Pmax of closest neighbor in TransformProfile
rouille opened this issue · 4 comments
A new instance of the Grid class is used in the TransformProfile class to retrieve the capacity of the closest neighbors when new renewable plants have been added to the network.
If the capacity of the closest neighbors are updated in the TAMU plant.csv file then the profile of added generators will be incorrectly calculated in a TransformProfile object for a scenario produced before the update. The reason is that the updated Pmax will be used in place of the Pmax that has been used to generate the profile.
Moving one observation during the discussion with @rouille here: in general, if we change the base grid, i.e. TAMU plant.csv, we will have the corresponding base profile scaled accordingly. If the new profile of the changed grid is scaled consistently with respect to the new Pmax, this won't be a problem, however, if that's not the case, the issue mentioned above will show up, i.e. we won't be able to recover the profile of the added plants of an old scenario with the update.
I can be solved by saving the Pmax of the neighbor in the change table when creating the scenario. The base grid won't be necessary anymore in the TransformProfile class anymore.
Is this something that could be fixed with normalized [0,1] profiles?
Is this something that could be fixed with normalized [0,1] profiles?
Yes, this would be the clean way. As you know, normalizing hydro is not easy and we would need to renormalize all the old profiles, i.e., knowing the Pmax that were used to produce them.