Modelling nonlinearities:
-
Losses (either charge/discharge efficiency or self-discharge rate) in batteries can be approximated as linear for some applications, e.g. LiIon for residential or grid (but not so big)
-
How does that apply to batteries: depending on battery system
§ LFP Li-Ion is quite linear
§ Lead Acid is highly non-linear
§ Degradation being one of the main problems
§ Offering of data by Holger Hesse -
Vehicles to grid - there the nonlinearity does count → you can linearise
-
TUM-EES building a detailed battery model - open source release to come - www.ees.ei.tum.de/simSES
Rainfall algorithm: state of the art approach and matches experimental results well (but no histeresys effects accounted for) -
PYPSA linked to a battery model
-
The multitude of influencing parameters makes it hard to interpret results on the effect of degradation and other stuff (VIBESUM)
-
Mention of study, that concludes not to use current LiIon for large energy storage, since the degradation is too prevelant (Seattle)
Long-term, seasonal storage:
- PSR Brazil 1979, Pereira SDDP - Stochastic dual dynamic programming
- Concrete: how to manage the problem of rolling horizon with taking the cumulated storage capacity to next interval
- Remention of minima at end of interval with framework to move data from one interval to next
- How to make the model choose between short- and long-term storage? Assigned higher virtual emissions to the least efficient one (long-term storage)
Load shifting as short-term storage; problem is lack of data; there’s one implementation with OSeMOSYS; UK market
- Open EGo: Acatech
Cost of Storage - Data sources for prices of different storage technologies:
- JRC ETRI, 2014
- Deloitte, Electricity storage, technologies, impacts and prospects
Heat storage: Group in TU Wien - probably energy economics group