18-07-2020, 04:14 PM (This post was last modified: 18-07-2020, 04:15 PM by svecjos.)
Hello,
Can the YRFR attribute be too small to cause dummy imports(IMPNRGZ)?
I am trying to increase the number of TimeSlices in a TIMES model and I came across a puzzling finding. I define the names of the TIMESlices in the SysSettings and then specify the YRFR attribute in the scenario excel. I have scenarios with 24 daynite timeslices and I vary the number of seasonal time slices -1,4,12. I also change the portion of the year that each time slice represent. There are scenarios with 4 and 12 seasonal timeslices that work fine and other use dummy imports(like the scenario attached).
The dummy imports are used even when the availability of the sources is the same only spread into different timeslices. Furthermore, when I have a scenario below and change the YRFR attribute to numbers from different scenario that is working, I get no dummy imports (all other things being the same).
How can only a length of a timeslice have an impact on the electricity production? Is there any attribute causing this, that I might overlook?
Dummy imports are a nice tool for resolving infeasibilities in models, and at least for me they have helped to fix any such problems quite quickly. In your case I guess you probably just have insufficient capacity in some timeslice. After all, you have apparently redefined the COM_FR and availability factors for the individual timeslices of each timeslice configuration. You could also try using the Cplex infeasibility finder for getting some more information.
Or, if your dummy import processes are on the ANNUAL level, you could try setting them at DAYNITE to see if that reveals the timeslice(s) where you get an infeasibility. The VEDA demo model shows how to do that.
20-07-2020, 08:46 PM (This post was last modified: 20-07-2020, 08:55 PM by Antti-L.)
Sorry for not providing a direct answer.
Q: "How can only a length of a timeslice have an impact on the electricity production?"
If you have defined demand load curves (by using COM_FR), the answer should be obvious: YES it can, and would even be expected have an impact, because then the same demands must be satisfied within different slices of time. Simple example: If you have a demand of 1 TWh of electricity within a single two hour timeslice representing, say, 50 days (in that season), the capacity required to be available for producing that 1TWh would be at least 10 GW. If you then change the lengths of the timeslices so that this single timeslice would represent only one hour in those 50 days (thus keeping the season length unchanged), the capacity required to be available for producing that 1TWh would now be at least 20 GW. However, if you can also change the COM_FR accordingly, such that all load levels would remain exactly as before, you can of course avoid such an impact, although in general it is not possible to have all the load levels and demand matching after changing timeslice lengths. But you were asking about only changing the lengths of timeslices (G_YRFR).
20-07-2020, 11:12 PM (This post was last modified: 20-07-2020, 11:18 PM by svecjos.
Edit Reason: wrong formatting 2
)
(20-07-2020, 08:46 PM)Antti-L Wrote: Sorry for not providing a direct answer.
Q: "How can only a length of a timeslice have an impact on the electricity production?"
If you have defined demand load curves (by using COM_FR), the answer should be obvious: YES it can, and would even be expected have an impact, because then the same demands must be satisfied within different slices of time. Simple example: If you have a demand of 1 TWh of electricity within a single two hour timeslice representing, say, 50 days (in that season), the capacity required to be available for producing that 1TWh would be at least 10 GW. If you then change the lengths of the timeslices so that this single timeslice would represent only one hour in those 50 days (thus keeping the season length unchanged), the capacity required to be available for producing that 1TWh would now be at least 20 GW. However, if you can also change the COM_FR accordingly, such that all load levels would remain exactly as before, you can of course avoid such an impact, although in general it is not possible to have all the load levels and demand matching after changing timeslice lengths. But you were asking about only changing the lengths of timeslices (G_YRFR).
Hello Antti-L,
Thank you very much for your answers. You pointed me into the right direction.I indeed had timeslices with demand way higher than the available capacity. I noticed a mistake in my definition of the COM_FR attribute.
