[Bug] Bug_short_description

Comparing the results of oemof solph and MVS of an energy system containing dispatchable and non-dispatchable sources the following issue came to light: When investment mode/optimization is turned off for all sources, both models give the exact same results (as desired).

When investment mode is turned on for dispatchable sources, but turned off for non-dispatchable sources, both models give the same results (as desired).

However, when investment mode is turned on for non-dispatchable sources, the investments made in these technologies differ between the two models.

This is only the case, when the variable costs (in solph) and the dispatch_price (in MVS) are not set to zero. When they are set to zero, the results match.

We found, that the reason for the mismatch lies in the definition of the variable costs in MVS. They are assigned as follows (in module D1, source_non_dispatchable_optimize): variable_costs=dict_asset[DISPATCH_PRICE][VALUE]/dict_asset[TIMESERIES_PEAK][VALUE]

Removing the division and just assigning variable_costs=dict_asset[DISPATCH_PRICE][VALUE] leads to matching results between solph and MVS

Input units were:

source-timeseries: kWh/kWp
- dispatch_price: currency/kWh (MVS)
- variable_costs: currency/kWh (solph)

Looking at those units, dividing the dispatch price by the timeseries peak would lead to (currency*kWp)/kWh^2. This doesn’t seem sensible.

It seems like one of these three points has to be true:

the division is incorrect and should be removed
the input-timeseries needs to be in a different unit
the dispatch_price needs to be in a different unit

Checklist to make sure that the bug report ist complete:

[ ] OS: your_operating_system, your_distribution
[ ] Branch: branch_name, updated on update_date
[ ] If applicable: Attach full error message
[ ] If applicable: Share screenshots/images of your problem
[ ] If applicable: Share used input data

rl-institut / multi-vector-simulator

[Bug] Bug_short_description #951