First time here? We are a friendly community of Power Systems Engineers. Check out the FAQ!

Ask Your Question
1

Calculation speed

asked Jan 30 '5

Busbar gravatar image

When performing dynamic calculations, I call the psspy.run() command a hundred times (start time 0 sec., end time 10 sec., step 0.1 sec.). This is necessary to extract data from the .out file during simulation. However, I notice that the first iterations are performed very quickly (tpause = 0.1, 0.2 ...), but over time, with each call to the psspy.run() command, the calculation speed decreases. Is there any possibility to speed up the simulation duration?

2 answers

Sort by » oldest newest most voted
1

answered Feb 2 '5

perolofl gravatar image

You are reading the out file at each tpause, In the beginning of the simulation the out file is small. The out file becomes larger and larger, and it takes longer time at each tpause to open and read the file.

Try a simulation without reading the out file at every tpause and see if the simulation time is more normal.

This is not an effective way to solve the simulation problem. It is better to write a user model that makes the necessary analysis at each time step. Then there is no need for pausing the simulation at every 0.1 s.

link

Comments

I've tried writing data into smaller .out files, but the calculations are still slow. I’ve also tried using the psspy.chnval command to extract the current channel value, but the calculations remain slow. As for creating a user model, I don't think it's an easy task since I have a large power system

Busbar gravatar imageBusbar (Feb 3 '5)

Is it also slow without any reading of out file? The size of the power system has nothing to do with writing a user model. It is the same dynamic model regardless of how many buses and machines there are in the system. What kind of analysis do want to do?

perolofl gravatar imageperolofl (Feb 3 '5)

This code is designed to determine the Critical Clearing Time (CCT). The algorithm simulates faults at different locations with different fault durations and analyzes their impact on rotor angle.

Busbar gravatar imageBusbar (Feb 4 '5)

If difference between generator angle and whole system median angle exceeds 180 degrees, the generator is considered as out of synchronism.

Busbar gravatar imageBusbar (Feb 4 '5)

In the end, CCT is defined as the maximum fault duration during which all generators have maintained synchronism.

Busbar gravatar imageBusbar (Feb 4 '5)
0

answered Jan 30 '5

jconto gravatar image

No, not possible during simulation. Before simulation, you may reduce the number of channels, increase the integration step, reduce number of models, etc. All those steps may lower the confidence on the simulation's results.

Processing speed between two time intervals is faster before time(fault) simply because nothing should change in the system so the simulation equations (dynamic models) resolve faster and then it would take longer depending on the amount of dynamic activities (load model forces a change in P & Q consumption, protection relay triggers something else to trip, which causes other model to react to it. Some models contain internal loops that might take more time to converged given the network conditions) that the fault induces into the simulation process.

link

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.
Want to format code in your answer? Here is a one minute demo on Youtube

Add Answer

[hide preview]

Question Tools

1 follower

Stats

Asked: Jan 30 '5

Seen: 231 times

Last updated: Feb 02