problem with kre storage

[frangb99]

  SubRES_STO_Techs_noELC.xlsx (Size: 955.69 KB / Downloads: 3)    
Hello again, here is my file where i am modelling some storages, as you can see here i am not using ELC as auxiliary input. When i run the hole model, there is one result that makes 0 sense. Which is this one:   
     

As you can see there is a negative flow in the OUTPUT of the KRE storage. Can anyone maybe have a look at it, and maybe see why this could be happening? Thank you so much

[Antti-L]

Sure, I will look at it.  

[Antti-L]

It would help understanding what is going on there, if you could provide the TIMES input files (*.DD and *.RUN) for the model run, to reproduce your results.  Would that be possible?

[frangb99]

Sure, I will look at it.  

Also, reading the documentation, i see that for the auxiliary inputs that i had problems before, the ELC for charging and discharging process, i need to use the attribute PRC_ACTFLO (Defines a conversion coefficient between the activity and the flow in commodity c. For storage processes, PRC_ACTFLO can be used for the commodities in the PCG in the standard way, but also for defining the amount of auxiliary flow of c per unit of activity), but i do not find any Alias to use it, if i just write the name of the attribute on the table, it would work?

[frangb99]

It would help understanding what is going on there, if you could provide the TIMES input files (*.DD and *.RUN) for the model run, to reproduce your results.  Would that be possible?

It does not allow me to send those 2 files.... it says those types of files are not allowed

[Antti-L]

Can you not pack them into a Zip file? That would get through. But note that I am sure there should be several *.DD files.

[frangb99]

Can you not pack them into a Zip file?  That would get through.

[Antti-L]

Ok thanks, the RUN file was there, but not the *.DD files (there was the *.vdd file instead).  Any chance of including also the *.DD files?

The files listed below:
base.dd
sto_techs_noelc.dd
tra.dd
syssettings.dd
limit_capacities.dd
subannual624_af-re.dd
subannual624_exp_cost.dd
subannual624_yrfr.dd
x_no_dummies.dd

[frangb99]

Ok thanks, the RUN file was there, but not the *.DD files (there was the *.vdd file instead).  Any chance of including also the *.DD files?

The files listed below:
base.dd
sto_techs_noelc.dd
tra.dd
syssettings.dd
limit_capacities.dd
subannual624_af-re.dd
subannual624_exp_cost.dd
subannual624_yrfr.dd
x_no_dummies.dd

i do not find any .dd Lets see if one of this other files is

[Antti-L]

Ok, if you don't have the DD files there, do you have the *DATA*.gdx file (zipped)? That would provide me the same information.

[frangb99]

Ok, if you don't have the DD files there, do you have the *DATA*.gdx file (zipped)?  That would provide me the same information.

Okay the page does not allow me to send any gams file, which i guess is what you want. I really do not know why. Maybe if you give me your email i can send everything to you

[Antti-L]

It should work when zipped... unless it is very big? But that *DATA*.gdx should be rather small.

[frangb99]

It should work when zipped...  unless it is very big?  But that *DATA*.gdx should be rather small.

I hope is this one

[Antti-L]

Yes, thanks! Will let you know when I have the explanation for your problem...

[Antti-L]

Very good, I was able to run the model, and I obtained exactly the same OBJ value as in your listing file (to the finest decimal).

But there were no such results reported, which you showed! 
More precisely, there were no negative VAR_Fout values for the GRDSTOHSTNOKREXTRMN0_PO process in any year.  So, it is a puzzling issue. One possibility is that perhaps VO could have some problem in the results processing.  Other possibility is that in your case Cplex somehow arrived at a different (degenerate) solution.  But because that process GRDSTOHSTNOKREXTRMN0_PO is actually a normal process (it is not a storage process), I cannot imagine how the output flow could ever become negative, because the only output is KRE and it is the PCG, and the variable is by definition non-negative. Therefore, I would consider some kind of a VO issue more likely...