generators_variability bound error in TDR_results

[rulyfanuel]

Hello everyone,

I have problem when trying to run GenX. the message says data frame bound error. Below is the error picture.

I tried to look at the Generators_variability.csv inside /TDR_results, and it only generate the 71 column without any data in the rows. You can see the result in the picture below

I already double check my Generators_variability.csv data inside the /system, but nothing seems wrong in it. the file below is the file from /system
Generators_variability.csv

please help me

[RuaridhMacd]

Hi @rulyfanuel , could you share your settings folder? You can share the files individually here, or zip them and send it as one item.

[rulyfanuel]

Hi @RuaridhMacd

I have found the problem, it seems the TDR process needs the vre to be varied, instead filling it with a single constant numbers. By doing that, /TDR_results/Generators_variability.csv's rows will be filled. However, I found another problem, which is as you can see in the picture

I tried to identify the data in /TDR_results/Generators_variability.csv again, and it seems there are problem in all data inside the row in the picture below. If I replaced all the data in the rows using "1" or data from other rows, it suddenly worked.

here is the settings folder and generator_variability.csv

winter_project_settings_and_GenVar.zip

[RuaridhMacd]

Hi @rulyfanuel

This is a slightly unusual bug. I think the TDR system is identifying too many extreme periods in your variability data. There appears to be a bug in GenX which allows this number to be greater than the total number of representative periods, so the number of clusters is negative and you're getting the error you see.

When you replace some columns with 1's and reduce the variability of the data, you surreptitiously reduced the number of extreme periods and everything ran fine.

To test this, please try running your case with UseExtremePeriods: 0 in your settings/time_domain_reduction_settings.yml file.

If that works, you can turn it back on (it will improve the quality of your answer) by setting UseExtremePeriods: 1 and then either:

1. Increasing the total number of representative periods you use by increasing the MaxPeriods setting, or
2. Fine-tuning the choice of extreme periods by setting some Max values in the ExtremePeriods dictionary at the bottom of settings/time_domain_reduction_settings.yml

[rulyfanuel]

Hi @RuaridhMacd

I tried to test 'UseExtremePeriods: 0', now it works. I turn it back on again 'UseExtremePeriods: 1' and test 'MaxPeriods' but it doesn't work. however if I set 'PV/Integral/Min: 2' in 'settings/time_domain_reduction_settings.yml', it works. But when I run gen, the highs solver return infeasible. I tried to look at the tutorial to see how to solve it, and there is a mention if I use Gurobi, then I can see which constraints are conflicted by adding 'ComputeConflict'. It generates this message

so based on that message, it seems there is problem in 'cMinCapEnergyDuration' constraint. However, I never set this contraint in my input. Do I need to deactivate this constraint or change some parameters in my inputs?

[rulyfanuel]

nevermind, I have found the problem, I need to set "Min_Duration=0" on "storage.csv".

Thanks for the help

[RuaridhMacd]

Glad to hear you resolved the issues. Feel free to reach out if anything else comes up