Nonlinear Modeling of Synchronous Generators Using Wavelet TransformExperimental Results

O.P. Malik
elec. And comp. eng. Dept.
university of Calgary Calgary AB Canada

M. Karrari
Elec. Eng. Dept.
Amirkabir university of technology Tehran Iran

finding the parameters of the known structure. Usually

the procedures involve difficult and time-consuming
tests. These approaches include short-circuit tests,
standstill frequency response (SSFR) and open circuit
frequency response (OCFR). These tests can mainly be
carried out when the machine is not in service.
To overcome the shortcomings of the traditional
methods, identification methods based on on-line
measurements have gained attention during the recent
years [4-7]. These methods can be divided into two
categories. In the first category [4-5], assuming a known
structure for the synchronous machine (as the traditional
methods), the physical parameters are estimated from
on-line measurements. The second category [6-7] deals
with black-box modeling of synchronous generators
using input-output data. In the black-box modeling the
structure of the model is not assumed to be known a
priori. The only concern is to map the input data set to
the output data set.
Identification of linear dynamic systems has
theoretically been well established and many good
approaches are available [8-9]. However, identification
of nonlinear systems such as synchronous machines, is
still an active research topic. Many different approaches,
like Nonlinear Least Squares, Volterra series, Weiner
series, Wavelets, Neural networks, Fuzzy logic and
Genetic algorithm have been developed for identification
of nonlinear systems. A survey of techniques prior to the 1980s is

Keywords:
Synchronous
generators,
Nonlinear
identification, Discrete Wavelet Transform, Modeling of
power systems
INTRODUCTION

ch

1.

Ar

With the increased complexity of the modern

interconnected power systems, analysis of the dynamic
performance of such systems has become very
important. For the analysis of the dynamic performance
and stability of the system, a valid dynamic model is a
basic requirement. For this reason, identification and
modeling of different parts of the power systems, has
attracted many researchers.

given in [10]. A good recent review of the nonlinear identification

approaches can be found in [11].

Synchronous generators play a very important role in

the stability of the power systems. A valid model for
synchronous generators is essential for a valid analysis
of stability and dynamic performance. Almost three
quarters of a century after the first publications in this
area [1-2], the subject is still a challenging and attractive
research topic.

Among many methods developed for black-box

modeling are neural networks [12] and the recently
introduced wavelet decomposition [13-15]. These tools
have become useful in many scientific areas, among
which are signal processing and system identification.
The wavelet functions are becoming popular for
identification of nonlinear systems because of their
multiresolution analysis (MRA) and global optimization
property. Using multiresolution analysis (MRA) of the
wavelet transform, the accuracy of the model can be
increased to the desired level.

The traditional methods of modeling of synchronous

generators are well specified in IEEE standards [3].
These methods assume a known structure for the
synchronous machine, using well-established theories
like Park transformation. They address the problem of

25

Journal of Iranian Association of Electrical and Electronics Engineers _ Vol.1 _ No.1 _Spring 2004

ive

of

SI

Abstract-- Application of discrete wavelet transform

for identification of a synchronous generator is
investigated in this paper. Discrete wavelet transform
guarantees global minimization and the accuracy of the
identified model can be increased to the desired level by
the multiresolution analysis (MRA) of the transform.
The proposed method has been tested on a micromachine synchronous generator. In this study, the field
voltage is considered as the input and the active output
power and the terminal voltage are considered as the
outputs of the synchronous generator. Experimental
results show good accuracy of the identified model
following changes in the operating conditions.

  _   _   _
 

     

www.SID.ir

Melgoza J., Jesus Rico, Heydt, G.T., Keyhani A., An

algebraic approach for identifying operating point
dependent parameters of synchronous machine using
orthogonal series expansions, IEEE Trans. on Energy
Conversion, Vol. 16, no. 1, March 2001, pp.92-98.

[5]

Melgoza, J.J.R., Heydt G.T., Keyhani A., Agrawal B.L.,

Selin D., Synchronous machine parameter estimation
using Hartley series, IEEE Trans. on Energy
Conversion, Vol. 16, no. 1, March 2001, pp.49-54.

[6]

Karrari M., Menhaj M.B., Application of different

neural networks for identification of power systems,
UKACC Conference on Control 2000, University of
Cambridge, U.K. 4-7 September 2000.

[7]

Shamsoollahi, P., Malik, O.P.,On-line identification of

synchronous generator using neural networks,
Proceedings of the 1996 Canadian Conference on
Electrical and Computer Engineering, CCECE96, Part
2, 1996, p.595-598.

SI

[4]

[8]

Ljung, L., System Identification, theory for the user,

Prentice Hall PTR, c1999.

[9]

Goodwin G.C. Dynamic system identification :

experiment design and data analysis, New York,
Academic Press, 1977

of

Fig. 4: Identification results with the wavelet model

and the measured variables for a step change in the
field

[10] Billings,S.A., Identification of nonlinear systems-a

survey, IEE Proceddings, Part D., Vol. 127,1980, pp.
272-285.

ch

[12] Chen, S., Billings, S., Grant P.M., Nonlinear

identification using neural networks, International
Journal of Control, no. 51, 1990, pp. 1191-1214.
[13] C.K. Chui, An introduction to wavelet, Academic
Press, Inc., 1992.
[14] J.J., Benedetto, M.W., Frazier, Wavelets ,
Mathematics and Applications, CRC Press Inc., 1994.

Ar

Journal of Iranian Association of Electrical and Electronics Engineers _ Vol.1 _ No.1 _Spring 2004

ive

[11] Nelles, O., Nonlinear system identification: from

classical approaches to neural networks and fuzzy
models, Berlin ; New York : Springer, c2001.

[15] A. Sureshbabu , J.A. Farrell, Wavelet-based

identification for nonlinear control, IEEE Trans.
Automat. Contr. Vol. 44 , Feb. 1999, pp.412-417.

Fig. 5: Identification results with the wavelet model

and the measured variables with a PRBS signal
applied to the field

[16] L. Ljung, System Identification, theory for the user,

Prentice Hall PTR, c1999.

Reference:
[1]

[17] Yu, Yau-Nan, Electric Power System Dynamics,

Academic Press, 1983.

Kilgore, L.A., Calculation of Synchronous Machine

Constants, AIEE Trans., Vol. 50, December 1931,
pp.1201-1214.

[2]

Wright, S.H., Determination of Synchronous Machine

Constants by Test, AIEE Trans., Vol. 50, December
1931, pp.1331-1351.

[3]

IEEE Standard 115-1995 IEEE Guide : Test Procedures

for Synchronous Machines, Part IAcceptance and
Performance Testing, Part IITest Procedures and
Parameter Determination for Dynamic Analysis.

  _   _   _
 

     

[18] Kundor, P., Power System Stability and Control, Mc

Graw Hill Inc., 1994.

3
0

www.SID.ir