INDIAN INSTITUTE OF TECHNOLOGY, KHARAGPUR 721302, DECEMBER 27-29, 2002 455

Design of TCSC damping controller to damp

Power swing oscillations
Sujatha Subhash, B.N.Sarkar and K.R.Padiyar

Abstract- Thyristor controlled series compensators show The proposed control structure is simple and easy to
promise in damping low frequency oscillations apart from implement and is illustrated on a SMIB system. The small
controlling power flow over the line. The effectiveness of the signal stability analysis (SSSA) gives an insight into the
control design depends on the choice of input signals. The tuning aspects and the selection of gain for satisfactory
damping controller proposed in this paper use ‘Thevenin
performance. Further the design is evaluated using detailed
Angle’ as input signal, which is simple and easy to implement.
The design is illustrated through a SMIB system . The design EMTP simulation. Cold start and fault simulations confirm
obtained from small signal stability analysis is evaluated the effectiveness of the control design in damping power
through EMTP simulations. swings. The results are discussed in this paper.

Index terms: low frequency oscillations, Damping, EMTP. II SMIB system model
I INTRODUCTION The SMIB system as shown in Fig1consists of a
synchronous machine connected to the infinite bus through
Low frequency electromechanical oscillations are a major two transmission lines in parallel with the TCSC placed on
concern in a large interconnected power system. These one of them. The TCSC is modeled as a variable
oscillations arise due to the dynamics of inter area power capacitance. For small signal stability analysis the gate pulse
transfer and often exhibit poor damping when the aggregate generation is not modeled.
power transfer over a corridor is high compared to the
transmission strength. In the recent years, Thyristor
controlled series compensators (TCSCs), developed
primarily to control power flow on a transmission line show
great promise for damping power swings too [1,2].

The TCSC controller modulates the line reactance in

harmony with the power swings to improve damping. The
choice of the input signals is of utmost importance in the
control design. Unlike Power system stabilisers, TCSC is not
cited near a generator to access the speed deviations as Fig 1. IEEE second benchmark system
control signals. For practical implementation, it is desirable
to use locally available measurements like line current or
power flow as input signal to TCSC. Advances in digital
control have now made it possible to synthesis desired
control signal based on the local measurements of
instantaneous currents and voltages. The synthesized signal Fig 2 Transfer function of TCSC
should be sensitive to the desired swing mode and
insensitive to other modes and also cause less interactions In general the system can be represented in state space form
with other controllers. as,
The present paper focuses on the design of TCSC damping
controller using ‘Thevenin Angle’ proposed by Padiyar &
[∆x& ] = [ A] [∆x] + [ B ] ∆X REF (1)
Kulkarni [3] as input signal. This angle is defined as the [∆I C ] = [C ][ ∆x] + [D ]∆X REF (2)
phase angle of the Thevenin voltage, as seen from the TCSC
terminals and can be synthesized using local current [ ]
t
with [ ∆x ] = ∆δ , ∆s m , ∆Ed , ∆E q ∆E fd , ∆xC as state
' '

measurements. variables
It can be assumed that the desired value of TCSC reactance
XC is achieved in certain time frame as depicted in the Fig 2.
Linearisation yield,
Sujatha Subhash is with Central Power Research Institute, India, Bangalore-
560094 (telephone- (080)- 3604447, email [email protected] )
B N Sarkar is with Central Power Research Institute, India, Bangalore-
∆x&C =
1
(− ∆xC + ∆x REF ) (3)
560094 (telephone- (080)- 3601755 email [email protected] )
TC
K R Padiyar is with Indian Institute of Science, Bangalore - INDIA
[email protected]
456 NATIONAL POWER SYSTEMS CONFERENCE, NPSC 2002

TCSC Damping Controller

The proposed control law for the design of TCSC controller

[3] states that, the change in the impedance of the TCSC
should follow the rate of change of the Thevenin Angle φth .
The control laws are derived using circuit analogy of the
electro-mechanical system [3].

d∆φ th
∆X C = K (4)
dt
For implementation the φth is approximately determined,
from the equivalent network as seen from the TCSC output
terminals shown in Fig 3.
Fig 4 Control structure with Thevenin angle
The remote voltages are related by
E∠φT H = ES∠φTH - ER∠0° ( 5) The parameters ak and k can be tuned using root locus
analysis.
Assuming both bus voltages are equal to unity, and angle is Case studies carried out on the test system illustrate the
small, we can approximately calculate the angle difference design of the TCSC damping controller using Thevenin
φTH as angle as the control signal and tuning procedure for optimum
sinφTH ≅ φT H ≅ I(XTH - XC ) ( 6) performance
where XT H = Thevenin impedance
Xc = TCSC impedance at the quiescent point III Results and discussions
I = current through the line
The IEEE second benchmark system is chosen as the test
system [4]. For small signal stability analysis, the generator
is represented by Model 1.1, with one field winding on the
d-axis and one damper winding on the q-axis. For EMTP
simulation the generator is represented as a four mass spring
system : - high pressure turbine(HP), low pressure
turbine(LP), the generator(GEN), and the exciter(EXC),
coupled on a single shaft. The TCSC system comprises of a
fixed capacitance FC in series with a TCSC. The data are
given in Appendix .
The open loop eigenvalues of the system are -21.05 ±
j21.134, 0.53 ± j9.722, -3.785 and –50.0. The system
exhibits an unstable condition with dominant swing mode at
1.55Hz.
Fig 3 Thevenin equivalent seen from the TCSC terminals The Thevenin impedance as seen from the TCSC terminals
is equal to 0.539pu. The root locus analysis show that the
Linearising we get, value of ak=0.55 is found to give good response. Fig 5
∆ φTH = ∆ I(XTH - XC ) - Io ∆XC (7) shows the root loci of the eigenvalues of both the swing
= ak ∆ I - Io∆XC mode and controller mode, for varying gain in the range 0 to
0.2. Within this range the swing mode continues to move to
In view of the assumptions made the Thevenin impedance the left, but for the controller mode beyond k=0.13, the
can be viewed as a tunable parameter say ’a k ’, (the optimal eigenvalues are on the RH plane, implying that the controller
value determined using eigenvalue loci) and XT H can be goes unstable beyond this value of gain. The optimum gain
used as a starting point. Thus it is very convenient to is found to be K=0.09.
synthezise the Thevenin angle as input control signal as it note : the optimum gain is defined as the gain for which the
requires only one measurement and only one parameter to be damping factor for both the modes –‘swing and controller’
tuned to the operating point. are equal , and damping factor is defined as

To implement the control law, we use a controller as shown

ζ = cos(tan−1 (ω i / σ i )) for an eigenvalue
in Fig 4. The transfer function of the controller has a phase λi = σ i + jω i
lead of 90° because of the requirement of differentiating the
φTH signal Verification using EMTP
Type -59 source is used for modeling of synchronous
machine in EMTP. Cold start simulation is used in EMTP
INDIAN INSTITUTE OF TECHNOLOGY, KHARAGPUR 721302, DECEMBER 27-29, 2002 457

for comparison with the results obtained from small signal

stability analysis(SSSA). Finally the controller’s (b)k=0.05
performance is studied for a single line to ground (SLG) 1.8
fault cleared after 17ms at the generator side of the
transformer. 1.7

Following a cold start simulation in EMTP with TCSC 1.6

operating in constant impedance mode,
1.5

1.4
2 4 6 8
© k=0.1
1.8

1.7

1.6

1.5

Fig 5 Loci of eigenvalue with ak=0.55 1.4

(a)swing mode (10rad/s) (b) controller mode 2 4 6 8
(d) k=0.15
the negative damping of the swing mode causes growing
oscillations in electrical torque as shown in Fig 6 1.8
-1.8 1.7

-1.7 1.6

-1.6 1.5
-1.5 1.4
-1.4 2 4 6 8 2 4 6 8

Fig 6. Cold start simulation – TCSC at constant impedance control Fig 7. Cold start simulation – TCSC with damping controller -
, varying gain
Electrical torque(MNm) vs time (s) Electrical torque(MNm) vs time (s)

Fig 7 gives the corresponding EMTP simulation results with than one frequency, implying that other modes are excited.
TCSC having damping controller, for different gains k=0.01, Also in the eigenvalue loci plot we see that the system is
0.05, 0.1and 0.15 resp. With gain k=0.01 the oscillations are unstable. Moreover within the range of gain predicted in the
stable, having a frequenc y of oscillation of 1.59Hz (swing SSA analysis , we can see from the EMTP results that the
mode). But the SSSA analysis indicates that the system is response of the controller is similar. That is, at higher gains
unstable for the swing mode. With gain k=0.05 the damping we have the controller mode going unstable in both cases,
of swing mode is observed, with a damping factor σ of 1.02 except that in EMTP we see not just the controller modes but
and frequency of oscillation of 1.675Hz. In the case of gain many other modes of oscillations. Again we observe there is
k=0.1, we can see that initially the swing mode is damped, no one -to-one correlation between the optimum gain given
but subsequently we observe a sustained oscillation of by SSSA analysis and EMTP analysis. This is again due to
frequency 2.43Hz. On the other hand for gain k=0.15, we the difference in the modelling.
see oscillations having more For the optimum gain k=0.05obtained from EMTP
(a) k=0.0.01 simulation, the closed loop eigenvalues are -0.585± j0.13, -
-1.8 21.34 ±j24.94, -8.29 ±j24.84 and –3.78.
Figs 8 & 9 show the response of the damping controller of
-1.7 TCSC at the optimum settings obtained from EMTP. The
controller shows good damping of the swing mode. In the
-1.6 simulations we observe oscillations with higher frequencies,
implying that other modes of oscillations are getting excited.
-1.5 The shaft torques are damped after sometime.
-1.4
2 4 6 8
458 NATIONAL POWER SYSTEMS CONFERENCE, NPSC 2002

ra= 0.0045, xa=0.14 , xd=1.65, xd'= 0.25, xd"=0.20, xq

.
1.8 =1.59 , xq' =0.46, xq" =0.20,
T'do =4.5 , T"do =0.04 , T'qo =0.55 , T"qo =0.09
1.7
Network data on 500kV, 100MVA base
1.6 Parameter Positive Zero
sequence sequence
1.5 RT 0.0002 0.0002
XT 0.0200 0.0200
1.4 R1 0.0074 0.0220
6 7 8 XL1 0.0800 0.2400
Fig 8. SLG fault simulation – TCSC with damping controller R2 0.0067 0.0186
optmum gain k=0.05 XL2 0.0739 0.2100
Electrical torque(MNm) vs time (s) RSYS 0.0014 0.0014
XSYS 0.0300 0.0300
IV CONCLUSIONS AVR of Generator
1. For the SMIB test system the TCSC controller is
designed with Thevenin angle (synthesised from line
1
k , k=200, T1 =0.025
current) as control signal using small (1 + sT1 )
2. signal analysis. The root locus analysis show that the Shaft inertia, damping & spring constants for generator
controller goes unstable at higher gains, similar to the mass inertia damping spring constant
case of PSS design. lbm-ft2 lbf-ft-sec/rad lbf-ft/rad
3. The cold start simulations carried out using EMTP Exc 1383 4.3 4.39x102
confirm the effectiveness of the TCSC in damping Gen 17620 4547.9 97.97 x10 2
power swings. Here again we see system going unstable LP 310729 966.2 50.12 x102
at higher gains. An exact one-to-one correlation is not HP 49912 155.2
seen between the results got from the two analysis.
In SSSA analysis the model is simple and adequate for Infinite bus voltage :
designing damping controllers for low frequency V ∞ ∠θ = 0.9028 ∠-30.32
oscillations (swing modes). But in EMTP the modeling TCSC :
is detailed one, with both the network dynamics and TCR XL =5Ω Capacitance XC =30Ω
torsional interactions included. Hence in EMTP Fixed capacitance,FC of 50Ω
simulations we can expect multi-mode oscillations
which is not predicted by the small signal stability
analysis.

V ACKNOWLEDGMENT
The authors wish to thank the Central Power Research
Institute for the opportunity to work in the area of FACTS
devices.
-2
-1
GEN - EXC
0
LP -GEN
-1 VI REFERENCES
-2
HP -LP
-3 [1] James F.Gronquist, William A.Sethares, Fernando L.Alvarado and
Robert H.Lasseter :”Power oscillation damping control strategies for
-4 FACT devices using locally measurable quantities”, IEEE Trans on
-5 Power systems, vol-10 no3 Aug 1995, pp1598-1605.
5 6 7 8 [2] M.Noroozian, L.Angquist, M.Ghandari , G.Andersson :”Improving
power system dynamics by series connected FACTS devices”, IEEE
Fig 9. SLG fault simulation – TCSC with damping controller Trans on Power delivery vol 12, Oct 1997, pp 1635-1641
optmum gain k=0.05 [3] A.M.Kulkarni and K.R.Padiyar: “Damping of Power swings using series
Shaft torque(MNm) vs time (s) FACTS controller” Intl Journal of Electrical Power and Energy
Systems 21 pp475-495 1999.
APPENDIX [4] IEEE SSR working group: “Second benchmark model for computer
simulation of subsynchronous resonance”, IEEE Trans. on PAS vol
IEEE SECOND BENCHMARK SYSTEM
104, 1985 pp1057-1066.
Generator data :
Rating : 22kV, 600MVA
Parameters in pu ::