EE 554 Course Project

1
EE 554 Course Project

• James D. McCalley, Siddhartha Khaitan
• Spring 2009

2
Introduction

• Your assignment is to perform an operating study
  for the Diablo Canyon Nuclear Power Plant using
  the power flow and stability data provided to you
  by the instructor. The objective of the study is
  to determine the safe operating limits for the
  plant, in terms of MW output power, under the
  NERC disturbance class B given in the NERC
  Planning Standards. You may assume that this
  criterion is condensed to the system must
  perform satisfactorily for a three-phase fault at
  the machine terminals followed by loss of a
  single circuit. To perform satisfactorily, the
  machine must be stable and no first-swing
  transient voltage dip can fall below 0.75
  per-unit.

3
Problem Statement

• There are three 500 kV lines emanating from the
  plant under normal conditions. You are to
  develop the MW operating limits for a weakened
  condition whereby one line is out on maintenance.
  Since two of the lines are identical, this means
  that you must develop two different operating
  limits (with lines A, B, C, and A and B
  identical, you must identify (1) the operating
  limit with A out and (2) the operating limit with
  C out.

4
Study Area
Fig. 2 Lower-level view of study area
5
Data and Model

• You are provided with 2 data files
• two_diab.sav The power flow saved case. I can
  also provide raw data (IEEE or PTI) if necessary.
• wscc_pti_dyr The machine data for all machines
  in the system.
• These data files are from a test system which is
  similar in structure to the transmission system
  of the western U.S. However, I stress similar
  for two reasons
• The system is a product of gross approximation.
  Many essential features are omitted in order to
  keep the system relatively simple and small in
  terms of number of buses, lines, and machines.
  For example, this system represents only 179
  buses. Actual models of the western US grid
  typically have on the order of 10000 buses.
• I do not have authority to provide accurate
  transmission models of the western US grid.
• So one should understand that any results
  obtained using this model pertains only to this
  model and has no relevance whatsoever in regards
  to the actual western US grid.

6
Reduced System model

• Two high-level 1-line diagrams of the overall
  system are given in Fig. 1 below, and Fig. 2
  provides a more refined view of the portion of
  the system to be studied.

Fig. 1 High-level system one-line
7
Fig. 2 Lower-level view of study area
8
Important Points

• You should be aware that there are two identical
  generating units at Diablo Canyon. Dynamic data
  for one of these plants is given (Assume that the
  data is correct.) However, the data in your
  system files (power flow and stability)
  represents only a single unit at the power plant.
  I have scanned the system data files and feel
  that the data for this plant is questionable. You
  need to check it to verify that it appropriately
  represents a single machine equivalent of the two
  machines given in the data below. Specifically,
  you need to check
• The power flow data, especially the MVA base and
  the transformer impedance (the transformer
  impedance should be approximately 0.10 pu, when
  given on the machine base, but of course
  converted to the 100 MVA system base).
• The inertia constant, reactances, and time
  constants of the machine.

9
Cont

• The data at the end of this file is in another
  format, but I have provided you with Q-cards that
  you can copy and paste over the different data
  cards in order to easily see what the data is.
  You need to provide the data used in your final
  report and identify any assumptions you used.
  Note that one very simple approach here is, since
  you may assume that the data for the two units
  are identical, input data for one unit as given
  below, and then identify the MVA base in the
  power flow data as twice the MVA base of a single
  unit. This will have the effect of forcing the
  program to interpret the machine data as if it
  were given for a machine on the higher MVA base.
  Alternatively, you may represent two different
  machines at the plant, each with their data as
  given below and their actual MVA base represented
  in the power flow model. In this case, you would
  need to represent two separate transformers as
  well, each with approximately 0.10 pu reactance
  given on the base of a single machine, or a
  single transformer with a 0.05 pu reactance given
  on the base of a single machine (converted to the
  100 MVA system base, of course).
• Note Since the transformer impedance is in the
  direct path of the generator circuit, it is very
  important to get it right. Getting it wrong will
  make a large difference in your results!

10
Q card for single machine
11
Cont

• Note You should perform the study with only the
  machine model, i.e., do not represent the
  excitation system, power system stabilizer, or
  the turbine-governor dynamics.

12
PSS/e Access

• The basic commands to access the needed software
  in that lab are (you may need to perform csh
  first on the unix machines) ? open psse command
  prompt
• - psslf4 (to get the power flow program) ? or
  psse
• - pssds4 (to get the time domain simulator)
• - pssplot (to get the plotting program) ? pssplt
• Off-campus students may have access to PSS/E via
  their employer, and if acceptable to the
  employer, you may certainly use it that way. If
  you do not, then you can remote access to PSS/E
  via the following procedure

13
PSS/e Manuals

• The manuals are available by opening within Adobe
  Acrobat the file CONTENTS.pdf, then click on
  Programs Operation Manual, and then Volume
  I.   Also, Volume II of the Programs
  Operation Manual will be helpful in identifying
  data formats used by the PSS/E programs. These
  can accessed from the folder C\Program
  Files\PTI\PSSE30.3\DOCS\contents.pdf

14
Part I Data Modifications

• A) Power Flow data
•  
• 1) Single machines MVA base was found to be 1340
  MVA, so for two machine equivalent MVA base is
  2680 MVA
• 2) The transformer impedance was 0.00980 p.u.
• If the impedance is 0.1pu at machine base of 2680
  MVA, then at 100 MVA system base,  
• 
  0.1/2680100 0.003731343 pu
• So the transformer impedance was changed to
  0.003731343 pu
• B) Dynamics data
• Inertia, time constant and reactances

15

• Equivalent single machine data of 2 machine
• Original data in diab_pti.dyr file
•  
• Read as IBUS, GENROU, I, Tdo, T"do, T"qo,
  T"qo, H, D, Xd, Xq, Xd, Xq, X"d, Xl, S(1.0),
  S(1.2)/
•  
• 103 'GENROU' 1 6.12000 0.05200 1.50000
  0.14400
• 3.46000 0.00000 2.12900 2.07400
  0.46700
• 1.27000 0.31100 0.25000 0.09000
  0.38000 /
•  
• Data to be changed
•  
• 1) Inertia, H Inertia Constant for 1 machine is
  4650 MWs. Therefore, H 4650/1340 MW.s/MVA 3.47
  MW.s/MVA
•  
• 1) The two machine are assumed to be swinging
  together.
• For one unit, H4650/13403.47, on the machine
  base.
• For two units, H9300/26803.47, on the machine
  base.
•  
• 2) Time constants Assuming that they will stay
  the same
•  

16
Open the psse load flow and change the MVA of the
machine and the impedance of the transformer.
17
(No Transcript)
18
Solve the power flow. Save and Quit
19
PSSE Dynamics

• Line outage to create weakened condition can be
  done in power flow or while dynamic simulations.
  We will do the second.
• Next we will load the case in PSS/e dynamics
  environment and initiate the power flow module to
  create the weakened condition
• Go to PSS/E command prompt ? type pssds4
• Click LOFL ? Select READ ? select base case
• Run power flow using Newton Raphson method (use
  the powerflow menu)

20
Load the case
21
Run the power flow
22
Save
23
Contingency to obtain weakened condition

• Removing a line (so as to develop the weakened
  condition) Do edit, then loadflow data, then
  branch, and select, for example, 102-104, with
  circuit ID1. Then select statusout. Then
  resolve the case.
•  

. Save this file in some name, basecase.sav (It
will be used again and again to find the
operating limit for generator output by
increasing the load in subsequent simulations in
steps).
24
Study 1 - BASE CASE No increase in Generation at
bus 103 and system load

• Preparing for a stability run You must perform
  several actions before the case is ready for a
  stability run. These are as follows
• Perform CONG. This converts the generators to
  Norton equivalents (constant current injections).
• Perform CONL, ALL. This assigns load
  characteristics to the loads. I suggest that you
  use 50 constant current and 25 constant
  impedance for both real and reactive loads
  (leaving the other 25 to be constant power).
• Perform ORDR. This re-orders the buses for
  sparsity (required because we converted the swing
  bus to a type PV bus).
• Perform FACT. This factorizes the A-matrix.
• Perform TYSL. This performs what you might think
  of as an simplified load flow calculation
  (basically just an IYV).
• Perform SAVE. This saves the converted case.\
• FACT/RTRN
• Picking up an already converted case Each time
  you pick up an already converted case, then you
  need do only the following commands LOFL (if
  you need to toggle from the time-domain simulator
  to the power flow program), then CASE, file,
  then FACT, and then RTRN.

25

• Performing a stability run Access the
  time-domain simulator environment using pssds4.
  The below command sequences are from the command
  line. Most sequences have corresponding actions
  that can be taken from the menu.
• Enter DYRE, and then enter the filename of the
  dynamic data, then a carriage return.
• Perform DYCH. Then
• Perform the consistency check (1)
• Chan, (3) and look at generator 103. You will
  see GENROU (machine data), IEEEST (stabilizer),
  and EXST1 (exciter). Toggle off the stabilizer
  and exciter so that you are modeling only the
  machine dynamics.
• (or)
• Take out the generator stabilizer and
  exciter data from the wscc_pti.dyr file
  beforehand
• Enter CHAN.
• Program responds with Enter starting channel or
  carriage return. Do a carriage return.
• Program responds with Enter output category.
  Choose 1 (angle).
• Program responds with Enter bus number, mach ID,
  identifier. Type 103,1
• Program responds with Enter bus number, mach ID,
  identifier. Type 0.
• Repeat the above b-d steps for output categories
  2 (Pelect), 4 (Eterm), and 7 (speed).

26

• Enter STRT. This will perform the initial
  condition calculation. Program responds with
  Enter channel output filename. Enter a filename
  with a .out suffix. Program responds with
  Enter snapshot filename. Enter a filename.
• Enter RUN. Program responds with Enter Tpause,
  NPRT, NPLT, CRTPLT. Tpause is the simulation
  end-time, NPRT is the frequency of time steps to
  write to the screen. NPLT is the frequency of
  time steps to write to the plotting file. Suggest
  entering 1,0,1,0. This will run the simulation
  from 0 to 1 second, writing nothing to screen and
  writing every time step to the plotting file.
• Enter ALTR. This is the command to make network
  changes. First you need to apply the fault, then
  run the simulation, then clear the fault and drop
  the line, the run the simulation until done. The
  step we are taking here is to apply the fault.
  Here is a suggested sequence
• After entering ALTR, program responds with
  Enter change code. Enter 0 for no more changes.
• Program responds with Network data changes?
  Enter 1 for yes.
• Program responds with Pick up new saved case.
  Enter 0 for no.
• Program responds with Enter change code. Enter
  1 for bus data.
• Program responds with Enter bus number. Enter
  102.
• Program responds with Enter code, G, B. Enter
  1, 0, 99999999. This puts a fault with a very
  large susceptance at the bus (effectively,
  putting a short-circuit at the bus).
• Program responds with Change it? Enter
  carriage return.
• Program responds with Enter load ID. Enter -1.
• Program responds with Enter bus number. Enter
  0.
• Program responds with Enter change code. Enter
  -1 to exit.
• Enter RUN. Program responds with Enter Tpause,
  NPRT, NPLT, CRTPLT. Enter 1.0666, 0, 1, 0 (this
  will apply the fault for 4 cycles).

27

• Enter ALTR. (Now you need to clear the fault
  and remove the line.)
• Program responds with Enter change code. Enter
  0.
• Program responds with Network data changes.
  Enter 1
• Program responds with Pickup saved case. Enter
  0.
• Program responds with Enter change code. Enter
  1 for bus data.
• Program responds with Enter bus number. Enter
  102.
• Program responds with Change it? Enter Y.
• Program responds with Enter change code, G, B.
  Enter 1, 0, 0.
• Program responds with Change it? Enter carriage
  return.
• Program responds with Enter load ID. Enter -1.
• Program responds with Enter bus number. Enter
  0.
• Program responds with Enter change code. Enter
  -3 for branch data.
• Program responds with Enter from bus, to bus,
  circuit ID. Enter 102, 108, 1 (This is if you
  want to remove one of the circuits from Diable to
  Midway.)
• Program responds by giving the data for the
  indicated branch and then asking Change it?
  Enter Y.
• Program responds by querying for new data. Enter
  0 to toggle status from in to out.
• Program responds by giving the shunt data for the
  branch and then asking Change it? Enter N.
• Program responds by asking to reverse the metered
  ends. Enter carriage return.
• Program responds with Enter from bus, to bus,
  circuit ID. Enter -1.
• Enter RUN. Program responds with Enter Tpause,
  NPRT, NPLT, CRTPLT. Enter 10, 0, 1, 0 (this
  will simulate the system response for 10
  seconds).

28
Plotting

• From the Unix command line, enter pssplt to
  bring up the plotting program. Your plot data
  will be in the file that you named in step B-4
  above. I suggest using the menu commands. The
  essential ones are as follows
• CHNF (select the output file ltfilenamegt.out)
• TINT (give start and end time for plotting)
• SLCT (select the output to be plotted from the
  available channels)
• PLOT

29
Base Case results
The plot below shows the generator 103s angle,
power output, terminal voltage, and speed. It
can be seen that after the disturbance is
cleared, there is some transients and gradually
the system is settling down to a stable state.
(Might have to simulate for more than 10 s to see
the stable case)
30
Study 2 - Generator Operating limit Search

• Open again pssds4
• LOFL
• CASE
• Open the weakened saved case basecase.sav
• Now, we can increase the generator output at bus
  103. To change generation at bus 103 (Diablo
  25.), you must change generation elsewhere or
  change load (if you just change Diablo generation
  without making any other change, you will be
  implicitly forcing the swing bus to take the
  adjustment). I suggest to just scale the total
  system load. You may do this using edit, then
  changing, then scale, then all buses, and
  then go. Then resolve the case.

31
Changing the Load
32
(No Transcript)
33
Changing the generation
34
Run powerflow
35
Case 620 MW increase of generation at bus 103
and system load increase of 620 MW

• For this study, I changed 1 of total system
  load, which is about 620 MW (base system load
  62000.1 MW), with keeping the system power factor
  constant (Q/P ratio constant). Then I increased
  the generator output at bus 103 by 620 MW. New
  output is 2620 MW.
• This is done by reloading the power flow case
  that was saved initially with the name
  basecase.sav, then running a powerflow, and then
  perform the changes, and again run the powerflow
  to get the new initial operating condition before
  perform stability study.
•  
• Alternatively, you can identify a generator
  distance from your study and use it to balance
  generation changes you make. Again, each new
  operating point will require a new power flow
  solution.
•  
• For stability study, same procedure is carried
  out as presented in step 3 Preparing for a
  stability run till, step 4 Performing a
  stability run, and step 5 Plotting. The results
  are observed.

36
  We see in the below plot that the system become
unstable. The generator angle, power output and
speed are shown in the plot, and everything
indicates instability.