Simulation time of Doing Power Flow Cyclically

asked 2024-02-26 23:41:03 -0600

YKC
50 ●8 ●9 ●14

updated 2024-02-26 23:43:18 -0600

Hello everyone,

I met a time issue when doing the power flow cyclically. I use python 2.7.13 (in VS Code) to control PSS/E executing the power flow process cyclically and export the overloading report. During testing, I read the csv file of branch and set each branch out-of-service in each run.The number of branch is 1790, and I set executing the process for 12 times (which means "half a day"). Thus, I expect that it will export (12*1790) overloading reports and tell me which branches out-of-service will cause other branches overloading.

Then, I find that it takes for over 18 hours to complete the entire 12 times simulation. In first run, it took around 5 mins, then second run took 7 mins, and the simulation time in each run is irregularly and extremely increasing. The followings are the simulation time in per run (min): [4.454, 6.7685, 13.7459, 25.1792, 39.6676, 59.3887, 85.0343, 112.3582, 143.5624, 178.5427, 205.1365, 219.4923]

Since each run do the same work (csv file is same to verify each run gives the same results), I could not figure out the problem during simulation: 1) why each run took different time to export the report, and 2) the time is extremely increasing? Does anyone meet this problem before? and know the reason?

edit retag flag offensive close merge delete

add a comment

0

answered 2024-03-05 11:42:25 -0600

jconto
2701 ●10 ●30 ●41

Can you post the code? The time to create a report should be the 'same' for every iteration, and the size of the report file be also the same. Are you logging? A log file would show if 'past' activities are being repeated.

edit flag offensive delete link

Comments

Hi Jconto, I could not post on forum, the entire code is too long. If you willing review my code, could you please tell me your email, or you could send the mail to "yukaichung@itri.org.tw" and inform me who you are.

YKC ( 2024-03-05 22:59:51 -0600 )edit

add a comment

0

answered 2024-03-05 21:36:14 -0600

KYC_Korea
13 ●2 ●3 ●6

updated 2024-03-05 21:37:29 -0600

In my opinion, there are two possible reasons for this

First, it could be due to the different number of iterations it took for the PF calculation to converge. When the PF calculation converges, a message like "Reached tolerance in 10 iterations" is output, which means the number of iterations it took for the PF calculation to converge. The larger this number of iterations is, the more time the calculation can take.

Secondly, there may be more overloading and it took a long time to save it as an excel file. I recommend checking the capacity of your excel file.

Alternatively, use psspy.pssehalt_2() to completely shut down psse and then use psspy.psseinit(150000)

I hope this helps in some way.

edit flag offensive delete link

Comments

Hi KYC, Thank you for suggestion, I also did try using this API, and it took more time in each run. And for excel files, I will take a look. If you willing review my code, could you please tell me your email, or you could send the mail to "yukaichung@itri.org.tw" and inform me who you are.

YKC ( 2024-03-05 22:58:35 -0600 )edit

add a comment

0

answered 2024-03-04 18:07:24 -0600