perolofl's profile - activity

The speed channel in PSSE is in per unit speed deviation. Nominal frequency (60 or 50 Hz) is zero pu deviation. Positive pu speed deviation means overfrequency and negative means underfrequency.

Example: 61 Hz is (61-60)/60 pu speed deviation.

When plotting, use a formula and calculate (1+spd)*60 to get the channel value in Hz.

The length parameter is for user information only. Use it as you want. Just enter the line length in km. PSSE doesn’t use the parameter.

The block diagram for DEMOEX model:

The first time constant block has steady state gain 1 and the second block has steady state gain K. The output of time constant blocks are the same as the state of the block. This means that Esensed = Ec and E = Efd at steady state (mode 1). The two states therefore becomes:

STATE(K) = ECOMP(I)
STATE(K+1) = EFD(I)

where ECOMP and EFD are the arrays with measured voltage and field voltage respectively. I is the index for the machine.

The relationship between input and output at steady state is:

(Vref-Ec+Vothsg)*K = Efd

Efd is known (calculated by generator model in mode 1). Ec is terminal voltage if no compensation model is used. Vothsg is zero at steady state. The formula is rewritten to get the necessary value of Vref at initialisation:

Vref = Ec + Efd/K

The initialisation is made backward from the output signal (Efd) and forward from the input signals in mode 1.

How did you enter the model? Any message in progress window?

Note that var_channel defines a channel. It doesn't return the VAR value.

It is better you show your code and what is returned by awndreal. It seems like you give up too fast...

Status is a list, see the manual. Use psspy.var_channel([-1,239],'channel id')

There is no such model available in standard PSSE. You have to write your own model.

The API returns the index of the first VAR of the model. After simulation, you can only get the actual value of the VAR, i.e. the value at the end of the simulation.

Use API psspy.windmind. See API manaul.

There is no library model available for active power control of a CBEST. You have two options:

1. Updata VAR(L) manually during the simulation.
2. Write a frequency control user model with control logic that updates VAR(L) of CBEST during the simulation.

Use API dsrval:

ierr, time = psspy.dsrval('TIME', 0)

The API returns the current time.

Give me your email.

1. Check progress window for messages during the simulation.

2. Use activity Compare (DIFF) to compare the load flow after 15 s simulation with the original saved case to see if there are any changes in bus voltages or line flow. This may indicate what is going on in the simulation.

Model CBESTdoesn't have frequency control loop itself. Pout will be kept constant by the model. The power may be changed either manually or automatically by a separate model.

You can change Pout manually during the simulation by updating VAR(L).

For automatic primary frequency control you need to add a model updating VAR(L) of CBEST, as @Yunzhi Cheng wrote.

Try to add the model to savnw case and present the model data in this forum. As it is now it is difficult to nail down the problem.

Both CURNT and VOLT are indexed by IB (bus sequence number).

You must calculate IB after ENTRY TEST. As it is now IB is not set before you use it in CURNT(IB) = statement.

Also, normally there is a RETURN before the ENTRY point, i.e. CONEC-part of the subroutine returns before the code for CONET.

Thank you very much! Can be useful..

Select:

Winding I/O Code: Winding voltage (kV)

Winding 1 Ratio: 165.025

Winding 2 Ratio: 13.5

Impedance I/O Code: Zpu (winding base)

Winding MVA Base: 68.325

Specified X: 0.11

See post "Monitoring VAR and STATES - Dyn simulation"

User models must be added with DYRE. (Or, at least the first instance of each model)

Model TIOCR1 is a time inverse relay. You want to use it with definite time of 5 s. See description in manual Program Application Guide, vol II, Chapter 20.9.

Try with same multiple (1 pu of threshold current) for all four current points defining the time inverse curve. Use 5 s for all relay closing times.

It doesn't matter. In both cases the measured voltage will be used.

Kc is only used in voltage control if you want to compensate the measured voltage with the branch reactive power flow.

As I understand it you are measuring the voltage at bus 5 and want to control the voltage at bus 1000.

Select RefFlag = 1 and VCFlag = 1. Rc and Xc shall be the impedance of the line (in pu system base) between bus 5 and bus 1000. Note, Kc is not used here.

The dictionary is indexed with the channel number (integer).

I.e. f = chandata_dict[24] to get the values for channel 24

There is an issue in PSSE 34.3 so it is not possible to load DLL's for version 34.4 or newer. Try to update PSSE to rev 34.4 or newer.

You have to change the corresponding VAR in the model.

The STATE values are updated by PSSE during the dynamic and should not be changed by the user. What’s the reason for updating a STATE? The actual values of the states can be shown with DLST.

The approach using a GREF channel is correct, but it can be done simpler. There is no need for RUN command since the channel is initialised at start of simulation. Use API chnval to get the channel value directly, i.e. the GREF value. There is no need to read it from the output file.

GREF_chan_number = 10
ierr, GREF = psspy,chnval(GREF_chan_number)

There is no need to recalculate data since all data are in pu. See post ”How to simulate 2 systems with different frequency in one file?”.

Similarly, there is no need to update the inertia constant H!

So it is only needed to update the base frequency on the first row in the file with additional system data (network no 2).

Select channel XADIFD.

That means that the rate is 1 MVA.

All data with units MW, Mvar and MVA will be doubled. All data in pu are unchanged.

1: reactance 0.18 pu.

2: No changes in dynamic data since they are in pu. Only changes in load flow data.

3: No difference.

RATEA=0 means that no current limit is defined for rate A. It is equivalent to infinite current capacity and the rate will be ignored in all overload reports.

The approach in Pandapower seems correct.

Use a VAR to store the generator speed for the previous time step.

The simulation time at each time step is found in variabel TIME.

Also, remember that the model is called twice at the same time step when a pause is made in the simulation. Read Model Writing in POM.

If rotor angle stability is the issue here then it is sufficient to just add system angles statistics channels and monitor the angle spread in each simulation.

If any generator goes out of step in the simulation the angle spread will increase. Angle spread is therefore a good indicator of rotor angle instability. If angle spread is high (<300-400 deg) the bus number of the most unstable generator is seen in the channel with bus number of maximum rotor angle.

It is thereby possible to identify unstable generators without adding channels for individual generators.

If relative angles are used in PSSE, the difference between the rotor angle and the reference angle is shown. The reference angle may be a single machine or the average angle for all machines in the system. Negative rotor angles for some machines are normal when using relative angles.

Yes, the relay timer will start when the voltage goes below 0.8 pu. The total trip time is 2 + 0.1 = 2.1 s

Relay pickup time is the time the voltage must be over/under the limit to activate trip signal to the breaker. Breaker time is the time the breaker needs to open.

What do you mean with "it doesn't work"? Does model VTGTPAT trip the generator?

As @jconto wrote, you have to use HYGOVDU. There is no dead band in HYGOV.

A python script can be executed by another python script with execfile. Use a main script to execute the two scripts. For example:

execfile('file1.py')
execfile('file2.py')

You can have an SVC modelled as a generator and induction motors as loads at the same bus. The bus type must be 2. It is ok to have loads at generator buses.

See the model manual for the DYRE record for the model.

Search for "parallel" in this forum.

Missing sequence data shouldn't cause PSSE to crash. It must be a fatal error leading to division with 0.

See post "What is the correct way to finish a simulation"

Google swing bus or slack bus.

We are not dealing with a real generator, it is a swing bus! A swing bus has no boundaries.