New Tools For Analysing Power System Dynamics

1
New Tools for Analysing Power System Dynamics

Ian A. Hiskens Department of Electrical and Computer Engineering University of Wisconsin - Madison (With support from many talented people.)
PSerc Research Tele-Seminar March 2, 2004
PSERC
Two main themes
Dealing with parameter uncertainty How much condence is there in load model parameters? Generator parameters? Simulation is too time-consuming to repeat studies for multiple parameter sets. Inverse problems What does this mean? Trajectory sensitivities form a common link between these two areas.
Inverse problems
Direct problem:
input x process K model output ?
Inverse problem (system identication):

input x process ? model output y
Inverse problem (causation):

input ? process K model output y
Inverse problems (continued)
Power system context Identication problems Load models and parameters Generator parameters What controller parameters ensure appropriate fault-recovery dynamics? What load changes maintain dynamic security? Under what conditions could a fault induce incidental protection operation? Protection operation caused by post-fault transients.
Trajectory sensitivities
Let (t; ) describe the trajectory (ow), at time t, due to parameters . Determined by simulation. The corresponding trajectory sensitivities are (t; ) Characteristics of sensitivities : Quantify the change in a trajectory due to a small change in parameters. Can be computed eciently, as a by-product of simulating the nominal trajectory. Well dened for nonlinear non-smooth systems. Provide gradient information that underlies Newton-type algorithms (shooting methods).
(t; ).
Trajectory approximation
Taylor series expansion of the ow gives, (t; + ) = (t; ) + (t; )
6
Example: Generator eld voltage.

Field voltage, Efd (pu)
1 Actual computed trajectory, tcl=0.23 sec, Efdmax=5.8 0 Approximate perturbed trajectory, tcl=0.21 sec, Efdmax=5.0 Actual perturbed trajectory (for comparison) 1 0 0.2 0.4 0.6 0.8 1 Time (sec) 1.2 1.4 1.6 1.8 2
Error bounds
Many system parameters are not known precisely, but are better described by probability distributions. Select parameter sets using a Monte-Carlo process. Form an approximate trajectory for each set.
1.2
Nominal trajectory Terminal voltage (pu) 0.8 Error bound 0.6
0.4
0.2
0.2
0.4
0.6
0.8
1 Time (sec)
1.2
1.4
1.6
1.8
Parameter estimation
Desire a systematic approach to: Determine which parameters are well conditioned (identiable). Estimate those parameters Algorithm: Minimize the nonlinear least-squares cost 1 J () = 2
m
(tk ; ) ms(tk )
k=1
via the Gauss-Newton iterative scheme (j )T (j )j+1 = (j )T (j ) ms j+1 = j j+1 j+1
Parameter estimation (continued)

Nordel example
450
Bus 12 voltage, V
12
(kV)
400
350 Measured Initial parameter values Estimated parameter values
300
5 Time (s)
10
10
Performance specication (grazing)
Trajectory is tangential to a performance boundary.
Parameter values g + Boundary
x+ xg
11
Performance specication (continued)
Example: Generator eld control. What value of Ef d,max ensures voltage does not overshoot abnormally following a fault?
Vref Vt 1 1+sTR + Efdmax 1+sTC 1+sTB Efdmin KA 1+sTA Efd
VPSS
Vmax Vmin
1+sT1 1+sT2
sTw 1+sTw
KPSS
12
Terminal voltage constrained below 1.2 pu during initial transient.

1.3 Hypersurface b(x,y): V 1.2=0 1.2
t
1.1
1 Terminal voltage, V (pu)
0.9
0.8
0.7
Initial parameter value Grazing parameter value
0.6
0.5
0.4
0.3
0.2
0.4
0.6
0.8
1 Time (sec)
1.2
1.4
1.6
1.8
13

Example: Distance protection. How far can load be increased before a fault disturbance induces incidental protection operation?
Initial trajectory 1.2 Grazing trajectory
b(x,y)=0
0.8
0.6 X 0.4 a 0.2 c

2 2
a1
b2 b
1
0.2
0.4 0.2 0 0.2 0.4 0.6 0.8 R 1 1.2 1.4 1.6 1.8
14

Example: Distance protection with timing. Replace conditions enforcing tangential contact by conditions specifying time inside mho characteristic.
Initial trajectory 1.2 Timeconstraint trajectory
b(x,y)=0
0.8
0.6 X 0.4 a 0.2 c2 0 b2 c

1
b 0.2
0.4 0.2 0 0.2 0.4 0.6 0.8 R 1 1.2 1.4 1.6 1.8
15

Initial, grazing, and time-constraint cases.
3 Initial trajectory 2.5 Grazing trajectory Timeconstraint trajectory
1.5 b(x,y)
0.5
0 =0.3 s 0.5 0 0.2 0.4 0.6 0.8 Time (s) 1 1.2 1.4 1.6 1.8 2
16
Dynamic embedded optimization
Minimize the cost J () = (tf ; ), , tf +

tf
(t; ), , t dt
t0
where (t; ) satises the dynamic model. Closely related to optimal control, but optimizing over nite dimensional design parameters. Some technical issues arise for hybrid (switched) systems if event order changes.
17
Dynamic embedded optimization (continued)
Example: AVR/PSS tuning What values of PSS output limits give best damping?
Vref Vt 1 1+sTR + Efdmax 1+sTC 1+sTB Efdmin KA 1+sTA Efd
VPSS
Vmax Vmin
1+sT1 1+sT2
sTw 1+sTw
KPSS
18
Dynamic embedded optimization (continued)

Optimization adjusted the lower PSS limit from -0.1 to -0.33 pu. Noticable damping improvement.
2.5
Generator angle, (rad)
1.5
0.5
Initial limit values Optimal limit values
0.5
1.5
2.5 Time (sec)
3.5
4.5
19
Conclusions
Parameter uncertainty is unavoidable in power systems, and should be considered in decision making. This is computationally feasible using rst-order approximations of trajectories. Many analysis and design processes are eectively inverse problems. Such problems can be solved using gradient-based iterative algorithms. In both cases, ecient computation of trajectory sensitivities underlies practical algorithms.
20
Extra Material
21
Simulation model
Dierential Algebraic Impulsive Switched (DAIS) model x = f (x, y) x+ = hi (x , y ),
x = f (x, y) +
i=1
(yr,i ) hi (x, y) x
s
yr,i = 0
0 = g(x, y) g (0) (x, y) +

j=1
g (j) (x, y)
where g (j) (x, y) g (j) (x, y) = g (j+) (x, y)
ys,j < 0 ys,j > 0
j = 1, ..., s
22
Trajectory sensitivity computation

Smooth sections of trajectories evolve according to x = f (x, y) 0 = g(x, y) Dierentiating with respect to x0 gives f x = (t)x + x g (t)x + 0= x Initial conditions x (t0 ) = I f (t)y y g (t)y y
23
Trajectory sensitivities at events

At an event (at time ), sensitivities evolve according to jump conditions, x ( + ) = x ( ) f + f
Nominal trajectory Perturbed trajectory
x0
x0 f x( ) + x x x( ) + x0 + f x( + )
Triggering hypersurface
24
Grazing and time-constraint formulation
For simplicity, the ODE form of equations is presented rather than the DAE form.
Grazing (tg ; ) xg = 0 b(xg ) = 0 bx (xg )f (xg ) = 0
Time constraint (t1 ; ) x1 = 0 (t2 ; ) x2 = 0 b(x1 ) = 0 b(x2 ) = 0 t1 t2 spec = 0

New Tools For Analysing Power System Dynamics

Uploaded by

Copyright:

Available Formats

New Tools For Analysing Power System Dynamics

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

New Tools For Analysing Power System Dynamics

Uploaded by

Copyright:

Available Formats

1

New Tools for Analysing Power System Dynamics

PSerc Research Tele-Seminar March 2, 2004

Two main themes

Inverse problem (system identication):

Inverse problem (causation):

Inverse problems (continued)

Example: Generator eld voltage.

Nominal trajectory Terminal voltage (pu) 0.8 Error bound 0.6

via the Gauss-Newton iterative scheme (j )T (j )j+1 = (j )T (j ) ms j+1 = j j+1 j+1

Parameter estimation (continued)

350 Measured Initial parameter values Estimated parameter values

Performance specication (grazing)

Trajectory is tangential to a performance boundary.

Parameter values g + Boundary

Performance specication (continued)

Performance specication (continued)

Terminal voltage constrained below 1.2 pu during initial transient.

1 Terminal voltage, V (pu)

Initial parameter value Grazing parameter value

Performance specication (continued)

0.6 X 0.4 a 0.2 c

Performance specication (continued)

0.6 X 0.4 a 0.2 c2 0 b2 c

Performance specication (continued)

Dynamic embedded optimization

Minimize the cost J () = (tf ; ), , tf +

Dynamic embedded optimization (continued)

Dynamic embedded optimization (continued)

Generator angle, (rad)

Initial limit values Optimal limit values

2.5 Time (sec)

Dierential Algebraic Impulsive Switched (DAIS) model x = f (x, y) x+ = hi (x , y ),

0 = g(x, y) g (0) (x, y) +

where g (j) (x, y) g (j) (x, y) = g (j+) (x, y)

ys,j < 0 ys,j > 0

Trajectory sensitivity computation

Trajectory sensitivities at events

Grazing and time-constraint formulation

Grazing (tg ; ) xg = 0 b(xg ) = 0 bx (xg )f (xg ) = 0

Time constraint (t1 ; ) x1 = 0 (t2 ; ) x2 = 0 b(x1 ) = 0 b(x2 ) = 0 t1 t2 spec = 0

You might also like