MW's profile - activity

Hi all,

Can someone enlighten me on the subject of the pro's and cons between the two approaches to receive a reduced network model in PSCAD. The two options are below (if you have a better 3rd please feel free to add):

1. By using SCEQ function in PSS/E reducing the network to a reasonable size. Compare V and angles in load flow as well as 3 ph and SLG faults with full and reduced network so they are close. Then use E-tran to convert the reduced network to a PSCAD network for PSCAD simulations (also validation performed against PSS/E data: V, angles, 3 ph, SLG, rotor angle response of generators if any included in .dyr-file)

2. Use E-tran directly to both reduce and convert the network from PSS/E to PSCAD. In an E-tran information sheet they say it can be validated by performing a "network frequency scan" at the relevant bus. First using a reduced network 2 buses away, then 3 then 4 etc. until a nice approximation is received. (Source: descargas.indielec.com/web/Applicationnoteetranconvertingpssetopscad.pdf)

I guess if you have specific ideas on that you want to model 4 buses away on one end and 2 buses away on the other, 2. is not as efficient but also more specifically; Have anyone seen noticable differences in the simulation result between 1 and 2? Anyone with a link to some comparative study etc.?

Best regards MW

Hi Jacde,

I've had the same problem, ended up just running my cases in smaller batches as a "quick fix". Otherwise, the resulting output had alot of ripples in them and unstable starting conditions (during steady state before trip). However, I'd be very interested in hearing an answer on this subject if anyone has it.

My script is a three layered loop running all the combinations of the data file types and specified contingencies (.sav, .dyr and then specified line-trips) with each file type having approximately 8 files. I also noticed that when running the same script with the same number of cases/contingencies/dynamic data files using the example file savnw I did not get the same problem, however the .sav-file I am using for my simulations are >200 times in terms of kB.

I've sent the question forward to PTI Support but they have been unable to answer. Sorry I cannot be of more help, anyone else here with an answer to this?

Hi,

As I understand it Using conec calls and setting VAR(L+1)=VAR(L+15) thus getting my model A output as input to my model B. Many standard library models have available aux signals. But I am not able to connect my user written to my library model (switched shunt). Are there any requirements on the user written model to connect it? I currently have a CONEC model.

I will update with conec calls when I have them available. I have verified the correct numbers on my VARs, CONs etc.

Also make sure you have closed the excel application (no open excel sheets) while you run it. I usually get an error similair to that when i forget to close them and then run the script.

2 cont.) You could just import your results into python lists and remove all the zeroes then plot it yourself with modules such as win32com or xlrd/xlwr. A bit of programming involved but not impossible at all. Just make sure all the 0 voltages actually are islands before though.

Hi,

Any ideas on how to model the control of a phase shifting transformer? Ideally I would like a bandwith between 100 and 300 MW and then run load flows and ACCC:s for different load/generation scenarios where the optimal angle would be used. I understand PSS/E can't do this as it would instead just lock on to for example 110 MW and then go with that only to change to have the solution converge.

My idea so far is to just run the different scenarios manually and find my own optimal Vmin (MW) and Vmax (MW) then just change Vmin and Vmax for each load/generation scenario. I then run my ACCC:s etc. on that scenario before going to the next load scenario and so on...

As an example: lets say I find 230 MW is a good target for one of my load scenarios. For that load scenario by "manually" changing Vmin=225 and Vmax= 235 the transformer would end up in that same range and angle more in line with how that potential transformer would have been run in reality.

Is there a more convenient way that anyone here can suggest?

Hi,

The fact that pssexcel succeded in exporting to excel when you had results with relatively low flows makes me wonder if it's wrong with how the .sav file is setup before the load flow is executed. It's hard to say however without seeing what kind of output you are getting (flows and voltages). You say that you have high flows and low voltages, are they all 0 or 0.2 pu something? What do you mean with high flows, way over 150 %? For me that makes me think of two things that you ought to double check:

1) Your solution might've converged to the low voltage high flow solution or as my power system lecturer used to say "went past the "nose" of the PV-curve". You can see on this picture: www. openelectrical.org/wiki/images/5/56/PV_Curve.PNG that for one active power flow solution there are two different voltages (in theory atleast). One with voltages around 1 pu the other with low voltages and really high flows like 200% and upwards.

2) Do your contingencies contain alot of islands? That would also give you low or 0 voltages (assuming no generators in the islands or higher loads than generators). have you selected to export or not export results from islands (cannot remember if it is possible to do but check the API).

For me I've had some trouble with PSS/E converging to the low voltage high flow solution when I do major changes to my system in one go instead of taking it in steps. For example changing a phase shifting transformer from -30 to 30 degrees, instead I did it in steps of 10 or 20 degrees and it was fine. Hope this serves as some help.

Thanks for your answer Shengen. However mvrecmax/min only shows the limits I have put in the .mon file ex: Vmax = 1.10, that is what will show up for all the buses. .mvbuslabel shows the names of the monitored buses so it doesn't really give me any values of voltage deviations for different buses.

I'm new to both PSS/E and Python but I'm starting to get the hang of it.

I'm doing several ACCC studies and I want to import my results from the different scenarios into different sheets according to what is being monitored, ie. all low voltage violations for all scenarios in one sheet, one sheet with all overloaded lines etc. To do this I want to have my violations for my scenarios in lists so I can plot the using excelpy. I'm a bit surprised that something which is fairly easy to get in an output report in PSS/E or in .txt is so much more work to get into a python list so you can plot it the way you want instead of using methods like pssexcel.

Question 1) Is there a good way to extract your flow/voltage violations from psspy.accc_solution or any other command? I'm reading through the API and I'm trying to use some of the code used in for example: "Help Refining Multiple ACCC Excel Report" but I can't find any commands or anything to do that.

Question 2) Is there a way of getting the voltage deviations from an .acc-file without subtracting pre-contingency bus voltage with post-contingency bus voltage?

Question 3) Right now I'm deciding whether to try and extract these result from a .txt file or try and import PSS/E results directly into a list (for example using accc_solution), but which way is more convenient/easier?

On a side-note: This forum has been very helpful in teaching the psspy module as well as python, thanks!