Bachibouzouk
commented
2 years ago Closed mstich0 closed 2 years ago
Bachibouzouk
commented
2 years ago Bachibouzouk
commented
2 years ago How often did you see this happen?
mstich0
commented
2 years ago When working on the central-heat-network uscase it happened consistently (also on the same timesteps), while in other cases it does not. Changing parameters/costs would lead to more or fewer of these spikes. As in the Link described this would be a "reasonable" behaviour if there was too much heat and no sink, but in our cases there is always a sink in the bus which would be cheaper to use. As we dont want to create a MILP out of it it seems to be an issue in oemof. I´d suggest to give an option to set the charging and decharging capacity as ratio of the storage capacity (I think currently it is always 1). Especially for batteries this is a relevant input parameter anyway, which we should offer. For the heat storage I could set it to a reasonable value which might be 1/1000 and smaller for big tanks and therfore the spikes would at least not "ruin" the plot
Bachibouzouk
commented
2 years ago The dispatch price does not get passed correctly to the MVS, it will be set to the output bus of the storage
Bachibouzouk
commented
2 years ago Closed by #206
When working with the storage component there is showing a strange bug from time to time which I found hard to reproduce. Over a period of a year it only happens in 3 single timesteps.
The input and output of the storage rise at the same time for a very high amount. Neither reducing the efficiency nor adding dispatch costs to the component would stop this happening.
As it leads to rather strange results I would consider it as quite a problem. Is it possible that this might result from an unfinished optimization? It seems to happen only in longer simulations.