Different results when running simulation with ES from oemof.network instead of oemof.solph

[henhuy]

This issue is for all other, which might run into same bug, which cost me days to debug!

I accidentally used EnergySystem from oemof.networkenergy_system to run simulation with oemof.tabular, which gave me similar but different results than using EnergySystem from oemof.solph (as EnergySystem from oemof.solph internally adds some groupings, which ES from network does not).
This behavior is quite difficult to track down, as both EnergySystem work with oemof.tabular!

[nailend]

Indeed EnergySystem fromoemof.network.energy_system is used in tabular in general (see here) .... actually the implemented multi-period therefore should not work at all with tabular as oemof.network.energy_system.EnergySystem doesn't have the feature. Same for:

• infer_last_interval
• tsam (to be integrated)

I think we might have to switch so solph.EnergySystem

[nailend]

I got it wrong. As long as we use EnergySystem from oemof.solph its fine. I totaly get why you said, it's difficult to track down.

[henhuy]

We are fine - the issue was just intended to give people a hint, if they are encountering similar problems...
I will close

[nailend]

Stumbled over it again. Lets discuss this soon.
This implicityly works as long as we import from solph. Not sure if we should leave it like this here as its super confusing.
@FelixMau and me struggled with it after importing EnergySystem from oemof.tabular.datapackage which uses oemof.network version and therefore does not support multi-period etc.