I wanted to double check the way i Ma modeling 4hour battery because I am seeing very little storage in the results. I know it may be related to the battery economics but i thought i'd run my modeling by you. Here are my modeling assumptions:
Storage efficiency: 85%. I am reflecting this as STG_EFF=0.85 and FLO_FUNC(ELC, AuxStoOUT, ANNUAL) =1/0.85=1.1765
4hours of capacity for day night operation. This translates to: AFAC(ELC)= 4/24=0.1667, NCAP_AFC(ACT, DAYNITE)=4/24/0.85=0.1961
6hours to full charge. This translates to AFC(NRG, DAYNITE)=4/6=0.67
Auxstoin and auxstoout are used for some downstream processing
CapAct is 29.89 because my energy unit is Tbtu
the constraints shown in the screenshot below are neutral and are used for reporting purposes.
Do these modeling assumptions look reasonable to you?
Like in that earlier case, I am not able to see anything suspicious here either, apart from the AuxStoIN/AuxStoOUT flows, which are now defined in such a way that AuxStoIN is equal to the output flow of ELC, and AuxStoOut is 1.1765 times the input flow of ELC. This is different from the earlier case, and looks a bit unexpected, but not too relevant if only used for reporting.
Indeed, the investment cost of the 4h battery storage is a bit high, about 350 USD/kWh. I would suggest to give it some "learning rate" based on some reference, such that by 2050 the cost would be notably lower. The fixed O&M costs look likewise a bit high, and maybe a learning effect would be justified there as well.
[Edit:] Sorry, looks like I may have mixed up the Commodity and Commodity_group columns (when looking at the FLO_FUNCs), because the headers were not fully visible.
Thank you for your reply, Antti! Super helpful! Good to know that things look good to you for a modeling perspective. I run this model from a python wrapper and do modify capex and fixed o&m using adddata.gms. I will double check things! Thank you!!
