perolofl's profile - activity

Maybe you should present the progress window output.

The bus voltage frequency is calculated from the voltage angle with a filter of normally 40 ms (four time steps). That may explain the lag you see.

Initialise PSSE before the loop and perform case and Predyr before the if-statements.

psspy.psseinit(10000) #initialize PSSE
for i in range(1,16):
    psspy.case(r"""C:\Users\bbbb4\Downloads\112L-150M.sav""")
    Predyr()
    if i==1:#case 1
        psspy.strt(0,r"""C:\Users\bbbb4\Downloads\112L-150M\case1.out""")
        #### here's the code of event 1 ####
    if i==2:#case 1-1
        psspy.strt(0,r"""C:\Users\bbbb4\Downloads\112L-150M\case1-1.out""")
        #### here's the code of event 2####

Unfortunately, the python recording of activity GDIF is not working (33.12.2). I propose you try activity DIFF instead. Start recording and perform the activity from GUI (File - Compare…).

Open and study the created python script in a texteditor, and read about diff in API manual.

What do you mean not visible in .dyre? Can't you see the model in dynamic spreadsheet after reading the dynamic raw data file with DYRE? If so, check for reading error messages in progress window. It might be that the model wasn't read due to an error in the model record in the .dyr file.

The DLL file has nothing to do with the visibility of model calls in the spreadsheets.

You have to write your own model or use two separate REPCA.

The idea is to update governor GMAX and GMIN to the same MW-value as PMAX and PMIN in the saved case.

Use amachint and amachchar to get lists with all generators. Loop the lists and get governor model name with mdlnam. Get CON starting index with mdlind. Use dsrval to get the CON value for GMAX and GMIN. Update those values, if necessary, with change_con if governor limits are too narrow.

The only auto-correct PSSE can do is to net in-service generators with missing generator model.

There is no problem to write a script for auto-correction. I have done something similar that updates the governor limits based on Pmax in saved case.

However, for exciter limiter violations I prefer to solve it manually, since almost always the problem lies in wrong load flow data, e.g., wrong Q-limits, solved without Q-limits, wrong Mbase, etc. For a correctly initialised exciter model, Emax should be well above the calculated field voltage (Efd).

There is no such API available. You can write your own function that returns the load or machine current calculated from the MVA load and the bus voltage.

Edit:

As an example, I wrote a script that plots the complex currents in a vector plot. See example below for bus 201 in savnw network. There are three lines, one 2W-transformer and one fixed shunt at the bus. The bus voltage is the reference angle (0 degrees).

No, you cannot control two sources with REPCA model.

You should set both setpoint individually.

What you see in PSSE is the bypass of the inverter, not the effect on the HVDC of a commutation failure. On the AC side the commutation failure is modeled with a shunt. You need a Pscad-study to see the dynamics of the link during a commutation failure.

Use API

var_channel

And

state_channel

Edit: The starting index for plant model are calulated with:

ierr, index = psspy.mdlind()

The starting index for wind model are calculated with:

ierr, index = psspy.windmind()

See the API manual for the arguments of the functions.

Just supervise the rotor angle and check if the generator goes out of step, i.e., as you mentioned above.

It is much better to disable the model…:-)

Isn't it just to change to write access for the file???

Activate the convergence monitor at initialization. You will then see which bus has the largest mismatch.

You can report all buses with voltages below a certain limit, for example 90% voltage:

psspy.vchk(0,1, 0.9, 2.0)

I don't know about OPF but in PSSE the cost curves are given in economic dispatch files. See Program Operation Manual and "Performing Unit Commitment and Economic Dispatch".

An economic dispatch file has the file extension .ecd.

Only the positive sequence is modelled in dynamic simulations. An unbalanced fault is then modelled with an equivalent positive fault admittance, so the positive sequence voltage will be correct during the fault.

In your case the positive sequence voltage is calculated to be 80% during the fault.

2. Yes, it is the reactance of the machine.

A DLL-file for rev 33 cannot be used in rev 34. The user model must be compiled and linked again.

You can try the following line directly in CONEC:

CON(L+25) = VOTHSG(MCHNINDX)

where L is the first VAR number of SVSMO1U2 and MCHNINDX is the machine index for the dummy generator. This should update the POD input of SVSMO1U2 with the signal from STBSVC.

Yes, but it is much easier to use bsnm as shown in the answer Aug 17, 2018.

1. Xsource must be low for an infinite bus, as ffl wrote. But, Zsource is not used in load flow so it doesn't affect the steady state voltage.
2. Transient reactance is only applicable in dynamics, not in load flow.
3. The hand calculation is wrong, because the PSSE results are correct.

The total PLOAD is in channel # 4.

Yes, any fault impedance is given in pu and a solid fault is entered with XL-G = 0.

The pu data system is described in manual "program Application Guide, Volume 1, Chapter 3.

Check the dynamic models that have STATEs with initial conditions suspect.

Use channels for "Subsystem Power Totals" and select "All buses". Five channels are added:

1: TOTAL PMECH

2: TOTAL PELEC

3: TOTAL accelerating power

4: TOTAL PLOAD

5: TOTAL PELEC - PLOAD

The total accelerating power is a good indicator of the deficit of turbine power in the system.

In a script, use psspy.chnval to retrieve the accelerating power during the simulation.

What's your email?

Local data can be transfered between models by using common blocks, i.e. model A and B can share and update the same variables.

I have read POM about Coordinated Call Models. It says "coordinated call models, both those from the PSS®E Model Library as well as user-written models, must have their primary entry point name beginning with a character other than T.".

If I want to write a current injections model, how should I do?

1: Write a model begin with SUBROUTINE Modelname (MC, ISLOT) and call it in other model by writing:call Todelname (MC, ISLOT)?

or 2:

SUBROUTINE Modelname (MC, ISLOT)
include 'comon4.ins'
local declarations
(mode 1,2...)
RETURN
ENTRY Todelname (MC, ISLOT)
(mode 1,2...calculate current injection)
RETURN
END SUBROUTINE

Alternative 2 is correct.

It seems you have initial conditions suspect after first STRT. It is better to correct the issues causing initial conditions suspect. Is the saved case properly solved? Are any generator models hitting limits?

Yes, you are correct. The second VAR is the Q reference for model REPCA. The output signal, also named Qref (!), seems to be hidden.

Qref is the second VAR of model REPCA1. Put the VAR into a channel.

Qref is the second VAR of model REPCA1. Put the VAR into a channel.

PSSE is using pu speed deviation, not Hz. At nominal speed (50 or 60 Hz) the pu speed deviation is zero. At under-frequency the pu speed deviation is negative.

What do you mean with Qext? Model REPCA1 sends signals Pref and Qref to REECA1.

I assume the mismatch became too high for Newton/Raphson to solve, so you did it right in your case.

In load flow the base frequency is only used for conversion of susceptance in pu system base to capacitance in microFarad in the GUI if line option is Ohms. So, there is no problem to have 50 Hz and 60 Hz systems in the same saved case.

In PSSE all network data are stored in pu.

It is possible to have one saved case combining a 50 Hz and 60 Hz system, but that saved case will only have ONE base frequency; 50 or 60 Hz. You can't assign base frequency to individual buses.

Before reading the raw data for the second system you must change the frequency at the first line in the raw data file to the frequency of the first system. Also, change the IC code to 1. Note, there is no need to recalculate any branch or transformer data.

The line capacitance in the second system will be shown wrong (in spreadsheets and reports) if transmission line input option is set to Ohms, since the capacitance is calculated with wrong base frequency. However, the pu value is correct and the load flow is correct.

That means it is just to pretend that, for example, the 60 Hz system is a 50 Hz system and add it to the 50 Hz system.

One way to list the buses is to export a bus location file.

From menu: File - Export - Bus locations...

There is no need for balancing the flow. Point 2 is not needed.

No, I mean what I wrote. You can ignore XTRAN and GENTAP if the step-up transformer is modelled in the saved case. If so, the transformer is already modelled and there is no need to enter the data in the machine record.