Fibre Optic Sensing Technology and

Applications in Wind Energy

Sandia Blade Workshop 2008

14-05-08

Phil Rhead
Presentation Contents

• Introduction to Insensys Limted

• Insensys Technology
• System Overview and Key Components
• Verification and Reliability
• Application 1 - Test and Measurement
• Application 2 - Individual Pitch Control
• Application 3 - Rotor Condition Monitoring
• Summary

Proprietary Insensys 2008 Slide 2

Introduction to Insensys
Insensys Introduction
- Company Overview

• Founded in 2002

• 40 staff across 3 offices located

in the UK

• Focussed on 2 key market areas

– Wind Energy
– Aerospace

• Oil and Gas division sold to

Schlumberger in 2007

• World class engineering skills

– Fibre optics
– Composite design
– Composite manufacture

Proprietary Insensys 2008 Slide 4

Insensys Introduction
- Wind Energy Applications & Market Status
Focus
Supply advanced load measurement technology
to the Wind Turbine Industry enabling improved
Wind Turbine performance and reliability

Key Application Areas

1) Individual Pitch Control (IPC)
2) Rotor Condition Monitoring
3) Test and Measurement Applications

Market Experience
•Insensys system is designed into production
turbines between 1.5 and 6 MW with multiple
turbine manufacturers
•Systems are currently being supplied in series
quantities
•System is currently under test in 14 different
turbine platforms
•Deployed in blades from 27m to 60m

Proprietary Insensys 2008 Slide 5

Insensys Technology
Time Division Multiplexed (TDM)
Sensor Interrogation
Insensys Technology
- Time Division Multiplexing Schematic

OEM-1030 Measurement Unit Blades / Sensor Array

Emitted light Reflected light

from light source from FBGs

Light
Source
Sensor1
Optical
Detector
t1 t2 t3 t4 t5 t6 t7 t8

Signal Reflected light

Processing received at detector
in timeslots t1 to t8
TDM electronics

Serial
Transmission Sensor6
to controller

Proprietary Insensys 2008 Slide 7

Insensys Technology
- Typical System Configuration

Sensor Arrays
installed in the blade
( 4 per blade)

OEM-1030
Measurement unit
located in turbine hub

The single system provides

information for both turbine
control and health
monitoring applications
Optical
Interconnection
Cables (3 per turbine)

Proprietary Insensys 2008 Slide 8

Insensys Technology
- Key Advantages of Fibre Optic Sensing
Sensors
• Optical fibre Bragg grating sensors – non disruptive to the laminate
• Absolute strain measurement with no drift or de-bonding
• Immune to EMI and lightning effects in blade environment
• Installed during blade build or retrofitted to operational machines
• Sensor quantities, locations and spacing can be custom designed to
suit exact turbine dimensions and sensing requirements

Measurement Unit
• Designed specifically for hub environment - No moving parts
• High speed & low measurement latency
• +/- 4500 microstrain measurement range
• Low power (3W typical) and low weight (< 3Kg)

Cable System
• IP65 cable system when connected
• All interconnection cables are replaceable by the field service team
without the need to recalibration

Proprietary Insensys 2008 Slide 9

System Verification & Reliability
Testing
System Verification Testing

Sensors for control or SHM applications must be highly reliable!

• Performance Testing
– Sensor patch testing > 45 million cycles of 0 – 1000 microstrain
– Dynamic coupon fatigue test > 2 million cycles +/- 5000
microstrain range
– Active blade testing – 1 x 10^6 cycles during an active blade tests
– Static blade test – sensors used for multiple GL certification tests
– No sensor failures, degradation or de-bonding in any of these tests

• Laboratory Testing
– Lightning strike tests, Impact tests
– Environmentally tested to IEC standards - Shock, vibration,
thermal cycling etc

• Design Verification
– MTBF in excess of 20 years (from calculation and hours in
service)

Proprietary Insensys 2008 Slide 11

Product Reliability
– In Field Testing
Long term comparisons have been carried our with conventional electrical strain gauges
instrumented by DEWI GmbH, WindTest GmbH and Garrad Hassan!

WindTest Grevenbroich DMS Sensor Insensys FBG Sensor Patch

Proprietary Insensys 2008 Slide 12

In Field Testing
- Data Comparison
Edgewise Data Sample Flapwise Data Sample

a) Time Series a) Time Series

b) Time Series (Difference) b) Time Series (Difference)

Proprietary Insensys 2008 Slide 13

In Field Testing
- Data Comparison (zoom)
Edgewise Data Flapwise Data

a) Time Series a) Time Series

b) Correlation b) Correlation

Proprietary Insensys 2008 Slide 14

Fibre Optic Sensor
Deployment Techniques
Sensor Deployment Techniques
- Overview
• Many people have proven that bare fibre installation doesn’t work!

• Insensys have developed specific deployment techniques for Wind Energy to

ensure: accuracy of installation, high yield, reliability and simplicity of installation

• Multiple sensor deployment techniques developed to suit different blade

manufacturing processes, materials and applications

• Embedded during blade infusion

• Retrofit to completed blades or assemblies (in-factory / up-tower)

• Blade manufacturing process - Pre-preg (including ATL), infusion, hand layup

• Blade materials - GRP, CFRP, Hybrids

• All deployment processes

– utilise standard materials and blade manufacturing processes
– are designed to minimise intrusion into blade production process

Proprietary Insensys 2008 Slide 16

Deployment Techniques
- Retrofit Sensor Installation

•Applied internally or externally

•Applied to shells, spars, webs & root sections
•Simple customisation of positions
•Standard arrays from stock or fully configurable
•Standard method for prototype test and measurement applications and blade testing
•Cost effective for series application in low labour rate countries

Proprietary Insensys 2008 Slide 17

Deployment Techniques
- Retrofitted to Shells (Secondary Infusion)

•Sensor applied to blade LE and TE post shell manufacture

•Rapid / reliable / low cost installation technique for series production

Proprietary Insensys 2008 Slide 18

Deployment Techniques
- Custom Sensor Application (deep install)

•Custom sensor patch designed for measurement deep inside laminate 30 mm

•27m long, 13 sensors (tree effect)
•Installed in shells prior to central belt being installed

Proprietary Insensys 2008 Slide 19

Deployment Techniques
- Bonding to Shells (Primary infusion)

•Applied during primary blade infusion

•Rapid deployment of multiple sensors – sensors treated as per any other layer
•Can be located near blade surface (deep) or near inner skin (shallow)
•Cost effective for series deployment in high labour rate countries
•Specific care must be taken when designing connection points!

Proprietary Insensys 2008 Slide 20

Example Wind Energy Applications
Applications Overview

• Prototype Turbine and Blade Measurement Campaigns

• Individual Pitch Control
• Structural Health Monitoring

• System designed as modular platform with common

architecture
• Enables dual functionality to be achieved

Proprietary Insensys 2008 Slide 22

Insensys Technology
- Blade Installation Schemes

Proprietary Insensys 2008 Slide 23

Application 1
Prototype Turbine and Blade Test
Test and Measurement
- Example Applications
• Used for in place of a conventional electrical strain gauge
– Simple installation and connectivity
– Immunity to lightning and EMI
– Highly reliable – no de-bonding or sensor fatigue
– Data use for design validation and correlation of loads with FEA models during the turbine
design phase

• Blade measurements
– On turbine data collection – multiple points per blade
– Static proof test
– Dynamic blade test
– Blade subsection / panel test

• Structural component measurements

– Low speed shaft (bending and torsion)
– Tower (bending and torsion)
– Hub casing (strain)
– Gearbox and bedplate (strain)

Proprietary Insensys 2008 Slide 25

Test and Measurement
- Dynamic Fatigue Test (Time Series Data)

•Time series data ( 24 sensors dynamic fatigue test)

•Generating data is the easy part!
•Analysis and reporting needs effort and software
tools!

Proprietary Insensys 2008 Slide 26

Test and Measurement
- Dynamic Fatigue Test (Time Series Analysis)

Measured strain profile along a blade at 7

sensor locations and under 3 different
load conditions

Proprietary Insensys 2008 Slide 27

Test and Measurement
- Ultimate Load Test

Measured strain profile along a blade at 15

sensor locations during a static fatigue test

Proprietary Insensys 2008 Slide 28

Prototype Turbine Measurement
- Low Speed Shaft Design Validation

• Used during design phase to validate FEA models / design of low speed shaft

Sensor Patch Located on

Main Shaft

Proprietary Insensys 2008 Slide 29

Application 2
Individual Pitch Control
Individual Pitch Control
- Current Status of the Market
• Turbines rotors are increasing in size and are being installed on more complex
terrains. This is leading to:
– Increased asymmetric loading across the rotor
– Increased yaw and tilt moments
– Due to wind speed and spatial variations - (stochastic process)

• Multiple load reduction strategies have been proposed

– Any successful load reduction strategy must be based on measurements

• Individual Pitch Control can reduce loads significantly!

– Blade loads typically reduced by 10 – 20%
– Main shaft loads typically reduced by 20 – 30%
– Reduced tower and yaw bearing loads, particularly with 2P based -IPC

• IPC is already being deployed in series production

Proprietary Insensys 2008 Slide 31

Individual Pitch Control
- Process Requirements
Z-1

Measured fore-aft Integrator Collective Pitch

Filters + Gain schedule
acceleration
Control Loop
Pitch demand
output for torque
controller
Measured PID with gain schedule,
Filters Collective pitch
generator speed pitch position and rate + angle demand
limits, and extra inputs

Inputs from torque

controller

Individual Pitch
d-axis Control Loop
Blade 1 pitch
Blade 1 My Filter PI
demand Edgewise and flapwise
Inverse
Blade 2 My
d,q-axis
d,q-axis
Blade 2 pitch moment input data
transformation demand
transformation
Blade 3 pitch
Blade 3 My Filter PI
demand Out of plane moments
q-axis
transformed to non-
rotating d-q axes
Limit
Rotor azimuth
schedule
Non-rotating d-q axis pitch
Maximum
demands
d, q-axis
pitch limit
Final pitch demands

Proprietary Insensys 2008 Slide 32

Individual Blade Pitch Control
- Example Data

Edgewise and flapwise

moment input data

Pitch angles,
rotor azimuth, filtering

Out of plane Moments

transformed to non-
rotating d-q axes

PI control

Non-rotating d-q axis

pitch demands

Rotor azimuth,
Collective pitch demand

Final pitch demands

Proprietary Insensys 2008 Slide 33

Individual Pitch Control
- Hardware Implementation
1) Fibre optic sensors installed in the blade root
2) Measurement unit installed in the turbine hub

3) Data communicated to PLC cabinet

4) IPC calculations completed in PLC
5) Turbine blades pitch and load signal changes completing
feedback loop
6) New pitch angle commanded to pitch actuators

Proprietary Insensys 2008 Slide 34

Individual Blade Pitch Control
- Options for a Turbine Manufacturer
The load reductions from IPC can be leveraged in multiple ways by a turbine
manufacturer;

1) Cost Optimisation
- Turbine’s structural components can be designed for lower loads
- Lighter, cheaper parts, reduced transportation and installation cost

2) Modified Wind Class or Installation Conditions

- Increase rotor diameter for higher energy yield
- Installation in more turbulent locations i.e. more densely packed on a wind farm

3) Improved Turbine Reliability

- Utilise improved rotor balancing from IPC
- Reduced loads on blades, bearings, gearbox and drive-train
- Increase MTBF

All options can improve the turbine performance!

Proprietary Insensys 2008 Slide 35

Application 3
Structural Health Monitoring
 Structural Health Monitoring
- As it is today!
• A large number of parameters are monitored on
modern wind turbines
– Drive train vibrations
– Generator oil condition
– Pitch motor torques / pressures
– Wind and machine parameters

• Very little if anything is monitoring on the blades

or rotor

• Rotor is subjected to:

– Instantaneous load variations
– Fluctuating load pattern
– Frequent peak loads
– Rotor torque and axial thrust forces
– Extreme environmental conditions
– Acts of God!

Proprietary Insensys 2008 Slide 37

Images courtesy of SKF, ISET, Gram and Juhl
Rotor Condition Monitoring
- Utilising Blade Load Data
• Measuring information from the blades can reveal a great deal of additional
information about the performance of the turbine that can not be gained from
conventional SHM techniques

• Insensys has developed algorithms to provide additional condition monitoring

from blade load information that is complimentary to existing information

• Blade strain data can be processed in many different ways to real information
about the turbine and blade performance

• Data can be issued to PLC, linked to an existing Condition Monitoring System

enabling direct correlation between cause and effect or logged for latter analysis

Proprietary Insensys 2008 Slide 38

Structural Health Monitoring
- Overview
(click image below to play video)
• Blade performance data
– Strain and bending moments
– Load histories and extreme loads

• Rotor performance data

– Imbalance / offset load
– Tilt moments / Yaw moments
– Mass / Aerodynamic
– L/D rations

• Performance history and defect detection

– Accumulative fatigue and residual
lifetime
– Material Loss
– Debond / delamination identification

• Lightning Strike Detection

• Blade Icing

Proprietary Insensys 2008 Slide 39

Structural Health Monitoring
- Lightning Detection and Measurement

• Insensys has developed a fibre-optic lightning detection and measurement system,

based on existing architecture
• Measures every lightning strike, on each blade, in real time, providing several key
intensity parameters Increases generating revenue by avoiding waiting for un-
necessary inspections, and scheduling required inspections
• Allows decision to be taken whether protection system, or blades, likely to have
been damaged in strike

Proprietary Insensys 2008 Slide 40

Structural Health Monitoring
- Ice Detection and Measurement
• Insensys is developing an ice detection and measurement system, based on
existing fibre-optic system architecture

• Key benefits:
– Enabling safe shutdown, preventing ice throwing
– Safe, automatic restart
– Minimising generation loss
– Avoiding rotor imbalance caused by icing
– Compliance with latest EU legislation on ice detection

Proprietary Insensys 2008 Slide 41

Summary
Summary

• Insensys fibre optic instrumentation is a proven, reliable technology for

blade strain & load measurement in Wind Turbines

• The benefits of using blade loads sensors for turbine control applications are
already well understood

• A number of turbine manufacturers today include IPC in their designs and

many others will shortly be following suit.

• Significant additional benefit can be achieved by monitoring the blades loads

and correlating the data with the data from the drivetrain monitoring system

• Àdvanced technologies and data processing techniques are being

developed toprovide manufacturers and operators with further functionality

Proprietary Insensys 2008 Slide 43

What Next?

• Whatever you guys throw at us next……….

– Multipart blades
– Blades with adaptive flaps / actuators
– Load shedding blades
– Next generation IPC

• The sensors are ready!!!!!!

Proprietary Insensys 2008 Slide 44

Thanks for listening!

www.insensys.com

[email protected]