Using the node TH_NP15_GEN-APND with power = 50MWs and energy = params$MWhMWhs would have delivered different PL based on the charging/discharging strategy. This assumes that everyday the battery will go through a cycle. There are no checks that the cheapest prices as sufficiently different from the more expensive ones. There is also no check that the battery is discharging only after being charged. This is just meant as a framework to build it with more features in the future.

Using a strategy of charging in the DA market during the duration hours with the lowest load forecast and discharging, also in the DA market, during the hours with the highest load forecast would have generated pl as follows:

Using the same strategy but doing all charging and discharging in the RT market would have yielded the following:

Picking the charge/discharge hours based on perfect foresight would turn out like this first Day-ahead only:

next realtime:

Both perfect scenarios together

Another strategy is the groundhog day strategy where each day you do whatever was best in the previous day