WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

A Simple Approach to Estimate the Steady-State Performance

of Self-Excited Induction Generator
K S Sandhu, Dheeraj Joshi
Department of Electrical Engineering
National Institute of Technology
Kurukshetra,Haryana
INDIA
kjssandhu@yahoo.com, dheeraj_joshi@rediffmail.com

Abstract: This paper presents a new and simple model for the steady-state analysis of single and
parallel operated self-excited induction generators (SEIG). In this paper an attempt has been made
to incorporate the unjustified assumptions in an existing (Watsons) model. This has resulted into
an improved model for the estimation of performance of SEIG. A close agreement of simulated
results using proposed modeling with experimental values on test machines proves the validity
and superiority of proposed model. Further proposed model is extended for the analysis of a
system comprising of number of such machines operating in parallel.

Keywords: Parallel operation, steady state analysis, self-excited induction generator (SEIG),
wind energy conversion.

Nomenclature C at rated frequency

a per unit frequency Xm magnetizing reactance per

b per unit speed phase at rated frequency
Rc Core loss resistance
C terminal excitation
Sub
capacitance per phase
Im magnetizing current per 1. Introduction

phase Self-excited induction generators have

RL load resistance per phase attained a lot of attraction in recent years,
due to the suitability of these machines in
Rs stator resistance per phase many applications including wind energy
and small hydro energy systems. Further,
Rr rotor resistance per phase these machines and specially cage induction
V load voltage per phase machines pocesses many advantages such as
low cost, brushless and rugged construction,
Xs stator reactance per phase self protection capability etc.
To estimate the steady-state performance of
Xr rotor reactance per phase, a SEIG, most of the researchers used the
referred to stator conventional equivalent circuit
representation of an induction motor [1-13].
XC capacitive reactance due to Some of the researchers used the impedance
model, and a few used the admittance-based

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

208
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

model for the treatment of these circuits.

Whereas [10, 12] developed a new circuit
representation, which includes an active
power source in the rotor circuit. 2. Steady-State Analysis
Apart from this, [14] developed the simple
most approach to describe the autonomous
and parallel operation of self excited
induction generators. Simplicity is the main
attraction of Watsons [14] model.
Different models are suggested by
researchers for analysis and control of
parallel operated self-excited induction
generators in steady state. Al-Bahrani et al
[15] suggests Newton Raphson technique
for voltage control of two or more parallel
operated SEIG. Various control parameters
were found simultaneously. The results are Fig.1 Equivalent circuit representation.
better but with lengthy and complex terms.
Chakraborty et al [16,17] suggests T and Fig.1 shows the equivalent circuit as
inverse models for the analysis and adopted by Watson [14].
control of parallel operated SEIG. More This circuit representation may be modified
than one iterative loops were used to find the as given in Fig.2 with the provision of
various parameters. followings, which were found to be missing;
Sandhu [18] suggests a simple model having Inclusion of stator reactance
balanced resistive loads for the analysis of Inclusion of pu frequency to make
parallel operated SEIG. Methodology all leakage reactances and
adopted needs an estimation of capacitance excitation capacitance more
sharing by individual machine. Classical and effective.
conventional control techniques are Further magnetizing branch has
described by [19-20]. P. Haiguo et al [21] been shifted to stator side, as
presents an Fuzzy-PID based approach to usually adopted in motoring case.
control the wind turbine system. It is found
that the vector control in combination with
Fuzzy-PID control enhances the system
stability. Such developments are the
indication of current research in the area of
wind energy generation through induction
machines.
In the present paper an attempt has been
made to improve the results of [14] with few
modifications but without loosing the
simplicity of the approach. Close agreement
of simulated results with experimental
results, confirms the validity of proposed
modeling. Model is found to be suitable for Fig.2 Modified equivalent circuit.
the steady-state analysis of single as well as
for multi machine systems. Analysis of circuit as given in Fig.2, in
terms of real power gives;

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

209
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

V 2 V 2 Rearrangement of (5) and (6) results in to

k the quadratic equation in terms of unknown
a a
P= + (1) voltage as;
R s + Rr R L
a as a
k1V 2 + (k 2 aC )V + k 3 = 0 (7)

Where k is a fractional value and for a single Values of k1,k2 and k3 are as per Appendix 1.
machine is (1) to (7) may be used to determine the
R s Rr generated frequency and terminal voltage
+
a as for any operating speed and excitation
k= 1 capacitance. Further analysis of the
R R
2 2 generator including core loss is given in
s + r + ( X s + X r )2
a as Appendix 2.
This approach can be extended for N self
In the absence of power source (1) may be excited induction generators operating in
written as; parallel as shown in Fig.3.
Where
Rr
s= (2) N represents the number of machines
R s + kR L operating in parallel.

It can be also written as follows;

a b
s= (3)
a

Rotor speed and slip determines the

frequency of the generated voltage.

b
a= (4)
1 s

At no-load ( R L ), therefore the

generated frequency is same as the driven
frequency. However, as evident from above
expressions, generated frequency falls with
load.
For self-excitation, magnetizing current is
supplied by the capacitor and is given as;

I m = VaC (5)

Further, magnetization characteristics of

SEIG (Appendix-I) may be expressed in the
form as;

Im = k1V 2 + k2V + k3 (6)

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

210
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

s
R ri
i = (9)
i =1
R si + k i R L

a bi
si = (10)
a

b
a = i
(11)
1 si

I
i =1
mi = VaC (12)

and

I
i =1
mi = k1iV 2 + k 2iV + k 3i (13)

(8) to (13) may be used to estimate the

generated voltage and frequency as given in
Appendix-3.

3. Results and Discussions

TABLE 1 Comparison of results for

single machine.

Fig.3. Equivalent circuit representation of

parallel operated self-excited induction
generators.

Equation (1) to (7) referred to multi machine

system comprising of N such machines may
be modified as;

2 2
Table 1 gives the comparison of simulated
V V results with experimental results on a test
N ki
a a
P =
machine, SEIG-1 (Appendix 1). Simulated
+ (8)
R si R ri RL results with modifications as suggested in
i =1 +
a as i a this paper are found to be more close to
experimental one in comparison to

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

211
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

simulated results as obtained using Watson

model. This proves the validity and
superiority of the proposed modeling. Fig.4
to Fig.7 gives the variation of terminal
voltage and frequency with load for different
values of excitation capacitance and
operating speed. It is observed that
excitation capacitance at stator terminals and
operating speed of the machine effect the
performance to a great extent and thus may
be used to control the terminal conditions.

Fig.5. Variation of generated frequency with

load for different excitation capacitance.

Fig.4. Variation of terminal voltage with

load for different excitation capacitance.

Fig.6. Variation of terminal voltage with

load for different operating speed.

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

212
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

Fig.7. Variation of terminal voltage with Fig.8 Variation of generated frequency and
load for different operating speed. voltage with load.
TABLE 2 Comparison of results for two
machines operating in parallel.

C=90F ,RL=90

Table 2 gives the comparison of simulated

results with experimental results on a set of
two test machines [Appendix-1] operating in
parallel. Results are found to be in close
agreement especially for low slip operations,
justified in case of induction machines.
Fig.9 Variation of load ,generated frequency
and voltage with excitation capacitance.

Fig. 8 to Fig. 11 shows the simulated results

for a system comprising of two self excited
induction generators. As observed terminal
voltage falls sharply with load. However fall
of generated frequency with load is small in
comparison with the terminal voltage. This
reflects the necessacity to control the

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

213
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

terminal voltage with load variations. Fig.9 of parallel operation of two machines. It is
and Fig.10 shows the effect of excitation seen that generated voltage is greatly
capacitance and operating speed on the influenced due to any change of rotor
generated voltage, frequency and load resistance of one of the machines. However
supplied by two machine system operating variations in generated frequency are
in parallel. Load capability of the system negligible in case both machines are running
increases with an increase in excitation at constant speeds. This provides the
capacitance. However it effects the opportunity to control the generated voltage
generated voltage and frequency of the system through rotor resistance,
simultaneously. provided operating speeds are maintained.

4. Conclusion
In this paper an attempt has been made
to prepare a new model to investigate the
steady-state performance of single as
well as parallel operated self-excited
induction generators. The main
attraction of the model is its simplicity to
obtain the final solution. Results
obtained are found to be close agreement
to the experimental results obtained on a
test machine/set of machines. This
proves the validity of model proposed
Fig.10. Variation of load, generated for the analysis of single unit or number
frequency and voltage with rotor speed of of units operating in parallel. Efforts are
machine-2. made to include the core loss component
which is generally neglected. Inclusion
of core loss branch makes the model
more realistic. It is found that, to meet
the power needs of the world, wind
energy is emerging s a potential
candidate among renewable energy
resources. Therefore future research
plans of the authors in the area of wind
energy extraction using induction
generators is as;

Analysis and control of power

quality of wind energy systems
using artificial intelligence.
Fig.11. Variation of load, generated Design modifications in the
frequency and voltage with rotor resistance induction generators according to
of machine-2. operating constraints of a specific
area.
Fig.11 gives the effect of variations of rotor
resistance of machine-2 on the performance

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

214
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

References method, IEEE Trans. Energy

Conversion,vol.10 no. 3,1995,pp. 502-507.
[1] S. S. Murthy, O. P. Malik, and
[10] K. S. Sandhu and S. K. Jain,
A.K.Tandon, Analysis of self-excited
Operational aspects of self-excited
induction generators, Proc. IEE,vol. 129,pt.
induction generator using a new model,
C, no. 6,1982,pp. 260-265.
Electric Machines and Power Systems, vol.
27, no. 2,1999, pp. 169-180.
[2] G. Raina and O. P. Malik, Wind energy
conversion using a self-excited induction
[11] I. A. M. Abdel-Halim, M. A. Al-
generator, IEEE Trans. Power Apparatus
Ahmar, and M. Z. El-Sherif, A novel
and Systems, vol. PAS-102, no. 12,1983,pp.
approach for the analysis of self-excited
3933-3936.
induction generator, Electric Machines and
Power Systems, vol. 27,1999, pp. 879-888.
[3] A. K. Tandon, S. S. Murthy and C. S.
Jha, New method of computing steady-state
[12] K. S. Sandhu, Iterative model for the
response of capacitor self-excited induction
analysis of self-excited induction
generator, IE (I) Journal-EL, vol. 65,1985,
generators, Electric Power Components
pp. 196-201.
and Systems, vol. 31, no. 10, 2003, pp. 925-
939.
[4] N. H. Malik and S. E. Haque, Steady-
state analysis and performance of an isolated
[13] A. K .Tandon, S. S. Murthy, and G. J.
self-excited induction generator, IEEE
Berg, Steady-state analysis of capacitor
Trans. Energy Conversion, vol. EC-1, no.3,
self-excited induction generators, IEEE
1986,pp.134-139.
Trans.Power Apparatus and Systems, vol.
103, no. 3, 1984,pp 612-618.
[5] N. H. Malik and A. H. Al-Bahrani,
Influence of the terminal capacitor on the
[14] D.B. Watson and I.P. Milner,
performance characteristics of a self-excited
Autonomous and parallel operation of self
induction generator, Proc. IEE, vol.137, pt.
excited induction generator for lighting
C, no. 2, 1990, pp.168-173.
loads in remote areas, Int. J. Elect. Enging.
Educ., (22), 1985, 363-373.
[6] L. Shridhar, B. Singh, and C. S. Jha, A
step towards improvements in the
[15] A.H. Al-Bahrani and N.H. Malik,
characteristics of self-excited induction
Voltage Control of parallel operated self-
generator, IEEE Trans. Energy
excited induction generators, IEEE Trans.
Conversion, vol. 8, no. 1,1993,pp. 40-46.
Energy Conversion, Vol. 8, No. 2, June
1993, pp. 236-242.
[7] T. F. Chan, Steady-state analysis of
self-excited induction generators, IEEE
[16] C. Chakraborty, S. N. Bhadra, S.P. Das
Trans. Energy Conversion, vol. 9, no.
and A.K. Chattopadhyay, Some studies on
2,1994, pp.288-296.
the parallel operation of self excited
induction generators, Proc. IEE,vol. 129,pt.
[8] L. Quazene and G. McPherson,
C, no. 6,1982, pp. 361-366.
Analysis of the isolated induction
generator, IEEE Trans. Power Apparatus
[17] C. Chakraborty, S.N. Bhadra and A.K.
and Systems, vol. PAS-102, no. 8, 1983, pp.
Chattopadhyay, Analysis of parallel-
2793-2798.
operated self-excited induction generator,
IEEE Trans. Energy Conversion, vol. 14,
[9] T. F. Chan, Analysis of self-excited
No.2, June 1999, pp. 209-216.
induction generators using an iterative

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

215
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

[18] K. S. Sandhu, Steady-state analysis of Rs = 3.35 ohm

multimachine system with self-excited Rr = 1.76 ohm
induction generators, Electrical Power Xs = Xr = 4.85 ohm
Components and Systems, vol. 34, April
2006, pp. 483-496.
APPENDIX 2
[19] K.C. Wu, R.K. Joseph, N.K. Thupili,
Evaluation of classical and fuzzy logic
controllers for wind turbine yaw control,
Aerospace Control Systems, Proc. IEEE
Regional Conference, vol. 1, May 25-27,
1993, pp. 254-258.

[20] M. Namazov, R. Samet, R. Huseynov,

Modeling and Simulation of the fuzzy relay
type controller for solving the double
integrator control problem, Proc. Of 9th
WSEAS International Conference on
Automatic Control, Modeling & Simulation,
Istanbul, Turkey, vol.1, May 27-29, 2007, Fig.12. Modified equivalent circuit
pp.7-11. including core losses.

[21] P. Haiguo, W. Zhixin, Simulation Analysis of circuit as given in Fig.12, in

research of Fuzzy PID synthesis yaw vector terms of real power gives;
control system of wind turbine, WSEAS
Trans. on System and Control, Issue 10, vol. V 2 V 2 V 2
2, Oct. 2007, pp. 469-476. k
a a a
P= + +
APPENDIX 1 Rs + Rr Rc RL
a as a a

SPECIFICATIONS OF SEIG-1

Line voltage=380V Where k is a fractional value and depends

Line current=1.9A upon the design of the machine.
In the absence of power source may be
Rating=1.0HP
written as;
Number of poles= 4
Frequency=50Hz
Base speed =1500 rpm Rr
Rs = 9.5 ohm s=
k
Rr = 8.04 ohm Rs +
1 1
Xs = Xr = 8.84 ohm a +

c
R R L

SPECIFICATIONS OF SEIG-2
Line voltage=230V
Line current=4.96A
Rating=3.0HP
Frequency=50Hz
Number of poles= 4
Base speed =1500 rpm

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

216
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

TABLE 2 Simulated results including

core losses. N6=n32*d31+n22*d41;

P6=kk*(d41*d12+d31*d22+d21*d32+d1
1*d42);

Q5=M5+N5+P5;

M5=n31*d22+n21*d32+n11*d42;

N5=n32*d21+n22*d31+n21*d41;

P5=kk*(d41*d02+d31*d12+d21*d22+d1
1*d32+d01*d42);

Q4=M4+N4+P4;

APPENDIX 3 M4=n31*d12+n21*d22+n11*d32;

Generated frequency for two machine N4=n32*d11+n22*d21+n12*d31;

systems including core losses may be
calculated using (8) and (10). This results P4=kk*(d31*d02+d21*d12+d11*d22+d0
into the solution of 9th order polynomial 1*d32);
equation in unknown a as follows;
Q3=M3+N3+P3;
9 8 7 6 5
Q9*a +Q8*a +Q7*a +Q6*a +Q5*a +Q4*
M3=n31*d02+n21*d12+n11*d22;
4 3 2 1
a +Q3*a +Q2*a +Q1*a
N3=n32*d01+n22*d11+n12*d21;
Where
P3=kk*(d21*d02+d11*d12+d01*d22);
Q9=P9;
Q2=M2+N2+P2;
P9=kk*d41*d42;
M2=n21*d02+n11*d12;
Q8=P8;
N2=n22*d01+n12*d11;
P8=kk*(d41*d32+d31*d42);
P2=kk*(d11*d02+d01*d12);
Q7=M7+N7+P7;
Q1=M1+N1+P1;
M7=n31*d42;
M1=n11*d02;
N7=n32*d41;
N1=n12*d01;
P7=kk*(d41*d22+d31*d32+d21*d42);
P1=k*d01*d02;
Q6=M6+N6+P6;
kk=(1/RL)+(1/Rc1)+(1/Rc2);
M6=n31*d32+n21*d42;

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

217
WSEAS TRANSACTIONS on SYSTEMS and CONTROL K S Sandhu, Dheeraj Joshi

n31=Rs1+Rr1;
K1c=K1b*K1b;
n21=-b1*(2*Rs1+Rr1);
K1d=a*a*(a-b1)*(a-b1)*(Xs1+Xr1)^2;
n11=b1*b1*Rs1;
K1e=K1c+K1d;
d41=(Xs1+Xr1)^2;
K2a=a*(a-b2)*(a-b2)*(Xs2+Xr2);
d31=-2*b1*d41;
K2b=Rs2*(a-b2)+a*Rr2;
d21=Rs1+Rr1+2*Rs1*Rr1+b1*b1*d41;
K2c=K2b*K2b;
d11=-2*b1*Rs1-2*b1*Rs1*Rr1;
K2d=a*a*(a-b2)*(a-b2)*(Xs2+Xr2)^2;
d01=b1*b1*Rs1;
K2e=K2c+K2d;
n32=Rs2+Rr2;

n22=-b2*(2*Rs2+Rr2);

n12=b2*b2*Rs2;

d42=(Xs2+Xr2)^2;

d32=-2*b2*d42;

d22=Rs2+Rr2+2*Rs2*Rr2+b2*b2*d42;

d12=-2*b2*Rs2-2*b2*Rs2*Rr2;

d02=b2*b2*Rs2;

(12) and (13) may be used to develop the

quadratic expression in terms of V as given
below;

0.00029*V2-(0.0753+w*a*C-K1-
K2)*V+6.8767

Where

K1=K1a/K1e;

and

K2=K2a/K2e;

K1a=a*(a-b1)*(a-b1)*(Xs1+Xr1);

K1b=Rs1*(a-b1)+a*Rr1;

ISSN: 1991-8763 Issue 3, Volume 3, March 2008

218