Professional Documents
Culture Documents
d1 J3 J4
R 1=00R = 0.0120 Inertia Damping Spring Constant
Mass lbm-ft2 lbf-ft-sec/rad in lbf-ft/rad
Table lC - Two Axis Rotor Circuit Model Table 1D - R-X vs. Frequency Model for Unit 1
Derived From Conventional Stability Data Using
Approximate Approach for Unit 1 The following model is derived from the rotor circuit
model of Table 1C using the following equations.
d-axis
Z (f s ) + Z ave (fr)
= Req +x
+
jxeq = R a
S
g
ad -
fr Where: S = fs - f
0
0 f
s
Zd(fr) afkd f
fr 'f
fs RReq x eq
f f
PI r 10.0 0.0034 0.0334
fd kd fr
0 15.0 0.0026 0.0501
ff
fo 20.0 0.0017 0.0668
fo= Frequency corresponding to Average Rotor Speed 25.0 0.0004 0.0836
f r = Rotor current frequency 30.0 -0.0012 0.1004
35.0 -0.0034 0.1173
40.0 -0.0067 0.1344
45.0 -0.0121 0.1521
ad =£ = 0.14 pu 50.0 -0.0223 0.1716
55.0 -0.0473 0.2002
= 1.51 pu 65.0 0.0658 0.2366
IafkdXd - X9. 70.0 0.0421 0.2403
75.0 0.0322 0. 2535
fd afkd (xg-x2)/ (Xd -
dX = 0.1186 pu 80.0 0.0269 0.2688
85.0 0.0237 0.2848
Ikd = Qafkd *fd (Xd X1) = 0.132 pu 90.0 0.0215 0.3011
95.0 0.0199 0.3175
kd fkd * fd ( afkd fd'(d) -d X) 100.0 0.0187 0.3340
105.0 0.0178 0.3505
I
Rfdf= fd + I afkd = 0.00096 pu 110.0 0.0170 0.3671
w T" 115.0 0.0164 0.3837
o do 120.0 0.0159 0.4003
K P K
q-axis GM 12 23 H
f 1 12 3
aq T Inertia Damping Spring Constant
0
Mass lbm-ft2 lbf-ft-sec/rad lbf-ft/rad
Rfq Rkq Gen 334914 208.2
f 156.1 x 106
zq(f r) Qafkq fr f LP 370483 230.4
qr§ f
0 ,
fq
r {kt
kq "O
r
HP -09922 68.38
198.7 x 106
Mode Shapes
aq= 0.14 pu, afkq= 1.45 pu, Ifq= 0.411 pu, kq = .0738. pu The calculated modal quantities follow.
f a H
Mode n n n
X = 0.12 pu R = 0.0045 pu f R x
a eq
xd = 1.54 pu X
q
= 1.50 pu
10.0 0.0031 0.0301
Xd = 0.23 pu Xt = .0. 42 pu 15.0 0.0021 0.0451
d q 20.0 0. 0009 0.0602
X'td = 0.18 pu X'= 0. 18 Pu 25.0 -0.0006 0. 0753
q 30.0 -0.0026 0.0906
T' = 3.70 s T'qo = 0.43 s
35.0 -0.0055 0. 1059
do 40.0 -0.0095 0.1217
T" = 0.04 s T" = 0.06 5
45.0 -0.0163 0.1383
do qo 50.0 -0.0288 0.1576
55.0 -0.0569 0.1896
Vt = 1.0 =0 (unloaded unit) 65.0
70.0
0. 0771
0.0511
0.2240
0.2207
75.0 0.0392 0.2304
80.0 0.0328 0.2433
Table 2C - Two Axis Rotor Circuit Model 85.0 0.0287 0.2573
Derived'From Conventional Stability Data for Unit 2 90.0 0.0259 0.2717
(See Table 1C for Conversion Method) 95.0 0.0240 0.2863
100.0 0.0225 0.3010
105.0 0.0213 0.3158
d-axis 110.0 0.0203 0.3307
ffr 115.0 0.0196 0.3456
0.12 r 120.0 0.0189 0.3605
0
Frequency Scanning
Frequency scan program (FSP) study resul ts are
based upon the following approximate relationship
from Ref. 3.
R
CT
n =
60- fn an 60 + fn Rbn
8 f H R2 + X2 60 - f
n n an an n n n
Rb
Zbn = +
i is the
impedance as viewed
0.00949 0.0166 from behind the study generator at the
f
frequency of 60 + f
Zq(fr ) 1.38
f0
rL
f fr
n
Only the results for this common torsional mode are From FSP
compared here. Figure 4 shows a plot of the results
from Table 4. C- .02 - From EVP
The FSP cannot represent the torsional interaction H-4'
C9
between the units and hence undamping for each unit MODE 2
has been calculated by representing the other unit as 0
a passive element by their respective R-X data of 40
20 60 80
Table lD for Unit 1 and Table 2D for Unit 2. Fig. 3 Undamping Vs. % Compensation Due
To Self-Excitation For Case 1
1062
Undamping vs Additional Damping Required for Stability
Table 4 - Comparisons of Results* for Case 2 Mode 1
Near the stability boundary, undamping is nearly
equal to the additional damping required for stabil-
FROM ity. However, as the undamping values increase, the
FSP MINIMUM Ac FROM EVP FOR AVERAGE difference between undamping and the additional damp-
ol REQUIRED MOST Aa ing required for the stability may not be negligible.
% UNIT UNIT UNDAMPED REMAINING FROM
COMP 1 2 MODE MODE EMTP The frequency scanning program performs an ap-
proximate calculation of the mechanical system damping
0 -0.013 -0.007 required for insipient stability. The eigenvalue pro-
gram, on the other hand, calculates the negative damp-
20 -0.005 -0.004 ing (undamping). Undamping is obtained from the EMTP
program by measuring the growth rate of the pertinent
40 0.024 0.006 variables. Other programs, such as Nyquist and Boun-
dary Eigenvalue, do not provi.de undamping as an out-
50 0.077 0.025 0.063 -0.0070 put. Therefore, to provide a conmnon basis for com-
paring results from various programs, damping for in-
55 0.143 0.048 0.091 0.0259 sipient stability is reported below for one specific
system condition.
60 0.287 0.099 0.133 0.1106 0.27
65 0.612 0.123 0.440 0.0863 Mechanical Damping Required for Insipient Stability
for Mode 1, Case 1 at 55% Compensation Level
70 0.887 0.308 0.951 0.0187 0.87
Program Damping Required in Red/Sec.
75 0.890 0.309 1.030 -0.0073
FSP 0.448*
80 0.650 0.224 0.622 -0.0201 0.62 EVP O.399**
EMTP 0.374**
85 0.313 0.106
* The value is directly from Table 3.
90 0.156 0.051 0. 1706 -0. 0316
** These values are obtained by representing just
enough mechanical damping in the respective
programs to obtain insipient instability.
*The following notes apply:
1. Undamping or damping required for stability Comparing these with values in Table 3 it is seen
measured in rad/sec. that they are slightly different for EVP and EMTP. Of
course, they are exactly the same for FSP since FSP
2. No mechanical damping represented in studies. gives the damping required for insipient stability.
3. FSP values account for supersynchronous effect. TORQUE AMPLIFICATION STUDY RESULTS
Torque Amplification has been studied using Case 1
as an example. A compensation value of 55% in line 2
1.5 is used since it produces the best tuning for EMTP
simulation. However, in order to separate the self-
excitation and torque amplification phenomena, the
mechanical system damping is represented. A damping
H-1
LL
of 0.4 radians/sec was chosen in order to keep the
1.2 unit steady-state stable for this case. This is
simply done by multiplying the self damping coeffi-
0- cients in Table IA by a factor of 8. Since the damp-
--J ing of 0.4 rad/sec is small compared to the first
LL
cc 0.9 mode torsional frequency of 155 rad/sec, the damped
natural frequency will be almost the same as the un-
cL damped natural frequency.
ct A 3-phase fault is applied on the high side of
P-
0.6 generator step-up (GSU) transformer and the faul t
clearing time is varied from 1 cycle to 4 cycles.
LL
cCD The resultant torques vary greatly with the faul t
clearing time. Table 5 shows a tabulation of perti-
<-1 nent torques, capacitor voltages, electrical torques,
0-
tD
etc., as a function of fault clearing time.
cr- The peak Gen-LP turbine torque vs. the fault
2- clearing time is plotted in Figure 5. The impact of
0 - fault clearing time on the peak torque is seen to be
40 50 60 70 8( very significant. The time simulation plots for cases
lA, lE and lR from Table 5 are shown in Figures 6, 7
% COMPENSATION and 8 respectively.
Fig. 4 Undamping Vs. % Compensation Due To
Self-Excitation For Case 2
1063
Table 5 - EMTP Torque Amplification Cases Results Summary
FAULT Toraues in P.U. (P-P/2N4 Peak Capacitor Voltages Peak Velocity Deviatior
CLEARING EX GEN LP in Kv in % on 377 rad/sec bas FAULT FAULT
CASE TIME IN TO TO TO PHASE PHASE PHASE CLEARING CURRENT
S# | SECONDS GEN LP HP A B C EX GEN HP SEQUENCE AMPS
1A 0.017 0.101 4.02 1.97 409 464 346 3.32 1.51 2.42 C,B,A 9,254
B 0.020 0.115 3.50 1.52 412 437 384 3.32 1.37 1.93 A,C,B 11,193
C 0.023 0.139 3.09 1.21 410 437 362 3.78 1.27 1.67 C,B,A 9,903
D 0.030 0.144 2.98 1.15 410 437 362 3.90 1.23 1.62 B,A,C 9,903
E 0.040 0.141 1.16 0.69 410 437 362 3.209 0.90 0.91 C,B,A 9.903
F | 0.050 O0.148 1.44 0.82 410 437 362 3.50 0.90 1.04 C,B,A 9,903
G 0.060 0.145 3.00 1.29 410 437 362 3.89 1.21 1.71 B,A,C 9,903
H 0.067 0.147 13.06 1.37 j 410 437 362 4.00 1.24 1.81 C,B,A 9,903
IT
,2
A
00 10/18/82 v1V1 8J1 2 \I3\
4.0 PLOT TYPE 9
GEN-LP TORQUE IN P.U. Fi.0 NODE NRMES MRCH 1 T442V2
Fig . 6 GEN-LP Torque For 55% Compensation and .017
sec. Fault Clearinq Time
2.0 *
0
ENSATIONIICASE TORQ E I I I
X:
>-('\J
0
01 20 A
0 0.5
L
[Al] P. Dandeno, et al., "Recent Trends and process in Synchronous
machines Modeling in the Electric Utility Industry," Proc. IEEE, .0 .04 .06 p"
vol. 62, pp. 941-950, July 1974. .02
tities and Exact Equivalent Diagrams of the Synchronous Fig. D3. Effect of rotor body to field mutual impedance on SEDC damp-
Machine," IEEE Trans., vol. PAS-88, No. 7, pp. 1114-1120, July ing of torsional modes.
1969.
[A3] R. Schultz, et al., "Dynamic Models of Turbine Generators Derived Manuscript received August 13, 1984.
from Solid Rotor Equivalent Circuits," IEEE Trans, vol. PAS-92,
pp. 926-933, May/June 1973.
sL
rc
s = fr/fo
Fig. D1. D-axis equivalent circuit for solid iron rotor synchronou.S
machines. R. G. Farmer: On behalf of the SSR Working Group, I wish to thank
Mr. Hammad for his discussion which makes a valuable contribution
to the working group paper.
We agree with Mr. Hammad that the classical two-axis rotor circuit
model may not be ideally suited for SSR analysis. A model containing
the coupling impedance between the rotor body and field circuit may
improve the model for SSR studies. However, the generator models for
efd the second benchmark model paper were selected to provide compatible
models for the various programs normally used for SSR. This was best
accomplished by the classidal fotor circuit model since EMTP and some
eigenvalue programs do not have provision for the rotor body-field cir-
cuit coupling term. Mr. Hammad's point here is quite valid and the SSR
Working Group will address the question of rotor circuit models for SSR
s = fr/fo analysis. If it is found that there are more suitable models, the working
group will encourage that these models be incorporated into programs
Fig. D2. Simplified d axis circuit.
-
used for SSR analysis.
1066
The discusser has correctly pointed out that Fig. 4 cannot be correlated
with Table 4 due to an error in the horizontal scale of Fig. 4. A cor-
rected Fig. 4R is included with this closure, and in accordance with Mr.
Hammad's suggestion the average damping from the EMTP studies has
been added to Fig. 4R.
Once again we want to thank Mr. Hammad for his valuable discussion.
1.5
0 AVERAGE UNDAMPING FROM EMTP
LUi
LU-
FJ
C) 1.2 MOST UNDAMPED EV
FROM EVP. TI
ry
-,
REPRESENTED
LJ
LUI
0.9
L.)
CC MODE 1 UNIT 1
FROM FSP
0-
0.6
CD
r-
0.3
QD
50 60 70 80 90
% COMPENSATION