemented as timer interrupt, whereby the inputs are checked on a periodic basis, every 5 milliseconds. Any value exceeding the threshold limit triggers an internal timer for which the sensors remain inactive, preventing multiple detection instances of the same mine. Following timer expiration the marking system is activated and the sensors put back on line.

4.1.4

Physical Marking System

The physical marking system incorporates an externally actuated spray paint canister, mounted through the hull. A hull mount is required to achieve a low release point within in the pressure cushion, minimizing ow disturbance whilst the paint is traveling towards the ground. A low release point reduces the overall air gap in which disturbance may occur and a release point within the pressure cushion results in minimal air movement compared to that directly outside the skirt. Flow disturbances are of particular importance, required to be kept at a minimum to achieve a large, highly

CHAPTER 4. DETAILED DESIGN

visible mark; achieved from a large portion of the paint reaching the ground. The physical mark is to take the form of a short line and should be within a 1 meter halo radius of the landmine, as specied in the project denition. The system is designed as a self contained unit (Figure 4.12). An outer sleeve is manufactured from 90mm PVC pipe with a custom release end cap. The end cap is manufactured from a standard 90 mm PVC pipe cap, with a self sealing release point, designed to prevent back ow from the pressure chamber through the marking system. A subsequent foam locating seal is used at the top of the system, to further prevent back ow and ensure the system remains central and stable.

Figure 4.12: Marking System Design

The system requires a 18N actuation force to operate and incorporates a self returning spring feature; allowing the system to quickly return to its initial position, for fast consecutive operations. The activation force of 18N is provided through a 34mm lever arm attached to a JR ES579 ballrace metal geared servo motor (Figure 4.13). The servo motor was selected based on availability and required calculated torque levels. The maximum theoretical torque of the servo is 8.3 kg.cm force over a 90 range of motion, providing a maximum force of 24N at the point of contact, above the required force (Figure 4.13). The nal conguration actuates the marking system through a 25 degree motion, for 1500ms; resulting in an approximte 200mm line. The system is powered through an external 12V power source and controlled via the MCU (see Section 4.6.2).

4.2. LIFT

Figure 4.13: Marking System Actuation Forces

4.2
4.2.1

Lift
Design Specications

The lift system was designed to supply air ow to the air cushion in order to lift the weight of the hovercraft, including the desired payload, giving a total lifting weight capacity of 100 kilograms to allow for battery weight and additional mounted devices. The system itself was required to meet a set of requirements, as detailed below.
It was to be of light weight, as this was one of the general requirements for all components. The system should be small given the spacial restraints present on the platform. The design should be simple, as a complex design would lead to further complications with the

potential to increase the overall size of the system.

4.2.2

Design Calculations

The required cushion pressure and ow rate were calculated using the equations shown below, which include losses accrued in the system. Pc = (Brooks, 2005) Q= 2Pc / h C Dc () (4.5) (4.6) M g Ac (4.4)

Dc () = 0.5 + (0.4 103 ) + (0.109 104 )2 (0.494 107 )4 + (0.345 109 )6 .

56 (Fitzpatrick, 2009)

CHAPTER 4. DETAILED DESIGN

This resulted in a required cushion pressure of 486 Pa, with a required ow rate in order to maintain this cushion pressure of 1.6 m3 /s. A full breakdown of these calculations can be found in Appendix F.

4.2.3

Fan Design and Selection

In order to select a suitable fan for this application, an analysis of the various fan congurations was carried out. Given the requirements of the lift fan as calculated previously to be 486 pascals of static pressure at a ow rate of 1.6 m3 /s, a suitable type of fan was required to be selected. Table 4.4 below shows the characteristics of a variety of fan types. Type of fan Propeller Tube Axial VAF, one stage VAF, two stage Mixed Flow Wide FC AF, BC, BI Narrow FC Radial Tip Radial Blades Turbo Blower Multistage Turbo Blower Air Volume (cfm) Large Large Large Large Medium Large Medium Medium Medium Low Small Small Static Pressure Low Low Medium Medium Medium High Medium High Medium Medium High High Eciency Low Medium High High Low Medium High Medium Medium Low Medium Low Manufacturing Cost Low Low Medium Medium High Medium Medium Medium Medium Medium Medium High Size and weight Small Small Small Medium Medium Large Medium Medium Medium Medium Medium Large

Table 4.4: Fan Types and Characteristics (Bleier, 1997) Inspection of this table shows that centrifugal fans, or radial fans, are best when a low ow rate and high static pressure is required. Given that a high ow rate and comparatively low static pressure was required in order to achieve lift, an axial ow fan was selected as the most suitable fan type for this application. A detailed analysis of the variables present in axial fan design was carried out to determine which characteristics best suited this application. Their are two main parameters in fan design which inuence the characteristics of a fan; the blade pitch, and the number of blades. In general, the rotational speed, diameter, ow rate and static pressure are all specied prior to selecting a fan. The blade pitch of a fan can be calculated using Equations 4.7, 4.8, where ACFM is the actual ow in cubic feet per minute.

Pbladepitch =

12 Va S RP M

(4.7)

4.2. LIFT Va = (Moore-Fans, 2009) 1.27 ACFM S 2 DR 2 ( 12 )

57 (4.8)

Equations 4.7 shows that for a xed rotational speed, an increase in the blade angle results in a higher velocity of air ow. The velocity of the airow is directly proportional to the ow rate of the fan, thus the higher the pitch of the blades, the higher the ow rate. The following equations relate to the number of blades present on a fan.

Nt =

M Pt Pb

(4.9)

P b = 3.8 108 r S RP M 2 (Moore-Fans, 2009)

(4.10)

The theoretical number of blades is directly proportional to the total pressure, and inversely proportional to the pressure per blade. The pressure on each blade is dependent on the aerodynamic hub of the fan and its rotational speed. This implies that at a constant rotational speed, the greater the number of blades present on a fan, the higher the total pressure generated by the fan will be. By consideration of the eect each design variable places on the performance of the fan, suitable fan characteristics were determined. As the ow rate required was quite high, a higher blade pitch was required, furthermore, given the relatively low static pressure requirement, a lower number of blades would be sucient. This analysis does however neglect the shape of the blade prole itself, which has a signicant impact on the performance of a fan. A blade designed with an aerofoil characteristic of a higher coecient of lift will produce a pressure, as shown in Equation 4.11.

CL = (Wallis, 1983)

L 0.5 a Va2 Ap

(4.11)

As a result, a fan with a high number of blades and low pitch could still fulll the air ow and static pressure requirements given a suitable aerofoil design. This eects the selection of a fan, as a more complex aerofoil blade would not be as cost eective as a less complex blade, and as such, with a more complex aerofoil, a lower number of blades may be required. Given the desired characteristics of the fan design, a fan which closely matched this conguration was

CHAPTER 4. DETAILED DESIGN

selected as an o the shelf product. This was to sure a decrease in the lead time to allow testing of the platform to begin as early as possible. As such, a 12 blade fan of 600 millimeter diameter, with a blade angle of 25 degrees was selected. The fan curve for this impeller is shown below in Figure 4.14. This fan is capable of producing a static pressure of 489 pascals and a volumetric ow rate of 1.61 m3 /s at a rotational speed of 2000 revolutions per minute. This fan curve also shows that the power requirement of the fan at this operating point is 2.2 horsepower. A detailed breakdown of the fans performance characteristics at this operating point is also detailed in Table 4.5.

Figure 4.14: Fan Curve (Multi-Wing, 2009) Table 4.5: Fan Data (Multi-Wing, 2009) Tip Speed 63 m/s Air Velocity 5.69 m/s Power 2.08 HP Torque 7.41 Nm Axial Force 144 N Air Flow 1.61 m3 /s Static Pressure 489 Pa Mass 2.27 kg

4.2.4

Engine Selection

A power of 2.2 horsepower at a rotational speed of 2000 revolutions per minute was required to achieve the desired fan characteristics. Three engines were considered to power the lift system. The rst was a CMG Engineering, HGA-100L electric motor. This motor was capable of outputting 3.5 horsepower at 2000 rpm, with a torque of 12.9 Nm (CMG-Engineering, 2009). A table of this engines data is shown in Table 4.6. This engine was capable of producing sucient power and torque for the lift system, however their were several drawbacks. The HGA-100L is large, with a length of 410mm, and also exceedingly heavy,

4.2. LIFT

Table 4.6: HGA-100L Data (CMG-Engineering, 2009) Input Current 5.6 A Rotational Speed 2000 rpm Power 3.5 HP Torque 12.9 Nm Height 245 mm Width 285 mm Length 410 mm Mass 37 kg with a mass of 37kg. Powering a motor of this size would also require quite large batteries, which would further add to this motors excessive weight. As all electric motors of this power will suer from the same weight limitations, it was determined that electric motors in general would not be suitable. The second option considered was a Yamaha KT100J petrol engine. This engine is commonly used in racing Go-Karts, and its specications are shown in Table 4.7. This engine has a maximum operational speed of 10,000RPM, at which it achieves a power of 9.4 HP. It is designed to run at higher speeds given that it is a racing class engine, and as such, running at a speed as low as 2000 rpm would be detrimental to this engines performance. The engine also does not have an onboard fuel tank, which would require a tank to be installed on the platform. This could prove dicult given the spacial limitations of the platform. The KT100J also requires an additional electric start system to be installed. Table 4.7: KT100J Data Power Height Width Length Mass (Yamaha-Australia, 2009) 9.4 HP 340 mm 330 mm 330 mm 9.1 kg

The nal engine considered was a Honda GSV190. This engine is a vertical shaft, 6.5 HP lawn mower engine. The characteristics of this engine are shown in Table 4.8, as well as a power curve in Figure 4.15. The GSV190 has several key advantages over the rst two options that were considered. This engine is heavier than the KT100J, however it includes a 1.1L fuel tank mounted as part of the engine, removing the need for an additional fuel tank to be installed. A vertical shaft conguration negates the need for a gearing system as the fan will be rotating about a vertical axis; the previous two engines were of a horizontal conguration. Futhermore, a recoil starter system allows the engine to be entirely self contained with an easy to use startup mechanism. This is signicantly simpler when compared to the KT100J.

CHAPTER 4. DETAILED DESIGN

Table 4.8: Engine Specications (Honda and power equipment, 2009) Model GSV190 Type 4 stroke Net Power 6.5HP @ 3600rpm Max Torque 13.2Nm @ 2500rpm Starting System Recoil Starter Cooling System Forced Air Fuel Consumption 1.3L/hr @ 3000rpm Fuel Tank Capacity 1.1L Dimensions 370.5x342.5x368mm Operating Weight 12.5kg

Figure 4.15: Engine Power Curve (Honda and power equipment, 2009)

4.2. LIFT

Inspection of Figure 4.15 shows that at 2000 rpm, this engine produces 3.25 HP and a torque of 11.2 Nm. This suuciently meets the power requirements, and is the most suitable conguration being an entirely self contained system. The Honda GSV190 engine was selected as the power source for the lift fan.

4.2.5

Inlet Bell Design

An inlet bell, as shown in Figures 4.16 and 4.17, was installed as part of the lift system in order to increase the eciency of the fan, and hence increase the ow rate, allowing for the engine to be run at a lower RPM, yet achieve the required lift.

Figure 4.16: Inlet Bell Example (ABC-Ventilation, 2009)

Figure 4.17: Diagram of Inlet Bell (Crane-Co., 1957) An inlet bell works by allowing the air owing into the fan duct to enter the duct more easily from greater angles from the duct itself. Figure 4.18 shows a plot of the loss coecient of a duct inlet, versus the ratio of the radius of curvature of an inlet to the duct diameter. A ratio of zero represents a completely vertical duct, without any curve on the inlet. As can be seen in Figure 4.18, as the r/d ratio tends towards greater than 0.15, the loss coecient begins to plateau, however before this value, the loss coecient decreases dramatically as the ratio increases. Given

CHAPTER 4. DETAILED DESIGN

Figure 4.18: Duct Diameter vs r/d (Crane-Co., 1957)

the spacial restrictions on the platform, and the neglibible decrease in losses beyond a ratio of 0.15, this ratio was selected for the inlet bell, equating to a total radius of curvate for the inlet bell of approximately 90 mm.

4.2.6

Mount Design

Several designs for the engine mount were considered and narrowed down to two possible solutions; a square shaped support plate fastened to the hull by hollow tube beams, or a round shaped support plate with three long T sections connected vertically to the support plate joining to the hull.The best solution was determined to be the round plate for two main reasons; the constant radius round plate would be easier to manufacture, and the thinner plates connecting to the hull would cause less disruption to the air ow. Further consideration lead to the addition of greater functionality to the engine mount, as it would be possible to utilise the engine mount to also support a the inlet bell for the fan. The mount was also designed in such a way that the engine and fan assemblies could be removed from the hull easily, without the need to remove the components of the mount which attach directly to the hull. Material selection for the mount was important in the overall design, as the structure had to be strong, but also had to be light. Initially, the beams joining the support plate to the hull, and the support plate itself was considered to be steel to support the loads, however it was too heavy. Aluminium was therefore considered. Basic deection calculations were carried out on the support beams of 3 mm thick aluminium, as these would be under the highest load. The maximum load was determined based on the mass of the engine with fuel, as when operational, the forces from the fan would act in the opposite direction. This load was determined to be 4.2 kg, acting at a single point equidistant from the two edges of the mounting plate. Equation 4.12 was used to calculate the deection of the beam.

4.2. LIFT

= (Kotousov, 2007)

W B 2 (2bB + 3aB ) 6EI

(4.12)

The maximum deection in these beams was 0.0351mm. Given negligible deection and its low weight characteristic, aluminium was selected for the mount. Details of the calculations carried out are shown in Appendix F.

Figure 4.19: Fan and Engine Mount

4.2.7

Shaft Extension Design

As a consequence of mounting the engine directly above the fan with sucient height to allow for the installation of an inlet bell, the maximum shaft length available for the chosen engine was not sucient to allow the fan to sit level with the hull top plate, and as such an extension to the shaft was required. This shaft extension was designed to be of the same material and dimensions as the current shaft, mild steel. This decision was made to ensure that any stresses present throughout the shaft would be equally transmitted through the shaft extension, eliminating the possibility of the extension failing through stress concentration. The extension itself was designed in two parts, the shaft and a large collar to connect the two shafts. This collar served the purpose of housing both the existing shaft and the extension, and in eect joining the two through keyways. A keyway was positioned at both the top and bottom ends of the collar to allow for keys to be positioned to lock both shafts into the same rotational position. The existing shaft had an internal thread to allow for attachment of various components. The extension of the shaft was hollowed out to allow a bolt to be inserted through the extension, and screwed into the existing shaft. A collar cap was placed between the bolt head and the shaft extension to ensure the assembly is securely fastened to the current shaft. This extension

64 assembly can be seen in Figures 4.20.

CHAPTER 4. DETAILED DESIGN

Figure 4.20: Engine Shaft Extension

4.2.8

Design Conclusion

The lift system design was able to achieve all requirements. The system was constructed entirely of aluminium in order to be light weight, a vertical shaft conguration was adopted allowing the engine to be positioned directly above the fan to reduce spacial impact on the platform, and the design was kept simple by incorporating o the shelf items. Figure 4.21 details the nal conguration and assembly of the lift system.

4.3

Propulsion

Propulsion is dened as the forces required to cause motion over a given terrain. The basic principle of hovercraft propulsion is momentum theory. Momentum theory in principle, describes motion of the platform based on a reactive force to the change in momentum of airow accelerated via a propulsion system. (Bliault and Yun, 2000). Using dimensionless coecients of thrust and power, a propeller and power house have been selected based on these principles.

4.3.1

Design Specications

Several design criteria were identied for the propulsion system. The main requirement was the need to maintain maneuverability of the craft by providing the ability to apply forces in the forward, back and sidewards directions. This ensures sucient maneuverability to control the hovercraft in accordance with operational scenarios identied for mine detection regimes. Additional requirements, beyond providing thrust, include obtaining the highest eciency system at the lowest weight and power costs.

4.3. PROPULSION

Figure 4.21: Assembly of Mount with Inlet Bell

CHAPTER 4. DETAILED DESIGN

As the spacial limitations of the platform are signicant this also becomes an important criteria for selection. The design specications are summarised below:
Propulsion systems must be independent of the lift system Provide thrust forces to overcome the inertial and drag loadings on the craft Accelerate at a rate of 0.14 m/s2 The platform must have the capacity to deccelerate to rest in the same time Maintain a constant velocity at operational of 5 m/s Provide thrust forcing in all directions outward of the platform to ensure sucient maneuver-

ability for autonomy

Be of high eciency, and t within the spatial bounds of the platform

As estimated in Section 3.3.2 the drag loading is made up of 4 primary components; aerodynamic drag, skirt drag, cushion momentum drag and terrain drag. Thrust forces can be estimated combining the required net thrust and drag forces, and the requirement was a minimum of 49 N.

4.3.2

Conguration Selection

Vectored Thrust Literature review of vectored propulsion systems identied two basic categories. Air jet propulsion systems are based on the ow of air from a duct or outlet resulting in a thrusting force. Primarily the energy source is removed from the outlet. Air-screw propulsions systems, describe the generation of thrust through the direct use of a propeller or ducted fan. Figure 4.22 shows air jet propulsion. The traditional advantages of air jet vectoring is that it is safe and quiet, as the fan is generally completely concealed. (Amyot, 1989) The disadvantes are that the system has low eciency, and thus to produce required thrust capacity requires signicantly bigger fans relative to other methods. (Amyot, 1989) Air-jet propulsion primarily includes fan jet vectoring of a concealed fan (centrifugal), where ows are ducted. It also include pu ports, by which the static pressure of the air cushion itself is converted to kinetic energy (Bliault and Yun, 2000). Fan jet vectoring is particularly inecient due to losses both at the inlet and outlet (Bliault and Yun, 2000). Figure 4.23 shows air-screw propulsion methods These methods are the most commonly used on conventional hovercrafts, and if spaced suciently are highly ecient (Amyot, 1989). Multiple fan

4.3. PROPULSION

Figure 4.22: Air Jet Vectoring congurations may cause signicant reductions in eciency as the inlet air source is shared (Bliault and Yun, 2000). Due to the exposure of these propellers, they are high in noise generation and also require large propellers to produce sucient thrust forces. These spacial issues may be problematic for small platforms.

Figure 4.23: Air-Screw Vectoring Conguration selection was undertaken by directly addressing the requirements outline in section 4.3.1. Due to the requirement to select the method of highest eciency, Air-jet propulsion methods were dismissed due to their low eciency characteristic. The capability for both dierential and vectored thrust was therefore selected. Although a relatively complex mechanism would need to be designed to control the thrust vector, this conguration was the only able to satisy all requirements of the propulsion system. Dierential thrust was desireable for low yaw angle disturbance correction as it is simple to implement and, relative to other methods, eective.

4.3.3

Propeller Selection

The propeller causes a change in momentum of the air by an eective screw action. This change in momentum results in a reaction force onto the craft and subsequent propulsion. There are three main

CHAPTER 4. DETAILED DESIGN

parameters which must be specied in propulsion selection. The eective screw area (dependant on diameter), the number of blades, and the pitch of each blade (Weick, 1930). The requirement of the propeller is that it must produce approxiamtely 30N thrust and must also be able to satisfy the spacial limitations of the platform. The design process associated with propeller selection is well published and has been adapted from McCormick (1979). The process includes calculation of preliminary requirements, estimation of thrust (CT ) and power (CP ) coecients, and calculation of propeller thrusts and power usage at dierent diameters and pitch. McCormick (1979) provides charts for estimating (CT ) and (CP ) however these are characterised for larger aircraft propellers. Alternative graphs relating to Remote Control (RC) aircraft propellers have been sourced to obtain more accurate estimates of these coecients. In order to place an initial estimate on the diameter of the propeller the spacial requirements of the platform were analysed. The rear of the craft was split into two circles whose area were dictated by the turning circle a propeller assembly of diameter Dmax . By splitting the craft width between these two circles a theoretical Dmax was calculated to be 17. A realistic limitation of 15 was placed on the diameter to account for propeller housing. In order to select a theoretical propeller a range of thrust values at varying diameters were calculated at a mid range Rotations Per Minute (RPM) of 10000 rpm of the electric motor. Figure 4.24 shows typical CT values for general propellers depending on the advance ratio. The advance ratio was calculated using equation 4.13 from McCormick (1979) and equals 0.02.

V np D p

(4.13)

For the purposes of determining CT , J can be assumed to be close to zero. Due to the low speeds of the hovercraft it was recognised a low pitch propeller required to minimise the power requirements of the system. A pd ratio of 0.4 was therefore selected for primary estimates. From Figure 4.24, the minimum pd ratio of 0.4, a CT value of 0.09 was determined. Equation 4.14 was then used to calculate the theoretical thrust of the propellers at dierent diameters (McCormick, 1979).

T = C T a np 2 D p 4

(4.14)

For a 12 propeller the pd ratio gives a 5 pitch and the calculated thrust was 28.7N. A 13 propeller was therefore selected giving a calculated theoretical thrust of 39.5N. It was recognised that testing would be required to determine the best propeller however this theory provided a starting point for

4.3. PROPULSION

Figure 4.24: CT estimatation curves (McCormick, 1979)

the propeller selection. At a ratio of 0.4 the propeller pitch is equal to 5.2. From this result we elected to test the available propeller sizes of 134, 136.5 and 1310 to maximise thrust. For the purpose of power calculations a 136.5 propeller was selected. This propeller is shown in Figure 4.25. In order to maximise ecieny a 2-blade propeller was selected.

Figure 4.25: Selected Propeller (ModelFlight, 2009)

To estimate the required power for this propeller a similar process was run with CP . Figure 4.26 was used to estimate CP relative to CT and J. From the graph a value of 0.019 was obtained. The power required could then be calculated using Equation 4.15 from McCormick (1979) and was calculated to be 460W.

CHAPTER 4. DETAILED DESIGN

P = C P a n p 3 D p 5

(4.15)

Figure 4.26: CP estimatation curves (McCormick, 1979) This power as calculated was a raw value and it was recognised that the eciency of the motor would increase the power requirements. The max eciency of the Typhoon HET was 0.85, however a value of 0.75 was assumed as the operational value of the motor. Additional to this, a safety factor of 1.8 was applied. This safety factor is justied in that these equations are basic guidelines based on testing of RC aircraft propellers and some safety in the calculations is desired (B, 2009).

4.3.4

Motor Selection

The propulsion motors are used to drive the propellers, providing thrust to accelerate the hovercraft. Two identical motors are required for a dual vector thrust design (see Section 3.2.5). The motors required a high degree of controllability, the ability to turn on/o quickly and operate over a variety of operational speeds. To achieve this an electric motor was required, as no other engine can achieve the high degree of controllability needed. Conceptual thrust design calculations have indicated a minimum 500 W power requirement, resulting in a design power of 1250 W, with a 2.5 safety factor. Such a high safety factor is required to ensure adequate thrust for a wide range of propeller diameters, pitches, blades and operational speeds. Due to the rotational properties of the vectored thrust design, a motor was required with a high power to weight ratio. Able to provide high power levels, whilst remaining light weight, so not to restrict the rotation of the propulsion system. Further the motor size must remain small enough not to restrict the ow into the propeller, in a direct drive conguration. Based on this it was determined

4.3. PROPULSION

a brushless electric motor was required over a brushed motor, able to provide more power and lower electromagnetic interference, which reduced landmine detection interference. Size restrictions further limited selection to an in-runner motor over an out runner, being signicantly smaller, allowing for higher ow and thrust eciencies. The propulsion system needed to be free to rotate therefore a low moment of inertia is desirable, restricting the weight of the mount. Subsequently limiting the size of the propeller to a small range of propellers, in the lower end of propeller sizes. This results in high required operational speeds, to achieve the desired thrust levels. Based on these limitations an in-runner, brushless electric, Typhoon 600-43 HET motor was selected (Figure 4.27).

Figure 4.27: Typhoon 600-43 HET This motor is the highest power motor in its range, providing 1400 W of power above the required 1250 W, whilst remaining very light weight with a total weight of 290 g. Capable of operating at speeds up to 20,000 rpm and at a max eciency of 85%, adequate to power propellers up to a diameter of 20 inches (as stated by manufactuer).

4.3.5

ESC and Battery Selection

The electronic speed controller is used as an electrical isolator between the motor and the MCU, whilst acting as a controller for the electric motor. An Electronic Speed Controller (ESC) is required in order to provide a high degree of controllability and protect the motor from incorrect signals and interference levels. Manufacturers recommendations suggest a 90 A ESC and based on this we selected an available Dual DualSky 90A brushless controller from model ight (Figure 4.28). Lower rated ESCs were available, however are undesirable due to incompatibility warnings with high powered motors operating near maximum performance limitations and known reliability issues.

CHAPTER 4. DETAILED DESIGN

Figure 4.28: Dual DualSky 90A brushless controller

The selected ESC can operate with up to a 10 cell Lithium Ion Polymer (LiPo), providing automatic brake settings, battery type recognition, low voltage cut-o protection, startup modes and variable signal timing. Due to high costs of a 10 cell battery, a 6 cell LiPo was used, in particular a Hacker HTFU30C-38006S with a charge capacity of 3800 mAh. This gives a minimum battery life of 5.19 minutes (calculated in Equations 4.16 and 4.17) when motor is run at calculated power of 975 W . It is expected that the endurance of the propulsion system will far exceed this however, as motors will not need to be run at full power for the entire time.

975 [W] = 43.92 [A] 22.2 [V]

(4.16)

3.8 [Ah] = 0.0865 [h] = 5.19 [min] 43.92 [A]

(4.17)

Figure 4.29: 3800 mAh 8 cell LiPo battery

4.3. PROPULSION

4.3.6

Mount Design

General Requirements There were several requirements determined for the design of the propulsion mount. These included the ability to limitlessly rotate the thrust vector 360 degrees, the ability to resist deection as a result of thrusting, and simple and easy manufacture and assembly. Furthermore, with safety was an imporant consideration due to the rotation of a fast moving propeller. It was to be ensured that the rotation of the two propulsion mounts could be achieved independently without interference. Based on the mission prole of the hovercraft some specic requirements were dened;
Rotational acceleration of the thrust vector was dened at 1 rad/s2 to achieve a suitable response

in autonomy.
Deection of the support plate must be within 2 degree of the thrust vector at rest. The full rotation must t within a circular area of 0.23 m2 based on spacial limitations at the

rear of the craft.

The mount must fully enclose the propeller to ensure operator safety. Minimising weight is an important limitation on hovercraft platforms.

Mount Conguration and Design The general conguration of the propulsion mount was selected in such a way that the outlined criteria for design were addressed. Two general congurations were considered for the support being a traditional tripod design and a bar support being directly mounted to the hull. The tripod design was chosen for two specic reasons. The rst was that it would considerably increase the rigidity of the structure providing the mount with sucient strength to resist loading. The second was that due to vibration transmission from the lift system, a straight beam conguration was deemed to posses poor vibrational characteristics, with high numbers of resonant modes (Young and R, 2001). It was therefore chosen that sucient complexity and rigidity should be added to increase the quality of vibrational resistance. This tripod type conguration was chosen to support an outer pipe which could be milled to house bushings to prevent material surface contact during rotation. In order to achieve the specied requirements, aluminium was chosen to maintain high structural stength while maintaining low weight characteristics. Aluminium was also advantageous in that it could be easily worked into a complex mount able to house bearings for rotation and support thrust

CHAPTER 4. DETAILED DESIGN

forces. This forms the basic support structure of the mount. An inner pipe with a top plate attached was selected to be inserted into the outer support and provide a rotational input to the system. This top plate was to house the components required to generate thrust. A fan protector guard was designed and attached to the end of the support plate to entirely encage the fan for safety purposes. Figure 4.30 shows the overall conguration of the mount.

Figure 4.30: General Propulsion Mount Conguration Material thickness were selected to satisfy the previously dened design criterion. For the at aluminium pieces standard sizes of 3, 5 and 7 mm were considered. Deection of the top plate itself was calculated using a simply supported beam with weight and moment loadings. The force loading at the tip of the plate is due to the weight of components, and the moment about the Centre of Gravity (COG) is due to thrust forces. A 5 mm plate was chosen, with a deection of <1 mm. ANSYS was performed on this plate to verify the hand calculations and is shown in Figure 4.31. As all the other components were non-load bearing, 3 mm thick aluminium was sucient. The top and bottom struts were added to increase vibrational rigidity of the fan guard. The inner pipe was selected at the available material size of 50 mm which was the largest size available to t within the bounds of the top plate. This diameter was also well suited to the standard sizes of static bushings.

Bushing and Thrust Washer Selection A bearing conguration was required to facilitate the turning of the propulsion system inside its housing shaft. The main requirement for bearing selection was simplicity in implementation and maintenance requirements. For this reason the use of roller or cylindrical bearings was discounted due to the need to machine complex housings with high tolerances. An alternative static bush (shown in Figure 4.32 was selected to achieve rotational functionality of the mount.

4.3. PROPULSION

Figure 4.31: ANSYS Analysis of Propulsion Deection

Figure 4.32: SKF Composite Bushing

The dimensions of the bearing were selected based on the limitations of the mount. Two types of static bearings, PTFE and POM bearings were compared. PTFE have good dry sliding characteristics and do not generally require lubrication. POM bearings will require lubrication on installation. PTFE bearings were selected for simplicity in installation and run in. Silicon lubricants were chosen to help facilitate turning should they be required after run in of the bearings.

The design process as stipulated in the SKF selection manual was undertaken to ensure the life of the bearings would satisfy the requirements of the platform. Due to the low loading levels and slow rotational speeds of the operational parts, the expected life of the bearings far exceeded any durability requirements placed onto them.

In order to match the bearings, a suitable PTFE thrust washer was selected to facilitate contact of the mounting plate with the rotational support shaft.

76 Stepper Motor Selection

CHAPTER 4. DETAILED DESIGN

The stepper motor provides the ability to rotate the propulsion mount for vectored thrust. Stepper motors are brushless, DC motors that work by having a set of four toothed electromagnets slightly out of phase with the rotors teeth (Acarnley, 2002). When the rst electromagnet is powered the teeth are slightly misaligned with the adjacent electromagnets teeth, and thus powering these electromagnets in a series result in stepping of the rotor (Acarnley, 2002). The general requirements of the selected motor were that it must hold the required torque to turn the mount system, be able to turn in both directions, must be adaptable to a suitable drive chain, and be less than 70mm in height. The Moment of Inertia (MOI) Ixp , of the propulsion mount was determined by inserting the material properties of the designed components into CATIA and calculating their individual moments of inertia using the parallel axis theorem. For components of unknown density distribution (such as the electric motor) a constant density was assumed based on their volumes and weights. Ixp was evaluated at 0.059 kgm2 . The torque required to then induce an angular acceleration of 1 rad/s2 claculated using the standard torque equation. This requirement is calculated to be 0.3732 Nm. In order to determine the required torque for the stepper motor a safety factor must be applied to account for the frictional loading that he been neglected. A safety factor of 60% was placed on the stepper motor torque giving a design torque of 0.622 Nm. Two types of stepper motors were considered, bi-polar and uni-polar stepper motors. The dierence in these two motor types is the way in which the windings are congured. A uni-polar stepper motor has two windings per pole, through which currents can be run independently to set direction. (Acarnley, 2002) A bi-polar stepper has one coil per pole by which the direction is parralel to current ow. These stepper motors are able to achieve a higher torque per stepper size as the the full coil is used. (Acarnley, 2002) Bi-polar motors are however, harder to control with H-bridge circuitry being required to reverse current ow direction. One stepper motor selection criterion was spacial requirements. This was to ensure sucient space for installation of a drive train to the rotating column of the propulsion assembly. The bi-polar stepper conguration was chosen to achieve the required torque in a smaller motor. The control issues associated with this motor conguration were addressed by the selection of a motor controller containing the H-bridge circuitry required. The selected stepper motor was an Ocean Controls MOT-124 FL57STH56-1006A Bipolar Stepper Motor A controller was matched to this motor and the Ocean Controls SMC-002 M325 Microstepping Bi-

4.3. PROPULSION

Figure 4.33: Ocean Controls Stepper Motor (Ocean-Controls, 2009) polar Controller was selected.

Figure 4.34: Ocean Controls Stepper Motor Controller (Ocean-Controls, 2009)

Gear Train Design A power transmission circuit was required to transfer the rotational force of the stepper motor into a rotational input to the vectored thrust. The general requirements of this transmission were that the position of the rotation must be controlled and known, meaning there must be minimal slip. Furthermore, low maintenance and adaptability of the system to the current stepper motor conguration was

78 required.

CHAPTER 4. DETAILED DESIGN

Two potential systems were considered for implementation. Timing belts and chain drives. The timing belt is the superior conguration with regard to accuracy in positioning as high tension in conjunction with the belt teeth ensures no slip (Evaluations, 1971). The disadvantages of the toothed belt are that this high tension will impart high moment arms onto the shaft of the stepper motor. Furthermore, the pulleys required are wider than the rotating shaft depth, making attachment complex. Although chain drives may experience a small amount of slip in operation they are able to be driven by much smaller cogs as the teeth are contained within the boundaries of the chain. This makes for simple attachment to the stepper. In addition to this, chain drives do not need to be driven at the tension characteristic of timing belts and the radial loads on the stepper are therefore signicantly reduced. Other advantages of the chain drive include high eciency and reliability (Evaluations, 1971).

For these reasons a chain belt drive system was selected. The specication controlling design was the ability to step from a 6mm shaft to a 50 mm shaft whilst maintaining the selected gear ratio. A gear ratio of 2:1 was chosen to increase the generated torque of the stepper motor, decreasing the work rate of the stepper to hold a position, and to allow for a signicantly larger gear to accommodate the step in shaft size.

The max torque generated by the stepper is 0.88 Nm meaning the torque at the propulsion shaft is equal to

= gen .T oothRatio = 1.76 [Nm] (Evaluations, 1971)

(4.18)

The size of each gear was selected based on suitability to be attached to the respective shafts. The driven pulley dictated selection as the hub was required to be of sucient size to house the 50 mm rotating shaft. A 48 tooth, catalogue pinion of diameter 70 mm was chosen to allow space to machine a 50 mm bore. The smaller gear was then selected to maintain a ratio of 2:1 with 24 tooth. An inside bush was designed to eectively step down the gear bore to 1/4 inch as required for fastening to the stepper motor drive shaft. A nominal centre distance of 100 mm was selected keeping spacial limitations in mind. This distance was validated using general design manuals.

4.3. PROPULSION

Figure 4.35: Propulsion System Parts List

CHAPTER 4. DETAILED DESIGN

4.4
4.4.1

Structural
Primary Design Criteria

Weight, energy absorbency, damageability, homogenous construction, corrosion and sound or vibration deadening were the factors considering when designing the platforms hull. Ease of manufacture, lead and lag times as well as cost were also factored into the preliminary design decisions. Weight being the most inuential factor in hovercraft performance (Amiruddin et al., 2008), and cost being the constraining factor in all decisions, the platform was designed with the primary aim of minimizing these variables. The proposed operational environment necessitated moderate levels of impact resistance as well as high levels of abrasion resistance of the hull. Hence energy absorbency and damageability were considered in the design and manufacture of the base of the pressure chamber. Corrosion resistance and intrinsic vibration isolative properties were also considered, to ensure longevity of platforms life and inherent sound deadening capability to reduce the environmental impact of the platforms hull. From the preliminary design section, the design of the hull was broken down into four distinct sections with the top and bottom plate contribution to approximately 67% of the hovercrafts overall volume.

Figure 4.36: Area Distribution

Consequently a majority of the design initiative was aimed at optimising the top and bottom plate structure.

4.4. STRUCTURAL

4.4.2

Design Specications

Top Plate (Pressure Chamber Ceiling) The top plates primary function is to house and support all sub systems as well as form the top seal of the pressure chamber, and hence requires a high level exural rigidity to avoid undue deection with the potential of damaging the lift fan which is mounted within the pressure chamber. Moreover a high compressive modulus is required to endure repetitive application of torque on the bolts which hold the primary engine mounts, safety tether points and sealing ring. Reliability; in particular the ability to replace and the ability to modify mounting points were assigned the greatest priority in the design considerations, given the high probability of modication of the position of subsystems being mounted on the top plate. A sandwiched structure was specically chosen to ensure a marginal penalty in weight for exponential gains in stiness and strength.

The Internal Ribs The primary function of the internal ribbing structure was to support the top plate and transfer load from the ceiling of the pressure chamber (the top plate) to the oor, bypassing the walls of the pressure chamber. This was done so as to reduce any load being carried through the walls of the pressure chamber given that structurally, they are intrinsically the weakest point in the structure given the location of bleed holes in the structure. The internal structure was designed with the aim of allowing for the most feasible amount of air ow in the internal structure without compromising the structural integrity or load bearing capacity of the internal structure High exural rigidity, high torsional rigidity, and resistance to buckling under operational loads were the key variables considered in the design of the internal structure

The Pressure Chamber Floor The pressure chamber oor formed the foundation of the entire structure. The primary aim of the component was to provide a solid base from which the structure could be erected. High torsional rigidity was required to ensure that the entire structure would not distort during operation which would lead to indenable handeling characteristics of the platform. The oor is designed to support the weight of the platform when grounded and eectively transfer the operational loads Low weight and high rigidity was the key variables considered in the design of this sub-component of the hull

CHAPTER 4. DETAILED DESIGN

4.4.3

Specic Loading Conditions

Initial design specications stipulated that the platform be designed with a nominal operational weight of no more than 100 kg with an operational speed of no more that 1.4 m/s. Although historically, hovercrafts are designed on the premise of operation in marine environments, the prototype platform was designed with the aim of terrain based operation only. Given these limitation, in particular the absence of aerodynamic loads and hydrostatic loads given low operational velocity and exclusion from a marine environment , the platforms structure was designed with the aim of accommodating static loads exclusivly Inertial loads and quasi static loads were also neglected given the restriction on operational environment and velocity. A summary of the loading conditions and constraints can be seen in the gure below. A nominal deection limit was set on all sub components of the hull as deection above this limit would increase the risk of interference with the main lift fan and the internal structure of the hull. A thickness limit was also employed based once again on special limitations and the increased risk of interference with sub components if the limit was breached. A weight limit was dened based on the preliminary statistical analysis and the area distribution which was derived in the conceptual design of the platforms geometry. This consequently dened the weight fraction assigned to each sub component. The platforms maximum laden weight was estimated at 97 kg and tare weight was estimated at 77 kg. To ensure structural integrity under all conditions, the platform is estimated to be loaded to 100 kg at all times. The platforms were designed to endure pressure loads exerted by the plenum chamber.

4.4.4

The Sandwich Principle

A Glass Reinforced Plastics (GRP) composite sandwiched panel construction was employed throughout the structure to accommodate the predened key design criteria and two specic variables; low weight and high exural and torsional rigidity. The skin in such structures is predominantly in compression and tension and the core separating the top and bottom skin in shear, as illustrated by Fig 4.37 Shear within the skin was ignored given its marginal eects on the overall structural integrity. From general engineering theory, it can be established that by employing a sandwiched construc-

4.4. STRUCTURAL

Figure 4.37: Specic Loading Conditions

Figure 4.38: Weight, Strength and Stifness in Sandwitch Panel

CHAPTER 4. DETAILED DESIGN

tion, exponential gains in strength and stiness can be gained with a marginal penalty in weight as illustrated by the gure above. Appendix H details the specic properties of a sandwiched pannel and its applicability to the platforms structure.

Figure 4.39: Force Distribution in a Sandwich Panel

4.4.5

Skin Selection

The choice of the reinforcing bre was given a high precedence in the overall design process given its direct impact on the overall structural properties of the platform , in particular the platforms strength, rigidity and resistance to impact. The selection of the reinforcing bre was based on seven criteria and each was weighted on a linear scale in relation to their overall impact in achieving project the goals. Cost and density of the bre were given the highest precedence respectively considering that project cost and the high sensitivity of platforms performance to weight were considered limiting variables. Tensile and compressive strength of the bre were given the next tier of precidence considering their inuence on platform strength. In-plane shear strength and exural strength were given the lowest precedence considering that the structure is being designed for minimal deection and in the case where the structure deects to the extent that its exural strength becomes a factor in maintaining structural integrity, the structure will be nominally considered to have failed. Three bres, Carbon, E glass and Aramid (Kevlar) were chosen from a selection of bres as outlined in Appendix H. The three bres can be compared against each other based on the project specic weighting and properties outlined by the SP Composite guide. The table below, details the bres

4.4. STRUCTURAL property versus the weighting and their cumulative sums. Property Low Cost Low Density High Tensile Strength High Compressive Strength High Impact Strength High In-Plane Shear Strength High Flexure Strength Total Weighting 3 2 1 Property High Suitability Moderate Suitablity Low Suitability Weighting 7 6 5 4 3 2 1 Aramid 1 3 2 1 3 2 1 63 Carbon 1 2 3 3 1 3 3 68 Glass 3 1 2 2 2 3 2 69

As illustrated by the gure above there was less than 10% dierence in the nominal weighting of the bres resulting in the decision to amalgamate the bres selection and make use of their composite properties.

Tensile Strength Vs Density A comparison between their respective densities and tensile strengths generated the following results.

Figure 4.40: Density Vs Tensile Strength When compared against carbon bre, Kevlar demonstrated a loss of approximately 70% in tensile

CHAPTER 4. DETAILED DESIGN

Figure 4.41: Density Vs Tensile Strength strength at a saving of 24% in weight. However when compared against E glass bre, Kevlar demonstrated an increase in tensile strength of 29% at a saving of 42% in weight. It is important to note however that the weight estimation gures are based on the average weight of the most commonly available bres. Larger selections of bres are commercially available, but proved to be uneconomical to source given the scale of manufacture. The data implied that a Carbon - Aramid bre combination would form the ideal solution in regards to weight and tensile strength, however a combination of Aramid and E-glass bres would provide a marginal loss in tensile strength but at a signicant saving in weight, making their combination a viable option.

Laminate Impact Strength

Figure 4.42: Laminate Impact Strength Laminate impact strength was compared using data published by SP systems based on DIN53453 standards. The data is based on the standard Charpy test. Ceteris paribus, when compared against Kevlar, E glass bres absorbed 12% less energy per square inch where as Carbon bre absorbed 46%less energy per square inch, implying that any combination other than that of E glass and Aramid bres would be suboptimal.

4.4. STRUCTURAL Matrix Interaction and Polymer Selection

Given the complementary role of the matrix in the overall underlying properties of the composite skin, an appropriate polymer was selected to compound and enhance the desired mechanical properties of the selected bres. It important to note however, that unlike the selection of the bre being used in the skin of the platform, selection of the polymer matrix was limited by compatibility with the selected bre. A detailed comparison of the polymers considered can be found in Appendix H. The polymer used in the platform, was selected on the basis that it would complement the bre and core selection. The key criterion used in the selection of the polymer was that the resin must be able to deform to at least the same extent as the bre. Modication of the polymers chemical structure was beyond the scope of the project, and hence the resin selection was based on commercially available grades. Adhesion, toughness and environmental resistance were also considered qualitatively in the selection of the resin system. The three main types of resin considered were Polyester, Vinylester and Epoxy resins. Phenolics , Bismaleimides and Polyurethanes were not considered due to specialised properties which were not needed for the platform; namely re resistance and high temperature resistance and their incompatibility with the available manufacturing facilities When comparing tensile strength and stiness Epoxies resins outperformed both Vinlyester and Polyester resins in both slow and fast cure times.

Figure 4.43: Tensile Strength and Tensile Modulus The low shrinkage properties of both epoxies and vinyl esters implied lower residual stresses if employed on the platform and less susceptibility of fatigue damage on the platforms walls and supporting plate Epoxy resins showed the best stress stain characteristics as well as the lowest moisture absorption over time, making it the ideal choice. Three types of epoxy resins were considered, the R180, R330 resin system from FGI and the 105 resin

CHAPTER 4. DETAILED DESIGN

from West Systems. The 105 West System resin with a 205 fast hardener was selected, given the relatively quicker cure times, lower viscosity and the ability to cure at lower than room temperature without the need for a pressurised environment. The actual amount of bre used in the composite, would be a factor that is ultimately determined by the manufacturing process. Reinforcing fabrics with closely packed bres give a higher Fibre Volume Fraction (FVF) and help in the distribution of the bre matrix interfacial loads. As a general rule the stiness and strength of a laminate will increase in proportion to the amount of bre present.

F V F = 1 + 1
Wf Wr r f

1 (4.19)

Where: Wf = Weight of Fiber Wr =Weight of Resin r =Density of Resin f =Density of Fiber However above 60% FVF, the tensile strength may continue to increase but the laminated strength will reach a peak and begin to decrease due to a lack of sucient resin to hold the bres properly. To account for this in the design and selection process, the overall exural rigidity was calculated per ply based on the 50-55% FVF as recommended by the American Society for Testing Materials 1990.

D = EI =

E f bt3 E f bt2 E c bc3 + + 6 2 12

(4.20)

From the data collated on a per ply basis, we can deduce that per ply Carbon bre exhibits a greater exural rigidity and increases logarithmically as the number of plies increase. Aramide and E-glass bres exhibit similar rigidities with Aramide bres being marginally stier.They also exhibit an exponential increase in rigidity as the number of plies increase, attesting to multi-ply comparability. From an amalgamation perspective, although it may be advantageous to have bres embedded in a matrix that provide cumulatively greater levels of rigidity, having bres with rigidity varying by a factor of 10 bring about a complexity when predicting the most probable mode of failure in the life cycle of the platform.

4.4. STRUCTURAL

Figure 4.44: Flexural Ridigity

CP T (mm) =

Wf f F V F 1000

(4.21)

FWF =

f F V F [m + ((f m ) F V F )]

(4.22)

Figure 4.45: Weight Per Ply The weight per ply can be calculated from the Fibre Weight Fraction (FWF) from the FVF and Cure Ply Thickness (CPT). From the data collected we can conrm the linear trend in weight vs ply. We can also deduce that when combined with a polymer matrix, carbon bre per ply will only be marginally lighter than E glass bre. This provided further evidence that, in the case of the platform a Glass-Aramid bre amalgamation would provide a favourable solution for skin selection.

CHAPTER 4. DETAILED DESIGN

4.4.6

Core Selcetion

To achieve desirable levels of exural rigidity to ensure complience with the predened parameters of deection are met without the addition of a core would require a disproportionate skin thickness. Moreover the weight penalty would outweigh any feasible advantage of increasing the skin thickness as illustrated in the graph below.

Figure 4.46: Comparitve Weight for Bending A literature review detailed in Appendix H, limited the cores being considered to Polyvinyl-Chloride (cross linked) and Ply-wood core. Two distinct cores with very dierent properties had to be selected given the specic loading conditions dierent sections of the hovercraft would be subjected to in operation.

Ply Wood Vs Structrual Foam A Klegcell core, a derivative of the Polimex range of structural foams was considered over a range of other structural foams given its availability in a range of densities and relatively low cost as well as being compliant with the criteria identied in Appendix H. The Klegcell range of foams came in densities from 45 kg/m3 to 100 kg/m where as the available plywood core densities ranged from 650 kg/m3 for light weight ply wood to 950 kg/m3 for heavy weight plywood. A comparison of shear modulus and compressive strength to density demonstrated that a signicant increase in density gave proportional gain in shear and compressive strength when considering foam cores but not when considering ply wood cores as illustrated in the gures below.

4.4. STRUCTURAL

Figure 4.47: Density and Compressive Modulud

Figure 4.48: Density and Shear Modulud

CHAPTER 4. DETAILED DESIGN

The lowest grade of available ply wood had a mean density of 650 kg/m3 a compressive strength of 20.7 MPa , a shear strength of 4.1 MPa and a shear modulus of 470 MPa. Its properties well surpassed even the highest grade of available foam. As identied earlier the primary function of the core in a sandwich structure is to separate the skins, enhances the stiness of the structure, whilst being able to withstand the shear forces which are as are a result of placing the structure through cycles of tension and compression. Hence in areas where a high compressive modulus was needed, a ply wood core was employed, and in all other cases, a foam core was used, to optimise the structural properties of the platform

4.4.7

Design Calculations

Assumptions The following section is based on the conclusion drawn in the preliminary design section. 1. The primary skin to be used on the platform is E-glass, 330 g/m( 3) 2. The secondary skin to be used in the platform is 190grms/m3 Kevlar where high impact and abrasion is anticipated 3. The two cores being considered in the platforms structure were (a) A Kelgcell R80 PVC structural foam core (b) Structural plywood compliant with AS/NZS 2269, bonded with Type A bond phenol formaldehyde and of stress grade F11 and F14 designated Light Weight. To estimate deections in the structure and resulting stresses the following simplications were made:
The entire weight of the structure would be isolated on the bottom plate (chamber oor) in

ight and at rest. Although in reality loads may be transferred throughout the structure during dierent operational manoeuvres, a majority of the loads would always be concentrated on the chamber oor as a result of the internal ribbing structure that was convinced during the conceptual design section. The top plate in particular would be subjected to loading, as it is the primary mounting surface of all the sub systems on board the platform. However, as indicated by the preliminary desingn criteria, the transfer of loads would be through the internal ribbing strucutre.

4.4. STRUCTURAL

If geometrical distortion of the chamber oor is isolated, we can assume that the overall geometrical deformation will be low, for as long as the internal ribbing structure is intact.
The chamber oor was simplied to a simply-supported at plate of dimensions 1.6 m by 1.1

meters. This simplication would allow for an overestimation in deection and an underestimation of the stresses in the structure. Acknowledging that this may have consequences regarding the longevity of the platforms operational life as residual stresses may lead to accelerated deterioration of the skin, the primary concern in the design criteria is to minimise deection and eliminate the chance of contact between fan and chamber ceiling.
The internal ribbing is assumed to be a simple beam and not a truss structure.

This assumption overestimates the structural properties of the internal ribbing. To compensate for this the load on the internal rib is overestimated. This assumption was made to reduce the computational expense of calculating the deection in the ribbing structure. Nevertheless it is acknowledged that, if the structure is to be optimised, the loads on the internal ribbing structure would have to be revaluated.

Core Designation A foam core was initially considered as an exclusive core for the entire structure, given its high strength and stiness and low cost however the need to incorporate a panel which could be easily bolted into implied the need for the addition of hard points throughout the top and bottom plate. In anticipation of the need to modify the position of components on the top plate, and the complexity of the addition of hard points once the top plate was sealed with a skin, the decision was made to use a ply wood core for the top given the high compressive modulus and low cost. The bottom plate on the other hand, was envisioned as having little to no modication, and hence a structural foam core was designated. The moderately high shear modulus of the core would be better suited to bear the internal stresses of the core in bending and tension without an excessive weight penalty. The top and bottom lips, were, as described in the section concerining specic loading of the platform, expected to take the loads of the pressure sealing bolts and skirt mounting eyelets, required a core with a high compressive modulus just as the top plate.

CHAPTER 4. DETAILED DESIGN

The internal ribs would be subjected to primarily bucking loads thus would require a core, with a high compressive modulus capable of supporting the specic loading condition. Hence a ply wood core was designated as core material of choice for the internal structure.

Structrual Modeling Chamber Floor As noted previously, the chamber oor formed the basis of geometrical and structural integrity based on the design criteria. Formula and theory is based on Plate Theory by Timoshenko:

Figure 4.49: Density and Compressive Modulud Where: t=Skin Thickness c=Core Thickness Ec =Core Tensile Modulus Ec =Core Shear Modulus Es =Skin Tensile Modulud

Figure 4.50: Loading 1

Overall deection was calculated by rst calculating the exural rigidity of the prescribed bre resin mix, based on the characteristics of the primary skin. As the cores contribution is near negligible, the

4.4. STRUCTURAL equation for exural rigidity can be simplied to:

Es b h3 c3 12

(4.23)

Q = Gc bC

(4.24)

The Shear stiness can then calculated which yield all variables necessary to calculate plate deection

P a2 D2

D2 a2 Q

(4.25)

The facing stress and core shear stress can then be calculated to ensure that the selected bre resin mix can accommodate the impressed load. Four iterations of core thickness were run to identify what core thickness would reduce panel deection to the nominal level of 5 mm and what penalty of weight would have to be endured.

s =

M Es h 2D

(4.26)

c =

S Es td Ec + D 2 2

c4 y2 4

(4.27)

Cost constraint restricted the design of the platforms structure to a maximum of 6 layers of skin. An R80 core was chosen over an R45 core as a result of a markedly better compressive modulus and over an R60 core as a result of the gains in mechanical properties at a low penalty in weight and cost. The cores ranged from 5 mm, the thinnest available core to a 20 mm core in increments of 5mm.All calculations were based on a skin oriented in [0,90].

A 5mm core proved insucient as at 6 layers the expected deection was in excess of 7 mm. The 10, 15 and 20 mm core all satised the criteria at 4, 2 and 1 layer respectively. The penalty in weight ranged from 3.64 kg for a 20 mm core with 1 layer to 4.72 kg for a 10 mm core with 4 layers marking a dierence of approximately 29.7% As the calculations were based on the primary skin only, and the addition of a secondary skin of Aramid bres was necessitated for abrasion an impact resistance, the use of a 10 mm core with 4 layers of breglass was commission for the chamber oor. This would bring the deection down to approximately 2.41 mm. The skin facing stress was calculated to be 4.8

CHAPTER 4. DETAILED DESIGN

Figure 4.51: Weight and Deection of R80 MPa and core shear stress to be 0.0235 MPa, both well within the limitation of the bres and cores intrinsic structural property limitations. Including a safety factor of 2 and doubling the distributed load to 200 kg over the chamber oor brought the overall deection with 4 layers of skin to 7.9 mm and with 6 layers to 4.8 mm with a core shear stress of 0.047 MPa and 0.044 MPa respective, once again within the intrinsic structural limitations of the materials. A table of deections, core shear, and facing skin stress for multiple core and skin combinations can be found in Appendix H.

Chamber Ceiling (Top Plate) A similar method was used to calculate the deection, core shear and facing skin stress of the top plate. However, considering that top plate would consist of a ply wood core and that the rigidity of the core would contribute of overall structural stiness of the plate, exural rigidity was calculated as

Es bt3 Es btd2 Ec bc3 + + 6 2 12

(4.28)

The contribution of the cores stiness brought down the overall deection to less than 1mm with a single layer of bre glass. The core shear was calculated to be approximately 0.015MPa with a facing skin stress of 0.23MPa. Doubling the load brought the deection to just over 2mm with a skin facing stress of 0.47MPa and a core shear stress of 0.02 MPa. The total weight of the top plate calculated to be 11.44 kg, with one layer of bre glass and 12.267 kg with 2 layers marking a 7.2% increase in weight for a reduction in deection of over 50%. Although the calculated deection was well within the prescribed limit with a single layer of glass, the possibility of particulate contamination of the bre and inconsistency in layup, with the particular risk of air trapped in the bre resin mix, two layers of 330 g/m3 bre glass was commission for the top plate.

4.4. STRUCTURAL Internal Rib

The internal rib was modelled as a beam with end loads. The beam was dimensioned as a simply supported beam 10 mm thick, 150 mm wide to emulate the full height of the plenum chamber and 1000 mm long to represent the longest section of rib in the internal structure of the hull. A key assumption in the analysis was that the end loads were taken by both edges of the skin and that an ideal transfer of loads was achieved. Hence it is assumed that the 100 kg load is to be spread equally over the 4 supporting ribs which form the basis for the internal structure.

Figure 4.52: Rib-Load

The facing skin stress and critical buckling load were calculated to verify the load bearing characteristics of the internal ribbing sub-structure.

f =

P 2tf b

(4.29)

Pb =

2D l2 +
2 D Gc hb

(4.30)

Where D is calculated as

Ef tf h2 b 2

(4.31) (4.32)

h = tf + tc

The facing stress was calculated to be 0.56 MPa which is considerable less that the skin, whose typical yield strength, is around 150 MPa.

CHAPTER 4. DETAILED DESIGN

The critical panel buckling load was calculated to be 16730 N giving a safety factor of 6.69. However it is important to recall the simplications and their negative ramications on the actual load bearing capacity of the internal ribbing structure.

Final Layup
The following layup was employed for the layup for the hull and its

4.5

Skirt

Skirts were rst implemented in hovercraft designs in the 1960s to assist with containment of the air cushion (Bliault and Yun, 2000). By acting as a exible sidewall to the air cushion, skirts led to a decrease in the horsepower requirements of hovercrafts and an increase in several performance parameters, namely; air cushion performance, stability and obstacle clearance (Bliault and Yun, 2000). Skirts have been implemented in several dierent congurations, each oering dierent advantages to the ACVs performance. There are many considerations to undertake when choosing the conguration

4.5. SKIRT

suitable for a crafts application, including drag, stability and manoeuvrability requirements of the craft.

4.5.1

Design Specication

Skirt conguration is an area of ongoing research in industry and dierent skirt designs oer varying performance characteristics. Benchmarking of small hovercraft under two metres has been done to obtain empirical data on the most common congurations used for small craft. Four main congurations were identied; the segmented skirt (74%), open-loop bag skirt (22%), the bag and segment combination and the pericell or jupe skirt. Furthermore, combinations of these sets could be used to splice performance characteristics of diering congurations. A set of design criteria were generated for skirt conguration selection:
The drag association of the skirt should be minimised due to the application of the hovercraft.

This is to ensure that the probability of landmine detonation through skirt contact is minimised.
As active control of the hovercraft in pitch and roll is not being attempted inherent stability of

the platform is of high importance. The skirt must therefore control stability of the platform to maximum disturbances in pitch or roll.
The skirt must be adaptable to the platform shape and able to be both fastened and shaped to

the preceding hull design.

The hover height of a hovercraft is achieved through obtaining the correct skirt depth and cushion

pressure (Bliault and Yun, 2000). The skirt must therefore achieve a hover height clearance of 150 mm to the bottom of the hull.

4.5.2

Conguration Selection

Bag Skirt The bag skirt consists of an open loop bag (cross section Figure 4.53) oering inherent high stability of the craft as it forms a high pressure ring of air around the outer perimeter of the air cushion (Fitzgerald and Wilson, 1995). Implementation of this conguration requires that the bag is held at a higher pressure than the air cushion so that forces resulting from the bags internal pressure exceed those being exerted by the hovercrafts air cushion (Fitzgerald and Wilson, 1995). The individual pressures could be controlled by limiting ow through the bag using an area relationship. A disturbance to either

100

CHAPTER 4. DETAILED DESIGN

the pitch or roll balance of the hovercraft results in an eective increase in area (as the skirt deforms) causing an increased upward force and correcting moment. The high pressure edge to the cushion increases this moment. The main disadvantage is that this deformation also causes an increase in contact area with the ground, meaning the bag skirt has high terrain interaction drag characteristics.

Figure 4.53: Bag Skirt Cross-section (McLeavy and Wood, 1977)

Segmented (Finger) Skirt The segmented skirt conguration consists of small ngers attached around the perimeter of the hull to make up the skirt wall. These segments are fastened at the bottom of the hull to prevent blow out under pressure. Figure 4.54 shows this conguration. One advantage of the nger skirt is that the contact area with the ground is lower than the traditional open-loop skirt (Amyot, 1989) This is due to the fact that each segment is able to deform independently, and as variations in platform orientation occur a lower contact area is aected. This implies a relatively low drag characteristic.

Figure 4.54: Segmented (Finger) Skirt Conguration (McLeavy and Wood, 1977) Although the stability mechanism of the nger skirt is the same as the open-loop conguration (change in area resulting in corrective moment) (Fitzgerald and Wilson, 1995), this particular conguration is less eective as the pressure in the ngers is equal to that of the cushion, not higher. Another stability consideration is that the point of contact of these ngers should be within the perimeter of the hull to ensure that imbalances result in a deformation inwards rather than outwards (blow out) of the skirt

4.5. SKIRT (Fitzgerald and Wilson, 1995).

101

Bag and Segmented Combination

The bag and nger skirt combination combines the advantages of the independent bag and nger congurations.(Wheeler, 1974) It oers increased stability by employing the high pressure loop of the open-loop bag and decreased drag through a contact area equivalent of the segmented conguration at the ground. Bag and segment is the current conguration of choice for large commercial hovercraft (Bliault and Yun, 2000). The reason this conguration is rarely implemented in small hovercraft is that a high enough hover height is required to have sucient space for eective bag and nger sections. Other disadvantages are built into the complexity of this design, such as high material use and complex joins leading to decreased reliability.

Figure 4.55: Bag and Segment Skirt Combination (McLeavy and Wood, 1977)

Pericell/Jupe

The pericell or jupe style skirt is not common but is an alternative conguration for small hovercraft. The jupes themselves must be angled inwards so that as pressure builds the vertical component of force (from pressure) on the skirt causes ination of the individual cells (Fitzgerald and Wilson, 1995). Trimming of the jupes can become a dicult exercise and is important to achieve the desired stability of this conguration (Fitzgerald and Wilson, 1995) Implementation of jupe skirt cells is eectively segmenting the air cushion to create separate areas of pressure. The stability mechanism of this conguration is slightly dierent to that of the other congurations. As imbalances occur in pitch or roll, closing of the air gap causes a build up of pressure, resulting in a correcting moment based on jupe area.

102

CHAPTER 4. DETAILED DESIGN

Figure 4.56: Pericell (Jupe) Skirt Conguration (McLeavy and Wood, 1977)

Conguration Selection Based on the design criteria outlined for the skirt a list of criteria to compare against each conguration was dened. The main criteria can be dened as follows:
Drag: Drag resulting from skirt interaction with the terrain. Stability: The transverse and longitudinal stability resulting from the skirt congurations man-

agement of the air cushion.

Obstacle clearance capability: The skirts ability to clear obstacles and manoeuvre the operational

environment
Plough-in: Eect of skirt conguration on incidence of plough-in

Table 4.9: Skirt Conguration Comparison, Characteristic Bag Finger Drag High Low Stability High Low Obstacle Clearance Low High Ease of Attachment Low High Plough-in Medium Medium Repairability Low High Durability High Low Aesthetics Low High

Adapted from (Fitzgerald and Wilson, 1995) Pericell Finger & Pericell Combination Medium Low Very High High Low Medium Low Medium Medium Medium Low Medium Medium Medium Medium High

The three most important considerations choice were drag, stability and obstacle clearance. Skirt interaction drag was placed of signicantly high importance (25%) to ensure that the probability of prematurely detonating a landmine was minimised. Stability was also placed at the same signicance level (25%) as active control of stability was not being undertaken. This implies that inherent stability of the platform is required. The ability to clear obstacles was placed at a lower signicance level (15%) due to the relatively obstacle free denition of the operational environment.

4.5. SKIRT

103

Other considerations in the decision making process included durability and repairability of the skirt, ease of attachment to the desired hull shape, susceptibility of congurations to plough-in and aesthetics. Repairability was considered of signicant importance in the second-tier criteria (10%) as the skirt is incremental in successful operation of the platform. In the case of failure it was important that repairs were possible. Ease of skirt attachment to curved hulls was also placed at 10% signicance to introduce the element of diculty of manufacture into the selection criteria. Susceptibility to plough-in was weighted slightly lower (8%). Although the reduction of plough-in is of importance to hovercraft stability, the skirt conguration doesnt aect a crafts tolerance to tuck under so much as the care taken in its manufacture (Fitzgerald and Wilson, 1995). A well built and balanced skirt will reduce uneven drag concentrations. Similarly, the operational speeds and environment signicantly reduce the risk of plough-in. Other considerations such as durability and aesthetics were considered important but weighted low relative to other criteria (5% and 2% respectively).

Using the weightings placed on these requirements and the benchmarking done for small hovercraft skirts, a comparison between dierent congurations was undertaken to determine which is most suitable for Project Hover. Due to the low hover height requirement of the sensors, the combination bag and nger was deemed non-practical and unsuitable for the platform. This was because there was insucient space to implement both congurations. Thus, the comparison only considered bag, nger and pericell style skirts. Each was ranked, based on their previously discussed eciencies, for the dierent criteria outlined.

Table 4.10: Skirt Conguration Comparison, Adapted from (Fitzgerald and Wilson, 1995, pg. 12) & (Amyot, 1989) Characteristic Weight(%) Bag Finger Pericell Finger and Pericell Combination Drag 25 1 3 2 3 Stability 25 3 1 3 3 Obstacle Clearance 15 1 3 1 2 Repairability 10 1 3 1 2 Ease of Attachment 10 1 3 1 2 Plough-in 8 2 2 2 2 Durability 5 3 1 2 2 Aesthetics 2 1 3 2 3 Total 100 168 232 190 252

The selected conguration was a nger and pericell conguration which incorporated both low drag and high stability. Ease of attachment and repairability weighed heavily for these styles, which could easily be adapted to the platform shape relative to other congurations.

104

CHAPTER 4. DETAILED DESIGN

4.5.3

Material Selection

General Requirements A range of general requirements have been identied to dictate the material selection of the skirt based on its material specic failure modes. These modes include failure through abrasion, cutting or tearing, fatigue or cracking of the material proong (coating), and delamination. Signicant wear may also occur through contact between the mechanical fastenings and the material (Wheeler, 1974). Consideration of these failure modes in combination with the desired performance prole of the platform leads to a set of specic requirements;
The material must be air-tight; Have a high strength to weight ratio; Have signicant tensile and tear strengths; Be of high abrasion resistance; Possess the inherent ability to manage constant cyclical straining from tensile and compressive

loads (Amyot, 1989); The material used in commercial hovercrafts is a rubber coated polymer weave most suited to address all of the aforementioned criteria. The advantage of such materials is that specic properties can be optimised for an application based on altering the density or weave of the constituent layers (Amyot, 1989). Material eectiveness must be analysed after selection.

Constituent Material Selection Many dierent materials have been used for the elastomer/rubber constituent of skirt materials. The properties provided by this layer of the material are primarily high resistance to abrasion and cutting, resistant to weather and high in strength. (Amyot, 1989) Although many of the suitable materials uphold similar properties, no single material is the best for all the desired requirements. A comparison of dierent elastomers and there relative advantages and disadvantages has been detailed. The main material layer is made up of a woven structure which ideally provides the bulk of the material strength without severely diminishing the exibility of the elastomer coat. One material generally used for this backing structure is nylon as it has a superior spread of properties. Material comparison was however been undertaken to determine the most suitable combination.

4.5. SKIRT

105

Table 4.11: Elastomer Material Comparison, Adapted from (Amyot, 1989) Elastomer Name Advantages Relative Disadvantages Natural Rubber Fatigue, cut and Poor oxidation abrasion resistance. resistance. EPDM Excellent oxidation resistance Fair cut resistance resistance and dicult to bond. Polyurethane Highest Specic strength Fair to good and excellent abrasion fatigue resistance resistance Neoprene Excellent oxidation resistance fair cut and and easy to bond fatigue resistance Table 4.12: Fabric Material Comparison, Adapted from (Amyot, 1989) Fabric Name Advantages Relative Disadvantages Nylon (9g/denier) Strong, tough and elastic Higher creep rate. Bonds well to rubber Water Absorption Fatigue resistant and cheap Polyester (9g/denier) Lower creep rate Bonds poorly to rubber Water Absorption Aramid Fiber (20g/denier) High strength tendency for self abrasion A nylon backing structure with polyurethane skin was chosen based on the direct address of the general requirements. The nylon is able to provide the required strength to resist both tensile and compressive loads. Furthermore the material bonds well to rubber elastomers making it ideal for this application. Selection of the elastomer is largely a trade o between properties of importance. Polyurethane was chosen for its ability to provide strength and abrasion resistance whilst maintain relatively fair resistance to fatigue (due to cyclic tensile and compressive loading).

Material Analysis In order to ensure the requirements of the skirt are satised, a theoretical model was developed to estimate loading on the skirt segments from the cushion pressure. A cylindrical model was used to represent the skirt geometry most accurately while maintaining simplicity in the model. Using standard equations, hoop and longitudinal stresses in a pressurised thin wall cylinder were calculated. There were two thicknesses of the material available (0.25 and 0.35mm) and the stresses were calculated for each. A theoretical comparison for the yield strength of Nylon 6,6 was obtained (44.8-58.6MPa) (Callister, 2003) and far exceeded both the longitudinal and hoop stresses for the 0.25mm (73kPa and 146kPa respectively) and the 0.35mm (52kPa and 104kPa respectively) materials. In order to minimise the weight of the craft the thinner material was chosen as its material properties far exceeded the theoretically expected loading.

106 Joint Selection

CHAPTER 4. DETAILED DESIGN

From literature three main types of joint were identied. These were gluing, rubber welding, and sewing (Wheeler, 1974). Rubber welding was immediately ruled out as the facilities were not available to perform it. Gluing and sewing were compared in regards to ease of manufacture and sewing was selected as the most viable method. In order to eectively glue the material together dicult fastening methods would be required, however with sewing a join is instantly made and the material can be manipulated immediately. In order to select a thread it was decided that dierent available threads would be tested prior to manufacture to determine the strongest thread, and to ensure the join integrity would be maintained under pressure loadings.

Attachment Selection Several methods of attachment were identied from literature. The main criteria for selection were that the attachment method must:
create an air tight seal with the hull; be non-permanent yet robust to operational loads; be simple and cause minimal assembly loads on the skirt segments;

Several attachment methods were reviewed from literature however were discounted as they were primarily adaptable to straight edge craft only. Appendix L details these methods. Clipping the skirt material at the top to a suitable extrusion prole was the solution selected for hull attachment. Two types of materials were considered for this extrusion, aluminium and plastic. The use of plastics was eliminated due to insucient rigidity to withstand operational loads of the skirt and maintain a seal. An o the shelf aluminium extrusion prole used for the attachment of garden plastics to various structures was selected. Redpath Duralock, pictured in Figure 4.57, creates an eective airtight seal through the insertion of the inner locking strip, which is clipped into place using plastic clips. The strips should match the width of the skirt segments, with slightly shorter locking strips to account for material thicknesses, particularly sew joints. To stop the skirt segments aring out at the bottom they should be fastened at their lower contact point as well. Cable ties were chosen as they are weaker than the material and should a nger get caught on something, the cable tie will fail before the material tears.

4.5. SKIRT

107

Figure 4.57: Skirt Attachment Bracket

4.5.4

Template Generation

Segmented Skirt Methods of Template Generation Literature identied two main methods of template generation (Fitzgerald and Wilson, 1995) of which the geometrical method was chosen. This method involves creating a scale cross-section of the hull and shaping the skirt based on an existing template. This is eectively a theoretical approach. The shaping approach incorporates cutting a piece of foam to the desired skirt shape underneath the hull, draping material over and cutting to shape. The disadvantage of this is that the hull must be manufactured prior to template generation.

Existing Skirt Template Designs Two existing skirt templates were benchmarked, the bi-conical and extended skirt segments. The bi-conical skirt segment (Figure 4.58) takes on a cylindrical shape under pressure and is the simpler of the two templates to manufacture. The template is a one-seam design; a characteristic generally desired to increase reliability. The seam is relatively simple to stitch in comparison to the extended skirt segment. Figure 4.59 show the cross section of this templates. It is necessary in segment design that the point of contact with the ground lies within the outer perimeter of the hull for stability. (Fitzgerald and Wilson, 1995) This is to ensure that the nger folds in and not out on the occurrence of a disturbance in pitch or roll. Furthermore the outside edge of either segment should be parallel to the adjoining

108

CHAPTER 4. DETAILED DESIGN

Figure 4.58: Bi-conical Skirt Segment

Figure 4.59: Skirt attachment bracket

4.5. SKIRT

109

Figure 4.60: Skirt attachment bracket edge of the hull. The bi-conical segment was chosen as a base template for the skirt based on its simplicity and manufacturability, limited by the time constraints of the project.

Template Generation & Optimisation The initial stage of template generation was of a theoretical nature and included the creation of a scale geometric model of the hovercrafts cross-section. Because the hovercraft hull was shallow relative to multi-terrain craft, an interior ap was included to prevent ap out of the bottom of the skirt ngers. Furthermore, in order to attach the ngers using Redpath Duralock locking systems, a pouch at the top of each skirt segment was required to ensure that a good grip was formed yet the sealing properties of each nger maintained. Ination of the skirt pieces would result in ination of the pouch around the locking system and subsequent sealing against the adjacent segments. Ease of assembly was the main consideration in the creation of this pouch, leading to an interchangeable nger design. Figure 4.60 shows the geometric model with included dimensions used to generate the base template design. These dimensions are based on the design of the hull. This cross-sectional template could then be unfolded into a two-dimensional template. Some further optimisation of the curves to ensure three-dimensional eectiveness was required. This couldnt be achieved in a quantitative fashion and small rig was made, representative of the hulls cross section. Pictured in gure 4.61, this rig had the ability to inate the segments through small air feed holes and optimise geometry based on the following criteria.
Sealing between segments along their entire depth.

110

CHAPTER 4. DETAILED DESIGN

Figure 4.61: Skirt Test Rig

Flat contact with the ground when under pressure Ability to easily attach segments at the attachment point.

The nal template of the skirt segments is shown in Appendix J.

Jupe Skirt

Jupe skirts are characterised by their diameter, height and the slope angle of the sides. The area of the jupe minus the area of the feed hole determines the area over which the pressure may act to produce lift. The angle the jupe makes with the ground controls the manner in which the skirt inates, while the height of the skirt should be equal to the desired height of the craft. Given that the perimeter segmented skirt oers little contribution to the pitch and roll stability of the craft, the primary purpose of the jupe skirt will not be to create lift, nor maintain the air cushion (this is provided by the segmented skirt), but rather to ensure the pitch and roll stability of the craft. As outlined in Section 3.7.1 the requirement for static stability, as supplied by the jupe skirt conguration, is that they lie outside of the projection of the centre of gravity on the hull for all possible pitch and roll angular displacements. As such, a symmetric arrangement of four jupe cells has been proposed, as shown in Figure 4.62, to ensure the static stability of the platform.

4.5. SKIRT

111

Figure 4.62: Skirt conguration arrangement

Additional Stability Considerations For stability in operation, it is desirable that the jupes be able to counter moments that arise from the operation of the craft. In particular, the most severe operational loading will arise from the propulsion system. As such, the requirement is set that the jupes be able to produce a corrective moment equal to the maximum moment induced by the propulsion system operating at full throttle (see Figure 4.63). Hence, the jupe reaction moment requirment can be expressed using Equation 4.33.

MT < MR = pj Aj xj

(4.33)

Given that the moment arm of the propulsion system in pitch approximately 0.50 m, and the maximum combined output of the propulsion system is 60 N, means that the two forward jupes are required to produce a correcting moment of approximately 31.2 Nm, or 15.6 Nm each.

Jupe Sizing and Positioning As the fore section of the craft is reserved for payload, any feed hole for the jupe cells just be located aft of this section. As such, the maximum moment arm of the jupe from the planform centroid is at the border of the fan, that is approximately 0.3 m. Jupes generally have a pressure of approximately 1.2 times the cushion pressure (Fitzgerald and Wilson, 1995), leading to a area requirement of 0.047 m2 , and hence a jupe diameter of 0.245 m (Equation 4.34), hence a nominal diameter of 0.25 mm was designed, a height that is equal the the segment (150 mm). The angle which denes the slope of the

112

CHAPTER 4. DETAILED DESIGN

Figure 4.63: Propulsion moment during operation

D1 D2

250 mm 208 mm

Table 4.13: R1 898 mm R2 748 mm

1 2

8 50

sides ( ) with a slope of 8 , as to ensure proper ination (Fitzgerald and Wilson, 1995).

Template Generation

Manufacturing of the jupe requires the generation of a two dimensional template that can be cut out of material. The material then needs to be bent into shape and a seam sewn to join the two edges. Generation of the template requires the construction of two concentric arcs. The radius of the upper and lower arcs can be determined from the geometry of the cone that the jupe would form should it be extended to a point, and can be determined by Equation 4.34. The quantities dening these construction parameters are summarised in Table 4.13. The nal template had two additional tabs of material, the rst an extension of the top edge, the second an extension of the side, to allow for attachment to an appropriate mount and a seam line respectively.

1 2D

sin

(4.34)

4.6. CONTROL SYSTEM

113

Figure 4.64: Jupe Template Generation Jupe Attachment The jupes were attached to the underside of the hovercraft with their centres 0.3 m from the centroid of the planform along both axes. A mounting plate for the jupes was designed from 6 mm plywood with a 250 mm diameter hole removed. The completed jupe was placed though the hole, and slits cut from the upper edge tab and folded over for attachment using an adhevise. The entire mount was then attached to the underside of the hull using an removable adhesive.

4.6

Control System

The control system interfaces the various physical sensors and actuators of the hovercraft, providing both manual and autonomous control of the craft.

4.6.1

Components

In order to interface and control the various sensors and actuators for both manual and autonomous operation several components were required as outlined in Section 2.3. To achieve manual control at minimum a microcontroller and remote are required. For autonomous control in addition to manual

114

CHAPTER 4. DETAILED DESIGN

control components an IMU, GPS and planar control system are required (detailed in Section 4.7). The interactions between the various components are shown in Figure 4.65.

Figure 4.65: Control System Components

Microcontroller Selection The MCU directly interfaces with the various systems on board the craft, as summarised in Table 4.14. At a minimum requirement the MCU requires three Analog to Digital (ATD) converters for the gyro and two metal detectors. Ten digital outputs, four for servo signals and three each for stepper motors. Four digital inputs for receiving signals from the remote and autopilot, however more would be desirable to allow for the use of various digital buttons and switches. Additionally due to the high frequency of servo signals (between 45-50 Hz) an Input Capture Timer (ICT) is required for each of the four servos. Five digital timers are also required for each servo output as well one for vectored thrust and one for a failsafe Watchdog. Total minimum requirements of the MCU are:
2 ATD Converters 10 Digital Outputs At least 4 Digital Inputs 4 ICTs

4.6. CONTROL SYSTEM

6 other General Purpose Timers

115

Table 4.14: Summary System Metal Detector Marking System Remote Receiver Lift System Vectored Thrust Dierential Thrust Digital Gyro Autopilot

of Inputs and Outputs for Microcontroller Input Output Analogue Signal Servo Signal Servo Signals Servo Signal Stepper Motor Signals Servo Signals Analogue Signal Servo Signals -

The MCU was required to be low cost, low weight and have low power requirements. Additionally it needs to meet total minimum requirements for the system. To achieve this a MiniDragon+ from Wytec (Figure 4.66) was selected which meets all these criteria, whilst having the advantage of interfacing with Matlab. The MiniDragon+ has eight general purpose timers (which can be congured to be ICTs or Output Compare Timers) meaning that all servo outputs must be coalesced into one timer. This is possible because the signal width of a servo signal varies from 1-2 ms at a period of 20 ms (see Figure 4.68), which means that at maximum ten servos can be controlled from one timer, whilst maintaining no signal overlap.

Figure 4.66: MiniDragon Revision D from Wytec

Remote and Receiver Four channels were the minimum required to manually control each actuator on the craft: lift system, thrust vector, dierential thrust. The remote and receiver selected was a spare unit held by the University, which had a total of seven channels. In particular the remote, or radio, was a JR X-2610 and the receiver was a JR RS77S (as shown in Figure 4.67).

116

CHAPTER 4. DETAILED DESIGN

Figure 4.67: Remote and Receiver (ModelFlight, 2009)

4.6.2

Sensors and Actuators

Marking System

The marking system is actuated by a servo (Section 4.1.4). Servos are controlled by varying the pulse width (typically from 1-2 ms) and then sending the pulse at a frequency between 45-50 Hz (as shown in Figure 4.68). For a rotational servo this pulse width determines position, typically 1 ms is 0 rotation, and 2 ms is maximum rotation (typically 180 ). The servo selected for the marking system uses 1.1 ms and 1.9 ms pulse width representing 0 and 90 rotatation respectively.

Figure 4.68: Typical Minimum and Maximum Servo Pulse Width When not active the marking system servo is required to be in a neutral position (45 ), when actuatedthe servo moves to 20 , which corresponds to a signal pulse width of approximately 1.278 ms as

4.6. CONTROL SYSTEM shown in Equation 4.35.

117

1.1 + ((1.9 1.1) (20/90)) 1.278 [ms]

(4.35)

Landmine Sensors The landmine detection sensors produce an analog voltage output requiring a certain threshold voltage for a positive mine detection (see Section 4.1.3). The voltage produced ranges between 0 to 10 V, whereas the microcontroller has analogue inputs limited from 0 to 5 V. Hence a voltage divider circuit was used to scale the output voltage (Figure 4.69). When the input voltage exceeds the detection level threshold the marking system is activated. Before actuation a delay of 0.92 seconds (calculated in Equation 4.36, as craft speed is 5 km/h and distance between systems is 1.1 m) to correct for the crafts movement.

(1.1 0.5) 3.6 = 0.92 [s] 5

(4.36)

Figure 4.69: Voltage Divider for Metal Detector Signal The marking system is then actuated for a period of 0.72 seconds, creating a visible mark per specications in Section 4.1.4.

Lift System
The lift system throttle is actuated by a second servo of the same type as the marking system and is thus actuated the same way. The lift servo operates under two modes, an idle and operational mode. An idle mode allows the motor to warm up before being set to operational speed, this is done

118

CHAPTER 4. DETAILED DESIGN

for approximately 1 min after start up. An operational mode allows the throttle to be increased or decreased by using the elevator stick on the remote control, which if held up for 1 second will increase throttle by 5% every second, and if held down will decrease throttle by 5% every second. If throttle is held down for less than 1 second the lift throttle will reduce to idle for an emergency stop.

Vectored Thrust
The thrust vector produced from a rotating propulsion system is controlled by two stepper motors used to provide a single rotation vector for the propulsion system, in order to control the yaw angle of the craft. The stepper motors are controlled using a stepper motor microdriver (see Section 4.34). This was used in order to simplify the control of the stepper motor and to optically isolate it from the microcontroller in order to prevent damage. The microcontroller produces a signal through an Open-Connector (shown in Figure 4 (Lea, 2000)), inverted through a BC548 NPN transistor. This was connected to each PUL (pulse signal), DIR (direction signal) and ENA (enable signal) pin. The stepper motor was connected in a full coil conguration, in order to maximise the torque (but decreasing the maximum speed). These connections are shown in Figure 4.70.

Figure 4.70: Stepper Control Circuit Lea (2000) species that PUL must be at least 3.0 s in duration, for the stepper motor to step correctly. To achieve this a for loop in the microcontroller was utilised to give a constant 3.0 s period, by calculating the number of clock cycles taken to execute the loop and then multiplying by the time needed. This delay can be produced by calling stepperDelay(DELAY3US), as shown in the C code below: #define DELAY3US (500L)

void stepperDelay(unsigned long constant) {

4.6. CONTROL SYSTEM volatile unsigned long counter; for(counter = constant; counter > 0; counter--); }

119

The stepper motor may be reversed by changing the level of DIR signal, this must be ahead of PUL by at least 5 s, again utilising a for loop. To enable or disable the stepper motors ENA must be ahead of DIR or PUL by at least 5 s. The full timing diagram is shown in Figure 4.71.

Figure 4.71: Stepper Driver Signal Diagram (Lea, 2000)

Dierential Thrust Dierential thrust is used to control yaw angle at low speeds. Under manual control the thrust motors are actuated directly from the remote and receiver (see Section 4.6.1). The left motor is connected to channel Throttle and the right motor is connected to channel Elevator. In order to achieve a more natural control for the operator (to use one control stick for both motors) the two channels are mixed into a single channel using the remote, along with the channel Gear. This means that the right control stick of the remote is moved vertically to set the magnitude of the throttle, and moved horizontally to set the dierence between the two motors. Pushing the control stick horizontally left increases the right motors throttle, in turn yawing the hovercraft to the left, while pushing the control stick right increases the left motors throttle yawing the hovercraft to the right, which is similar to how many RC cars are operated. Under automatic control the two thrust motors are both given the same thrust magnitude, and are

120 controlled by only vectored thrust.

CHAPTER 4. DETAILED DESIGN

4.6.3

Yaw Control

To achieve closed loop control of the hovercraft a yaw control scheme was implemented. A gyroscope is used as input, to provide an estimate for yaw angle, and the stepper motors as output, to control the thrust vector of the hovercraft. The control scheme selected was Proportional Integral Derivative (PID) control, used in the standard form as shown in Equation 4.37. In this case the error term e(t) is the dierence in yaw set point to the current yaw estimate, determined from the integral of the angular velocity (measured by the gyro). Gains KP , KI and KD are proportional, integral and derivative gains respectively and are determined using the Ziegler-Nichols tuning method. The control scheme is as follows: 1. Read gyro value 2. Low pass lter of gyro value 3. Calculate bearing relative to start position 4. Calculate error between current heading and set point 5. Use PID to generate output value 6. Rotate thrust vector to new set point 7. Repeat step 1

u(t) = KP e(t) + KI

e(t)dt + KD e (t)

(4.37)

Prototype A prototype yaw system was developed as a test rig while the hovercraft was being designed and constructed. It consisted of two opposing brushless electric ducted fans, rotating on an axis, simulating the yaw of the hovercraft (as shown in Figure 4.72). The specications for the controller were:
Variable set point (0-270 ) Settling time of less than 10 s Less than 10% overshoot

4.6. CONTROL SYSTEM

Rise time less than 5 s

121

The system was mounted on a potentiometer to provide rotational feedback, which had a limited rotating shaft of 270 , which is dierent to the hovercraft which has no such limitation.

Figure 4.72: Yaw Control Prototype The controller was designed in Simulink and then targeted to a dSPACE controller. After this a step response was recorded which is used to obtain control tuning parameters, which were subsequently inputted as the gains. The system was then tested and it was found to be less than satisfactory so a low pass lter, anti-windup loop, manual tuned gains and signal saturation was added to the controller (the nal controller shown in Appendix C). The nal closed loop response of the system is shown in Figure 4.73, indicating that the nal controller met all requirements. This was then used as a basis for yaw control on the hovercraft itself.

Implementation Unlinke the prototype controller the yaw controller could not be implemented in Simulink, due to the complex signal interface needed for controlling the stepper motors (as shown in Figure 4.71). Hence the controller was implemented in C code on the MiniDragon. When power is applied to the MiniDragon the initialise function for PID is called, this sets the gains and other constants needed. void PID_Init(){ angle_setpoint = 90; pid_dt = 0.044; previous_error = 0; Ki = 3;

122

CHAPTER 4. DETAILED DESIGN

Figure 4.73: Yaw Prototype Closed Loop Response Kp = 5; Kd = .5; anti_windupgain = 10; } Yaw control updates at a constant frequency of 23 Hz (interrupt frequency of the MiniDragon). This greatly simplies the implementation of integral and derivitive terms, as they are determined at a constant time step. The desired set point is then used to calculate the thrust vector angle of the stepper motors, which is an angle between -180 and +180 . The code to acheive this is shown below. void PID_Loop(float angle_measured){ float output; float angle_error = angle_setpoint - angle_measured; /* I term */ integral_error += angle_error*pid_dt; /* D term */ derivative_error = (angle_error - previous_error)/pid_dt; /* Calculate overall error */ previous_error = angle_error; output = (Kp*angle_error) + (Ki*integral_error) + (Kd*derivative_error); Stepper_Desired(output); }

4.6. CONTROL SYSTEM

123

In order to change the setpoint a simple function is used, which can be accessed any time in order to update the new setpoint, this be controlled with the remote or the autopilot (see Section 4.7.3). void PID_Setpoint(float angle){ angle_setpoint = angle; }

Tuning

Ziegler-Nichols tuning was used to determine an initial estimate of the control gains. This involved inputting a step command into the hovercraft yaw system and measuring the response of the yaw. The tuning procedure is: 1. Turn the thrust vector to +90 2. Set the thrust magnitude to various levels (25%/50%/75%) 3. Record the yaw angle with a gyroscope 4. Wait until the hovercraft rotated 360 5. Repeat from step 1 with -90 thrust vector The result of the normalised gyro output is shown in Figure 4.74, along with the integrated value which is the approximated yaw angle for clockwise rotation1 at 75% throttle .

Low Pass Filter

From Figure 4.74 it is evident that the gyro measurements are subject to a large level of noise. This noise is most likely due to high frequency vibrations from the lift system travelling through the hovercraft. To attenuate this noise a low pass lter was designed and implemented. The maximum yaw rate of the hovercraft was measured to be approximately 40 /s, resulting in a low pass lter with a pole at 50 /s. With a sample period of 44 ms, this results in a time constant of 0.468 s (=44/(50+44)), producing the following low pass lter:

Low Pass Filter =

1 1 = 1 + s 1 + 0.468s

(4.38)

Clockwise rotation results in a negative gyro reading

124

CHAPTER 4. DETAILED DESIGN

Figure 4.74: Gyro Step Response with both Normalised and Integrated Readings

Closed Loop Response

The results from testing the closed loop yaw response are shown in Figure 4.75. The small drift in the rst three seconds is likely attributed to either slope of the ground or symmetry of the hovercraft skirt. It can be seen after three seconds the controller has corrected for this, but a large overshoot is exhibited, until integral gain has increased suciently to drive it back towards 0 by 7 seconds. Again an overshoot is exhibited as it passes 0 due to the lack of an agressive dierential gain.

The poor performance of the controller may be attributed to an unequal response in clockwise and counter clockwise rotations of the craft. This means that a dierent set of gains would need to be set, depending on which direction the hovercraft is yawing. This was an unexpected complication in the implentation of control and due to time contraints this was not fully implemented.

Further improvements could be made with augmenting the system with a digital compass to reduce the proportional error in the estimator. Although a digital compass does not have as high an updating frequency as a gyro, the long term hysteresis of a gyro would be mitigated with this solution.

4.7. NAVIGATION SYSTEM

125

Figure 4.75: Closed Loop Yaw Response

4.7

Navigation System

To meet the autonomous requirements of the platform a two dimensional (planar) navigation and control system was used. This would be a commercial o-the-shelf system, as it was deemed outside the scope of the project to develop one for the hovercraft. Additionally a custom built system is unlikely to be as reliable as a commercial system.

4.7.1

Selection Criteria

In order to meet the requirements of the project design goals, a navigation system was found necessary to include the following components as shown in the feasibility study (see Section 2.3) these are a measurement system and a sensor or sensors for obstacle avoidance along with control system to combine these. A measurement system typically consists of a DGPS or a GPS for positioning. A DGPS is used when a more accurate position (within a few cm) is required, but this comes at a much higher cost. To provide more accurate estimates of velocity and acceleration this system is often augmented with either or both an IMU or KF. An IMU often contains a gyroscope and an accelerometer for direct measurements of velocity and acceleration respectively, which gives a more accurate measure because there is no need to integrate the data. Additionally these can be used when a GPS signal can not be acquired, due to inclement weather or obstacles. A KF is used to better estimate the states of a system from noisy measurements which will also improve the accuracy of the system. It is not a requirement for the project to exhibit obstacle avoidance, therefore a seperate system that

126

CHAPTER 4. DETAILED DESIGN

acheives this was not investigated. However, a system that has obstacle avoidance as a feature will still gain favour over other systems. The control system usually consists of one or several MCUs or Personal Computer (PC)s depending on the type of sensors used. Since a MCU is both lower cost and smaller size a system that uses one will be given more preference. A futher criteria is that the system be customisable, as it is unlikely that an autonomous system for a hovercraft will be available.

4.7.2

Selection

Autonomous planar navigation systems had a very low availability, thus an autopilot system was selected. The types of autopilots considered were ones often found in a small Unmanned Aerial Vehicle (UAV), as this was deemed to be the next most suitable system. The main criteria when selecting an autopilot was for it to be highly congurable. This was due to not using a typical platform that an autopilot is designed for, since most commercial autopilots are designed solely for use with a particular UAV airframe (or small set of airframes) and certainly not a hovercraft. In order to further distinguish a suitable autopilot the following extra requirements were used:
Remote Monitoring: Allow the hovercraft to be remotely operated, controlled by software on

a host computer.
Low cost: Found that an autopilot can range in cost from three-hundred dollars to over ten-

thousand dollars.
Communications Bus: To allow communications between autopilot and separate micro-

controller.
Several Outputs: In order to accurately control the hovercraft many actuators are deemed

necessary.
Small: As with all components on the hovercraft low mass and size are desirable, though not

as important as other factors. In order to eectively quantify the dierent types of autopilots they were individually benchmarked against our selection criteria:

4.7. NAVIGATION SYSTEM

127

Table 4.15: Selection Criteria for Autopilot Priority Maximise Minimise Lowest Outputs Power Consumption GPS Accuracy Size Communications Bus Weight Highest Flexibility Cost

Tiny v2

Table 4.16: Tiny v2 Features Criteria Values Outputs High (8xPulse Width Modulation (PWM)) Power Consumption Low (3600 mW) GPS Accuracy Moderate (Within 3 metres) Weight Low (24 grams) Size Medium (71x40x4mm) Flexibility High (Fully Customisable) Cost Moderate (US $685.00) The Tiny v2 developed by Paparazzi is free and open-source hardware and software project intended to create an exceptionally powerful and versatile autopilot system by allowing and encouraging input from the community (Project, 2008). The most attractive feature of this autopilot is the high level of customisation oered, as full source code is supplied (released under the GNU Public License (GPL)), both for the autopilot itself and for the software on the host computer. By including the full source code it is possible to tailor the system to meet the exact requirements of the navigation system. Another feature is that the Central Processing Unit (CPU) on the autopilot has two communications buses to be able to communicate with an external micro-controller, using either Serial Peripheral Interface (SPI) or Inter-Integrated Circuit (I2C) protocols. The GPS accuracy of this autopilot was signicantly lower than most other autopilots investigated, but this could be combined with an IMU system to improve accuracy. Due to the open-source nature of this autopilot this could be achieved with much less diculty than other autopilots.

Kestrel Autopilot v2.2 The Kestrel Autopilot v2.2x is designed for use with small UAVs, and it is the smallest and lightest full-featured autopilot on the market (Pro, 2008). The Kestrel has a high level of exibility, such as four serial ports to interface with external CPUs. These serial interfaces allow the use of standard modems and GPS units.

128

CHAPTER 4. DETAILED DESIGN

Figure 4.76: Tiny v2.1 Top and Bottom View (Project, 2008) Table 4.17: Kestrel Autopilot Features Criteria Values Outputs High (4xPWM) Power Consumption Low (4800 mW) GPS Accuracy Moderate (Within 3 metres) Weight Low (16.65 grams) Size Small (51x35x12 mm) Flexibility High (Many Congurations) Cost High (US $9,000) One negative feature of this autopilot is the cost, totaling US $9000 for both the hardware and the software.

Figure 4.77: Kestrel v2.2 with Magnetometer (Pro, 2008)

Microbot APS The Microbot APS Series Platform is an autopilot that is designed with emphasis on tight integration, high reliability, and maximum exibility (Microbotics, 2008). The Microbot is the largest autopilot studied, but even this is not overly large in comparison with the rest of the hovercraft. The most

4.7. NAVIGATION SYSTEM

129

Table 4.18: Microbot APS Features Criteria Values Outputs High (40xPWM) Power Consumption Low (3300 mW) GPS Accuracy High (Within 2 metres) Weight High (250 grams) Size Medium (67x53x67 mm) Flexibility Medium (Expandable) Cost High (AUS $16,150) notable feature of the Microbot is its wireless transmitter range, with a maximum range of 32 km. This autopilot has also been designed to oer much exibility by including many I/O lines, but this is not necessary for this project. This autopilot is also quite costly, and to include an IMU and GPS costs $16,150.

Figure 4.78: Microbot APS (Microbotics, 2008)

Ardupilot

Table 4.19: Ardupilot Features Criteria Values Outputs Low (4xPWM) Power Consumption Unknown GPS Accuracy Low (Within 3 metres) Weight Low (50 grams) Size Low (30x47x2 mm) Flexibility Medium (Customisable) Cost Low (US $250) The ArduPilot is the most basic of the autopilots studied, and the cost of the system reects this. One

130

CHAPTER 4. DETAILED DESIGN

advantage to this system is that the hardware and software are fully open source, but to reach the level of operation needed for this project many of the features need to be designed and implemented. Even though the ArduPilot is fully open source, its exibility was not regarded as high because the autopilot itself is limited in the hardware already chosen.

Figure 4.79: Ardupilot (Sparkfun, 2008)

4.7.3

Autopilot Selection

Each autopilot was critiqued using the selection criteria and then compared to each other. From this it was clear that the Tiny v2 was the best choice, because of the moderate cost (US $685.00) but being the most congurable (which was given the highest priority) and performing on par with the other autopilots in the other categories (except the Ardupilot which it exceeded). One area that this autopilot is considerably weaker in, that was not considered in the selection criteria (see Section 4.7.2), was the lack of support if diculties arose. However, due to the open-source nature of the Tiny v2 documentation can be found from their Wiki (Project, 2008) which contains many articles to help the user, and their discussion board. Another area it was slightly below the Microbot AP was in GPS accuracy, but this was only by one metre and could be improved by substituting in a DGPS instead. The Tiny uses common PID control for stability and navigation routines, using Equation 4.37. The control gains of which must be tuned for the specic airframe that is used, or in the case of this project the hovercraft. In order to simplify the navigation, Tiny uses a carrot system. The carrots are positioned in the direction of travel, at a distance that the craft could reach in a set time of ve seconds. The control

4.7. NAVIGATION SYSTEM

131

loops setpoints are then adjusted to reect this new position, and the craft is controlled towards this position. The carrot position is then updated at a period of one second to get improve the response. In order to negotiate curves, the carrots are placed at a tangent to the curve as shown in Figure 4.80.

Figure 4.80: Carrot Placement

Paparazzi Center Paparazzi Center was the software that comes with Tiny, which runs on a Linux based Operating System (OS) is used for conguring most aspects of the autopilot through the use of .xml les. This program is divided into ve main sections. The airframe conguration le contains all the hardware and software settings for an aircraft. The ight plan conguration le sets the way points and routines that will be used in an autonomous ight. The setting conguration le is used to alter ight parameters and modes of ight (such as launch, GPS reset). The radio conguration le sets the variable for each channel and type of pulse (positive or negative). The telemetry conguration le is used to congure which messages are sent back from the autopilot and the period of which they are sent (in seconds). The conguration les used for the project are shown in Appendix A.

Airframe Many adjustments of the airframe le were required. In the servo module the three servos outputs to be used were dened relative to the actuators controlled, namely the two thrust fans and thrust vector, the code used is shown below: <servos> <servo name="THROTTLE_LEFT" no="4" min="1100" neutral="1100" max="1900"/> <servo name="THROTTLE_RIGHT" no="0" min="1000" neutral="1500" max="2000"/> <servo name="YAW_OUTPUT" no="2" max="1000" neutral="1500" min="2000"/> </servos>

132 <rc_commands> <set command="THROTTLE_LEFT" value="@THROTTLE"/> <set command="THROTTLE_RIGHT" value="@ROLL"/> <set command="ROLL" value="@ROLL"/> </rc_commands>

CHAPTER 4. DETAILED DESIGN

<command_laws> <let var="aileron" value="@ROLL"/> <let var="elevator" value="@PITCH"/> <set servo="THROTTLE" value="@THROTTLE"/> <set servo="AILERON_LEFT" value="$aileron"/> <set servo="AILERON_RIGHT" value="$aileron"/> </command_laws> The nominal airspeed was then reduced from 20 m/s to 1.4 m/s (operational speed of the hovercraft), in order to adjust the nominal speed that the autopilot expects the vehicle to travel. Adjusting this allows for better control of the system, as it will adjust the placement of carrots. The takeo block was also modied, enabling the autopilot to initialise variables that are more consistent with the performance of the hovercraft. <section name="Takeoff" prefix="Takeoff_"> <define name="Height" value="6" unit="m"/> <define name="Speed" value="1.4" unit="m/s"/> <define name="MinSpeed" value="1" unit="m/s"/> </section>

Radio Conguration The receiver used (shown in Figure 4.67) had 7 channels, thus the radio le needed to be congured to reect this as each channel sets a dierent command in the airframe, but only the ones shown in the airframe conguration are used. Min, max and neutral for each channel is the length of the pulse in ms and the neutral position by the remote. <radio name="JRRS77S" pulse_type="NEGATIVE"> <channel ctl="THRO" function="THROTTLE" <channel ctl="AIL" function="ROLL" min="1100" neutral="1100" max="2000"/> min="1000" neutral="1500" max="2000"/>

4.7. NAVIGATION SYSTEM <channel ctl="ELE" <channel ctl="RUD" function="PITCH" function="YAW" max="1000" neutral="1500" min="2000"/> min="1000" neutral="1500" max="2000"/> min="1000" neutral="1500" max="2000"/> min="1000" neutral="1500" max="2000"/> min="1000" neutral="1500" max="2000"/>

133

<channel ctl="GEAR" function="HATCH" <channel ctl="AUX" function="MODE"

<channel ctl="AUX2" function="MODE2" </radio>

Flight Plan The hovercraft was to be tested for an autonomous ight on the Torrens Parade grounds. A simple map for initial tests is shown in Figure 4.81 . The ight plan was simply to navigate in a line from start position (HOME) to way point 1, point 1 to point 2 and nally point 2 to point 3. The rst step was to nd the GPS position of the Torrens Parade ground. These were then inputted in the rst line of ight plan le.

Figure 4.81: Torrens Parade Ground Test  <flight_plan alt="75" ground_alt="0" lat0="-34.917475" lon0="138.599839" max_dist_from_home="1500" name="Basic" security_height="25"> The way points were dened, which is done by the x and y displacement from home position (in metres). <waypoint name="HOME" x="0" y="0"/>

134 <waypoint name="1" x="25" y="-35"/> <waypoint name="2" x="70" y="-40"/> <waypoint name="3" x="90" y="-65"/>

CHAPTER 4. DETAILED DESIGN

Blocks of navigation routines are dened, to move the hovercraft from point to point. After completing each navigation block Paparazzi automatically starts the next this is detailed in Appendix A.

Yaw Setpoint In order to control the setpoint for the active yaw control (see Section 4.6.3), the autopilot must be congured to output the correct signal. To produce this signal the geometry is derived, a graphical depiction of the angles is shown in Figure 4.82. The angles are found from the following equations where CP is Current Position and DP is Desired Position.

CPy CPx DPy = atan DPx DPy CPy = = atan DPx CPx = atan

(4.39) (4.40) (4.41)

This was then implemented as the roll command in the Paparazzi source code, due to yaw angle control not being currently implemented (as of October 2009). All the routines present for roll were removed, and custom code was inserted in the function h_ctl_roll_loop in fw_h_ctl.c (located in paparazzi3/sw/airborne/). The code is shown below2 . First the angles were calculated, then the command (cmd) for the servo was determined, which must be normalised between -9600 and +9600. Finally the new setpoint for roll (h_ctl_aileron_setpoint) was set, which sets the output for the servo on the next cycle. inline static void h_ctl_roll_loop( void ) { // This code is called at approx 60Hz float alpha_yaw, beta_yaw, gamma_yaw;

alpha_yaw = atan2(estimator_y, estimator_x); beta_yaw = atan2(desired_y, desired_x);

Current Position is given by estimator x and y, where Desired Position is given by desired x and y

4.7. NAVIGATION SYSTEM // Get reference to hovercraft frame gamma_yaw = PI - beta_yaw - alpha_yaw;

135

// Normalise gamma between -9600 and 9600 float cmd = (((PI+gamma_yaw)/(2*PI)) *(2*9600)) - 9600;

// -9600 makes the plane turn to right CW, +9600 makes it turn to left CCW h_ctl_aileron_setpoint = TRIM_PPRZ(cmd);

// Save new readings oldreading_x = estimator_x; oldreading_y = estimator_y; }

Figure 4.82: Yaw Angle Calculation

4.7.4

Integration

The autopilot is able to communicate yaw angle by connecting the customised roll servo output directly into the MCU as a servo input. This then controls the setpoint for the hovercraft yaw.

136

CHAPTER 4. DETAILED DESIGN

Unfortunately due to time constraints and the yaw control diculties this was not nalised. Further integration would allow for bidirectional communication between these two systems. The Minidragon is able to use Serial or SPI communications, where the Tiny is able to use either I2C or SPI protocols. Thus a SPI protocol would need to be implemented on both devices.

4.8
4.8.1

Platform Analysis
Platform Weight

After the design process was completed a nal weight was desired to re-evaluate the performance characteristics of the platform. In order to do this the design weights of all components were summated to determine the total design weight (Wd ). Components designed on CADs material properties were entered in and weights determined. Purchased components weights were obtained from data sheets. The summary of total component weights is summarised in Appendix I. Wd was determined to be 77 kg.

4.8.2

Trim and COG Calculations

The design COG of the craft was calculated to help specify trim conditions using the specic locations and design weights of each component. The detail of this process is documented in Appendix L. The calculated COG conditions without trim are: 1. xpos = 81 mm Aft of the centre 2. ypos = At the centre 3. zpos = 44 mm above the centre The condition on static trim is that the centre of gravity sits on or above the centre of pressure. In order to statically trim the craft a balancing moment is required to adjust for this variation in COG in the x-direction. It was calculated that to balance these moments the trim weight (set at a moment arm of 800 mm from the local co-ordinate system) is 7.5 kg.

4.8. PLATFORM ANALYSIS

137

Figure 4.83: 3 View Drawing

This Page Intentionally Left Blank

Chapter 5 Manufacturing
The manufacturing methods utilised were strongly dependent on the design requirements of the platform; lightweight, yet high strength to minimise the eects of object collisions. The platform was manufactured using a variety of composite materials and methods, as well as aluminium and several metal working processes. The construction consisted of the following 6 main systems ; the lift system, the inlet bell, propulsion system, hull, the skirt and the fairing. The components were assembled as they were manufactured in order to test their functionality as quickly as possible.

5.1

Manufacturing Methods

A variety of manufacturing methods were utilised in the construction of the platform in order to meet the requirements of each component. These methods were selected based on availability, cost, and the ease at which the methods could be adopted without prior experience. The lift system and propulsion system were constructed entirely of aluminium. A variety of manufacturing methods were available, however some were required to be outsourced. Welding, cutting, grinding, lathing, and drilling methods were readily available at the Mechanical Engineering Workshop, as well as a CNC milling machine. Components consisting of complex geometry and high tolerances required the use of a laser cutter. This level of machinery was not available, and as such an external manufacturer was sourced. The manufacture of the inlet bell, hull and fairing consisted of a foam mould that was constructed and shaped using a hot wire cutter. The hot wire cutter consisted of a tightly strung length of stainless steel wire connected to a high amperage power supply. The current applied caused the wire to heat up, allowing the wire to cut through the foam. This method was readily available, and as such, other methods were not considered. This mould created the surface to allow for the composite lay-up to take place. The lay-up method utilised was a hand lay-up technique, as this was dened by the selection of the resin and cloth. Vaccuum bagging was also used in manufacture of the hull to ensure adhesion of the cloth to the klegcell and plywood. 139

140

CHAPTER 5. MANUFACTURING

Manufacture of the skirt was achieved through the use of a standard sewing machine, as well as an eyelet kit to embed eyelets into the material.

5.2

Lift System Construction

Mounting Plate - The aluminium mounting plate was laser cut as four bolt holes were required to align precisely with the existing mounting bolt holes of the engine, and a large circular hole was required to be cut to allow for the shaft to freely rotate. Eighteen additional bolt holes were drilled in the plate, twelve of which were four millimetre diameter and the remaining six were six millimetre diameter. These holes were positioned to align centrally on either side of the T-section, and to give even torsional strength along the joints. Aluminium T-sections - The T-sections were cut to the required length using a metal cutter. Six holes were then drilled in appropriate locations in order to attach the mounting plate. A further two holes were drilled in the upright of the T-section to facilitate the attachment of the support plates. Plate Aluminium - Plate aluminium was used to make the support plates, the space plates and the duct. This material was simply cut to the desired dimensions using a guillotine, and then bolt holes were drilled for attachment. In the case of the duct, the material was also bent using a wooden mould of the correct diameter, and joined through welding to create a ring. Mild Steel Bars - The mild steel bars used to manufacture the shaft extension were cut to the required length again using the metal cutter. This was followed by the use of a drill press to bore a hole through the bar to allow it to be positioned over the existing engine shaft. Slots were milled inside this steel bar as keyways. Key Steel - This steel was cut to the appropriate sized keys and led at one end to create a rounded tip to allow them to t in the existing engine shaft and the fan hub.

5.3

Inlet Bell Construction

The inlet bell was constructed by rst shaping a foam ring to the desired radius of curvature. The foam ring itself was cut from a foam block of the required height using plywood templates. As the radius was too dicult to achieve using the hot wire cutter, sand paper was used in order to handform the required shape. A cardboard template was cut out with a hole in the middle which was the same size and shape as the required foam radius. The foam was then sanded down until this cardboard

5.4. PROPULSION SYSTEM CONSTRUCTION

141

cutout was able to be placed onto the foam and moved around the entire length unimpeded. This process is shown in Figure 5.1 A layer of breglass was applied to the foam and sanded down to give a smooth nish. Following this, a generous amount of PVA release agent was applied to the layer of breglass, follow by a further two layers of breglass, again sanded to the required nish. The bell was then able to be removed the mould by peeling away the last 2 layers of breglass from the PVA release agent.

Figure 5.1: Bell Manufacture

5.4

Propulsion System Construction

This system consisted of an aluminium mounting plate, an aluminium fan ring, two aluminium supporting plates, a hollow aluminium tube section, three aluminium support struts, four aluminium braces, mesh, sprockets and stepper motor brackets. Mounting Plate - The mounting plate outline was laser cut for similar reason to the lift system mounting plate. Several small bolt holes were required to be precisely positioned to allow for the mounting of the electric motor. This plate was then bent in a metal bender to an angle of 90 degrees at the front to allow for the motor to be mounted horizontally. Fan Ring - The aluminium fan ring was cut from a piece of sheet aluminium using a guillotine. This material was then bent to the required diameter using a wooden support brace, and the two sides of the sheet were joined together through welding in order create a permanent ring. Supporting Plates - Support plates for the fan housing were also required to be laser cut so as to

142

CHAPTER 5. MANUFACTURING

achieve the required complex shapes and the required high tolerance for the bolt holes. One of these mounting plates was attached directly to the vertical angle of the mounting plates and to one side of the fan ring, while the other was attached to the opposing side of the fan ring, creating housing for the propeller. Aluminium Tube Section - An aluminium tube section was simply cut to the required length using a metal cutter, and welded onto the horizontal surface of the mounting plate. This served as a pivot location for the propulsion system, and was chosen to be hollow to decrease weight and also allow required signals to be transmitted from the base of the mount to the top. Aluminium Support Struts - The support struts were made of hollow square cross section aluminium, and were cut a suitable length using a metal cutter. One end of the strut was cut at a precise angle to allow a 45 degree angle between the tube section and the strut upon joining. A circular piece of sheet aluminium was also cut using a bandsaw to act as a foot at the base of the strut. This strut was joined to the tube section, and the foot was joined to the strut through welding. Braces - Aluminium braces were cut to the required length using a hacksaw. These braces were constructed using a length of aluminium which was 10mm wide and 2mm in thickness. As these braces were to be used to give added structural strength between the support plate and the fan ring, they were required to be twisted such that they could be bolted both into the support plate, as well as into the fan ring. This was achieved manually by clamping the length of aluminium in a vice, and using a pair of pliers to twist the aluminium until the required angle was achieved. Mesh - The mesh was to be used as a nger guard for the propellers and to prevent large objects from coming into contact with the propeller. The mesh was cut to be of a slightly larger diameter than the fan ring using wire cutters. This mesh was then sandwiched between the supporting plates and the fan ring, before any excess mesh was trimmed away, again using wire cutters. Sprockets - A pair of sprockets were required to provide the rotational properties of the system. These sprockets were purchased as of the shelf items, however the vendor was required to make an alteration. As the two sprockets would be mounted to dierent diameter shafts, the larger of the sprockets was required to be drilled out using a drill press to allow it to t over the tube section, whereas the smaller sprocket required an insert held in place by grub screws to be manufactured using a lathe and drill press to t the smaller shaft of the stepper. Stepper Motor Brackets - Six aluminium 90 degree sections were cut to the required length using a metal cutter to allow attachment of the stepper motor to the hull. One pair was attached to the hull, while the remaining four were attached directly to the stepper. The stepper motor was required

5.5. HULL CONSTRUCTION

143

to slide along these brackets to allow for tensioning of the chain attached from the stepper to the sprockets. A milling machine was used to create a pair of slits in each length attached to the hull such that bolts connecting the stepper brackets to the brackets on the hull could be loosened, allowing the stepper to slide within the slits.

5.5

Hull Construction

Construction of the hull was achieved by working from a ground up approach. The base of the hull was constructed rst, followed by the rib system, the top lip, and nally the top plate. The entire hull was then painted using 2pac paint. The Base - The base of the hull was constructed entirely of composite material, with the exception of a plywood outer ring to provide a fastening point for the skirt segments. The foam mould was cut to shape using a hotwire cutter and several templates which were cut to the desired shape through the use of a CNC milling machine. Upon completion of the mould shown in Figure 5.2, and MDF board was cut to shape and positioned on the base of the mould to provide a smooth surface for layup. Gaps present between the foam mould and the MDF were lled using phenolic micro balloons. A layer of breglass was then applied to the mould and sanded down to a smooth nish in order to achieve an ideal mould shape. The mould was then coated in wax to allow for the mould to be removed from the composite hull with ease. Two layers of breglass were applied over the mould, followed by two layers of kevlar as shown below in Figure 5.3. A piece of Klegcell R80 structural foam was then cut to size using a stanley knife and placed on the base of the hull, followed by the plywood outer ring which was cut to shape using a jigsaw. A further two layers of kevlar and two layers of breglass were applied to form the complete base of the hull. The mould was then removed, and the base was sanded to achieve the required nish. A hole saw was then used to cut the required diameter feed holes in the side of the hull. These holes were then sanded to remove any loose breglass or kevlar cloth. Ribs - The rib system present in the hull was constructed entirely of plywood. These ribs were cut to the required shape using a CNC milling machine as the geometry was complex and required accurate tolerances to facilitate the positioning within the hull. A layer of breglass was added to each of the ribs to increase strength. The rib system was attached to the hull using a mix of phenolic micro balloons and carbon bre strands, as shown in Figure 5.4. This provided an extremely strong glue that was applied along all joins, securing the ribs in place. Top Lip - A top lip was manufactured for the hull to provide a mounting point for the top plate. This lip was constructed of plywood, to allow for the attachment of bolts. The lip was cut to shape

144

CHAPTER 5. MANUFACTURING

Figure 5.2: Hull Mould using a jigsaw, and positioned inside the outer edge of the hull, supported by the rib system. Phenolic micro balloons and carbon bre were again used as a glue to attach the lip securely to the hull and rib system. Following the securing of the lip, two layers of breglass were applied along the top of the lip and down the side of the hull to again increase strength, and to ensure the entire hull acted as a single structure. Top Plate - Construction of the top plate was achieved through the use of a jigsaw to cut out the required outline of the plate, followed by the hole in the middle to facilitate the lift system. The top plate was again constructed of plywood; however two layers of breglass were applied to both sides to ensure adequate strength. Attachment of the top plate to the lip was achieved using bolts. The holes for these bolts were drilled on assembly once the plate was positioned in the correct location. The image below in Figure 5.5 shows the completed top lip, top plate and rib system.

5.6

Skirt Construction

Skirt construction was a straight forward process once a suitable method for sewing the segments was achieved. Polyurethane coated nylon was used as the material for the skirt segments, aswell as an aluminium attachment mechanism. Segments - Two paper templates were prepared for the shape of the skirt. These paper templates were used to outline the segment shapes on the skirt material using a piece of chalk. The two pieces were then cut out and sewn together around the edges to create one full segment.

5.7. FAIRING CONSTRUCTION

145

Figure 5.3: Hull Glassing Attachment Mechanism - The skirt attachment mechanism was purchased as an o the shelf item, and as such was only required to be cut to suitable lengths using a hacksaw.

5.7

Fairing Construction

Construction of the fairing was carried out in a similar fashion to that of the base of the hull. A foam mould was cut to the desired shape, and sanded to achieve a smooth nish. Two layers of breglass were then applied and again sanded down. Given that the fairing was to be the most visible portion of the platform, a superior nish was required. To achieve this, Pro Builders Bog was applied over any imperfections that were present and again sanded down to ensure all areas of the fairing were of a consistent height and surface nish. A coating of paint primer was then applied to the surface, and any imperfections overlooked on the rst application of bog were corrected, followed by a second coat of primer. 2pac paint was applied, followed by a light sand with 1200 grade wet sandpaper to remove any concentrations of paint. A second coat of 2pac paint was then applied to give the required nish. The rear section of the fairing was cut into two pieces to allow for easy installation around the propulsion systems.

5.8

Final Assembly

Having completed the manufacture of the major systems present on the platform, the nal assembly could then take place. This process was undertaken in several stages outlined in this section. All

146

CHAPTER 5. MANUFACTURING

Figure 5.4: Hull Installation

installations excluding the process related to the skirt were carried out with the top plate removed from the hull to allow access to the underside for the fastening of bolts.

5.8.1

Lift System Installation

Installation of the lift system onto the hull was required to take place in a precise order such that all components of the system could be tted correctly. The fan duct was attached to the hull rst using aluminium L brackets. These were rst bolted to the ring, then drilled on assembly and bolted to the top plate. The six vibration isolators were then bolted to the spacer plates and support plates, before being bolted to the hull. The position of these supports was measured so as to allow the engine to face forward on the platform for aesthetic purposes. The inlet bell was then placed inside the fan duct, and slits were cut into the bell to allow it to t over the supports attached to the isolators. A piece of rubber with a hole cut out to a diameter slightly smaller than that of the bell was glued to the underside of the top plate in order to better locate the bell within the fan duct. The mounting plate was then attached to the T-sections, followed by the engine being bolted to the mounting plate. The shaft extension was then tted to the engine, followed by the attachment of the fan to the shaft extension. This subassembly was then positioned within the supports already mounted to the hull, and bolted into place.

5.8. FINAL ASSEMBLY

147

Figure 5.5: Top Lip, Top Plate and Rib System

5.8.2

Propulsion System

The propulsion system was assembled entirely separate from the top plate until the nal stages. The support plate was rst bolted to the vertical section of the mounting plate, followed by the placement of the mesh between the support plate and the fan ring. The fan ring was bolted to this support plate again through the use of aluminium L brackets. The motor and propeller were then attached to the vertical surface of the mounting plate and the motor shaft respectively, followed by another piece of mesh and another support plate, attached using L brackets. The mesh was then held in place on the support plate using cable ties to ensure contact with the spinning propeller is avoided. The braces were then simply bolted to both the mounting plate and the fan ring. The sprockets were positioned on their respective shafts and fastened in place using grub screws. The infra-red circuits were then placed inside the tube section. At this point the propulsion system was ready to be mounted to the hull top plate. The three support struts were bolted to the hull at the feet, with 2 bolts holding each foot in place. The stepper motor brackets were also bolted to the hull, with their position measured such that the chain could be both loosened and tightened suciently for both removal and operation.

5.8.3

Payload

The payload system consisted of the marking subsystem and the sensors. In order to facilitate installation of the marking system, a hole slightly larger than the diameter of the paint can was cut in the top plate and base of the hull using a hole saw. A piece of 90mm diameter pipe was then microballooned to the base of the hull to hold it rmy in place, and hence seal the marking system from the air ow

148

CHAPTER 5. MANUFACTURING

inside the hull. A cap was placed on the underside of the pipe, with a slit cut out to allow the paint can to sit rmly in place. Upon reassemly of the hull, a piece of foam was to be inserted between the top of the pipe and the top plate to create an airtight seal around the can. The actuator arm was cut using tin snips from an aluminium sheet, and screwed to the servo motor. An aluminium bracket for the servo motor was constructed from a length of hollow square section aluminium and screwed to the hull. The payload system was installed in the front payload bay in the base of the hull. This was achieved through the positioning of wooded blocks, the same height as that of the sensor, in order to prevent translational motion. These blocks were microballooned to the hull. Once the sensors were in place, another wooden block was positioned across the top of the sensor and screwed into the wooden blocks previously fastened to the hull, to prevent vertical motion of the sensors. A small hole was drilled in the top plate to allow for the data cables from the sensors to be attached to their circuit boards on the top plate.

5.8.4

Skirt

Installation of the lift propulsion and payload systems was now complete, and as such the top plate was again fastened to the hull. The skirt attachment mechanism was then screwed into the top plate to hold it rmly in place. Eyelets were screwed into the outer lip of the base of the hull to allow for the skirt to be attached on the underside. The segments were then clipped into place within the attachment mechanism, and cable tied to the eyelets underneath.

5.8.5

Fairing

The fairing attachment was quite simple in comparison to the assembly of the other systems. The fairing itself sat between the skirt attachment mechanism rmly, such that this was able to hold the fairing in place. Figure 5.6 shows the fully assembled platform.

5.9

Quality Control

A Quality control system was implemented during the manufacturing process to ensure that all components and subsystems were manufactured and assembled to within a specied standard. The quality control system chosen for the duration of this project was the PDCA system, or plan-do-check-act system (Langley, 2009). This system consisted of four main stages, the rst of which was the plan phase, during which all objectives and processes were decided upon in order to deliver the results

5.9. QUALITY CONTROL

149

Figure 5.6: Fully Assembled Platform required as expected outcomes. Secondly, these processes were put into action, initially on smaller components rst. A checking process was then implemented. This consisted of at least two team members physically inspecting the components and making comparisons to the expected outcomes. The obtained outcomes were then analysed in the act phase in order to determine the reasons why these components may not have achieved the desired results, or in the case that the results were suciently close to the desired outcomes, how this process could easily be repeated.

This Page Intentionally Left Blank

Chapter 6 Testing

6.1
6.1.1

Component Testing
Metal Detector Range Testing

Metal detector range analysis tests were carried out to determine the exact performance characteristics the Bounty Hunter Tracker IV 10 inch coil metal detector. Desired performance characteristics included, vertical and horizontal range. Vertical rage testing used a purpose built apparatus, incorporating a hull structure prototype with a variable height adjustment (Figure 6.1). Allowing a simulation landmine to be placed at the bottom of the rig and the signal strength measured over a range of heights.

Figure 6.1: Test apparatus The signal strength was recorded for a variety of heights, at 4 sensitivity levels, as shown in the Figure 6.2. The results indicate a direct relationship between sensitivity and overall range, as expected. The maximum vertical range achieved from the metal detector was 370mm, which is within 3% of the 151

152
V e r t i c a l D e t e c t i o nR a n g ev s . S i g n a l S t r e n g t h

CHAPTER 6. TESTING

Figure 6.2: Metal detector vertical range testing results theoretical detection range of 381 mm. At an operational sensitivity of 75% the maximuum detection range is 300 mm, above the minimum desired 250 mm detection range. Horizontal range tests were carried out using a wooden straight edge placed between the simulation landmine and the metal detector coil, as a guide which the sensor can run along. The straight edge was the moved away from the landmine in increments recording, the signal strength over a range of sensitivities and distances. The signal strength was recorded for a variety of distances, at four sensitivity levels, as shown in Figure 6.3. The graph indicates the displacement from the edge of the coil to the simulation landmine, for the overall detection range, half the coil width must be added, 127 mm.
H o r i z o n t a l D e t e c t i o nR a g n ev s . S i g n a l S t r e n g t h

Figure 6.3: Metal detector range horizontal testing results As with vertical detection range, tests results indicated a direct relationship between sensitivity and

6.1. COMPONENT TESTING

153

detection range. A maximum horizontal detection range of 242mm from the center of the coil was achieved, far exceeding the theoretical 190 mm. Such a large dierence is not surprising due to the simplistic methods used to estimate the theoretical detection range. At an operational sensitivity of 75% a maximum detection range of 212mm was achieved, still above the theoretical value. The overall range detection results correlate well with the theoretical range estimates and are adequate for operation in the hovercraft.

6.1.2

Metal Detector Signal Testing

Signal testing was undertaken to analyse respective detection and interference signals. The test apparatus used simulates the theoretical metal detector mounting positions (Figure 6.4).
I n p u t S i g n a l O u t p u t S i g n a l

M C U 3 8 1mm

P r o c e s s i n gU n i t

2 0 0mm

Figure 6.4: Signal testing apparatus Free testing, with no landmines present, demonstrated the presence of a interference switching signal (Figure 6.5). The switching signal takes the form of a small impulse, with an average period of 200 ms and magnitude of 2.5 Volts. This interference was not expected and is detrimental to the systems performance, creating a high level interference signal. The interference is produced from an electrical charge through the coil upon start up, whereby in essence the coil detects itself. Subsequent tests were undertaken with the presence of a simulation landmine. Figure 6.5 illustrates the positive detection of a simulation landmine, with a period of 400 ms and magnitude of 7 Volts. High frequency scanning tests indicated dead period, of approximately 50 ms, in which the metal detector is non-operational (Figure 6.6). The dead period is caused from the high level discharge of magnetic ux, rendering the receiver unit of the coil inactive for a period of time.

154

CHAPTER 6. TESTING

P o s t i v eD e t e c t i o nS i g n a l

S w i t c h i n gI n t e r f e r e n c e

Figure 6.5: Periodic switching interference and positive signal

L a r g eO b j e c t S ma l l O b j e c t
5 0 ms

Figure 6.6: Positive detection and Object Interference Levels Interference signals from various sized objects were assessed (Figure 6.6), indicating various levels of interference based on the object size. A direct relationship was determined between the magnitude of the interference and the object size and distance from the coil. Interference levels indicate a large resolution between small objects and a positive detection signal, allowing for some level of ltering. However interference from larger objects, has an almost identical signal to that of a landmine, which cannot be ltered.

6.1.3

Landmine Detection Operational Scenario Testing

The inetergrated landmine detection sensors were tested together with the physical marking system under operational conditions. Whereby a number of landmines and decoy objects were placed on a given trajectory and the hovercraft maneuvered over them (Figure 6.7). Landmine detection accuracy was tested by running the hovercraft along the given tracjectory, at operaational speed, a number of times; whilst moving the position landmines relitive to the trajectory.

6.1. COMPONENT TESTING

155

D e c o y L a n d mi n e

T e s t L a n d mi n e s

Figure 6.7: Operational landmine detection scenario

Such that the landmines pass under a dierent point on the hovercraft in 50 mm increments, over the entire width of the hovercraft (Figure 6.8). Resulting in a detection accuracy of 52% over the entire width of the hovercraft and an accuracy of 68% over the specied detection width, 832mm.

Figure 6.8: Detection accuracy over hovercraft width The marking system was tested under restricted operational conditions. The hovercraft was propelled over the landmine on a straight trajectory, at operational speed and the ground marked accordingly, on detection of the landmine. Over the 10 tests ran 7 marking tests were sucessful in detecting the landmine and all were within the specied 1 meter halo radius, for accuracy. Figure 6.9 below shows

156 a number of marks made and their locations relitive to the landmine.
T e s t L a n d mi n e

CHAPTER 6. TESTING

2 5 0mm
( 2 5 0 , 2 0 ) ( 2 2 0 , 5 0 )

5 1 0mm

( 1 4 0 , 5 1 0 ) ( 3 5 0 , 5 8 0 )

Figure 6.9: Landmine physical marking results Overall intergrated sensor and marking system tests were sucessfull, meeting all desired specications.

6.2
6.2.1

System Testing
Lift System Airow Testing

The air ow generated by the lift system to conrm that the design ow rate was achieved while also determining an appropriate operating point for the system. The performance of the lift system was tested under two boundary conditions: 1. 0.05 % tip clearance, 80 mm bleed holes 2. 5 mm fan tip clearance with 15 % radius inlet bell, 80 mm bleed holes The velocity of air exiting each bleed hole was measured using a wind vane anenometer, and the result multiplied by the area of hole to determine the volumetric ow rate. With ow losses due to leakages from the hull as well as recirculation of air around the fan tips assumed to be neglegible, the operating ow rate for the fan is equal to the sum of the ow rates from the bleed holes. Measurements were taken over a range of engine speeds to construct appropriate graphs. These results where then compared with the manufacturers data for similar inlet conditions (no data was available for the inlet bell, only the reduced tip clearance).

6.2. SYSTEM TESTING

157

4 .0 5 % 3 .5 T ip 1 2 3 3 3 E C le a r a n c e 7 % 5 % 1 % 4 % 6 % x p e r im e n ta l

4 .0

3 .5

3 .0

/s )

2 .5

/s ) R a te (m F lo w
3

2 .5

R a te (m

2 .0

F lo w
1 .5

1 .5

5 m m

1 .0

T ip C le a r a n c e 2 3 % 3 8 % 5 4 % 5 6 % E x p e r im e n ta l 2 5 0 0

0 .5 1 5 0 0 2 0 0 0 2 5 0 0

0 .5 1 5 0 0 2 0 0 0

E n g in e S p e e d ( R P M )

Figure 6.10: Volumetric ow rate as a function of RPM and eciency

Figure 6.10 displays the ow experimental ow rates for each entry condition, as well as the manufacters data for various eciecies. Using the initial 5% tip clearance, (Figure 6.10 (left)) an average fan eciency of 34% was obtained. Reducing the tip clearance to 5mm (Figure 6.10 (right)) and installing the bell raise the ecency of the system to an average of 55%. By comparing both the sets of manufacters data in Figure 6.10, the reduction of the tip clearance resulted in little change in ow rate for each eciency level. Given this, the large jump in operating eciency can be attributed to the installation of the bell reducing losses at the inlet. Furthermore, this improvement can also be attributed to the fact that the installation of the bell changes the geometric layout of the system. In particular, the fan is not located along the plane of the hull top plate, and as such recirculation was susspected to be occuring around the edge of the fan. The installation of the bell eectively encased the fan within a duct, which would remove these eects and hence further reduce loses. The change in operating eciency of the system with the improvment of the inlet conditions, eectively reduced the operating point of the system, i.e. the point at which the design ow rate of 1.7 m3 /s is reached, is reduced from approximately 2100 RPM for inlet condition 1, to 1650 RPM for condition 2 (Bell installation). This increases the endurance of the lift system, reduces engine strain, prolongs operational life, reduces vibrations and reduces the strain and wear on the lift system mount.

158

CHAPTER 6. TESTING

Figure 6.11: Thrust Testing Setup

Table 6.1: Propeller thrust output at 75% throttle # Blades Diameter Pitch Power (W) Thrust (kg) 2 13 6.5 940 2.67 2 13 10 1027 1.90 2 13 4 1035 3.00 2 11 7 976 2.33 3 11 7 1002 2.49 2 11 8.5 964 1.94

6.2.2

Propulsion System Testing

Final selection of a suitable propeller for application to the craft was performed on the basis of experimental evalution. The thrust output of a variety of propellers allowing for variations in diameter, pitch and the number of blades was experimentally determined.

Thrust testing was performed on the rig shown in Figure 6.11. The motor was remotely throttled in 12.5 % increments with the generated thrust measured using the load cell. The current and potential of the motor was monitored thoughout the test to ensure that the power levels were kept under 1kW to protect the motor. The output of the load cell is shown in Figure 6.12. A summary of the thrust at 75 % of maximum throttle is shown in Table 6.1.

Testing showed that a 2 bladed 13 diameter, 4 pitch propeller produced 3 kg of thrust at a power level of approximately 1 kW (75 % maximum throttle) and hence two such propellers would be able to produce adequate thrust to propell the hovercraft.

6.2. SYSTEM TESTING

159

3 .0

2 .5

2 .0

T h ru s t (k g )

1 .5

1 .0

0 .5

1 3 " 1 3 " 1 3 " 1 1 " 1 1 " 1 1 "

0 2 0 0 4 0 0 6 0 0 8 0 0

x 6 .5 " x 1 0 " x 4 " x 7 " x 7 " x 3 b la d e x 8 .5 "

1 0 0 0

0 .0

P o w e r (W )

Figure 6.12: Thrust Test Results

6.2.3

Static Hover Testing

Drift of the craft in the absence of applied thrust can arise from a variety of sources. This includes:
Precession of the craft due to the rotation of the lift fan Dierential air ow from the perimeter of the skirt due to irregular skirt geometry and ow

conditions
Non-zero pitch and roll conditions

Such drift can therefore be controlled by using controlling the position of the centre of gravity, as well as by inducing skirt ground contact. A series of static hover tests were designed to deteremine an appropriate trim condition that would minimise the drift of the craft under the static hover scenario. As per the scenario of operation, during static hover, the hovercraft is required to have a maximum, Cartesian, uniaxial displacement of 20 mm over 30 seconds, and to have a change in orientation of under 15 degrees in yaw. Two series of tests were performed, the rst was to determine an appropriate trim condition, at operational engine speed, that would minimise drift. The second was to determine the sensitivity of

160 the craft drift to engine speed using this trim condition.

CHAPTER 6. TESTING

Figure 6.14 illustrates the procedure used to determine an trim condition that would minimise the drift of the hovercraft under the static hover scenario. The trim condition of the hovercraft was manipulated by changing the position of three ballasting weights in the fore section of the craft. These weights were arranged, and their position recorded based on a specic local coordinate system (Figure 6.13). The uni-axial displacement of the craft was measured along a two dimensional, global co-ordinate system in line with the lateral and transverse axses of the craft in its initial condition, with their origin set at the perimeter of the initial position. Change in orientation was measured, using a digital compass, as the change in orientation of the transverse axis with respect to magnetic north. All tests were performed inside, on a at, level surface away from signicant external disturbances.

Trim Point

The hovercraft was tested under four dierent trim conditions, for 30 seconds, until the drag requirements were met, and the results are summarised in Table 6.2. The trim condition which minimised drift required a slight wight imbalance along the longnitudinal axis such that the rear skirt was dragging. This was was desirable in that the friction of the skirt with the ground allowed the hovercraft to maintain a steady position during static hover, and would provide some compensation for the downwards pitch angle induced by forward thrust. The nal trim condition (as displayed in Figure 6.14) was set as the trim condition for subsequent testing and operation of the hovercraft.

Figure 6.13: Local Coordinate System for Static Hover Test

6.2. SYSTEM TESTING

161

Figure 6.14: Static Hover Test Arrangement Table 6.2: Drift Results for Trim Conditions Displacement (mm) Test Weight Number 1 2 3 1 2 3 1 2 3 1 2 3 Mass (kg) 1.5 3 2.5 1.5 3 2.5 1.5 3 1.5 1.5 3 1.5 Location on Grid (x,y) 3,4 5,5 7,5 3,4 5,4 7,4 3,4 5,5 7,4 3,4 5,5 7,4 RPM Longnitudinal Transverse Change in Orientation (deg) 73

1843

1000

1857

1870

1859

<10

Drift as a function of Engine Speed

This aim of this test was to determine the sensivity of drift to the lift engine speed. The nal trim point obtained in the previous test was used, while the engine speed was altered. Engine speed was measured using a tachometer, the uniaxial drift and orientation was measured as in the preceeding test. The results of this test are summarised in Table 6.3, and the drift is shown graphically in Figure 6.15. These graph illustrated two distinct regions, a relatively stable region between 1150 RPM and 1880 RPM, where the drift of the craft displays little variation with engine speed, and a region of every increasign instability (above 1880 RPM), where stationary hover is unsustainable. At this point, the air ow rate is such that it provides too much lubribation to the platform, such that the drag of the skirt on the ground is insucent to maintain a steady position. From these results, the rst region (i.e. 1150 - 1800 RPM) was identied as a stable operating range of the lift engine such that

162 the hovercraft may maintain a stationary hover position.

CHAPTER 6. TESTING

Table 6.3: Drift response to lift engine speed during static hover Displacement (mm) RPM Longnitudinal Lateral Yaw (Deg) 1100 0 0 NDC 1520 5 2 0 1718 5 3 0 1813 10 3 0 1827 20 10 0 2200 150 150 > 15 2600 150 150 > 15

Figure 6.15: Variation is drift due to lift engine speed

6.3

Longnitudinal and Transverse Behaviour Testing

The design of the hovercraft allows for no control input for either pitch and roll. As such, the stability of the hovercraft in operation must arise passively in response to disturbances along the longnitudinal and transverse axes. To ensure static stability in both pitch and yaw, the hovercraft was tested to ensure the ability to recover from the maximum possible roll and pitch angles that are permitted by the geometric and operating parameters of the craft. To test this response, the hovercraft was manually pitched and rolled by displacing the fore, aft, port and starboard edges of the craft until the hull made contact with the ground. The hovercraft was then released and allowed to correct itself. The pitch and roll response arising from this displacement was sampled every 0.01 s using an inclinometer. To reduce the high frequency noise of the response that resulted from engine vibration, the output was smoothed

6.4. PERFORMANCE TESTING

163

using the Savitsky-Golay method. The smoothed and normalised output of the inclinometer is shown in Figures 6.16 and 6.16 along the longnitudinal and transverse axes respectively.

Figure 6.16: Longintudinal and Transverse Response These gues indicate that the hovercraft is capable of producing a restoring moment sucent to return the hovercraft to the equilibrium position given a maximum possible initial disturbance. The hovercraft was shown to settle in the longnitudinal direction within 1.5 s and in the transverse direction within 5 seconds. There was little overshoot in the response, and minimal ossillatory behavour. Given that the hovercraft can correct for the maximum possible initial disturbance along the hovercraft is statically stable in both pitch and roll.

6.4

Performance Testing

The overall platfrom was subjected to a battery of tests while where designed to assess its performance charateristics inclusing:
Acceleration Deceleration Climb Gradient

All performance testing was performed with a trimmed hovercraft and at half payload weight (20 kg).

6.4.1

Acceleration Testing

Two independant tests where designed to assess the acceleration charateristics of the craft. The rst aimed to provide a low resolution analysis of the maximum accleration capabilities of the craft. The craft was held in place while the propulsion engines were throttled to a contant level for the test, then

164

CHAPTER 6. TESTING

released. The test was performed indoors, on a smooth, level concrete surface aways from weather disturbances such as wind or precipitation. Minimal control inputs including dierential thrust and vectored thrust were used to maintain a linear course. Stopwatchs were used to measure the time taken to pass 5 m checkpoints,up to a total length of 15 m. Hence, basic linear equations of motion could be used to determine the average velocity and acceleration of the motion. The results of this test are summarised in Table 6.4. Note that the throttle has been scaled by a factor of 0.6 to limit to maximum power level of the propulsion motors. Table 6.4: Split (s) 1 2 3 5.9 1.3 2.4 4.7 2.5 1.8 3.9 1.2 2.7 Acceleration Weight Velocity m/s 1 2 3 0.847 3.846 2.083 1.064 2.000 2.778 1.282 4.167 1.852

Throttle 80% 100% 100%

Time(s) 1 2 3 5.9 7.2 9.6 4.7 7.2 9 3.9 5.1 7.8

Acceleration m/s2 1 2 3 Average 0.144 2.307 -0.735 0.163 0.226 0.374 0.432 0.185 0.329 2.404 -0.857 0.247

Figure 6.17 displays the average acceleration of the tests accross the 5 m intervals. The theoretical values for the acceleration over this range are summarized in Table 6.5, allowing for 3.5, 2.9 and 1.275 N of aerodynamic, cushion and skirt drag respectively. The average theoretical quantities where found to be 40-60% under the theoretical values. This deviation can potentially be accounted for by deviation from the linear path, reduction in straight line acceleration due to course corrections and additional drag than that proposed due to dynamic pitching of the craft. These results indicate that the hovercraft marginally exceed the minimum acceleration requirement of 0.14 m/s2 for this environment. However, it is unlikely that it will achieve this result in more challenging environments as the increased drag will sharply increase the amount of the force required to overcome drag. Despite this, raising the scaling factor for the propulsion system to approximately 80% would make far more thrust available, close to the maximum 3 kg available per propeller, which the engines still running at approximately 1 kW of the 1.4 kW power rating for the engines. While this would lead to reduced life for the engines and reduced endurance, this reserve thrust could be used to overcome initial resistance, or small variations in slope for applicaiton to more challenging environments. Table 6.5: Theoretical Acceleration Thottle 80% 100% Scaled Throttle 0.48% 60% Thrust per engine 16N 24N Drag 7.675N 7.675N Net Thrust 24.325N 40.325N Acceleration 0.286m/s2 0.416m/s2

6.4. PERFORMANCE TESTING

165

2 .5 2 .0 1 .5
8 0 % 1 0 0 % 1 0 0 % (1 ) (2 )

A c c e le r a tio n ( m /s )

1 .0 0 .5 0 .0 -0 .5 -1 .0 0 2 4 6 8 1 0 1 2 1 4 1 6

D is p la c e m e n t ( m )

Figure 6.17: Acceleration over 5 m intervals

6.4.2

Deceleration

The deceleration performance of the platform was tested in a similar fashion to the acceleration testing. Testing in the same environment, the hovercraft was manually accelerated to a set velocity and released. Upon release, the propulsion systems where activated at, at full throttle (100% scaled by 0.6), in the full reverse thrust conguration (180 vectored thrust). The initial velocity of the craft was determined by measuring the time taken for the platform to cover 5 m of gound, then measure the time to stop, and the linear displacement to stop. Again, the entire process was approximated using linar motion, will small corrections to course accounted for using dierential and vectored thrust. The deceleration results for three test runs are summarised in Table 6.6. Table 6.6: Deceleration performance of the hovercraft Entry Speed Stopping Distance Stopping Time Deceleration (m/s) (m) (s) (m/s2 ) 2.5 4.5 3.2 0.44 2.4 5 2.10 1.13 2.5 5 2.10 1.13 2.47 4.83 2.47 0.90

1 2 3 Average

The net theoretical deceleration can adapted from Table 6.5, however the full thrust of fans in available for deceleration, and is enhaunced by the drag, giving a theoretical deceleration of 0.364 m/s2 . The actual decelertion of the craft is approximately 2.5 times this value, which can be attributed to underestimation of the drag, and increased drag due to the pitching of the craft arising from the the

166

CHAPTER 6. TESTING

thrust force. This experimental data meets the minimum project requirement for deceleration of 0.14 m/s2 . As this is a relatively ideal environemnt, it can be claimed, with a high degree of condence, that is requirement will be meet in more challenging environment due to the increased drag in associated from these environments. The only case in which this requirement may not be met is in the presence of negative slope.

6.4.3

Additional Acceleration Performance Testing

Additional acceleration performance testing was attempted using 2 dimensional data sampling, in the aim of producing higher resolution and potentially more accurate results, as well as to test the performance of the craft at throttles ranging from 37.5-80% range. This testing was to be performed outside, on a at level surface, with the response of the hovercraft to be measured using the GPS and logged by the ground station. Despite performing several successful test runs, the accuracy of the GPS was insucient to ensure accurate results, with accelerations that would often be an order of magnitude higher than theoretically possible. As such, this method of testing was abandoned.

6.4.4

Climb Gradient Tesing

The ability of the craft to transverse sloped terrain was tested on a large inclinded plane. The plane was raised to a set angle, and the hovercraft set on it and held in place with ropes. Once the propulsion system was activated and increased to full throttle, if the ropes went slack showing that the plaform could maintain its own weight on the slope, a positive response was recorded. The incline plane was raised from 0 degrees in half degree increments. The hovercraft was found to be able to maintain its position on a maximum slope of 2.5 . Theoretically, the hovercraft haveing a weight of 87 kilograms (50% payload) and having a maximum thrust of 48N (60% scaling), should be able to transverse a slope of approximately 3 , which is in range of the theoretical limit given the resolution of the sclope increments.

6.4.5

Additional Functionality Testing

Endurance Endurance testing was done in conjunction with the performance testing of the hovercraft. As per the project specications the hovercraft was required to have a minimum endurance of 10 minutes. Before the commencement of performance testing, the propulsion batteries were completely charged,

6.4. PERFORMANCE TESTING

167

as where the stepper motor batteries. In addition the lift engines full tank was lled. The time that the hovercraft was in continuous operation, that is the time for which the hovercraft was in motion, was noted. The tests ended with the depletion of the propulsion batteries, as hence these where identied as the limiting factor for endurance. The sum of the time for which the hovercraft was in motion, and powered under its own propulsion, was approximately 12 minutes 30 sections.

6.4.6

Turning circle

As a side to propulsion testing, the turning radius of the hovercraft was assessed in relation to the minimum requirements of two times the craft length, or 3.2m. To test this capability, a 3.2 metre radius was marked around the perimeter of the craft, and the thrust angle of the propulsion systems set at 90 . Upon throttling of the propulsion system, the platform was required to complete a 180 rotation and remain within the set boundary. The hovercraft was able to meets these requirements in the clockwise direction is all tests, however was never able to achieve this for counterclockwise motion. When turning in this direction, the hovercraft would undergo surge motion and tend to move side wise until is exceeded the specied operational area. This tendency is most probably resultant from the precession of the fan. While a comprehensive set of turning radius tests were planned, including rotation rates and turning radii for dierent thrust values, they relied on the GPS to log the motion of the hovercraft for later analysis. Unfortunately, the accuracy of the GPS feel outside of acceptable levels.

This Page Intentionally Left Blank

Chapter 7 Project Management and Finance

7.1

Practical Project Management Methodology

Project management methodology was dictated by the student environment within which the group was required to operate. Characteristic to the university environment is limited and uncertain access to resources and other group members due to highly variable timetables, and the need balance them with the part-time work and home commitments. These constraints dictated that the working environment and time had to be exible, and a traditional management methodology involving set working hours and positions under direct supervision was not appropriate. Rather, a methodology centered on individual work with emphasis of individual roles and responsibilities was adopted and evolved as the project progressed.

7.2

Resource Management

In keeping with the aforementioned project management methodology, individuals were assigned various managerial portfolios. Each portfolio entailed a series of corresponding responsibilities and roles with related tasks as outlined in the following sections.

Mechanical/Technical Manager

The Mechanical Manager has overarching responsibility in all mechanical aspects of design, as well as their integration with the Mechatronic counterparts. They are responsible for mechanical design integration. In addition to this role, the Mechanical Manager is the Technical Manger of the group. The Technical Manager is responsible for the day to day management of the projects human resources. They are responsible to ensure that project milestones and deadlines are delivered on time by. Tasks include running ocial meetings, taking and distributing minutes, as well as assigning and reviewing tasks. 169

170

CHAPTER 7. PROJECT MANAGEMENT AND FINANCE

In addition, to their short-term managerial responsibilities, the technical manager is responsible for the harmonious operation of the group as a whole, and is the rst point of contact for any conicts that may arise.

Mechatronics/Assistant Technical Manager The Mechatronics Manager is required to oversee the design and integration of all Mechatronic aspects of the project. As the Mechatronics Manger, they also have an Assistant Technical Managerial role, whereby they are required to assign and review Mechatronic related tasks. In the absence of the Technical Manager, they assume leadership of the group and all corresponding roles and responsibilities.

Drafting Manager The drafting manager is the primary design drafter of the group. They are the rst point of contact for all CAD issues. They oversee all CAD models and drafts produced by the group and are charged with quality control of all technical drawings produced.

Manufacturing Manager The Manufacturing Manager oversees all product manufacture. They are required to work closely with the workshop sta and/or external workshops and suppliers such that manufacturing is performed on time and to a satisfactory standard. They provide manufacturing advice upon request and work with designers to simplify and standardise manufacturing design aspects when appropriate.

Test Manager The test manager designs component and system evaluation procedures to ensure that purchased projects achieve their required level of performance and reliability, and that integrated systems operate eectively. In addition, they are responsible for designing eective means by which the goals of the project denition may be measured.

Safety Ocer The Safety Ocer has the role of ensuring the health and welfare of the group and others which may be aected by the construction, testing or operation of the nal platform. They are required to get approval for manufacture and testing of the platform by assessing and where possible, mitigating any

7.2. RESOURCE MANAGEMENT

171

safety risks related to the manufacture, testing or operation of the hovercraft. They should be present at all tests to police prescribed safety measures.

Treasurer

The Treasurer manages the scal resources of the project by keeping detailed records of all sources of income and expenditure.

7.2.1

Organisational Hierarchy

The general organisational hierarchy is shown in the organisational chart (Figure 7.1).

Figure 7.1: Organisational Chart

7.2.2

Work Distribution

In addition to their overarching managerial responsibilities, group members were assigned particular topics such as Lift System, Skirt and Autonomy, that were primary responsibility. This responsibility entailed research, design and acquisition or manufacture of relevant items, as well as the testing and evaluation. While such topics were the responsibility of a single person, from time to time the workload would require the allocation of additional resources, and responsibility of managing these resources would ultimately fall under the jurisdiction of the topic leader. The technical manager would assign relevant tasks and deadlines to all group members upon consultation with the supervisor and group members. These task were tracked using a master list, and were only deemed complete upon the production and submission of related documentation for review.

172

CHAPTER 7. PROJECT MANAGEMENT AND FINANCE

7.3

Review Process

Upon its completion, documentation of work was submitted to the technical manager, for review. The technical manager was responsible for the review of all work, and either reviewing it themselves or designated another group member to perform the review. The review process itself was two-fold. The rst review was technical, and assessed the content of the work for validity, accuracy and completeness. The second review process was only invoked if the work was required for ocial project documentation. This review was language and format orientated, assessing the structure, grammar, spelling and format of the submitted work. Upon completion of the technical review, the document was returned with comments to the author for editing before submission for proong. Should the initial technical review require that signicant changes be made, another technical review may be recommended by the reviewer. In the event that the initial technical review concluded that there was no need of technical modications, it was immediately proofed and returned for further modication (linguistic, not technical), and further proong, else for nal submission to the technical manager.

7.4
7.4.1

Time Management
Deadline Specication

The overall progression of the project was measure relative to both external and internal milestones. External Milestones in Appendix D list the major external milestones that centered around the project deliverables. Table in Appendix E lists the major internal milestones the were dened early on to ensure that deliverables are met on time, and were tracked using a Gantt Chart.

7.5

Risk Management

Due to the nature of this project it was considered to be relatively high in risk and as such the guidlines put forth by the university on risk identication and management were followed closely. The scope of risks could be classied into three main categories. Operational risks largely adresses risks associated with the operation of the platform and testing. This included risks such as harm to bystanders through failure during testing. Platform risks were risks associated with the platform and the ability to complete the project. This risk category incorporated risks involved with sourcing

7.5. RISK MANAGEMENT

173

components and manufacture. The third chategory was general safety risks. Each risk category was managed using tailored mitigation strategies. For operational and safety risks, detailed processes on risk management were outlined by the school and followed closely. With regards to platform risks, no formal management structure was given, and so a general project management approach to risk management was undertaken. Operational risks included all risks associated with testing and operation of the platform. The standard university procedure for management of such risks was the development of a Safe Operating Procedure (SOP) which considered the associated risks in advance, by developing a procedure for operation and testing. The full SOP is detailed in Appendix K of the craft. It was highlighted that the main operational risk was the failure of a component due to insucient maintenance and checking prior to initiation. In order to minimised this eect a more detailed checklist was developed to ensure all high risk components were checked prior to each start up routine. This checklist is also detailed in Appendix K, and highlights tasks such as ensuring grub screw are tight, and moving parts are fastened correctly. Platform risks inluded risks to the platform itself, through component failure or poor design, and risks threatening completion of the project to the deadlines required. This incorporated risks such as the arrival of components from interstate or overseas, unexpected delays to manufacture and integration issues. A risk register compiling the main risks to the the project was created which identied the risk level of each risk and suitable mitigation strategies to prevent its occurence. Included in these risks was the failure of fast rotating components, such as the propulsion propellers, which was mitigated through constant checks and testing (in isolation) prior to operation. Other risks, such as delays in component deliveries were identied early on, and critical components were ordered promptly to ensure contingency if delays should occur. Manufacturing complications were anticipated based on the problems faced by previous projects undertaking composite manufacture, and thus building was started in May to ensure time was available for mistakes and proper application of processes. A proactive risk management process was undertaken for platform risks to ensure any problems could be identied early and anticipated. The nal category of risks was safety and incorporated all risks which posed a threat to operater or bystander safety. General university guidelines were followed with regard to the management of these risks, with the SOP and risk assessment being developed in conjunction with the project supervisor and Mechanical Engineering school safety ocer. The risk assessment for Project Hover is detailed in Appendix K and identies the key safety risks and issues associated with operation of the vehicle. The document further identies strategies that must be undertake to reduce the impact of these risks

174 should they occur.

CHAPTER 7. PROJECT MANAGEMENT AND FINANCE

7.6

Finance

Final year projects undertaken in 2009 were allocated $200 per student from the School of Mechanical Engineering. This monetary assistance was accompanied by the labour of the Mechanical Engineering Workshop. Hours spent on projects by the sta of the workshop were charged at a rate of $50 per hour, however at the conclusion of the project, information regarding the total workshop ours spent was unavailable to us. The following section will detail the total funds available to the project, as well as outlining the total cost and time spent developing the project.

7.6.1

Sponsorship

It was clear upon commencement of the project that a total funding of $1,200 as supplied by the School of Mechanical Engineering was insucient in order to complete a project, and as such additional support was saught. A series of contacts were initiated by several of the group members in an attempt at securing further monetary assistance. In total, over 15 businesses and institutions were contacted through either email, telephone or in person. As a result of these endeavours, the project was able to secure a total of $13,000 in additional nancial support from three contributors; The Defence, Science and Technology Organisation (DSTO), AMCOR and Soil Testing Services (STS). Table 7.1 below outlines the support received from each of these contributors.

7.6.2

Financial Position

As of the conclusion of the project; including all sourced funds, in-kind university support, and total project expenditure, the following table details the nal nancial position of the project. Table 7.1: Financial Position Item Contribution DSTO $10,000 AMCOR $2,000 STS $1,000 School of Mechanical Engineering $1,200 Expenditure -$10,495.4 Current Standing $3704.6 As shown, this project has successfully completed all processes while keeping to a strict budget. Below in Figure 7.2 is a detailed outline of the total expenditure on each system, as well as a monthly

7.6. FINANCE breakdown of expenditure also shown in Figure 7.2.

175

Figure 7.2: Expenditure and Monthly Breakdown

7.6.3

Labour Costs

In order to assess the total labour costs associated with this project, each team member recorded the number of hours spent on the project on a daily basis. A detailed monthly breakdown of the hours spent per student can be found in Appendix P. The total labour cost for the project are shown on a monthly basis below in Figure 7.3.

Figure 7.3: Monthly Breakdown of Hours Spent To obtain a realistic gure for the total cost of the project including labour hours, a rule of thumb of
$26 an hr was applied to every hour of labour time spent. Also, an additional 130% of this value was

added as an estimate for indirect costs such as internet, phone, and technical support, and a further 30% as an estimate for direct costs associated with superannuation, workcover and alike. This resulted in a total project cost of $435,583.64. Appendix W again details the breakdown of this gure.

This Page Intentionally Left Blank

Chapter 8 Conclusions
The chapter summarises the project overcomes in relation to the project objectives that were outlined in Section 1.2. This section concludes with recommendations for the future direction that the project should take.

8.1

Primary Project Objectives

1. To design and manufacture a remotely controlled hovercraft

An operational hovercraft has been designed and built, with control supplied by remote. The platform has demonstrated through extensive testing multi-terrain capability, with optimal performance recorded on predened operational terrain. Performance parameters including a acceleration and deceleration have exceeded their design values under controlled conditions, and the platforms specic vectored thrust design has allowed for precision control in rotation; demonstrated by the crafts ability to turn on the spot, limited only the pre-cessive eects of the main lift fan. Nevertheless, in its current state the platform is capable of completing its design mission prole and is fully compliant with all specications outlined under the projects Standard Goals. Furthermore, the demonstrated endurance of the craft has been documented at least 12 minutes. This project is considered on target. The performance requirements are not so overly superior to the design specications, and the inability to extend to non-ideal operational environments prevents this objective from exceeding targets. There is no outstanding work to be done for meeting this objective. 2. To provide adequate control and stability to the hovercraft such that it can hover while stationary.

Control of the static trim of the craft was identies in the design process as the means by which the hovercraft may maintain a constant position with an activated lift system. Stationary hover 177

178

CHAPTER 8. CONCLUSIONS testing identied an appropriate weight and ballasting point for the hovercraft that was limited the uniaxial drift of the craft to under 10 mm with a minimal orientation change over 30 seconds, which is well within the two metre halo limit that was specied across this time period. This objective has been met, with no additional work required. However, this requirement is highly sensitive to the ballasting of the platform and will require review in the event of structural alterations to the platform that may result in the future.

3. To manufacture a hovercraft to accommodate an o-the-shelf metal detector array

Two o the shelf metal detectors were selected and successfully arranged in an array within the craft. Despite interference between the metal detectors, a solution that involved operational switching of the coils allowed for a demonstrated detection eciency of the payload system of 60% accross the width of the craft. In addition, there where no measured false positives that arose from the physical structure of the craft due to the limitation of metallic components in the vicinitiy of the payload bay. This objective has been achieved, and providing no additional metallic components are installed within the fore section of the craft, shouldnt require review. 4. Hovercraft must be able to identify and physically mark the location of mines.

A physical marking system that leaves a semi-permanent line of high visibility paint in response to a positive output from the sensor payload. This feature of the craft has been tested during functionality craft of the combined sensor and marking system, in a controlled environment. This test produced marks within 0.6 m of the test mine during all seven tests that exhibited a positive detection signal. This is well within one metre halo radius requirement. While the capability of the physical marking system has been demonstrated, future work should expand on this objective by testing the hovercraft under its entire operational scenario, that is in the eld and with many test mines at various positions. All primary objectives have been met in accordance with their minimal requirements, however if future work is undertaken all of these objective should be re-evaluated at the design stage of the modications, with the aim of maintaining these minimum standards in accordance with the requirements of the application. In addition, several objective should be conrmed under more taxing conditions, such as those aimed for as part of the extended goals.

8.1. PRIMARY PROJECT OBJECTIVES

179

8.1.1

Extended Project Objectives

1. The hovercraft will autonomously follow a predened path over at terrain with no obstacles

Two stages were done to achieve this goal. First assisted autonomous control was designed, where yaw angle is controlled, but thrust magnitude is controlled by an operator. To do this active yaw control was implemented, with allowance for a setpoint input, allowing the hovercraft to move to a desired angle. The autopilot is used to plot way points and then navigate the hovercraft between these points, by controlling the angle that the hovercraft is moved to. After this the thrust motors could be controlled from the autopilot which would allow completely autonomous control. Future work would be to improve the active yaw control to be more robust, possibly by implementing a digital compass to have a better estimate of the yaw angle. In addition, ight testing of the hovercraft under the way point navigation ight regime should be attempted and the outcomes used to iterate the design of this system for another design pass. 2. GPS tracking of the hovercraft combined with graphical user interface To achieve this goal a commercial autopilot was purchased, and then integrated into the hovercraft systems. The autopilot comes with a ground control station software, which allows live GPS tracking of the hovercraft. No further design is necessary for future work as this system is implemented. 3. Electronically mark the position of mines Two metal detection sensors output their signals into the on board microcontroller, which allows successful metal detection. This system once integrated with the autopilot will allow the electronic GPS marking of mines. Future work would implement this integration, using a communication protocol to provide this mechanism. Once this is completed the ground control station software would then be congured to overlay the found positions of mines over a map. 4. Operation in non-ideal environments

Operation of the hovercraft in non-ideal environments will be severely limited. Test of the hovercraft found that the minimum specications of the hovercraft have been met at the upper

180

CHAPTER 8. CONCLUSIONS end of the hovercrafts operational envelope under controlled conditions. Of particular limitation is the propulsion system, which being limited to a 2.5 incline will be unlikely to cope in non-ideal environments, particularly in the presence of more drag which will arise from uneven surfaces and vegetation. Furthermore, have currently been designed at marginal levels, and will be aected by moments that will arise due to the increased drag of these environments. In order to achieve operation in advanced environment, the propulsion capabilities will need to be signicantly augmented, perhaps with the replacement of propellers or by the addition of small ducted fans at the rear of the hovercraft. In addition, the weight of the craft could be signicantly reduced. By sacricing the ease of modication of hull top plate, a replacement composite top plate, minus the large area of plywood insert, could be manufactured. In addition, the lift engine is overspecied and could be replaced with a smaller, lighter one. Finally, the metal detectors applied here do not require the full 20kg payload capacity. A narrowing of the planform at the front of the craft, or an extension at the rear could shift the centre of pressure of the craft backwards, more incline with the centre of gravity and hence minimise the ballasting weight of the craft.

8.2

Concluding Remarks

To date, the authors of this report has successfully designed and manufactured an operational hovercraft. Remote control of this hovercraft has been achieved and a comprehensive set of performance tests implemented, conrming the design specication under ideal operating conditions. Metal detectors has successfully be integrated and tested on board the platform, as well as being combined with a physical marking system that reacts to positive sensor responses. While the autonomous functionality of the hovercraft, including active yaw control and way point navigation, hasnt reached the nal stage of testing and troubleshooting due to time constraints, the extensive design eort applied to these systems provides a solid foundation for further work in this area. In conclusion, the product prototype shows signicant promise for application to mine detection in the eld, however signicant work is still required to extend the operational environment of the platform from controlled conditions to the eld, and to establish autonomy. Despite the limitations of the craft, this prototype lends signicant support to the concept of an autonomous mine detecting hovercraft.

Bibliography
ABC-Ventilation. Inlet/outlet accessories, 2009. [Online] available at www.abcventilation.com/en/ fans_options.shtml [Accessed: 23/10/09]. Paul Acarnley. Stepping Motors a guide to theory and practice. The Institution of Electrical Engineers, London, United Kingdom, 2002. Hovercraft Air-Commander. Air commander hovercraft, 2009. [Online] available at http://www. aircommander.com/craftmodels/ac4hovercraft.html [Accessed: 5/10/09]. A.K. Amiruddin, S.M. Sapuan, and A.A. Jaafar. Analysis of glass bre reinforced epoxy composite hovercraft hull base. Materials and Design, 29:14531458, 2008. Hovercraft Amphibious. Amphibious marine - vanguard, 2009. [Online] available at http://www. amphibiousmarine.com/ [Accessed: 24/10/09]. J R Amyot. Hovercraft Technology, Economics and Applications. Elsevier Science Publishers, New York, 1989. Garner W B. Model airplane propellers. Columbian Radio Control Club, 2009. Frank P. Bleier. Fan Handbook: Selection, Application, and Design. Mcgraw-Hill, 1997. A Bliault and L Yun. Theory and design of air cushion craft. Arnold, London, 2000. Ian Brooks. A method for estimating the thrust and lift performance of an integrated hovercraft. Technical report, Hovercraft Club of Great Britain, March 2005. Claudio Bruschini and Bertrand Gros. A survey of research on sensor technology for landmine detection. The Journal of Humanitarian Demining, 2, 1998. Margaret S. Buse. What you should know about landmine victims, 1999. [Online] available at http: //maic.jmu.edu/JOURNAL/3.3/focus/how_many_victims.htm [Accessed: 2/10/09]. W D Jr Callister. Materials Science and Engineering an Introduction. Wiley and Sons, Inc., USA, 2003. CMG-Engineering. Hga motors catalogue, 2009. [Online] available at http://www.cmggroup.com. au/pdfs/Motors/HGA/HGA_Cat.pdf [Accessed: 24/10/09]. 181

182

BIBLIOGRAPHY

Crane-Co. Flow of uids through valves, ttings, and pipe. John Wiley & Sons, Inc., 1957. Crane Valves. Y Das, J.E McFee, and G Cross. Soil properties database for humanitarian demining: A proposed initiative. Technical report, International Union of Soil Science, 2002. Y. Das, J. E. McFee, K. Russell, G. Cross, and T. J. Katsube. Soil information requirements for humanitarian demining: The case for a soil properties database. Technical report, Terrascan Geophysics, Vancouver BC, Canada, 2004. J.L. Davis and A.P. Annan. Ground penetrating radar for high-resolution mapping of soil and rock stratigraphy. Geophysical Prospecting, 37:531551, 1989. Mike Wilson Dr. Ivan Hrvoic, Greg M. Hollyer. A new very high sensitivity potassium magnetometer for near surface geophysical mapping. Technical report, York University, Toronto, 2002. Macmillan Engineering Evaluations. Mechanical Power Tranmission. The Macmillan Press Limited, 1971. C Fitzgerald and R Wilson. Light Hovercraft Design. Hoverclub of America, USA, 1995. P Fitzpatrick. Principles of hovercraft design. Technical report, Hovercraft Club of Great Britain, Britain, 2009. MALA Geoscience-AB. MALA Shielded Antennas, 2009. C. P. Gooneratne, S. C. Mukhopahyay, and G. Sen Gupta. A review of sensing technologies for landmine detection: Unmanned vehicle based approach. In 2nd International Conference on Autonomous Robots and Agents, Massey University, New Zealand, December 2004. Institute of Information Sciences and Technology. Terence Arthur Graham. Pitch-heave dynamics models for an air cushion vehicle. Technical report, University of Toronto. Institute for Aerospace Studies, 1992. Steve Grin and Timothy Pippett. Ground penetrating radar. Technical report, Geophysical Technology Limited and Alpha Geoscience Pty. Limited, 2001. Erika Gudmundson. Nuclear quadrupole resonance, 2008. [Online] available at http://www.it.uu. se/research/syscon/signalprocessing/topics/NQR [Accessed: 14/9/09]. motorcycles Honda and power equipment. Technical Data for a GSV190 Four-Stroke Engine, 2009.

BIBLIOGRAPHY

183

Hovercraft Hovertechnics. Hovertechnics hovercraft - hovercraft for recreation and commuting, 2009. [Online] available at http://www.hovertechnics.com/recreation/hoverjetfeatures.htm [Accessed: 26/10/09]. Inc HowStuWorks. How landmines work, 2009. [Online] available at http://science.

howstuffworks.com/landmine.htm [Accessed: 12/8/09]. Alonzo Kelly, Anthony (Tony) Stentz, Omead Amidi, and Randon Warner. The international journal of robotics research. Toward Reliable O Road Autonomous Vehicles Operating in Challenging Environments, 25(1):449483, May 2006. Dr. Andrei Kotousov. Stress analysis and design. Course Notes for Stress Analysis and Design of Mechanical Engineering at the University of Adelaide, 2007. G J Langley. The Improvement Guide: A Practical Approach to Enhancing Organisational Performance. Jossey-Bass, 2009. Users Manual for M325. Leadshine Technology Co., Ltd., 2000. Micro Microstepping Driver. Rommel E. Mandapat. Development and evaluation of positioning systems for autonomous vehicle navigation. Masters thesis, University of Florida, December 2001. B W McCormick. Aerodynamics, Aeronautics and Flight Mechanics, 2nd Edition. Wiley, USA, 1979. R McLeavy and J Wood. Hovercraft and hydrofoils in colour. Arco Publishing Co., USA, 1977. Geocan 3D Media. Introduction to ground penetrating radar, 2009. [Online] available at http: //www.geoscan3d.nl/gpr_intro.html [Accessed: 3/10/09]. INC Microbotics. Microbot autopilot platform for unmanned vehicle, 2008. [Online] available at http://microboticsinc.com/rc_uav_autopilot.html [Accessed: 15/04/09]. ModelFlight. Jr receiver, 2009. [Online] available at http://modelflight.com.au/jr_rc/jr_rs77s_ receiver.htm [Accessed: 30/11/09]. Moore-Fans. Theory of rating, 2009. [Online] available at http://www.moorefans.com/TMC_656P_ .PDF [Accessed: 24/10/09]. International Multi-Wing. Technical Fan Information, 2009. B R Munson, D F Young, and T H Okiishi. Fundamentals of Fluid Mechanics, Fourth Edition. John Wiley and Sons, USA, 2002.

184

BIBLIOGRAPHY

Fox News. Deadly conventional weapon of the day: Anti-personnel mines, 2009. [Online] available at http://science.howstuffworks.com/landmine.htmhttp://commonsenselogic.blogspot.com/ 2009/01/deadly-conventional-weapon-of-day-anti.html [Accessed: 3/10/09]. Ocean-Controls. Ocean control stepper motors, 2009. [Online] available at http: [Accessed:

//www.oceancontrols.com.au/motors/stepper/hp_bipolar_stepper_motors.htm 10/9/09].

United States General Accounting Oce. Dods research program needs a comprehensive evaluation strategy, 2001. Pacic-Hovercraft. Pacic hovercraft - slider gallery, 2009. [Online] available at http://www.

hovercraft.co.nz/slider/gallery.htm [Accessed: 14/09/09]. Kestrel Autopilot System. Procerus Technologies, 2008. Description and Information of the Product. Applied Engineering Products. Unexploded ordinance removal, 2007. [Online] available at http: //www.aepcoinc.com [Accessed: 15/8/09]. Paparazzi Project. Welcome to paparazzi, 2008. [Online] available at http://paparazzi.enac.fr/ wiki/Main_Page [Accessed: 10/03/09]. qsine. Military land mine detection vehicle, 2005. [Online] available at http://qsine.ca/?page= InProjects/IldsRdv [Accessed: 24/10/09]. Stephan A Roth and Parag Batavia. Evaluating path tracker performance for outdoor mobile robots. In Automation Technology for O-Road Equipment, July 2002. Wayne Schmidts. Metal detecting, 2006. [Online] available at http://www.waynesthisandthat. com/metaldetecting.html [Accessed: 23/9/09]. Electronics Sparkfun. Ardupilot - arduino compatible uav controller w/ atmega328, 2008. [Online] available at http://www.sparkfun.com/commerce/product_info.php?products_id=8785 [Accessed: 16/04/09]. Sue Stull. Making landmine detection and removal practical. Science and Technology Review, November 97:1820, 1997. P. A. Sullivan, M. J. Hinchey, and R. G. Delaney. Static roll stiness characteristics of two multicell type air cushion systems. Journal of Terramechanics, 15(1):1541, 1978.

BIBLIOGRAPHY

185

James Trevelyan. Metal detector description, 2008. [Online] available at http://school.mech.uwa. edu.au/~jamest/demining/tech/detect/md-intro/text.html [Accessed: 3/10/09]. Unicef. International guidelines for landmine and unexploded ordnance awareness education. Technical report, Unicef, New York, 2000. R. Allan Wallis. Axial Flow Fans and Ducts. John Wiley & Sons, Inc., 1983. J.A. Waschl. A review of landmine detection. Technical report, Defence Science and Technology Organisation, New Zealand, 1994. F E Weick. Aircraft Propeller Design. McGraw-Hill, UK, 1930. R L Wheeler. Hovercraft skirts. International Hovering Craft, Hydrofoil and Advanced Transit Systems Conference, pages 441454, 1974. Wikipedia. File:m4a4 ail cfb borden 1.jpg, 2009. [Online] available at http://en.wikipedia.org/ wiki/File:M4a4_flail_cfb_borden_1.JPG [Accessed: 24/10/09]. Yamaha-Australia. Engine specications, 2009. By correspondance. W C Young and Budynas R. Roarks Formulas for Stress and Strain. McGraw-Hill, 2001. Y. L. Zhang, S. A. Velinsky, and X. Feng. On the tracking control of dierentially steered wheeled mobile robots. Journal of Dynamic Systems, Measurement, and Control, pages 455461, 1997.

This Page Intentionally Left Blank

Appendix A Autopilot Conguration Files

Airframe Conguration File
ProjectHover_final.xml: <!DOCTYPE airframe SYSTEM "airframe.dtd"> <airframe name="ProjectHover"> <servos> <servo name="THROTTLE_LEFT" no="4" min="1100" neutral="1100" max="1900"/> <servo name="THROTTLE_RIGHT" no="0" min="1000" neutral="1500" max="2000"/> <servo name="YAW_OUTPUT" no="2" max="1000" neutral="1500" min="2000"/> </servos> <commands> <axis name="THROTTLE_LEFT" failsafe_value="0"/> <axis name="THROTTLE_RIGHT" failsafe_value="0"/> <axis name="ROLL" failsafe_value="0"/> </commands> <rc_commands> <set command="THROTTLE_LEFT" value="@THROTTLE"/> <set command="THROTTLE_RIGHT" value="@ROLL"/> <set command="ROLL" value="@ROLL"/> </rc_commands> <command_laws> <let var="aileron" value="@ROLL"/> <let var="elevator" value="@PITCH"/> <set servo="THROTTLE" value="@THROTTLE"/> <set servo="AILERON_LEFT" value="$aileron"/> <set servo="AILERON_RIGHT" value="$aileron"/> </command_laws> <section name="AUTO1" prefix="AUTO1_"> <define name="MAX_ROLL" value="0.6"/> <define name="MAX_PITCH" value="0.6"/> </section> <section name="BAT"> <define name="MILLIAMP_AT_FULL_THROTTLE" value="11000" unit="mA"/> <define name="CATASTROPHIC_BAT_LEVEL" value="9.3" unit="V"/> </section> <section name="MISC"> 187

188

APPENDIX A. AUTOPILOT CONFIGURATION FILES

189 <define name="POWER_CTL_BAT_NOMINAL" value="11.1" unit="volt"/>  <define name="ALTITUDE_PGAIN" value="-0.03"/>  <define name="ALTITUDE_MAX_CLIMB" value="2."/>  <define name="AUTO_THROTTLE_NOMINAL_CRUISE_THROTTLE" value="0.35"/> <define name="AUTO_THROTTLE_MIN_CRUISE_THROTTLE" value="0.30"/> <define name="AUTO_THROTTLE_MAX_CRUISE_THROTTLE" value="0.80"/> <define name="AUTO_THROTTLE_LOITER_TRIM" value="1500"/> <define name="AUTO_THROTTLE_DASH_TRIM" value="-1000"/> <define name="AUTO_THROTTLE_CLIMB_THROTTLE_INCREMENT" value="0.15"/> <define name="AUTO_THROTTLE_PGAIN" value="-0.01"/> <define name="AUTO_THROTTLE_IGAIN" value="0.1"/> <define name="AUTO_THROTTLE_PITCH_OF_VZ_PGAIN" value="0.05"/>  <define name="AUTO_PITCH_PGAIN" value="-0.05"/> <define name="AUTO_PITCH_IGAIN" value="0.075"/> <define name="AUTO_PITCH_MAX_PITCH" value="0.35"/> <define name="AUTO_PITCH_MIN_PITCH" value="-0.35"/> <define name="THROTTLE_SLEW" value="0.1"/> </section> <section name="Takeoff" prefix="Takeoff_"> <define name="Height" value="6" unit="m"/> <define name="Speed" value="1.4" unit="m/s"/> <define name="Distance" value="3" unit="m"/> <define name="MinSpeed" value="1" unit="m/s"/> </section> <section name="NAV"> <define name="NAV_PITCH" value="0."/> <define name="NAV_GLIDE_PITCH_TRIM" value="0"/> <define name="DEFAULT_CIRCLE_RADIUS" value="5" unit="m"/> </section> <section name="AGGRESSIVE" prefix="AGR_"> <define name="CLIMB_THROTTLE" value="0.95"/> <define name="DESCENT_THROTTLE" value="0.1"/> <define name="CLIMB_PITCH" value="0.3"/> <define name="DESCENT_PITCH" value="-0.25"/> <define name="BLEND_START" value="20"/> <define name="BLEND_END" value="10"/> <define name="CLIMB_NAV_RATIO" value="0.8"/> <define name="DESCENT_NAV_RATIO" value="1.0"/> </section> <section name="FAILSAFE" prefix="FAILSAFE_">

190

APPENDIX A. AUTOPILOT CONFIGURATION FILES

<makefile> CONFIG = \"tiny_2_1_1.h\" include $(PAPARAZZI_SRC)/conf/autopilot/tiny.makefile FLASH_MODE=IAP ap.CFLAGS += -DFBW -DAP -DBOARD_CONFIG=$(CONFIG) -DLED -DTIME_LED=1 ap.srcs = sys_time.c $(SRC_ARCH)/sys_time_hw.c $(SRC_ARCH)/armVIC.c main_fbw.c main_ap.c main.c ap.srcs += commands.c ap.CFLAGS += -DACTUATORS=\"servos_4017_hw.h\" -DSERVOS_4017 ap.srcs += $(SRC_ARCH)/servos_4017_hw.c actuators.c ap.CFLAGS += -DRADIO_CONTROL -DRADIO_CONROL_TYPE=RC_JR ap.srcs += radio_control.c $(SRC_ARCH)/ppm_hw.c ap.CFLAGS += -DDOWNLINK -DUSE_UART1 -DDOWNLINK_TRANSPORT=PprzTransport -DDOWNLINK_FBW_DEVICE=Uart ap.srcs += downlink.c $(SRC_ARCH)/uart_hw.c datalink.c pprz_transport.c ap.CFLAGS += -DINTER_MCU ap.srcs += inter_mcu.c ap.CFLAGS += -DADC -DUSE_ADC_0 -DUSE_ADC_1 -DUSE_ADC_2 ap.srcs += $(SRC_ARCH)/adc_hw.c ap.CFLAGS += -DGPS -DUBX -DUSE_UART0 -DGPS_LINK=Uart0 -DUART0_BAUD=B38400 -DGPS_USE_LATLONG ap.srcs += gps_ubx.c gps.c latlong.c ap.CFLAGS += -DINFRARED -DALT_KALMAN -DWIND_INFO -DWIND_INFO_RET ap.srcs += infrared.c estimator.c ap.CFLAGS += -DNAV -DAGR_CLIMB -DLOITER_TRIM ap.srcs += nav.c fw_h_ctl.c fw_v_ctl.c ap.srcs += nav_line.c ap.srcs += nav_survey_rectangle.c ap.srcs += OSAMNav.c ap.srcs += bomb.c ap.srcs += snav.c ap.CFLAGS += -DUSE_MODULES sim.CFLAGS += -DUSE_MODULES # Config for SITL simulation include $(PAPARAZZI_SRC)/conf/autopilot/sitl.makefile sim.CFLAGS += -DBOARD_CONFIG=\"tiny.h\" -DAGR_CLIMB -DLOITER_TRIM -DALT_KALMAN -DTRAFFIC_INFO sim.srcs += nav_survey_rectangle.c traffic_info.c nav_line.c OSAMNav.c sim.srcs += bomb.c </makefile> </airframe>

191

Telemetry Conguration File

default.xml: <!DOCTYPE settings SYSTEM "settings.dtd">

STEP="1" VAR="pprz_mode" module="estimator" values="MANUAL STEP="1" VAR="launch"> STEP="1" VAR="kill_throttle">

STEP="1" VAR="telemetry_mode_Ap_DefaultChannel" shortname= STEP="1" VAR="gps_reset" module="gps_ubx" handler="Reset"

Settings Conguration File

basic.xml: <?xml version="1.0"?> <!DOCTYPE telemetry SYSTEM "telemetry.dtd"> <telemetry> <process name="Ap"> <mode name="default"> <message name="ALIVE" period="5"/> <message name="GPS" period="0.25"/> <message name="NAVIGATION" period="1."/> <message name="ATTITUDE" period="0.5"/> <message name="ESTIMATOR" period="0.5"/>

192

APPENDIX A. AUTOPILOT CONFIGURATION FILES

193 <message name="COMMANDS" <message name="FBW_STATUS" <message name="ACTUATORS" </mode> <mode name="debug"> <message name="PPM" <message name="RC" <message name="COMMANDS" <message name="FBW_STATUS" <message name="ACTUATORS" </mode> </process> </telemetry> period="5"/> period="2"/> period="5"/>

period="0.5"/> period="0.5"/> period="0.5"/> period="1"/> period="5"/>

Flight Plan Conguration File

projecthover_basictest2.xml: <!DOCTYPE flight_plan SYSTEM "flight_plan.dtd">

<flight_plan alt="75" ground_alt="0" lat0="-34.917475" lon0="138.599839" max_dist_from_home= <header> #include "nav_line.h" #include "datalink.h" #include "radio_control.h" </header> <waypoints> <waypoint name="HOME" x="0" y="0"/> <waypoint name="STDBY" x="49.5" y="100.1"/> <waypoint name="1" x="30" y="-35"/> <waypoint name="2" x="60" y="-40"/> <waypoint name="3" x="90" y="-65"/> <waypoint name="MOB" x="137.0" y="-11.6"/> <waypoint name="S1" x="0" y="200"/> <waypoint name="S2" x="500" y="400"/> <waypoint alt="30.0" name="AF" x="177.4" y="45.1"/> <waypoint alt="0.0" name="TD" x="28.8" y="57.0"/> <waypoint name="_BASELEG" x="168.8" y="-13.8"/> <waypoint name="CLIMB" x="0" y="600"/> </waypoints> <exceptions/> <blocks> <block name="Wait GPS"> <set value="1" var="kill_throttle"/> <while cond="!GpsFixValid()"/> </block> <block name="Geo init"> <while cond="LessThan(NavBlockTime(), 10)"/>

194

APPENDIX A. AUTOPILOT CONFIGURATION FILES

<call fun="NavSetGroundReferenceHere()"/> </block> <block name="Holding point"> <set value="1" var="kill_throttle"/> <attitude roll="0" throttle="0" vmode="throttle"/> </block> <block key="t" name="Takeoff" strip_button="Takeoff (wp CLIMB)" strip_icon="takeoff.png"> <exception cond="estimator_z > ground_alt+25" deroute="Line 1to2"/> <exception cond="rc_status == RC_OK" deroute="Line 1to2"/> <set value="0" var="kill_throttle"/> <set value="0" var="estimator_flight_time"/> <go from="HOME" throttle="1.0" vmode="throttle" wp="CLIMB" pitch="15"/> </block> <block name="Line Hometo1" strip_button="Line (wp Home-1)" strip_icon="line.png"> <exception cond="datalink_time > 22" deroute="Line 2to3"/> <call fun="nav_line_init()"/> <call fun="nav_line(WP_HOME, WP_1, nav_radius)"/> </block> <block name="Line 1to2" strip_button="Line (wp 1-2)" strip_icon="line.png"> <exception cond="datalink_time > 22" deroute="Line 1to2"/> <call fun="nav_line_init()"/> <call fun="nav_line(WP_1, WP_2, nav_radius)"/> </block> <block name="Line 2to3" strip_button="Line (wp 2-3)" strip_icon="line.png"> <exception cond="datalink_time > 22" deroute="Line 2to3"/> <call fun="nav_line_init()"/> <call fun="nav_line(WP_2, WP_3, nav_radius)"/> </block> <block name="Line 1-2" strip_button="Line (wp 1-2)" strip_icon="line.png"> <exception cond="datalink_time > 22" deroute="Standby"/> <call fun="nav_line_init()"/> <call fun="nav_line(WP_1, WP_2, nav_radius)"/> </block> <block name="Survey S1-S2" strip_button="Survey (wp S1-S2)" strip_icon="survey.png"> <survey_rectangle grid="50" wp1="S1" wp2="S2"/> </block> </blocks> </flight_plan>

Radio Conguration File

X2610.xml: <?xml version="1.0"?> <!DOCTYPE radio SYSTEM "radio.dtd"> <radio name="JRRS77S" data_min="900" data_max="2100" <channel ctl="THRO" function="THROTTLE" min="1100" <channel ctl="AIL" function="ROLL" min="1000" <channel ctl="ELE" function="PITCH" max="1000" <channel ctl="RUD" function="YAW" min="1000"

sync_min ="5000" sync_max neutral="1100" max="2000" neutral="1500" max="2000" neutral="1500" min="2000" neutral="1500" max="2000"

="15000" pulse_typ average="0"/> average="0"/> average="0"/> average="0"/>

195 <channel ctl="GEAR" function="HATCH" <channel ctl="AUX" function="MODE" <channel ctl="AUX2" function="MODE2" </radio>

min="1000" neutral="1500" max="2000" average="1"/> min="1000" neutral="1500" max="2000" average="1"/> min="1000" neutral="1500" max="2000" average="1"/>

This Page Intentionally Left Blank

Appendix B Segmented Skirt Stability Equations

This Appendix contains details of the calculation of a segmented skirt cushion footprint as it changes with pitch angle. These calculations related purely to the case of partial skirt-ground contact, as described by (Graham, 1992).

Variables
beta = Skirt contact angle Hs = Height hull above ground (at zero alpha) Hp = Height of CG above hull (at zero alpha) Hsp = Hs + Hp

Hsp = hst + H p A1

= Hsp cos +

Ls Ls sin hcg A2 = Hsp cos hcg 2 2

The reaction moment is expressed by the equation:

M p = pc
i=1

Af i XM P i

(B.1)

197

198

APPENDIX B. SEGMENTED SKIRT STABILITY EQUATIONS

Af 1 = (B st 2Rs )(Ls cos + A1 cot( )) Af 2 = 2Rs (Ls 2Rs )cos 2 Af 3 = Rs cos 2 Af 4 = (A1 Rs sin())2 cotcot Af 5 = 0if |A1 | < |Rs sin| = (Rs + (A1 Rs sin)cot )(Rs cos + A1 cot( )) 2

SymmetryCondition : Afi () = Afi (+)

(B.2)

XM P 1 = Hsp sin XM P 2 = Hsp sin

A1 cot( ) 2

Ls 4Rs Rs + )cos 2 3 1 2 XM P 4 = (Hsp hcos) (A1 Rs sin)cot sin 3 Ls 4 XM P 5 = Hsp sin ( Rs )cos (Rs cos + A1 cot( )) 2 3 XM P 3 = Hsp sin + (

SymmetryCondition : XM Pi () = XM Pi ()

(B.3)

Appendix C Simulink

199

This Page Intentionally Left Blank

Appendix D External Milestones

Milestone due dates. Table D.1: External Milestones Milestone Due Date Project Denition, Specication and Contract 20 March Draft Gantt Chart 27 March Final Gantt Chart 3 April Draft Preliminary Report 8 May Preliminary Report 22 May Manufacturing Drawings 5 June Seminar Abstract 18 September Draft Final Report 9 October Exhibition 22 October Final Report 30 October Workbook 30 October Equipment Return Form 6 November

201

This Page Intentionally Left Blank

Appendix E Internal Time Line

Table E.1: Internal Time line Preliminary Design Commence By Conceptual Platform Design 2 Mar Conceptual Design of Autonomous System 2 Mar Systems Integration 17 Apr Manufacturing 11 May Platform Ready Detailed design of Autonomous Systems 6 Jul Autonomous System Integration 13 Jul Project Ready Product Testing 4 Sep Draft Report Report

Complete By 8 May 8 May 5 Jun 3 Jul 10 Jul 24 Jul 14 Aug 4 Sep 18 Sep 25 Sep 30 Oct

203

This Page Intentionally Left Blank

Appendix F Lift Calculations

Pc = Where: Mhc = 100 [kg] g = 9.81 [m/s2 ] Ac = 1.79 [m] Pc = 485.64 [Pa] Brooks (2005) shows that: Q= and Dc () = 0.5 + (0.4 103 ) + (0.109 104 )2 (0.494 107 )4 + (0.345 109 )6 Where: = 1.22 [kg/m2 ] HH = 15 103 [m] Cst = 5.023 [m] = 45 (for a segmented skirt, this value is a common approximation) Dc () = 0.537 Q = 1.14 [m3 /s] A safety factor of 1.4 was applied to this value. 1.4 was selected as this was determined to give a sucient safety margin, without an unneccesary increase in the load. This equates to a total design ow rate of 1.6 m3 /s (Fitzpatrick, 2009). (F.3) 2Pc / H H C st Dc () (F.2) M hc g Ac (F.1)

205

This Page Intentionally Left Blank

Appendix G Lift Mount Strength Calculations

A cross section of the T-Section used in the fan and engine mount structure is shown below:

The moment of inertia of the T-section is calculated as follows: (Kotousov, 2007) 1 3 It = (tts yts + bts (dts yts )3 (bts tts )(dts yts sts )3 ) 3 Where: bts = 40 [mm] dts = 50 [mm] sts = 3 [mm] hts = 47 [mm] tts = 3 [mm] yts = 25 [mm] It = 9.26 108 [m4 ] 207 (G.2) (G.1)

208

APPENDIX G. LIFT MOUNT STRENGTH CALCULATIONS

The maximum deection of a beam can be calculated using the following equation, with the load of the engine assumed to act at a single point equidistant from the two edges of the mounting plate. (Kotousov, 2007) = Where: W = 40.9 [N] bB = 238 [mm] aB = 19 [mm] E = 70 109 [Pa] I = 9.26 108 [m4 ] = 0.0351 [mm] W bB 2 (2bB + 3aB ) 6EI (G.3)

Appendix H Hull Appendices

The Sandwich Principle
A sandwich structure consists of four main components, each of whose individual material properties are maximised and cumulatively work to enhance the overall structural properties.

The Faces
The faces carry tensile and compressive strengths in the sandwich as well as local pressure. Local exural rigidity is negligible and hence ignored.

The Core
The primary function of the core is to separate the facing skins of the sandwich structure. The secondary function of the core is to absorb the shear. In the event that the core is not rigid enough to keep the distance between the facing skins constant or its shear modulus too low that the faces slide over each other, the overall integrity of the structure is compromised. The core must also full a range of other ancillary demands including resistance to bucking, insulation, resistance to moisture absorption and aging resistance. 209

210

APPENDIX H. HULL APPENDICES

Adhesive
An adhesive layer acts as an interface between the skin and the core. Its primary function is to transfer the shear forces between the skin and the core. The general rule of thumb is that the adhesive in use must be able to carry the same shear force as the core .The adhesive layer being subjected to the same compressive and tensile forces as the face must be able to carry the forces without failure. The core to skin adhesive rigidly joins the sandwich components allowing them to act as one unit with high torsional and bending rigidity. In order for the craft to remain responsive during manoeuvre, it must maintain its geometric integrity. Moreover in order to avoid interference of the rotational components on board with their respective structural supports in order to ensure the integrity of sensor resolution, the deection of the hull oor and its supporting structure must be kept to a minimum. The sandwich principle utilises the general engineering theory that the exural stiness of any panel is proportional to the cube of its thickness. A sandwich beam of the same width and weight as a solid beam has a much greater stiness as a result of its higher moment of inertia. By sandwiching panel, at a penalty of 6% in weight the structures strength increases by 250%. Moreover there is a 600% increase in stiness of the structure. By doubling the core thickness at an additional 3% penalty in weight taking the cumulative penalty to 9%, the structures strength increased by a further 160% taking the cumulative increase to 825%. The end result is a 36 fold increase in structural stiness at a 9% penalty in weight. (SP Guide, 2000) To verify the applicability of this general engineering principle to the design and manufacture of the platform, a cross-section of materials identied in the literature review of hull design were considered for a comparison. With a 5mm steel beam as the control, a comparison of the dierence in weight and overall thickness of a range panels that would give a similar exural rigidity was sourced and yielded the following results.

A carbon sandwich panel yielded the greatest saving in weight and volume, followed by a glass sandwich panel. An aluminium panel and solid glass panel yielded signicant savings when compared to the control but a marginal saving when compared to a glass sandwich panel.

211

Fiber Selection

The role of the bre is to complement and compound the physical properties of the neat resin in essence giving the material mechanical properties that would otherwise be unattainable. Given the general requirements of the platform, high temperatures bres such as ceramic, boron and quartz bres were discounted completely. Polyethylene bres were discounted on the grounds that they only provided a marginal improvement in tensile strength and although their impact resistance is signicantly better that glass and amide and carbon bres (their cost makes their use restrictive Given that rigidity was a key criteria in the design of the platform polyester bres were discounted from review given the bres low modulus (Hexcel Guide,2000) The bers considered for the conceptual design were carbon, glass and armide bers. Aluminum was used as a control in the comparison.

Fiber Density

In terms of absolute density, armide bres (low and high modulus) have the lowest density and per cubic centimetre and give a saving of 86.2% when considered against the control. In contrast S and E glass bre when considered per cubic centimetre give a saving of 22.7% per cubic centimetre, making the use of any bre more attractive than the control.

Tensile Strength

In terms of tensile strength intermediate modulus carbon bre demonstrated the highest absolute value followed by the high strength and high modulus carbon bre. Nevertheless even the lowest scoring bre possessed signicantly better tensile strength making the use of bres over the control a viable structural option.

212

APPENDIX H. HULL APPENDICES

Resins
The matrix in a composite serves three primary functions:
It supports the bres and bonds them It transfers any loads to the bres and keeps them in their given orientations It gives the composite environmental resistance and determines the maximum operational temperature.

The resins considered are; Epoxy, Phenolic , Bismalemi and Vinylester. Epoxys provide excellent mechanical properties and easy processing. They also provide high toughness to the laminate and good environmental resistance making them an ideal choice for the platform. Phenolic resins provide excellent re resistance as well as low smoke and toxic emissions when combusted, and hence are seen extensively in the airline industry. They also have good temperature resistance and rapid cure. However considerign that the probability of re is low and its occurrence easily mitigated, the use of Phenolic resins is not critical The use of Bismalemi resins was discounted on the grounds that Bismalemi resins are best suited for high temperature operations, a condition not envisaged in the platforms operational life. The interlace in controlled by the extend of bonding that exist between the matrix and bre. This is inuenced by the treatment given to the bre surface as well as the intrinsic properties of the matrix itself. It is important to note however that the smaller bre diameter the greater the contact surface area of the matrix resulting in a better distribution of the interfacial loads. Nevertheless as the bre diameter decreases the cost of the bre increases However the smaller the bre diameter the more expensive the bre.

Core Selection
The hull oor and pressure chamber oor are most susceptible to geometric deection by virtue of their design. Together they constitute over 60% of the structural volume. Under these circumstances and given the need to minimize weight and deection, a sandwich construction for these components is prudent. The hull oor and pressure chamber oor will serve as mounting surfaces for various components and hence have to intrinsically be good acoustic and vibration isolators. Given the nature of the operational terrain, the pressure chamber oor must have a reasonable resistance to puncher

213 and impact in general. These factors serve to reinforce the decision to use a sandwich construction for the hull oor and pressure chamber oor.

Core Classication
Three basic types of cores were considered; foams, wood, honeycomb structures.

Foam Cores Foam cores can be manufactured from a variety of synthetic polymers including polyvinyl chloride (PVC), polystyrene (PS) and styreneacrylonitrile (SAN). They are available in a range of densities form 30 kg/m3 to 300 kg/m3 and a variety of thicknesses increasing the ease at which the theoretical design parameters can be satised using standard o the shelf parts. Major concerns with the core included maintaining a low density but with a relatively high shear modulus as well as good environmental resistance, and in particular, exceptional water resistance given the nature of operation. These concerns were met and surpassed with the use of a foam core. Several foams were investigated and yielded the following results. PVC foams PVC foams are a chemical hybrid of PVC and Polyuethane and simply referred to as PVC foam. They oer a good balance between static and dynamic properties and excellent resistance to water absorption making it ideal for the platforms pressure chamber oor. The foam although ammable retains its general structural properties over a large temperature range (-240deg to 80 deg Celsius) making it suitable for the hull oor, on which the engine is mounted. The foam exhibits good resistance to styrene therefore suitable to use with a rage of resins making the core easy and economical to work with. PVC foams come in two types: 1. Cross linked 2. Linear (uncross linked) Linear foams are tougher, more exible and easier to heat form around curves than cross-linked foams. However they exhibit lower mechanical properties to equivalent density cross linked PVC. Cross linked foams are harder but more brittle than linear foams making them less suitable as cores which will be subjected to repeated impact.However they are less susceptible to softening and creeping making them ideal for extended use and repeated loading. Klegcell and Divinycell H and HT are two brands commercially available in Adelaide.

Polystyrene Foams
Polystyrene foams although extremely light (40 kg/m3 exhibit low mechanical properties and hence are rarely used in high performance components. They are cheap and easy to work with in terms of cutting and sanding but cannot be used in conjunction with polystyrene resin systems as they are dissolved by the styrene present in the resin.

Polyurethane foams
Polyurethane foams exhibit only moderate mechanical properties and have the tendency for the foam to surface at the resin core/ interface to deteriorate with age leading to skin delamination. Even though it provide exceptional acoustic absorption, it is primarily used as a stiener and not as a load bearing panel, making it unsuitable for the hull oor and the pressure chamber oor

214

APPENDIX H. HULL APPENDICES

Polymethyl methacrylamide (Acrylic) Foams

Acrylic foams give the highest overall strength and stiness for a given density moreover they show exceptional thermal resistance making them ideal for all structural components on the platform. However they are extremely expensive, to the extent that they are used primarily used in the aerospace industry in applications such as the cores for helicopter blades.

Styrene Acrylonitrile (SAN) co polymer Foams

SAN foams behave in a similar way to cross linked PVC foams. They are comparable in terms of mechanical properties but have a much higher elongation and toughness. Their appeal for incorporation into the platform is their ability to absorb higher impacts that would fracture PVC foams. Their toughness characteristics are based on their molecular structure and hence are inherent and do not deteriorate with time increasing the overall life of the platform if incorporated into the design

Honeycombs
The load bearing cells in a honey comb structure give it an exceptional strength to stiness ratio. The extreme light weight and exceptional mechanical properties make it ideal for both the pressure chamber oor and hull oor, however due to the small bonding area they are primarily used with resin system s such as epoxies and prepreg bres, to ensure adequate contact adhesion, which make it a costly option. Two types of honeycomb structured cores were investigated: 1. Aluminum which although were extremely light and showed high mechanical properties has to be discounted due to the interference with the onboard sensors. 2. Nomex honeycomb cores consist of Kevlar paper based cells dipped in phenolic resin to produce a high strength core with good re resistance. The re resistance and high mechanical properties make it ideal for use on the hull oor and suitable for the pressure chamber oor. However given the extremely high cost of the core coupled with the fact that the core must be used with prepreg bres makes the use of Nomex prohibitive.

Wood
Wood is considered to be a natural honeycomb as it is structure on a microscopic scale similar to cellular hexagonal structure of a synthetic honeycomb. In marine environments balsa wood is one of the most common cores in use today. The main advantages of a balsa core stem from its high compressive strength, its thermal isolative properties its positive otation and most notable its ability to withstand repeated impact when lines with an aluminium skin. These properties coupled with the fact that it is a relatively good acoustic isolator and is available in a range of thicknesses make it suitable for both the pressure chamber oor and the hull oor. The disadvantages of using end grain balsa is primarily its cost and the fact that the minimum density of readily available balsa is around 100 kg/m3 .

215

Summary

This Page Intentionally Left Blank

Appendix I Bill Of Materials

The Bill of Materials (BOM) is a comprehensive list of all the components on the hovercraft. The BOM is separated into the main systems of the platform for which each is numbered. The numbering system is summarised as follows: 885-XX-YY-ZZ(D)(T)(A)(V) 1. 885 - the project number 2. XX - The System Number (a) 01 - Structural (b) 02 - Lift (c) 03 - Skirt (d) 04 - Propulsion (e) 05 - Electronics/Mechatronics (f) 06 - Marking System 3. YY - The System Subassembly Number 4. ZZ - The Item Number 5. D - The item number shoould be followed with a D if design is required 6. V - The item number should be followed with a V if its a vendor item Drawing Numbers follow the same covention. 217

218

APPENDIX I. BILL OF MATERIALS

Assembly Part Number

885-01-00-DA

885-02-03-DTA

885-02-02-DTA 885-02-01-DTA 885-02-00-DTA

885-03-00-DTA

885-04-02-DTA

ID Number Section Item Detailed Parts List for Hull Assembly 885-01-01-18V Structure Skirt Attachment Brackets 885-01-01-18V Structure Skirt Attachment Eyelets (specification) 885-01-01-17V Structure Qcell (1kg) 885-01-01-16V Structure (Curox) MEKP M100 Cure 885-01-01-15V Structure H180 Hardener 885-01-01-14V Structure R180 Epoxy Resin 885-01-01-13V Structure Kevlar Plain 170g/1.27m pSheet 885-01-01-12V Structure In. Std. E-glass 190g/1m pSheet 885-01-01-11V Structure 4.5 mm Natural Rubber (1x1m) 885-01-01-10V Structure 6x50mm Flathead Bolts 885-01-01-9V Structure 6mm Flat Round Washers 885-01-01-8V Structure 6mm Hex Nut 885-01-01-7V Structure 6 gauge x 25mm wood screw 885-01-01-6D Structure Composite Wall Structure 885-01-01-5D Structure Foam Mould 885-01-01-4D Structure Wooden Frame 885-01-01-3D Structure Bottom Plate Insert 885-01-01-2D Structure Bottom Plate 885-01-01-1D Structure Top Plate Structure Hull Assembly Detailed Parts List for Lift Assembly 885-02-02-08V Lift Lift Fan 885-02-03-07V Lift Lift Engine 885-02-03-6DA Lift Fan Guard 885-02-03-5DA Lift Vibration Isolator 885-02-03-4DA Lift Hull Attachment Bracket 885-02-03-3DA Lift Support Upright 885-02-03-2DA Lift Support Arm 885-02-03-1DA Lift Round Engine Support 885-02-02-7DA Lift Lift Mount 885-02-02-6DA Lift Shaft Extension 885-02-02-5DA Lift Collar 885-02-02-4DA Lift Extension 885-02-02-3DA Lift Fan Key 885-02-02-2DA Lift Shaft Key 885-02-02-1DA Lift Shaft Extension Cap 885-02-01-02DA Lift 90 deg bracket 885-02-01-01DA Lift Inlet Bell Lift Lift System Detailed Parts List for Skirt Assembly 885-03-00-05V Skirt Cable Ties 885-03-00-06DA Wooden Jupe Bracket 885-03-00-05DA Jupe Segment 885-03-00-04DA Skirt Back Finger Segment 885-03-00-03DA Skirt Front Finger Segment 885-03-00-02D Skirt Standard Finger Segment 885-03-00-01D Skirt Skirt Attachment Material (1m length) Skirt Skirt System Detailed Parts List for Propulsion Assembly 885-04-21VA Propulsion 4mm Nut 885-04-20VA Propulsion 4mm Washer 885-04-19VA Propulsion 4mm Bolt 885-04-02-18DA Propulsion Mount Bracket 885-04-02-17DA Propulsion Support Bracket 885-04-02-16DA Propulsion Ocean Controls 0.88Nm Stepper 885-04-02-15DA Propulsion IR Reciever Circuit 885-04-02-14DA Propulsion IR Transmitter Circuit 885-04-02-13VA Propulsion M3 Thread 885-04-02-12DA Propulsion PVC Pipe (21.6mm) 885-04-02-11DA Propulsion IR Support Plate 885-04-02-10DA Propusion IR Support Plate IR Mount 885-04-09VT Propulsion Battery (6cells) 885-04-08VT Propulsion Stepper Motor 885-04-07VT Propulsion Teflon Washer 885-04-06VT Propulsion ESC 885-04-05VT Propulsion Teflon Interior Bearing (50-55mm) 885-04-04VT Propulsion Electric Motor 885-04-03VT Propulsion Engine Hub 885-04-02VT Propulsion Propeller

Vendor or Design Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Design Design Design Design Design Design Design Vendor Vendor Design Vendor Design Design Design Design Design Design Design Design Design Design Design Design Design Design Vendor Design Design Design Design Design Design Design Vendor Vendor Vendor Design Design Vendor Design Design Vendor Design Design Design Design Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor

Amount 4m 32 1 As Req As Req As Req 4 10 1 20 40 20 8 1 1 1 1 1 1 1 1 1 1 6 6 6 3 1 1 1 1 1 1 1 1 6 1 1 64 4 4 2 2 28 10 2 32 32 32 4 4 2 2 2 4 2 2 2 2 4 2 2 2 4 2 2 2

Drawing No. N/A N/A 885-01-01-06 885-01-01-06 885-01-01-06 885-01-01-06 885-01-01-06 885-01-01-06 885-01-01-07 885-01-01-AS 885-01-01-AS 885-01-01-AS 885-01-01-AS 885-01-01-06 885-01-01-05 885-01-01-04 885-01-01-03 885-01-01-02 885-01-01-01 885-01-01-AS N/A N/A 885-02-03-06 885-02-03-05 885-02-03-04 885-02-03-03 885-02-03-02 885-02-03-01 885-02-03-AS 885-02-03-AS 885-02-02-05 885-02-02-04 885-02-02-03 885-02-02-02 885-02-02-01 885-02-03-AS 885-02-01-01 885-02-01-AS N/A 885-03-00-02 885-03-00-02 885-03-00-01 885-03-00-01 885-03-00-01 N/A N/A 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-03-01 885-04-03-01 885-04-03-01 N/A N/A 885-04-02-02 885-04-02-02 885-04-02-01 885-04-02-01 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS 885-04-01-AS

219

885-04-01-12DA 885-04-01-11DA 885-04-01-10DA 885-04-01-9DA 885-04-01-8DA 885-04-01-7DA 885-04-01-6DA 885-04-01-5DA 885-04-01-4DA 885-04-01-3DA 885-04-01-2DA 885-04-01-1DA 885-04-01-DTA 885-04-00-DTA Detailed Parts List for Electronics/Mechatronics 885-05-01V 885-05-01V 885-05-03V 885-05-04V 885-05-05V 885-05-06D 885-05-07V 885-05-08V 885-05-09V 885-05-10V 885-05-00-DTA 885-06-04VA 885-06-03VA 885-06-02VA 885-06-01DA 885-06-00-DTA

Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Propulsion Electrical/Systems Electrical/Systems Electrical/Systems Electrical/Systems Electrical/Systems Electrical/Systems Electrical/Systems Electrical/Systems Payload Payload Electrical/Payload Markin System Markin System Markin System Markin System Marking System

Lip Locking Collar A-frame Leg Support Pipe Rotating Shaft Support Bracket Top Supports Fan Protector Front Plate Fan Protector Back Plate L Bracket Fan Protector Body Support Plate Propulsion Mount Propulsion System Autopilot Microcontroller Mini Dragon Cabling Power Circuit Housing Box 12V Battery, 7.2AH 12V Battery, 1.3AH Metal Detector Signal Processing Unit Mechatronics Components Paint Can Servo Actuator (8.3kg) JR ES579 PVC Pipe (90mm) Actuator Bracket

Design Design Design Design Design Design Design Design Design Design Design Design Design Design Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor Vendor

2 2 3 2 2 2 4 2 2 16 2 2 2 2 1 1 1 1 1 1 2 1 2 2

885-04-01-AS 885-04-01-06 885-04-01-05 885-04-01-01 885-04-01-01 885-04-01-04 885-04-01-04 885-04-01-03 885-04-01-03 885-04-01-04 885-04-01-02 885-04-01-07 885-04-01-00 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A

Design Vendor Vendor Design

1 1 1 1

N/A N/A 885-06-01-01 885-06-01-01 885-06-01-AS

This Page Intentionally Left Blank

221

Appendix J

222

APPENDIX J. MANUFACTURING DRAWINGS

Manufacturing Drawings

223

224

APPENDIX J. MANUFACTURING DRAWINGS

225

226

APPENDIX J. MANUFACTURING DRAWINGS

227

228

APPENDIX J. MANUFACTURING DRAWINGS

229

230

APPENDIX J. MANUFACTURING DRAWINGS

231

232

APPENDIX J. MANUFACTURING DRAWINGS

233

234

APPENDIX J. MANUFACTURING DRAWINGS

235

236

APPENDIX J. MANUFACTURING DRAWINGS

237

238

APPENDIX J. MANUFACTURING DRAWINGS

239

240

APPENDIX J. MANUFACTURING DRAWINGS

241

242

APPENDIX J. MANUFACTURING DRAWINGS

243

244

APPENDIX J. MANUFACTURING DRAWINGS

245

246

APPENDIX J. MANUFACTURING DRAWINGS

247

248

APPENDIX J. MANUFACTURING DRAWINGS

249

250

APPENDIX J. MANUFACTURING DRAWINGS

251

252

APPENDIX J. MANUFACTURING DRAWINGS

253

254

APPENDIX J. MANUFACTURING DRAWINGS

255

256

APPENDIX J. MANUFACTURING DRAWINGS

257

258

APPENDIX J. MANUFACTURING DRAWINGS

259

260

APPENDIX J. MANUFACTURING DRAWINGS

Appendix K Risk Treatment

261

262

APPENDIX K. RISK TREATMENT

SAFE OPERATING PROCEDURE: GENERIC

LOCATION DETAILS School/Branch: Mechanical Engineering TASK/ACTIVITY Hovercraft Operation Date: 13/8/09

PREPARED BY Name, Position and Signature (insert names of the supervisor, HSR, HSO and operator involved) Name Paul Hocking Maziar Arjomandi Richard Patemen Group Member Project Supervisor HSO Position Signature

HAZARD IDENTIFICATION: Project Hover - Hovercraft See Risk Assessment dated / /

RISK ASSESSMENT

SAFE OPERATING PROCEDURE DETAILS

Preparation operational area check:

Ready access to and egress from the hovercraft (min of 2m clearance required). Hovercraft and operational area is free from debris and objects, which can be tripped over or impede hovercraft motion. All removable components on hovercraft adequately secured. Area is clear of unauthorised people before commencing operation.

Personal Attire & Safety Equipment:

Approved closed toe type shoes must be worn at all times. Approved safety ear equipment must be worn at all times. Approved safety spectacles/goggles must be worn at all times. Clothing must be tight fitting. Long hair must be confined close to the head by an appropriate restraint. Finger rings and exposed loose jewellery (eg bracelets and necklaces) must not be worn. Medic Alert bracelet must be taped if exposed. Pockets and clothing must not contain any lose articles, which may fall out.

Operation Of Emergency Stop Button:

Switch power on at the main switch. Open the fuel line. Start machine using the pull start cord. Check that the Emergency Stop Button is working. Release the Emergency Stop button in preparation for next machine use.

Machine Pre-operational Safety Checks Safety Precautions that MUST be Observed:

C:\Users\Adoe\Documents\My Received Files\Hovercraft SOP V2.docx Page 1 of 3

SOP No:

Issue No:

263

SAFE OPERATING PROCEDURE: GENERIC

Visual inspection of hovercraft to verify it is in good operational order, ensuring no damage to any stationary or moving parts, electrical cords etc. Any unsafe components are to be reported to an authorised staff member and must not be used. Ensure power is switched on. Ensure the fuel line is open. Be aware of other activities happening in the immediate area. Ensure that no slip and/or trip hazards are present. Check that all fan guards are locked in place and in proper working condition. Locate and be familiar with the operation of the ON/OFF and operational controls. Locate and be familiar with the operation of the Emergency Stop button. Check Emergency Stop button is working. Wear appropriate PPE for task being performed (as specified above).

Operation:
Place hovercraft in an open area, with no obstructions (minimum 2m clearance) Ensure hovercraft is stable on level ground. Carry out Pre-operational Safety Checks (as specified above). Release the emergency stop button, if necessary. Switch power on to all electrical components. Ensure hovercraft is in stationary mode. Start the petrol engine, using the pull cord. Allow the engine to warm up to ideal operational temperatures. Carry out desired tests. NEVER LEAVE HOVERCRAFT RUNNING WHILST UNATTENDED. Before shutting down, ensure hovercraft is in stationary mode, the thrust propellers have stopped and the lift motor is in idle. Shutdown lift motor using manual stop button. Switch power off to all electrical components.

C:\Users\Adoe\Documents\My Received Files\Hovercraft SOP V2.docx

Page 2 of 3

SOP No:

Issue No:

264

APPENDIX K. RISK TREATMENT

SAFE OPERATING PROCEDURE: GENERIC

General Safety
Visual inspection of hovercraft prior to use. Unsafe components to be reported and must not be used. Keep all parts of your body and attire safely clear of the rotating and moving parts, at all times. Ensure scheduled maintenance for components has been carried out; including scheduled testing of Emergency Stop. Only authorised individuals with knowledge of operation are permitted to operate the hovercraft. Guards must be securely fitted to the hovercraft at all times during operation. NEVER LEAVE HOVERCRAFT RUNNING WHILST UNATTENDED. Safety glasses must be worn at all times during the operation of this machine. Closed Toe Type Shoes must be worn during the operation of this machine. Approved hearing protection must be worn during the operation of this machine. Loose hair to be securely tied back, loose clothing to be rolled up and/or secured, loose jewellery to be removed.

Note: This Safe Operating Procedure must be reviewed: a) after any accident, incident or near miss; b) when training new operators; c) if adopted by new work group; d) if equipment, substances or processes change; or e) within 1 year of date of issue.

C:\Users\Adoe\Documents\My Received Files\Hovercraft SOP V2.docx

Page 3 of 3

SOP No:

Issue No:

265

Pre-Flight Check (Sign & Date Boxes)

1. Phase 1: secure platform
a. Check power disconnected b. Check platform rest position

2. Phase 2: Initiate check sequence

a. Check engine mount6 bolts engine to mount No loose nuts or washers 6 bolts mount to isolator No loose nuts or washers 6 bolts isolator to hull No loose nuts or washers b. Check engine Fuel check min 20% Check oil Check fuel line Intact and clean Check throttle cable Intact and clean c. Check lift fan Check blades Blades visually intact and not damages Maintaining clearance with the bell Check attachments Visual on shaft extension grub screws

d. Check Bell Surface check No visible cracks or breaks

266

APPENDIX K. RISK TREATMENT

e. Propulsion check Check mount Firmly fastened to hull Propeller clear of obstructions Stepper Motor Check chain tension (movement approx 1cm from rest position) Stepper Motors firm f. Skirt attachment check Skirt fingers all securely fastened (cable ties in tact) No significant tears or damage

3. Phase 3: Initiate call sequence

a. Start up procedure as outlined in SOP to be followed prior to starting the hovercraft.

267

Risk Uncontrollable Hovercraft Controller failure Power-system failure Leaves determined path/area Harm to third parties Crash

Project Impact

Risk Detail/Mitigation In the case of a subsequent failure the craft shouldnt cause significant damage to itself due to the nature of the craft and low maximum speeds.

HIGH

Ineffective Thrust System Unproven thrust system HIGH

To prevent damage to third parties and minimise damage to the craft a safety watchdog will be designed to ensure cut off of the vehicle in the event of uncontrollability. Since the thrust system is an unproven system there is a risk that it will not provide the thrust required for operation. The amount of thrust required will be overcompensated with release valves installed to ensure sufficient thrust availability. Several rotating parts on the Hovercraft may cause grievous bodily harm upon failure. Several considerations will be incorporated into the design to mitigate against this occurring. Welds should be avoided in areas with high variational loading to reduce the risk of failure.

Rotating Component Failure The failure of high speed rotating components on the hovercraft.

HIGH

Shielding of all rotating components should be designed to stop projectiles on failure. All high speed rotating components should be function tested in an isolated environment prior to field testing to ensure safety. Fastenings should be checked thoroughly and regularly to reduce the risk of maintenance failure.

Project Time Frame

Large requirements of the project in a relatively short time frame. MEDIUM All group members will contribute the recommended 10-15 hours per week to ensure project goals are met. Gantt charts will be used to schedule/monitor progress and regular meetings held to maintain progress in accordance with critical paths. Some parts such as the autopilot for the craft may need to be ordered from international suppliers. Any parts required will be identified as critical path items and ordered early to ensure prompt arrival.

Part Delivery MEDIUM

268

APPENDIX K. RISK TREATMENT

Risk Sensor Ineffective Detection

Project Impact

Risk Detail/Mitigation There is a risk that the sensor will be ineffective at detecting targets and will have unreliable signal return.

LOW

GPS Accuracy

Sufficient testing will take place to determine the accuracy of the sensor with known locations on the field. A back-up wireless transmitter will be researched for implementation in the case of inefficient signal transmission. There is a risk that the accuracy of the GPS will not be sufficient for the craft to accurately follow the predetermined path. Alternative methods of local location tracking will be researched for possible implementation/discussion should the GPS be inadequate. In the event of damage to the skirt during testing the vehicle will become unstable and unfit for testing.

LOW

Skirt Damage LOW

Fire Engine Failure Fuel Fire

The skirt will be designed such that segments are easily replaceable and extra segments manufactured to ensure availability if required. A fuel fire could cause significant damage to the craft. Fuel and engine compartments will be designed to be contained within firewalls to prevent spread of fire and subsequent damage to the vehicle.

LOW

Appendix L Centre of Gravity Calculations

This table details the positions and weights of all components relative to a co-ordinate system set up at the centre of the craft. The gure below shows this co-ordinate systes. The z-axis is aligned with the rotation of axis (and geometrical centre) of the platform. The x-axis is along the axis of symmetry of the craft. Y is out of the page.

Item Hull & Top Plate Engine Lift Mount Fan Shaft Extension Inlet Bell Propulsion (Port) Propulsion (Starboard) Drive Train (Port) Drive Train (Starboard) Skirt Front Fairing Rear Fairing Electronics Main Battery Left Main Battery Right Total

Design Weight (kg) 32 14 2.7 2.3 2 1.2 3.9 3.9 1.4 1.4 3 3 1.1 3 1.1 1.1 78.3

X(mm) 22.4

W*X 716.8

Y(mm)

W*Y

Z(mm) -68.2 205 81.7 -50 -20 20 285 285 40 40 -100 60 60 30 30

W*Z -2183.4 2870 219.1 -113.5 -40 24 1282.5 1282.5 56 56 -300 180 180 33 33 3465.1

-519 -519 -420 -420 400 -500 -650 50 50

-2335.5 -2335.5 -586.6 -586.6 1200 -550 -1950 55 55 -6317.4

290 -290 290 -290

1305 -1305 406 406

450 -450

495 -495 0

Using these values the centre of gravity can be calculated in each direction (x, y and z) using Equation L.1. 269

270

APPENDIX L. CENTRE OF GRAVITY CALCULATIONS

mi xi xpos =
i=1

(L.1)

The nal positions for centre of gravity relative to the co-ordinate system are: 1. xpos = 81mm Aft of the centre 2. ypos = At the centre 3. zpos = 44mm above the centre

Appendix M Landmine Detection Code

Figure M.1: Landmine detection code, main method

271

272

APPENDIX M. LANDMINE DETECTION CODE

Figure M.2: Landmine detection code, timer interrupt

Appendix N Landmine Statistical Analysis

In developing a prototype landmine, a statistical analysis of both anti-tank and anti-personnel landmines was carried out. The width, length, height, diameter, weight, explosive content and activation pressure were recorded and compared. Further the total voulmes and case weights were determined from the data.

Anti-tank mines, 103 models considered

Length - 273.5 mm Width - 225.7 mm Height - 127.7 mm Diameter - 287.7 mm Weight - 7.8 kg Explosive Content - 5.4 kg Lower Activation Pressure - 178.9 kg Upper Activation Pressure - 270.5 kg Volume Circular - 8986700 mm3 Volume Square - 7991800 mm3

The total distribution of landmine volumes is shown in the Figure below: 273

274

APPENDIX N. LANDMINE STATISTICAL ANALYSIS

Anti-personel mines, 38 models considered

Height - 141.0 mm Diameter - 77.4 mm Weight - 1.9 kg Explosive Content - 0.15 kg Lower Activation Pressure - 4.7 kg Upper Activation Pressure - 9.0 kg Volume Circular - 1488000 mm3

Appendix O Landmine Detection Sensors

Sensors Types
Landmine detection sensors are used to increase the speed and accuracy of landmine detection and may be used in both direct and mechanically assisted detection methods. There are a variety of landmine detection sensors currently available (table O.1), some of which are detailed subsequently, with unique advantages in varying environmental and operational conditions. The almost endless combination of landmine types and environmental conditions, means there is no single sensor capable of maintaining a high level of performance in all operational scenarios; thus often a combination/hybrid of sensors are used. Hybrid sensors use comparative algorithms to analyse the output from many sensors, requiring a positive detection signal from a combination of sensors before indicating the presence of a landmine, greatly reducing the number of false positives. Table O.1: List of current landmine senor types Category Sensor Metal Detector Methods Analog/Digital Array Magnetometer Electromagnetic Methods Ground Penetrating Radar Nuclear Quadrople Resonance Microwaves Electrical Impedance Tomography Infrared X-Ray Backscatter Sound and Ultrasound Neutron Methods Acoustic/Seismic Methods Laser Doppler Vibrometer Biological Methods Dogs and Rats Chemical Methods Mechanical Methods Prodders and Probes Mine Clearing Machines

Ground Penetrating Radar

GPR are commonly used for construction surveying, however they have been used extensively in landmine detection for the past three decades (Das et al., 2004) and are known for their detailed imagery. The GPR operates by emitting radio waves into the ground and analysing the reected signal, generated from the reection of waves at the boundaries of material with dierent dielectric constant (Media, 2009). The operational principle is shown in Figure O.1, whereby a pulse is emitted from the transmitter and reected back from the soil interfaces, S1, S2, S3 and S4. The dielectric constant is a value specic to each material, however various materials may have similar constants making it hard to distinguish between them. Observable boundaries may occur from objects in the 275

276

APPENDIX O. LANDMINE DETECTION SENSORS

soil and strong interfaces between soil types, such as dry sand to wet sand. Metallic objects produce complete reection and no signal can pass through, making them highly observable (Media, 2009).

Figure O.1: Basic GPR operation diagram There are a wide variety of operational frequencies for GPR, ranging from 25 MHz to 2.3 GHz (Grin and Pippett, 2001). The lower frequencies oer the best depth penetration, but higher frequencies have proven to be necessary to achieve the required detail and an acceptable signal to noise ratio required to accurately detect landmines (Gooneratne et al., 2004). Although higher frequency GPR are capable of producing far more detail than low frequency GPR, they are limited by a small detection area due to their physical size. Their small physical size is restricted by weight, power and processing limitations. Further both high and low frequency GPR are limited by ground clearance requirements, with ideal operational conditions requiring a clearance in the order of 10s of millimeters. This places large restrictions on the practical use of GPR, especially in mechanically assisted methods. The data produced from a GPR is able to provide a detailed cross-sectional view of subsurface features, such as boundaries between dierent soil types, water tables and underground structures such as pipes, cables and cavities (Davis and Annan, 1989). The imagery provided from a standard GPR is shown in Figure O.2, where each arc represents an object with a distinct dielectric constant, diering from that of the soil. For landmine detection specically, small individual objects are targeted, with a predetermined signature detection shape. This detection shape varies based on the landmine properties such as size, depth and material. Traditional landmines are easily identied, with the metallic case providing full reection and a strong signal, whilst composite mines are harder to detect, with objects such as rocks producing a similar detection shape.

Magnetometer
Magnetometers are used to measure the strength and/or direction of a magnetic eld. There are two main types of magnetometers, vector and total eld, examples of each are given in table O.2. Vector magnetometers measure a single component of the magnetic eld produced, whist total eld magnetometers measure the overall eld strength in no particular direction (Waschl, 1994). The magnetometer detects the magnetic eld associated with the dipole moment of ferrous objects, such as a metallic landmine. The dipole moment of a ferrous object may be a permanent moment, based on the thermal and magnetic history of the object, or an induced moment from the object ambient magnetic eld (Waschl, 1994). Magnetometers are capable of operating over a a range of 3

277

Figure O.2: GPR pulse detection and object output signatures (Media, 2009) Table O.2: Various Vector Magnetometers Fluxgate Superconducting Quantum Interference Devices (SQUID) magnetometer types Total Field Magnetometers Proton Precision Caesium Vapour

meters above or below ground, with a sensitivity as low as 0.035 nT (Dr. Ivan Hrvoic, 2002), which is sucient to detect small objects, such as the ferrous ring pin in a composite landmine. The high sensitivity of the magnetometer results in a large degree of background noise from magnetite bearing soil, shrapnel or scattered ferrous objects. Large amount of noise can limit the eectiveness of the magnetometer and result in large numbers of false positives when searching for landmines, especially when searching for low metallic content composite landmines (Waschl, 1994). The application of magnetometers to both direct and mechanically assisted detection have solid advantages but many complications. One distinct advantage in the application to mechanically assisted detection is the large detection range eliminating any issues with sensor ground clearance and penetration of a substantial air gap. However its application is further complicated by its high sensitivity, in unwanted interference, false positives and more so isolation from moving metallic components, such as a petrol piston engine. Moving components are particularly problematic as whilst background noise from ferrous deposits in the soil may be ltered using a low pass lter, based on estimated interference levels, interference from moving metallic components cannot be isolated so easily. Hence to prevent interference from moving components physical shielding is often required, usually in the form of a Faraday Cage or an extended arm creating considerable distance between the sensor and the moving components.

Metal Detector
Metal detectors operate using electromagnetic conduction to detect conducting metallic objects. The metal detector coil produces a time varying magnetic eld, from an alternating current, which then induces an eddy current in the conducting object. The induced current subsequently produces a secondary magnetic eld which may then be detected from a receiver coil (Waschl, 1994). The operation of an emitter and receiver coil is shown in Figure O.3, where the primary eld is the emitted magnetic eld and the secondary eld the induced magnetic eld. All metal detectors operate under the same principle described previously, however there are two main types of detectors, continuous wave and pulse detectors. Continuous wave metal detectors operate in the kHz range and superimpose the secondary magnetic eld on the primary one. Which

278

APPENDIX O. LANDMINE DETECTION SENSORS

Figure O.3: Metal detector coil operation, with an outer primary emitter coil and an inner secondary receiver coil (Trevelyan, 2008) is subsequently detected, and the secondary eld separated and analysed using electronic methods. The pulse metal detector also operates in the kHz range, however its secondary eld is only detected after the primary has decayed suciently, removing subsequent complex electronic analysis (Waschl, 1994). In general and specically for mine detection, pulse metal detectors are considered the better of the two metal detector types. As unlike continuous wave metal detectors they require no primary eld balancing and are more versatile to environmental changes (Waschl, 1994). Metal detectors are the most widely used demining device, besides the mechanical prodder, (Das et al., 2002) and have been since World War 2. However they are not ideal to use, as they are extremely slow, dangerous and labour intensive to use. Placing the operator in close proximity to the landmines and having to check the location of each potential landmine, with a high percentage of false positives (Oce, 2001). False positives may be produced for any ferrous material, such as mineral deposits, shrapnel and unexploded ordinance. There are a number of factors considered when producing a detection signal, such as the objects size, shape, orientation, depth, material and soil properties. The most important of which, being the soil properties, in particular the soils magnetic susceptibility and to a lesser degree, electrical conductivity, aecting all of the subsequent parameters, especially depth. To the extent that in some soil types the metal detector can not penetrate the soil to the target depth (Das et al., 2002). Metal detectors are further limited from the introduction of composite landmines. Although highly sensitive to metallic landmines, they are not able to reliably locate composite landmines, due to their only metallic component being a small ring pin (Stull, 1997). There is no soil theory behind the operation of metal detectors, rather a number of guidelines. The range of a metal detector is known to be proportional to the coil diameter and size of the detection object, with a larger coil and large object penetrating further then a smaller coil and a large object. However the smaller coil is known to have more accuracy, able to detect smaller objects than the larger coil. The eect of the target object size on detection range my be seen in Figure O.4, where each coloured line represents the maximum detection range of various objects ranging in size. A larger object is not considered one with more mass, but one with more surface area, able to produce large induced magnetic elds (Trevelyan, 2008). Depth and accuracy are also known to be aected by many operational factors, such as speed, ground clearance, angle and discrimination.

279

Figure O.4: The maximum detection range from a 9 inch coil, with varying object size, where the red line is the largest object and the blue line the smallest (Schmidts, 2006)

This Page Intentionally Left Blank

Appendix P Breakdown of Hours and Costs

Team Member March April May June July August September October Costs Salary Indirect Direct Total Labour Total Expenditure Project Total Jarrad 69 111 126 44 120 170 179 174
$25,818 $33,563.4 $7,745.4 $67,126.8

Paul 71 113 124 45 111 151 180 175

$25,220 $32,786 $7,566 $65,572

Adrian 60.5 101 135 43 116 169 178 181

$25,571 $33,242.3 $7,671.3 $66,484.6

Rahim 96 122 150 50 130 180 190 191

$28,834 $37,484.2 $8,650.2 $74,968.4

Beau 97.5 128.5 162 55 143 196 201 207

$30,940 $40,222 $9,282 $80,444

Reuben 68 105 140 47 125 175 181 187

$26,728 $34,746.4 $8,018.4 $69,492.8

Total 462 680.5 837 284 745 1041 1109 1115

$163,111 $212,044.3 $48,933.3 $424,088.6 $10,494.04 $434,583.64

Table P.1: Hours and Costs Breakdown

281

This Page Intentionally Left Blank

Appendix Q Sensitivities of Lift and Thrust

Power Requirements
Weight minimisation is a key feature of hovercraft design. While ideally the overall weight of a hovercraft should be as low as is possible, in reality, there are only so many resources that should be invested to this end. As such, sensitivity analysis is a valuable design tool that aids the designer when comparing the selection of certain congurations and items. The following section considered two such analyses, being the sensitivity of propulsion and lift requirements to the overall weight of platform.

Lift Power
The lift system must produce a steady state cushion pressure that will provide enough force to counteract the weight of the craft, shown by equation: Pc = f (M, g ) (Q.1)

Hence the lift system must provide sucient ow rate and pressure to establish the design cushion pressure and maintain it as air is lost around the perimeter of the craft. Equation Q.14 converts the ow rate and pressure to a single measure of power for the lift system. Pressure is specied by the weight of the craft as in Equation Q.2 and ow rate by Equation Q.3. Combining these two equations with Equation Q.14 produces Figure Q.1 (Dc = 0.5, hDL = 15 mm, Ac = 1.6, Perimeter = 5.0 m, safety factor of 1.5 applied to Q) for various masses. This graph is approximately linear, giving a sensitivity of 29 W/kg for the lift system.

Pc = Q=

Mg Ac 2Pc /hDL Cdc

(Q.2) (Q.3) (Q.4)

Sensitivity of Propulsion Requirements

The gross thrust of the system is the sum of the net thrust and the aerodynamic, cushion, terrain and skirt-terrain interaction drag of the system is shown in Equation Q.5. TGross = TN ET + DA + DC + DT + Df (Q.5)

For the rest of this analysis, the skirt-terrain interaction drag will be neglected as the interaction is too complex to examine analytically. 283

284

APPENDIX Q. SENSITIVITIES OF LIFT AND THRUST

2.2

2.0

Power Lift System (kW)

1.8

1.6

1.4

1.2

1.0

100

Figure Q.1: Sensitivity of Lift Power to Mass

As such, the gross thrust is a function of the mass of the craft, acceleration, slope of the terrain, wind speed, daylight clearance and platform area, shown by: TGross = f (M, a, , V , h, S ) Hence, the gross thrust can be rewritten: 1 TGross = (ma) + ( V 2 ) + DC + (mgsinT ) 2 (Q.6)

(Q.7)

Cushion drag is relatively constant and maxes out at 2.75 N, as neither the ow rate or velocity of escaping air from the cushion is highly sensitive to thrust (see Section Q.1 for these relations). Cushion drag is given by: DC = QV (Q.8)

Thrust sensitivity to Weight Taking the derivative of thrust with respect to weight gives Equation Q.9. dT dTN ET dDA dDC dDT = + + + dm dm dm dm dm = a + gsinT Using the small angle approximation, Equation Q.9 becomes:

(Q.9)

285

dT = a + gt dm = 0.139 + 0.171T ; [t ] = Degrees T = (0.139 + 0.171T )m

(Q.10)

Equation Q.10 displays the sensitivity of the thrust to weight in terms of the incline (expressed in degrees).

Sensitivity of Thrust to Wind Speed Assuming that the craft is operating in a head wind, the drag force associated with this condition is expressed by the equation: 1 DA = S(V + V )2 2 dDA = S(V + V ) dV

(Q.11)

Based upon operational velocity of 5 km/hr (1.39 m/s), a frontal area is estimated as a rectangle of area 0.88 m2 (1.1 by 0.8 m). With the fairing, the aerodynamic shape should be somewhere between a wedge (S = 0.5) and a block (S = 1.05), favoring the block shape so a shape factor of 0.8 was determined to be appropriate but still conservative. Substituting these values into Equation Q.11 gives Equation Q.12. dDA = (1.36 + 0.98V )V dV

(Q.12)

Figures Q.2 and Q.3 display the thrust for various slopes, and wind speeds respectively. Using these graphs for various conditions allows for rapid identication of cushion thrust. It is important to recognise the sensitivity of thrust to weight, but a more important design parameter for propulsion is that of installed power. Based on the hovercraft propulsion system of a single ducted fan, the thrust power requirements for the platform were determined using Equations Q.13 to Q.15.

T = QV PA = Q (Ps + Pd ) 1 PM = f Q Ps + V 2 2

(Q.13) (Q.14) (Q.15)

This makes engine power a function of both mass ow rate and velocity of outgoing air. Assuming isotropic expansion of the ow out of the fan duct, the velocity can be calculated from Equation Q.16. 2(Ps + Pd )

V =

(Q.16)

Combining Equation Q.15 with Equation Q.16 the power requirements for the propulsion system can be calculated by the sum of the individual thrust requirements. Based on a static pressure of 100 Pa,

286

APPENDIX Q. SENSITIVITIES OF LIFT AND THRUST

1 0 0 9 0 8 0 7 0

N o S 1 d e 2 d e 3 d e 4 d e 5 d e g g g g g

lo p e

T h ru s t (N )

6 0 5 0 4 0 3 0 2 0 1 0 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0

W e ig h t ( k g )

Figure Q.2: Sensitivity of Thrust to Weight and Terrain Slope the sensitivity of power to thrust is 0.0105 hp/N thrust or 7.8 W/N. Figure Q.4 displays the power requirement of a propeller operating at a given total pressure, to produce a certain thrust.

287

7 0 6 0 5 0

A e r o d y n a m ic D r a g ( N )

4 0 3 0 2 0 1 0 0 0 1 2 3 4 5 6 7 8 9 1 0

W in d S p e e d ( m /s )

Figure Q.3: Sensitivity of Thrust to Head Wind Speed

2 .0

1 .5

P o w e r (H P )

5 0 P 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0

a P a P a P a P a P a

1 .0

0 .5

0 .0 0 2 0 4 0 6 0 8 0 1 0 0

T h ru s t (N )

Figure Q.4: Eect of Mass of Propulsion Power Requirements

Technical Analysis from A to Z, 2nd Edition
From Everand
Technical Analysis from A to Z, 2nd Edition
Steven B. Achelis
4.5/5 (19)
Unmanned Aircraft Design: A Review of Fundamentals
100% (7)
Unmanned Aircraft Design: A Review of Fundamentals
211 pages
Blueprint Reading: Construction Drawings for the Building Trade
From Everand
Blueprint Reading: Construction Drawings for the Building Trade
Sam Kubba
No ratings yet
THE LTSPICE XVII SIMULATOR: Commands and Applications
From Everand
THE LTSPICE XVII SIMULATOR: Commands and Applications
Gilles Brocard
5/5 (1)
Instruction Operation and Maintenance Manual Unit Control Panel
100% (5)
Instruction Operation and Maintenance Manual Unit Control Panel
1,060 pages
6th Central Pay Commission Salary Calculator
100% (436)
6th Central Pay Commission Salary Calculator
15 pages
Marisol Mcdonald Doesnt Match-Long Form
No ratings yet
Marisol Mcdonald Doesnt Match-Long Form
2 pages
Fritzing for Inventors: Take Your Electronics Project from Prototype to Product
From Everand
Fritzing for Inventors: Take Your Electronics Project from Prototype to Product
Simon Monk
No ratings yet
NDRT CDR Compressed
No ratings yet
NDRT CDR Compressed
260 pages
Basic Research and Technologies for Two-Stage-to-Orbit Vehicles: Final Report of the Collaborative Research Centres 253, 255 and 259
From Everand
Basic Research and Technologies for Two-Stage-to-Orbit Vehicles: Final Report of the Collaborative Research Centres 253, 255 and 259
Dieter Jacob
No ratings yet
Developing Intelligent Agent Systems: A Practical Guide
From Everand
Developing Intelligent Agent Systems: A Practical Guide
Lin Padgham
3/5 (1)
Teardowns: Learn How Electronics Work by Taking Them Apart
From Everand
Teardowns: Learn How Electronics Work by Taking Them Apart
Bryan Bergeron
No ratings yet
Structural Health Monitoring
From Everand
Structural Health Monitoring
Daniel Balageas
No ratings yet
Pollution Prevention: Methodology, Technologies and Practices
From Everand
Pollution Prevention: Methodology, Technologies and Practices
Kenneth L. Mulholland
No ratings yet
Re-Cording Lives: Governing Asylum in Switzerland and the Need to Resolve
From Everand
Re-Cording Lives: Governing Asylum in Switzerland and the Need to Resolve
Ephraim Pörtner
No ratings yet
2009 Final Report Compressed PDF
No ratings yet
2009 Final Report Compressed PDF
298 pages
2009 Final Report Compressed
No ratings yet
2009 Final Report Compressed
298 pages
Nuclear Commerce: Its Control Regime and the Non-Proliferation Treaty
From Everand
Nuclear Commerce: Its Control Regime and the Non-Proliferation Treaty
Thomas Berndorfer
No ratings yet
Benefits of semantic data models. A study in the European goods transport industry
From Everand
Benefits of semantic data models. A study in the European goods transport industry
Andreas A. Pelekies
No ratings yet
TFG Joan Manero Álvarez Report
No ratings yet
TFG Joan Manero Álvarez Report
67 pages
FULLTEXT02
No ratings yet
FULLTEXT02
51 pages
Launch Canada 2024 Final Report Competing Team 16 Oronos Polytechnique
No ratings yet
Launch Canada 2024 Final Report Competing Team 16 Oronos Polytechnique
91 pages
859 Design and Build of A UAV With Morphing Config
No ratings yet
859 Design and Build of A UAV With Morphing Config
341 pages
AS5213 Group16 Submission6
No ratings yet
AS5213 Group16 Submission6
86 pages
Collapse analysis of externally prestressed structures
From Everand
Collapse analysis of externally prestressed structures
Jens Tandler
No ratings yet
Team - 4 - Project Report
No ratings yet
Team - 4 - Project Report
66 pages
Report FSC Ae648a
No ratings yet
Report FSC Ae648a
38 pages
CNC Machining Handbook: Building, Programming, and Implementation
From Everand
CNC Machining Handbook: Building, Programming, and Implementation
Alan Overby
No ratings yet
Hyperloop Design Report Makers UPV Team
No ratings yet
Hyperloop Design Report Makers UPV Team
107 pages
Programming the Intel Galileo: Getting Started with the Arduino -Compatible Development Board
From Everand
Programming the Intel Galileo: Getting Started with the Arduino -Compatible Development Board
Christopher Rush
5/5 (1)
Launch Canada 2024 Final Report Competing Team 16 Oronos Polytechnique
No ratings yet
Launch Canada 2024 Final Report Competing Team 16 Oronos Polytechnique
223 pages
ZEGLIN 1999 Bow Legged Hoppin Robot THESIS
No ratings yet
ZEGLIN 1999 Bow Legged Hoppin Robot THESIS
163 pages
Glow Discharge Plasmas in Analytical Spectroscopy
From Everand
Glow Discharge Plasmas in Analytical Spectroscopy
R. Kenneth Marcus
No ratings yet
AMelgosa BScThesis PDF
No ratings yet
AMelgosa BScThesis PDF
139 pages
Programming the Photon: Getting Started with the Internet of Things
From Everand
Programming the Photon: Getting Started with the Internet of Things
Christopher Rush
5/5 (1)
Final Report UCBC
No ratings yet
Final Report UCBC
27 pages
IAES-Rapport
No ratings yet
IAES-Rapport
63 pages
Gray Hat Hacking the Ethical Hacker's
From Everand
Gray Hat Hacking the Ethical Hacker's
Çağatay Şanlı
5/5 (1)
Human Powered Aircraft Report
No ratings yet
Human Powered Aircraft Report
83 pages
Continuous Auto UAV Inspections For FPSO Vessels
No ratings yet
Continuous Auto UAV Inspections For FPSO Vessels
100 pages
2018 Critical Design Review (CDR)
No ratings yet
2018 Critical Design Review (CDR)
149 pages
Smart Card Applications: Design models for using and programming smart cards
From Everand
Smart Card Applications: Design models for using and programming smart cards
Wolfgang Rankl
No ratings yet
Quasi-Monte Carlo Methods in Finance: With Application to Optimal Asset Allocation
From Everand
Quasi-Monte Carlo Methods in Finance: With Application to Optimal Asset Allocation
Mario Rometsch
No ratings yet
Facility Management: Business Process Integration
From Everand
Facility Management: Business Process Integration
Alexander Redlein
No ratings yet
Design of Automated Drone Chaser I
100% (1)
Design of Automated Drone Chaser I
37 pages
A Mobile Self-Leveling Landing Platform for Small-Scale UAVs
No ratings yet
A Mobile Self-Leveling Landing Platform for Small-Scale UAVs
75 pages
Letter of Transmittal
No ratings yet
Letter of Transmittal
51 pages
2018 PDR Report
No ratings yet
2018 PDR Report
95 pages
Structural Dynamics in Industry
From Everand
Structural Dynamics in Industry
Alain Girard
No ratings yet
UC Berkeley Space Technologies and Rocketry Preliminary Design Review Project U.R.S.A
No ratings yet
UC Berkeley Space Technologies and Rocketry Preliminary Design Review Project U.R.S.A
86 pages
Sample Project Report
No ratings yet
Sample Project Report
49 pages
ATC Digital
No ratings yet
ATC Digital
135 pages
Open Data Structures: An Introduction
From Everand
Open Data Structures: An Introduction
Pat Morin
4/5 (4)
Erickson 2
No ratings yet
Erickson 2
37 pages
Disturbance rejection control for bipedal robot walkers
From Everand
Disturbance rejection control for bipedal robot walkers
Jaime Arcos Legarda
No ratings yet
Biennial Report on Operations Evaluation
From Everand
Biennial Report on Operations Evaluation
The World Bank
No ratings yet
Biennial Report on Operations Evaluation: Assessing the Monitoring and Evaluation Systems of IFC and MIGA
From Everand
Biennial Report on Operations Evaluation: Assessing the Monitoring and Evaluation Systems of IFC and MIGA
The World Bank
No ratings yet
Quality and Performance Measurement in National Sport-Governing Bodies
From Everand
Quality and Performance Measurement in National Sport-Governing Bodies
Frank Daumann
No ratings yet
Striving for Clean Air: Air Pollution and Public Health in South Asia
From Everand
Striving for Clean Air: Air Pollution and Public Health in South Asia
The World Bank
No ratings yet
Tips & Traps for Building Decks, Patios, and Porches
From Everand
Tips & Traps for Building Decks, Patios, and Porches
R. Dodge Woodson
No ratings yet
Programming Arduino: Getting Started with Sketches, Third Edition
From Everand
Programming Arduino: Getting Started with Sketches, Third Edition
Simon Monk
No ratings yet
4-AS-1.3.3-Project 2022-23
No ratings yet
4-AS-1.3.3-Project 2022-23
115 pages
A Guide to Polymeric Geomembranes: A Practical Approach
From Everand
A Guide to Polymeric Geomembranes: A Practical Approach
John Scheirs
No ratings yet
Control Inst
No ratings yet
Control Inst
481 pages
Yokogawa Centum CS3000 Overview
No ratings yet
Yokogawa Centum CS3000 Overview
71 pages
Centum CS 3000 R3
No ratings yet
Centum CS 3000 R3
11 pages
Valves Pres
No ratings yet
Valves Pres
287 pages
Ovation Controller
No ratings yet
Ovation Controller
12 pages
Columns
No ratings yet
Columns
12 pages
Thrwat Books
No ratings yet
Thrwat Books
705 pages
Chap II - Statics
No ratings yet
Chap II - Statics
41 pages
Chap I - Statics
No ratings yet
Chap I - Statics
37 pages
Chap III - Statics
No ratings yet
Chap III - Statics
55 pages
Building A CycleGAN Model Using Tensorflow 2
No ratings yet
Building A CycleGAN Model Using Tensorflow 2
9 pages
Solutions For Internal Arc Protection According To IEC 62271-200 in MV Switchgear
100% (1)
Solutions For Internal Arc Protection According To IEC 62271-200 in MV Switchgear
5 pages
AGM Parent Letter 2015 (English)
No ratings yet
AGM Parent Letter 2015 (English)
5 pages
Course Schedule LMCE3071 - Student's PDF
No ratings yet
Course Schedule LMCE3071 - Student's PDF
2 pages
Part 2 Vitish
No ratings yet
Part 2 Vitish
15 pages
Laptop User Agreement Form
No ratings yet
Laptop User Agreement Form
1 page
Auditing Computer-Based Information Systems
No ratings yet
Auditing Computer-Based Information Systems
15 pages
Case Study 34 Closed Head Injury
No ratings yet
Case Study 34 Closed Head Injury
11 pages
(Ebook) Radiology Made Easy by Arpan K. Banerjee ISBN 9780521676359, 0521676355 All Chapters Instant Download
No ratings yet
(Ebook) Radiology Made Easy by Arpan K. Banerjee ISBN 9780521676359, 0521676355 All Chapters Instant Download
82 pages
UG Reading Comprehension
No ratings yet
UG Reading Comprehension
47 pages
Reflection Paper - Waste Segregation
No ratings yet
Reflection Paper - Waste Segregation
1 page
Ips Thesis Guideline Usm
100% (2)
Ips Thesis Guideline Usm
5 pages
Bridge Technical Note 2018 - 010 Integral and Semiintegral Bridges Jan 2018
No ratings yet
Bridge Technical Note 2018 - 010 Integral and Semiintegral Bridges Jan 2018
2 pages
160231514 Design a Layout Plan for Speciality Units to Print 2012
No ratings yet
160231514 Design a Layout Plan for Speciality Units to Print 2012
13 pages
Overcoming Voltage Fluctuation in Electric Vehicles by Considering Al Electrolytic Capacitor-Based Voltage Stabilizer
No ratings yet
Overcoming Voltage Fluctuation in Electric Vehicles by Considering Al Electrolytic Capacitor-Based Voltage Stabilizer
27 pages
Design and Optimization of Dimethyl Ether Production From Crude Glycerol in A Reactive Distillation
No ratings yet
Design and Optimization of Dimethyl Ether Production From Crude Glycerol in A Reactive Distillation
9 pages
MORETTI FORNI - USA 2020 (BROCHURE 2)
No ratings yet
MORETTI FORNI - USA 2020 (BROCHURE 2)
20 pages
Brain-computer Interface Documentation
No ratings yet
Brain-computer Interface Documentation
29 pages
Ibk 2021 e - 5
No ratings yet
Ibk 2021 e - 5
131 pages
Pricing Strategy Modules
No ratings yet
Pricing Strategy Modules
34 pages
Online Quiz 1 Final
No ratings yet
Online Quiz 1 Final
2 pages
1A AcR Midterm - Sample 2
No ratings yet
1A AcR Midterm - Sample 2
7 pages
21st Century Literature - Quarter 1 - Weeks 1 and 2
100% (1)
21st Century Literature - Quarter 1 - Weeks 1 and 2
32 pages
READING PASSAGE-WPS Office
No ratings yet
READING PASSAGE-WPS Office
6 pages
M.Tech DCN 2014 SCHEME
No ratings yet
M.Tech DCN 2014 SCHEME
47 pages
The Fast Medium Wave Krelus Infrar Ed-Heaters
No ratings yet
The Fast Medium Wave Krelus Infrar Ed-Heaters
8 pages
Holding Certificate-04-12-2024
No ratings yet
Holding Certificate-04-12-2024
1 page
FESCO ONLINE BILL - Compressed
No ratings yet
FESCO ONLINE BILL - Compressed
1 page
Phenotypic Characterization of White Fulani (Bunaji) and Bunaji X Friesian Breed of Cattle From National Animal Production Research Institute (NAPRI) Cattle Herd From Nigeria
No ratings yet
Phenotypic Characterization of White Fulani (Bunaji) and Bunaji X Friesian Breed of Cattle From National Animal Production Research Institute (NAPRI) Cattle Herd From Nigeria
5 pages