James D. McCalley, Siddhartha Khaitan Spring 2009

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.

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.

Study Area Fig. 2: Lower-level view of study area

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 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.

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

Fig. 2: Lower-level view of study area

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.

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!

Q card for single machine

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.

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:

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”

Part I: Data Modifications: A) Power Flow data 1) Single machine’s MVA base was found to be 1340 MVA, so for two machine equivalent MVA base is 2680 MVA 2) The transformer impedance was p.u. If the impedance is 0.1pu at machine base of 2680 MVA, then at 100 MVA system base, = 0.1/2680*100 = pu So the transformer impedance was changed to pu B) Dynamics data Inertia, time constant and reactances

Equivalent single machine data of 2 machine Original data in diab_pti.dyr file Read as IBUS, ’GENROU’, I, T’do, T"do, T"qo, T"qo, H, D, Xd, Xq, X’d, X’q, X"d, Xl, S(1.0), S(1.2)/ 103 'GENROU' / 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, H=4650/1340=3.47, on the machine base. For two units, H=9300/2680=3.47, on the machine base. 2) Time constants: Assuming that they will stay the same 3) Reactance: Equivalent reactance of a parallel combination So new dynamic data to be used looks like: Xd, Xq, X’d, X’q, X"d, X"q, Xl, H, and D are in pu, machine MVA base. Data for equivalent single generator for 2 generators 103 'GENROU' / The above data is used in the modified version of diab_pti.dyr file.

Open the psse load flow and change the MVA of the machine and the impedance of the transformer.

Solve the power flow. Save and Quit

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)

Load the case

Run the power flow

Save

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, , with circuit ID=1. Then select status=out. 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).

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 I=YV). 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.”

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).

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, 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 , 0, 1, 0 (this will apply the fault for 4 cycles).

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). Save Snapshot if you want to start from here in future Enter “STOP.”

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.out) TINT (give start and end time for plotting) SLCT (select the output to be plotted from the available channels) PLOT

Base Case results: The plot below shows the generator 103’s 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)

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.

Changing the Load

Changing the generation

Run powerflow

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 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.

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.