Idle storage, charging and discharging in the same timeslice

[NicoK]

Hei all,

I came about a weird phenomenon: Many storage technologies in my model are "idling". They both charge and discharge in the same timeslice. Here a random example for a thermal energy storage process EH_STG_TES2_1 in Germany in 2030. It is not really used to move any significant energy amounts between timeslices. Instead, it charges and discharges about 0.01 TWh in every timeslice, with the associated losses. Any idea why? The amounts are small, but too large to be rounding errors in my opinion. Has anybody seen this before?

Thanks in advance, 
Nico

[Antti-L]

> Has anybody seen this before?

Yes, it is relatively common, and so nothing too unusual about it.
But of course, due to the storage losses such would not happen unless either 1) the electricity heat price in those timeslices is zero or even negative, or 2) you have added some constraints that can be best satisfied by using the device in this way, or 3) the storage is directly serving a demand that has a load profile covering all timeslices (such as cell phone batteries, for example).

However, it would indeed be strange if the storage is only used in that way, without any true storage operation. I cannot say having seen such an issue. Therefore, I'd say your issue looks indeed peculiar.

[NicoK]

Thanks for the fast reply, @Antti-L!

In the optimal solution, the process is only used for shifting tiny amounts of energy around. So the "wasted" energy from this idle charging and discharging is negligible in the context of the entire model. Still, I found it weird.

Particularly, because the simultaneous charge/discharge goes on in every timeslice across the entire modelling horizon. There is not a single timeslice where either VAR_FIn or VAR_FOut is numerically zero.

But of course, due to the storage losses such would not happen unless either 1) the electricity price in those timeslices is zero or even negative, or 2) you have added some constraints that can be best satisfied by using the device in this way, or 3) the storage is directly serving a demand that has a load profile covering all timeslices (such as cell phone batteries, for example).

It is not serving any demand directly. It stores district heat serving the end-use sectors. I was also thinking whether some constraints push the model to use this storage for demand shedding. But the amounts are tiny and there are other processes with zero associated costs that are better suited to shed energy, I'd wager.

[Antti-L]

Ah... I missed that it was a thermal storage, but that doesn't change my comments. Anyway, I think the amounts look large enough to affect the optimal solution (if the heat has a positive price), and therefore I think you must have something in your model imposing the use of the storage in this way.

[NicoK]

I'll play around with de-activating this and other storage technologies (and constraints) and see what it does. Thanks! :-)