Licensed to University of Bath

Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

SAE TECHNICAL
PAPER SERIES 2006-01-3051

Closed Loop Controlled High Speed Induction

Generators Using Adaptive Control Technique
Ottman Elkhomri and Jay G Vaidya
Electrodynamics Associates, Inc

Earl Gregory
AFRL/PRPG, Power Generation Branch

Power Systems Conference

New Orleans, Louisiana
November 7-9, 2006

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-0790 Web: www.sae.org
Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

The Engineering Meetings Board has approved this paper for publication. It has successfully completed
SAE's peer review process under the supervision of the session organizer. This process requires a
minimum of three (3) reviews by industry experts.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of SAE.

For permission and licensing requests contact:

SAE Permissions
400 Commonwealth Drive
Warrendale, PA 15096-0001-USA
Email: [email protected]
Tel: 724-772-4028
Fax: 724-776-3036

For multiple print copies contact:

SAE Customer Service

Tel: 877-606-7323 (inside USA and Canada)
Tel: 724-776-4970 (outside USA)
Fax: 724-776-0790
Email: [email protected]

ISSN 0148-7191
Copyright © 2006 SAE International
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE.
The author is solely responsible for the content of the paper. A process is available by which discussions
will be printed with the paper if it is published in SAE Transactions.

Persons wishing to submit papers to be considered for presentation or publication by SAE should send the
manuscript or a 300 word abstract to Secretary, Engineering Meetings Board, SAE.

Printed in USA

Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

2006-01-3051

Closed Loop Controlled High Speed Induction Generators

Using Adaptive Control Technique
Ottman Elkhomri and Jay G Vaidya
Electrodynamics Associates, Inc

Earl Gregory
AFRL/PRPG, Power Generation Branch

Copyright © 2006 SAE International

ABSTRACT The adaptive control technique has been successfully

generalized for real time DSP implementation to regulate
High speed generators offer very high power density the DC voltage for high speed induction generators rated
solution for electric power requirements in airborne from 5 kW to 200 kW.
applications. Induction generators are suitable for the
high speed environment because of the ability to provide INTRODUCTION
controlled voltage and power output with a reliable rotor
construction. An important issue of power control for the In order to maintain a constant voltage at the generator
high speed induction generator is maintenance of the terminals, its excitation current needs to be adjusted
steady state output voltage within the specified limits during load variation from no load to full load. This is
over the entire range of speed and load variations. This accomplished with the use of a PI controller. When a
paper discusses the development of a closed loop sudden step load is applied, the PI controller can
control system for a 200 kW induction generator under become unstable. The transient voltage response was
different load conditions. not well controlled with a conventional PI control
scheme. Two issues have been identified: (1) overshoot
response caused by a large error between the DC
Field Oriented Control (FOC) schemes are implemented voltage command to the sensed DC voltage. (2) The
to both operate the generator in the maximum torque voltage response between step loads is very slow due to
conditions available and to decouple the maximum the fixed proportional and integral gains. The source of
torque from the field under transient and steady state this issue is that PI controller gains cannot be set to
operation. FOC uses Classical Proportional and Integral satisfy both the overshoot and load variation
(PI) controllers for regulation because of their simple simultaneously.
implementation. However, PI controllers do not perform
well when controlling high order non-linear dynamic To overcome this issue, the proportional gains need
plants - such as the high speed induction generators - adjustments with respect to the load variation as well as
due to the overshoot response and the output saturation the overshoot response in real-time. A gain scheduling
when generator is loaded. To resolve this issue, the control algorithm has been developed. This algorithm
proportional gains need adjustments with respect to the represents a set of linear controllers, each of them
load variation as well as the overshoot response in real- designed for a specific load condition. Thus, when the
time. generator is under a certain load condition, the algorithm
control signal determines which linear controller to
A gain scheduling control algorithm has been developed activate. The gain scheduling algorithm was
to select the appropriate controller gains with respect to implemented successfully to regulate the DC voltage at
the generator load. Further, a relationship between the load conditions where the gains were predetermined.
generator loads and the controller gains have been
established. This relationship was modeled using Later a relationship between the proportional, integral
adaptive control technique to vary the gains gains and the load conditions has been established. The
automatically under any load condition. relationship allowed the development of an adaptive
Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

control technique to vary the gains automatically at any current feedback loop and also from the difference
load condition. The adaptive control technique has been between the output voltage and the voltage command.
tested from 0 to 5 kW.
Both currents are compared and based on the error; the
GENERATOR CLOSED LOOP CONTROL (PI) current regulator generates the necessary switching
commands for the six switches in the three-phase full
For high performance control, FOC algorithm is bridge inverter using space vector modulation technique.
implemented. The algorithm transforms the three phase
sinusoidal currents from the stationary reference frame GAIN SCHEDULING METHOD
to DC quantities in rotating reference frame. The rotating
reference frame is composed of the direct current Finding the gains for each load condition was the key to
component and the quadrature current component which appropriately regulate the voltage. The closed loop
are to be orthogonal to each other. algorithm was developed with the capability to change
K p and K i as the generator load varies in real-time.
To maintain a constant DC output voltage, the excitation
current needs to be adjusted according to the generator Transient tests were conducted to find the gains at
load condition. The closed loop algorithm includes a 44000 RPM at no load then at 32.8 kW and 48.6 kW.
special case of the indirect FOC scheme [1]. It is based
on selecting the D-Q axes so that the direct current and ZIEGLER-NICHOLS TUNING METHOD
the quadrature current are equal. Equation 1 shows the
developed torque. The search for K P and K i in the unstable regions has
been conducted through transient tests. The gains were
1 obtained using Ziegler-Nichols tuning method [2], since
T K T I DQ (1)
2

2 the generator is not been modeled mathematically. The

procedure of this method is conducted by setting K P to
Equation (1) summarizes the mechanism of the a low value and K i to zero. Then a voltage step is
generator closed loop algorithm. The closed loop
applied at no-load and under 32.8 kW and 48.6 kW to
algorithm block diagram is shown on Figure 1.
investigate the DC voltage transient. When the PI output
oscillations decayed, K p had to be increased. Then,
when the oscillations increased in amplitude (unstable
system), K p had to be reduced. At the point of stability,
K i is set to a small value to eliminate the steady state
error. This method is iterated until the stable range over
which K p and K i varied is obtained. A constant value
of K i is found to be adequate but K p increased by 36
% under the step load of 32.8 kW to 48.6 kW. Table 1
shows the values for K p and K i under 0, 32.8kW, and
48.6 kW load conditions.

POWER, REG VOLTAGE, Volt Kp Ki

kW DC
0 270 0 9000
32.8 270 1100 9000
48.6 270 1500 9000

Table 1 : K p and K i under two load conditions.

Figure 1: Generator closed loop algorithm
After gathering K p and K i with respect to their load
The commanded voltage, measured generator bus conditions, these values were programmed into a closed
voltage, measured shaft speed, and measured three loop scheduling algorithm. Figure 2 shows the
phase ac generator currents are the inputs to the scheduling algorithm block diagram.
algorithm. The total I DQ
2
current is calculated from the

Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

ADAPTIVE CONTROL UNDER ANY LOAD

CONDITION

The gain scheduling method discussed previously is

further developed to maintain a constant DC voltage of
the generator to any load condition. The idea was to
implement an algorithm that automatically regulates K p
according to a load condition. To develop this kind of
algorithm, a relationship between K p , K i and power
needs to be established. However, in the case of the
200 kW generator, a constant value of K i was found to
be adequate. Only K p and power relationship needed to
be established. Based on the gain scheduling transient
tests data, proportional gains were fitted through
polynomial curves to represent this relationship. Figure 4
illustrates K p versus power.

Figure 2: Generator closed loop scheduling algorithm

The product between the DC voltage and the DC current
signals is established to obtain the power. The power
value is processed in a look up table to search for the
pre-selected gains. These gains are reserved in an array
and stored until triggered by the power value. Based on
the load, the PI regulator receives the corresponding
gains for that condition. Figure 3 illustrates the
responses of the DC voltage, DC current and space
vector command respectively under the tested load
conditions.

Figure 4: K P versus power

Analysis of curve fitting has been conducted for first and

second order polynomial to find the best fit. Figure 5
illustrates the curve fit for the first and second order
polynomial.

Figure 3: DC voltage, DC current and space vector

command versus time

The gain scheduling method has been successfully

implemented to regulate DC voltage of the 200 kW
generator. The technique was demonstrated under
various loads from no load to 48.6 kW.

Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

RESULTS

The relationship between the proportional gains and

power was established with a second order polynomial.

The 200 kW generator DC voltage regulation was

successfully tested at 12000 RPM with four load
conditions from 0 to 5.1 kW. The proportional gains were
automatically computed using the second order
polynomial. Table 2 illustrates four load conditions from
0 to 5.1 kW and figure 7 shows the responses of the DC
Current, DC Voltage.

POWER, REG VOLTAGE, Volt DC

kW DC CURRENT
0 60 0
2.7 60 45
4.02 60 67
4.74 60 79
5.1 60 85

Figure 5: First and second order polynomial Table 2: K p and K i under different load conditions.

SECOND ORDER POLYNOMIAL

The second order polynomial was implemented and

tested. Figure 6 below illustrates the curve fitting and
offsets.

Figure 7: DC Current, DC Voltage versus time

Figure 6: Second order fit

Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53
 Licensed to University of Bath
Licensed from the SAE Digital Library Copyright 2010 SAE International
E-mailing, copying and internet posting are prohibited
Downloaded Sunday, August 22, 2010 1:23:26 PM

CONCLUSION

The objective of this work was to develop closed control

system to control the DC voltage output of the 200 kW
induction generator using Field Oriented Control.
Difficulties occurred in the beginning of the
implementation using the classical proportional
integrator controller. A gain scheduling control algorithm
has been developed to solve this issue by selecting the
appropriate controller gains with respect to the generator
load. The gain scheduling control algorithm facilitates
the understanding of the generator behavior which
resulted in developing a better approach to control the
DC voltage adaptively. Both methods were implemented
successfully. However, the adaptive method is better
approach because it can be tailored to work for high
speed induction generators rated from 5 kW to 200 kW.

ACKNOWLEDGMENTS

The funding of this work was provided by the U.S Air

Force Research Laboratory under SBIR phase II
program, contract number: F33615-00-C-2018.

REFERENCES

[1] Jay Vaidya, Earl Gregory, “High speed induction

generator for application in aircraft power systems,”
2004 Power Systems Conference, Session: Advanced
Power Systems Technologies I, Reno, NV, Nov. 2004.

[2] Karl J. Åström and Björn Wittenmark, Adaptive

Control, Addison-Wesley, 1990.

[3] Ottman Elkhomri, “DSP implementation of DC

Voltage Regulation using Adaptive Control for 200 kW
62000 rpm Induction Generator” Thesis for MSEE,
University of Central Florida, 2006.

CONTACT

Ottman Elkhomri

Senior Program Engineer, Electrodynamics Associates,

Inc

Phone: 407-3845957

[email protected]

Jay G Vaidya

President, Electrodynamics Associates, Inc

Phone: 407-9771825

[email protected]

Author:Gilligan-SID:4970-GUID:35211820-138.38.0.53