Scribd
Upload
51 views

Robust Control Using Quantitative Feedback Theory: Prof. P. S. V. Nataraj

Robust control using Quantitative Feedback Theory (QFT) provides a method for designing controllers that ensure stability and performance in the presence of plant uncertainty. QFT involves shaping the open-loop transfer function to satisfy stability and performance specifications over a range of plant templates. The document presents two case studies applying QFT: a Boeing 707 aircraft longitudinal dynamics model and an F-16 fighter aircraft pitch axis model. QFT is shown to design controllers that meet specifications for both systems despite plant parameter uncertainty.

Uploaded by

jiksh
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
51 views

Robust Control Using Quantitative Feedback Theory: Prof. P. S. V. Nataraj

Robust control using Quantitative Feedback Theory (QFT) provides a method for designing controllers that ensure stability and performance in the presence of plant uncertainty. QFT involves shaping the open-loop transfer function to satisfy stability and performance specifications over a range of plant templates. The document presents two case studies applying QFT: a Boeing 707 aircraft longitudinal dynamics model and an F-16 fighter aircraft pitch axis model. QFT is shown to design controllers that meet specifications for both systems despite plant parameter uncertainty.

Uploaded by

jiksh
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PPT, PDF, TXT or read online on Scribd
You are on page 1/ 37

Robust control

using
Quantitative Feedback Theory

Prof. P. S. V. Nataraj

Systems and Control Engineering


IIT Bombay
The single loop feedback system

reference Input Output


disturbances disturbances disturbances
W V D
U
R E Y
F G P

reference error
signal signal Manipulated
pre-filter Control law signal Plant
Control hardware dynamics

N
H

Sensor
noise
Why Robust Control ?
 Fundamental issues in the design, analysis and
evaluation of control systems.
 Stability
 Performance
 Stability
 In the absence of external excitation, all signals in the
system decay to zero.
 It is an absolute requirement.
 Its absence causes signals to grow without bound,
eventually destroying and breaking down the plant.
Why Robust Control ? contd..

 Performance
 Ability to track reference signals closely
 Reject disturbances.

 Robust Control system


 Should preserve stability in the presence of various classes
of uncertainties.
 Should be capable of tracking all reference signals
belonging to a class of signals, without excessive error,
despite various types of uncertainties.
Why QFT ?
 One approach to robust control is QFT, started in
1960
 QFT is a natural extension of classical frequency
domain approaches.
 QFT is an engineering method for practical design of
feedback control systems.
 In QFT, controller can be obtained to satisfy
performance and stability specs over a range of plant
uncertainty.
Main objectives of QFT

 Get simple low order controllers


 Get minimum bandwidth controllers
 To avoid noise amplification
 To avoid resonance
 To avoid high frequency dynamics
Characteristics of QFT

 Amount of feedback is tuned to


 Amount of plant
 Magnitude of disturbance
 Level of performance specifications
 QFT offers direct insight into tradeoff between
 Stability and performance specs
 plant uncertainty and disturbance uncertainty
 controller complexity
 controller bandwidth
QFT design methodology

 Synthesis of Tracking, Disturbance, Stability


Models
 Specification of Plant Parameter Uncertainty
 Generation of Plant Templates
 Selection of Nominal Plant Po(s)
 Determination of Stability Contour
QFT design methodology (contd.)

 Generation & Integration of QFT Bounds


 Shaping of Nominal Loop Lo(s)=G(s)Po(s)
 Prefilter Design
 Analysis & Validation
Feedback Design with QFT
Define frequencies
Define Create frequency response data
problem data Define sampling time
Define nominal response
Define controller type
Define phase array for bounds.

Compute and
Compute bounds
manipulate
bounds.

Loop shaping
Design

Compare design to
specifications
Analysis

End
How is QFT different from LQG and H ?

 QFT is based on classical idea of shaping the open-


loop transfer function.
 It also differs the way in which uncertainty is
characterized – as gain-phase variations or templates
in Nichols chart
 Inclusion of phase uncertainty gives less conservative
solutions than with H or µ- synthesis methods.
Wide range of applications

 Handle plant uncertainty due to wide range of


operating conditions
 Use Plant models from experiments
 Design a single controller for different locations in a
flight envelope
 Design controller to meet several performance specs
 Meet hardware constraints
Wide range of applications contd..

 In last ten years, QFT has been successfully applied


to Flight Control System (FCS).
 for air-to-air automatic refueling
 for rotorcraft, remote pilot-less vehicle
 MIMO FCS for YF-16 using nonlinear QFT
Case Studies

 Boeing Commercial Aircraft

 AFTI/F16 fighter Aircraft


Basic Diagram of coordinate axes and
forces acting on an aircraft
 Angle of attack
 q : Pitch rate
 θ Pitch angle
 e : Elevator deflection
angle
 e : Flight path angle
Case Study 1

 Boeing 707 Commercial Aircraft


[http://www.engin.umich.edu/group/ctm/examples/pitch/Mpitch.html]
 Flight conditions
 Steady cruise at constant altitude and velocity
 Pitch angle is independent of speed of aircraft
Plant Specifications

SISO transfer function between pitch angle and


elevator deflection angle c
 ( s) ( as  b)
 3
 c ( s) ( s  cs 2  ds )
Parametric uncertainty :

a : [1.0935 1.2086]
b : [0.1685 0.1862]
c : [0.7020 0.7760]
d : [0.8750 0.9671]
Performance Specifications

 Rise time < 2 sec


 Settling Time < 10 sec
 Overshoot < 10 %
 Steady state Error < 2%
Generation of Templates
Intersection of Bounds
Loop Shaping
Loop Shaping contd.

 Controller transfer function

 s 
908.2  1
G (s)   18.77 
 s 
2
 s s
  1 2
  1
 300  400 400 
Filter Design
Filter Design contd..

 Filter transfer function

1
F ( s) 
 s  s 
  1  1
 6.842  14.36 
Controller design using
conventional methods
 PID technique:
 Kp = 2.01
 Ki = 2.99
 Kd = 4.00

 LQR technique (pole placement design):


 state-cost matrix (Q) = C’*C where C = [0 0 1]
 weighting factor (p) = 50
 performance index matrix (R) = 1
Controller Performance
Case Study 2

 F-16 Fighter Aircraft


[Brian J. Pawlowski, “Multivariable Flight for control design with parameter
uncertainty for the AFTI/F-16”, Air force Institute of Technology, Mar 1989]

 SISO case (Longitudinal dynamics)

 Flight conditions
 Speed 0.6 Mach
 Altitude 30,000 Feet
Plant Specifications

q= Pitch rate e= Elevator deflection angle

Speed 0.6 Mach, Altitude 30000 feet :


q( s)  117.24s ( s  0.010028)( s  0.5503)

 e ( s ) ( s  0.006473  j 0.07802)( s  1.167)( s  2.0276)( s  20)

Speed 1.6 Mach, Altitude 30000 feet :

q( s)  481.2 s ( s  0.012642)( s  1.510076)



 e ( s ) ( s  0.0075694  j 0.054)(s  0.96451)( s  3.2234)(s  20)
Plant Specifications contd.

Speed 0.9 Mach, Altitude 20000 feet :


q( s)  657.8s ( s  0.029979)( s  1.096926)

 e ( s ) ( s  0.01517  j 0.02368)( s  0.8011  j 6.5925)( s  20)

Nominal plant :
q( s)  120 s ( s  0.0233)( s  1.1229)

 e ( s ) ( s  0.0121)( s  0.737)( s  2.43)( s  20)
Performance Specifications

 Rise time = [1.23, 2.5] sec


 Settling time < 5 sec
 Steady state error < 0.1 %
 Gain margin = 5 dB
 Phase margin = 45

 No overshoot
Tracking specifications

0.5822 ( s  80)
TRU  2
( s  9.3084 s  (4.6542 ) ) ( s  2.15)
2

7.0745
TRL  2
( s  5.0265 s  (2.5233 ) ) ( s  1.12)
2
Loop Shaping
Loop Shaping contd.

 Controller transfer function

 s  s 
279.6  1  1
 2.962  55.43 
G( s) 
 s2 s 
s 2
  1
 (3194) 3194 
Filter Design
Filter Design contd.

 Filter transfer function

1
F ( s) 
 s  s 
  1  1
 1.132  4.581 
Time domain validation
Conclusions

 Controller designed by QFT has robust


stability

 The controller works well compared to the


PID and LQR Controllers for Boeing

 QFT technique is successfully applied to two


case studies.

You might also like