A Practical Approach for Faster Power System Transient Simulation · 2015. 6. 16. · V K sT Ts V...
Transcript of A Practical Approach for Faster Power System Transient Simulation · 2015. 6. 16. · V K sT Ts V...
A Practical Approach for Faster Power System Transient Simulation
34th Power Affiliates Program Annual Review
May 3, 2013
Soobae Kim, Thomas Overbye
Outline • Motivation
• Numerical Integration Method – Explicit method : Second Order RK method – Multi-rate method
• Practical problem definition
• Proposed Method
• Case study
• Conclusion
2
Motivation
• Power system dynamic simulation involves a wide range of different time constants [stiff]
• In order to avoid numerical instability, explicit integration methods require very small time steps for stiff power systems
• Only a small fraction of system states shows fast dynamics
• Therefore, simulation of an entire power system with small integration time steps is not efficient
• This work aims to reduce the transient simulation time by removing fast modes from a system.
3
Numerical Integration
Explicit method : Second Order RK method
1 1 21 1*( )2 2n nx x h k k-= + +
Region of Stability
where 1 1 1( , )n nk f x t- -=2 1 1 1( , )n nk f x k t h- -= + +
2( ) 11( ) 2zh z zz
rls
= = + +
1z <
2 0hl- < <Stable Region
if 2400l = -0.05h cycle<
Real hλ
Imag
hλ
-3 -2 -1 0 1 2 3-3
-2
-1
0
1
2
3
Stable Region
4
5 Figure from Jingjia Chen and M. L. Crow, "A Variable Partitioning Strategy for the Multirate Method in Power Systems," Power Systems, IEEE Transactions on, vol. 23, pp. 259-266, 2008.
Multi-rate Method
1. At each macro step(H), the slow variables(ys) are integrated.
2. Interpolate the slow variables to provide approximation for each micro step(h)
3. Fast variables can be found using the interpolated slow variables as needed
Practical Problem : BIG negative eigenvalues • WECC case (17,710 buses) SMIB result
Bus # 42711 Min. eigenvalue
-2601.8
• SMIB Participation factor
EXST1 exciter model
NEED Very Small Time Step
6
7
EXST1 exciter model
<EXST1 Exciter model> Figure from PowerWorld Manual
12
2
(1 ) (1 )( )( ) 1 ( )( )
A F A F
A F A F A F A F
V K sT K sTT sV T T s T T K K s T T s p s p+ -
+ += = =+ + + + - -
2( ) ( ) 42
A F A F A F A F A F
A F
T T K K T T K K T Tp
T T± - + + ± + + -
=where
When two lead-lag compensators are neglected,
V2
V1
T(s)
when
10-4
10-3
10-2
10-1
100
101
102
103
104
-40
-20
0
20
40
60
Mag
nitu
de [d
B]
Bode Diagram
10-4
10-3
10-2
10-1
100
101
102
103
104
-100
-80
-60
-40
-20
0
Frequency [Hz]
Pha
se [d
eg]
Original EXST1Reduced EXST1
8
1
2
(1 )( )( )( )
A F
A F
V K sTT sV T T s p s p+ -
+= =- -
1
2
(1 ) (1 )( )( )( ) 1 /
A F A Freduction
A F
V K sT K sTT sV T T p s p s p- + +
+ += = =- - -
EXST1 and its reduction of small eigenvalue
200, 0.01,0.5, 0.4
A A
F F
K TK T
= == =
0.01, 2512p p+ -= - = -
When the exciter input voltage shows only low frequency range,
the reduced model can be used
Only abrupt voltage change introduces high freq. component
Voltage change is dependent on fault given
9
Fault Type Dependency
Generator or Load fault Bus or Line to ground fault
When Load 1 open at bus 1
When Bus 1 to ground fault
109876543210
1
0.995
0.99
0.985
0.98
0.975
0.97
0.965
0.96
0.955
0.95
0.945
0.94
109876543210
1.014
1.013
1.012
1.011
1.01
1.009
1.008
1.007
1.006
1.005
1.004
1.003
1.002
1.001
1
0.999
0.998When Gen ID 1 open at bus 1
Bus 1 Voltage Magnitude
9876543210
1
0.95
0.9
0.85
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
When Line 1-2 to ground fault
9876543210
1
0.95
0.9
0.85
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0.3
10
Fault Type Dependency
0 10 20 30 40 50 60 70 80 90 1000
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09Single-Sided Amplitude Spectrum
Frequency (Hz)
|Y(f)
|
Bus to Ground faultLine to Ground faultGenerator OutageLoad Outage
FFT results of voltage magnitude
0 10 20 30 40 50 60 70 80 90 1000
0.01
0.02
0.03
0.04
0.05
0.06Single-Sided Amplitude Spectrum
Frequency (Hz)
|Y(f)
|
Bus 5Bus 4Bus 3Bus 2Bus 1
11
Bus Fault Location Dependency
21.91.81.71.61.51.41.31.21.110.90.80.70.60.5
1
0.95
0.9
0.85
0.8
0.75
0.7
0.65
0.6
0.55
0.5
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
When Bus to ground fault at Bus 5,
Electrically Close Electrically Far
Bus 5 Bus 1
Bus 1
Bus 5
Voltage Magnitude at different bus
location FFT results
of voltage magnitude
Bus 1
Bus 5
12
Proposed Approach I : Fault Dependent Model Reduction
13
Case Study : GSO 37 bus Case
JO345 EXST1 exciter min. eigenvalue : -2102
1. JO345 Gen ID1 OPEN
2. DEMAR69 Solid Bus to Ground fault
3. JO345 Solid Bus to Ground fault
0 1 2 3 4 5 6 7 8 9 10-10
-8
-6
-4
-2
0
2
4
6
8
Time [sec]
Gen
erat
or M
var T
erm
inal
[Mva
r]
Generator JO345 ID=2 Terminal Output when JO345 ID=1 Gen Outage
Original ModelReduced Model
0 1 2 3 4 5 6 7 8 9 10150
160
170
180
190
200
210
Time [sec]
Gen
erat
or M
W T
erm
inal
[Mva
r]
Generator JO345 ID2 Terminal Output when JO345 ID=1 Gen Outage
Original ModelReduced Model
14
GSO 37 bus Case : Simulation results
Generator Outage
Bus Voltage Magnitude at JO345
JO345 Gen ID1 OPEN
JO345 ID2 Gen MW output
JO345 ID2 Gen Mvar output
0 1 2 3 4 5 6 7 8 9 101.02
1.025
1.03
1.035
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 ID=1 Generator Outage
Original ModelReduced Model
0 1 2 3 4 5 6 7 8 9 10-10
0
10
20
30
40
50
60
70
80
90
Time [sec]
Gen
erat
or M
var T
erm
inal
[Mva
r]
Generator JO345 ID1 Terminal Output when DEMAR69 Bus Fault
Original ModelReduced Model
0 1 2 3 4 5 6 7 8 9 10140
142
144
146
148
150
152
154
156
158
160
Time [sec]
Gen
erat
or M
W T
erm
inal
[Mva
r]
Generator JO345 ID1 Terminal Output when DEMAR69 Bus Fault
Original ModelReduced Model
15
GSO 37 bus Case : Simulation results
Bus Fault - FAR
Bus Voltage Magnitude at JO345
DEMAR69 Solid Bus to Ground fault
JO345 ID1 Gen MW output
JO345 ID1 Gen Mvar output
0 1 2 3 4 5 6 7 8 9 10
0.92
0.94
0.96
0.98
1
1.02
1.04
1.06
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when DEMAR69 Bus Fault
Original ModelReduced Model
17
GSO 37 bus Case : Computation Time Benefits
Exciter Model
Numerical Integration
Max. Time Step
(cycle)
Time to Solve (sec)
Max. Angle Difference
(deg.)
Newton Solution Results
Total Newton
Solutions
Total Newton
Iterations
Number of Jacobian
Factorizations
Number of Forward
/Backward Substitutions
Original Single Rate 0.05 47.99 26.854 13975 12773 40 12773
Multi Rate 2.4 1.08 27.011 443 433 40 433
Reduced Single Rate 2.4 1.01 26.901 431 404 36 404
Bus Fault - FAR
DEMAR69 Solid Bus to Ground fault
Simulation Time : 10 sec
0 1 2 3 4 5 6 7 8 9 10-60
-40
-20
0
20
40
60
80
Time [sec]
Mva
r Ter
min
al O
utpu
t [M
var]
Generator JO345 ID=1 Terminal Output when JO345 Bus Fault
Original ModelReduced Model
0 1 2 3 4 5 6 7 8 9 100
20
40
60
80
100
120
140
160
180
Time [sec]
Term
inal
Out
put [
MW
]
JO345 ID1 MW output when JO345 Bus Fault
Original ModelReduced Model
Bus Voltage Magnitude at JO345
16
GSO 37 bus Case : Simulation results
Bus Fault - CLOSE
JO345 Solid Bus to Ground fault
JO345 ID1 Gen MW output
JO345 ID1 Gen Mvar output
0 1 2 3 4 5 6 7 8 9 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
EXST1 eigenvalues are changed
21
0 1 2 3 4 5 6 7 8 9 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
EXST1 eigenvalue = -2102
JO345 Solid Bus to Ground fault
0 1 2 3 4 5 6 7 8 9 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
EXST1 eigenvalue = -1102 EXST1 eigenvalue = -501 0 1 2 3 4 5 6 7 8 9 10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
0 1 2 3 4 5 6 7 8 9 100
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
EXST1 eigenvalue = -261 0 1 2 3 4 5 6 7 8 9 10
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Time [sec]
Vol
tage
Mag
nitu
de [p
u]
Bus Voltage when JO345 Bus Fault
Original ModelReduced Model
EXST1 eigenvalue = -27
Right eigenvector
22
1 1
( ) (0)i i
n nt t
i i i ii i
x t c e x el lf f y= =
D = = Då å
State \ Eigenvalue -2102 -1102 -501 -261 -27
Machine Angle 3.19E-10 -2.20E-09 -2.30E-08 -1.58E-07 0.000558929
Machine Speed w -1.78E-09 6.43E-09 3.06E-08 1.09E-07 -3.89E-05
Machine Eqp -6.81E-05 0.00013 0.000287 0.000554 -0.004242761
Machine PsiDp 8.22E-07 -3.04E-06 -1.53E-05 -6.02E-05 -0.022657701
Machine PsiQpp 1.11E-08 -4.10E-08 -2.04E-07 -7.88E-07 0.000444479
Machine Edp 0 0 0 0 0
Exciter Va 0.999999991 -1 -1 -1 0.999728538
Exciter VF 0.000119065 -0.00023 -0.00016 -0.00031 0.003303785
Governor Filter Output -1.71E-11 1.19E-10 1.28E-09 9.17E-09 -0.000218188
Governor Desired Gate 5.42E-15 -7.20E-14 -1.70E-12 -2.34E-11 5.55E-06
Governor Gate 5.42E-14 -7.22E-13 -1.71E-11 -2.36E-10 5.96E-05
Governor Turbine Flow -6.94E-17 1.76E-15 9.19E-14 2.46E-12 -6.81E-06
Revised Approach
23
Conclusion
24
• Practical approach is proposed in order to reduce a computational burden for power system transient simulation
• The method dynamically switches to original EXST1 or to reduced one depending on SMIB analysis and right eigenvector information
• The approach shows quite good performance in simulation time and accuracy. While maintaining the accuracy, the method reduces the simulation time about 98%
Questions ?
My mistake was
19
<EXST1 Exciter model>
V2
V1
T(s)
<SEXS Exciter model> Figure from PowerWorld Manual
Limiter block function is neglected
Tred(s)
The Reduced model
20
EXST1 model is used to define the reduced model transfer function by changing model parameters.
<EXST1 Exciter model>
V2
V1
Tred(s)
Questions ?
4
QSS(Quasi-Steady-State) approximation • Replaces short-term dynamics with their equilibrium
conditions • Thus transform the differential equations to algebraic ones
that are solved together with network equations
4
Numerical Integration • Approximate computation of an integral using numerical
techniques ( , , )
0 ( , , )
x f x y t
g x y t
·=
=
Explicit Method Implicit Method
1 1 1 1* ( , , )n n n n nx x h f x y t- - - -= + 1 * ( , , )n n n n nx x h f x y t-= +[Forward Euler] [Backward Euler]
For solving a stiff ODE,
Need very small time step
Larger time step is okay, But computational burden