Hello,
I can succesfully load a model into VEDA FE, but when I try to solve it I get "Unexpected termination of run", and I (with some help from colleagues) cannot figure out exactly what the problem is... In the BASE.lst file there are two error messages:
**** Exec Error at line 22202: division by eps
**** Exec Error at line 36721: division by eps
Unfortunately, I do not understand exactly what these messages refers to... Is there any way of tracing this?
Best regards
Jonas Forsberg, LTU
The listing file indicates that you most probably have defined some zero PRC_ACTFLO parameters.
That would be a modelling error: PRC_ACTFLO can never be zero.
Unfortunately, this modeling error has not been included in the QA checks, and so it may not be so easy to find where these zero parameters have been defined. But you can list all PRC_ACTFLO parameters in the VEDA Browser, and check it out there.
I'd suggest to the maintainer of TIMES that a QA check be added for this.
28-03-2023, 01:15 PM (This post was last modified: 28-03-2023, 01:17 PM by Hadinuryana.)
Hello, I also have a similar problem where I end up with an unexpected termination of run. I checked the LST file and got a "**** Exec Error at line 34891: division by zero (0)" and also do not know where to trace it, could you kindly help? The zipped LST file is attached.
28-03-2023, 01:40 PM (This post was last modified: 28-03-2023, 05:34 PM by Antti-L.)
It looks like you have COM_FR defined all zero for some commodity, which causes the problem. The sum of COM_FR over all timeslices of the year must always be 1, and the code ensures that by normalizing the sum to 1 if the sum is different, but if the sum happens to be zero, that would cause a divide by zero. It would thus indicate a serious modeling error.
Please can you show all COM_FR you have defined? In the Browser, select attribute=COM_FR, then click on Browse. Then, in the data spread, expand all dimensions, and click "Export data to Excel". Finally, upload that Excel file here, so that we could see what might be causing the error.
(28-03-2023, 01:40 PM)Antti-L Wrote: It looks like you have COM_FR defined all zero for some commodity, which causes the problem. The sum of COM_FR over all timeslices of the year must always be 1, and the code ensures that by normalizing the sum to 1 if the sum is different, but if the sum happens to be zero, that would cause a divide by zero. It would thus indicate a serious modeling error.
Please can you show all COM_FR you have defined? In the Browser, select attribute=COM_FR, then click on Browse. Then, in the data spread, expand all dimensions, and click "Export data to Excel". Finally, upload that Excel file here, so that we could see what might be causing the error.
Dear Antti
I have corrected the COM_FR to have result of 1, there was indeed some COM_FR SUM that resulted >1.
One question, since the SUM must be 1, is it okay if there's a data that is 0, but the SUM of the COM_FR is still 1?
Please find attached the COM_FR file that I have corrected. I have checked the export result, I'm also wondering why there are some blank COM_FR even though I have filled it with data in my B-Y template?
I created my B-Y template based on Veda-FE Demo 5, and so far already deleted Reg1 & Reg2 data, but it seems there are still some Reg1 & Reg2 data that shows, including in this attached COM_FR file, is it okay? Will it affect my model?
There is no need to normalize the COM_FR values by yourself. The TIMES code does it automatically.
>One question, since the SUM must be 1, is it okay if there's a data that is 0, but the SUM of the COM_FR is still 1?
Yes, of course any individual timeslices can have a zero value, as long as the sum is positive (and all values are non-negative).
I am not able to see why you have all those blank entries in the table. If you are still seeing the divide by zero issue, I guess it might be related to those. And, as you noted yourself, there are still some Reg1 & Reg2 data showing up, including this COM_FR file, and obviously, that may well be causing problems.
I am sorry that I cannot see the direct cause from your table, but it only raises various suspicions. Where are all those blank rows coming from? And why do you still have Demo 5 timeslices and regions in your model, if your timeslices are Annnual1, Annnual2, ... Annnual24, and your regions are JAKBAR, JAKPUS, JAKSEL, JAKTIM and JAKUT?
