Levelised cost of Storage

[Anjali]

Hi,
I am using a new storage technology as STGTSS in my model. It is having timeslice level DAYNITE and PrimaryCG NRG.
It is having investment cost and fixed operating cost. Variable operating cost is defined as zero.
Other defined parameters are STG_EFF = 0.85, AFC(NRG) = 0.99, AFC(ELC)=0.98, AFC(ACT)=0.1667
Input and output commodities are the same as ELC.

I have activated levelised cost reporting switch also.

I am facing two problems with this:
1. In VEDA BE, I am getting a very high value for VAR_NcapR corresponding to storage type.
2. In few timeslices (not all), it shows VAR_FIn and VAR_FOUT simultaneously.


Please help me to resolve this.

Thank You

[Antti-L]

Indeed, the reporting of levelized costs has been designed only in view of standard processes, producing some main output product. As it has not at all been designed to handle storage processes, the high cost for storage is actually just an artefact (caused by the non-existence of the necessary VAR_FLO denominator when the process is storage). Therefore, the levelised cost reporting should probably be suppressed for storage processes, but that issue has been apparently overlooked thus far.

I am sure you can expect that the levelized cost reporting for storage, which is now clearly misleading, will most likely be removed in the next version, thanks to your problem report.

Concerning the other problem, I have no idea why you have simultaneous VAR_SIN / VAR_SOUT, unless the price of electricity is zero (or negative) in those timeslices.  But if the price is indeed zero or negative, it is perfectly normal to see such simultaneous flows in the optimal solution.

[Anjali]

Thank You for clarifying my doubt regarding levelised cost of storage.

Regarding simultaneous charge and discharge, yes it is happening whenever only renewables are participating to meet the demand in that region. And the variable cost of renewables is zero. However, they have some fixed operating cost.
Even in VEDA BE, if I check EQ_CombaLM of ELC corresponding to that particular timeslice, It shows zero or negative value.
Is there any other way to handle this situation as it would be difficult to analyze 100% RE scenario with this.

Thank You once again.

[Antti-L]

Well, the condition that either Input=0 OR Output=0 (or both) is nonlinear, and thus cannot be modelled without using either an NLP or MIP formulation.  (For example, using NLP, one might simply require Input×Output = 0).

If we could add such a constraint into a TIMES model, it would make it more difficult to achieve a 100% RE scenario (because adding new constraints to a model can only make it more costly to achieve target policy constraints).  The model is obviously utilizing the energy losses from simultaneous charging and discharging for curtailing zero-value electricity production, when that is more economical compared to investing into additional storage capacity and charging the excess production there for use in some other timeslice.

So, it is not quite clear to me why you say that it makes it more difficult to analyse a 100% RE scenario, if simultaneous charge and discharge is allowed to take place in the same timeslice? Do you see some technical problem with it in the real world, such that with some types of storage it should not be technically possible to charge and discharge in the same model timeslice, and/or that curtailing excess variable generation in the scale that the model does would not feasible in the real world system?