Probabilistic Forecasting for Battery Managment
The purpose of this research project led by RISE is to investigate the conditions for how household batteries should be sized and managed for best usage. The aim is to deepen the knowledge for local storage; how it shall be dispatched and what external services it can provide to the grid. Load profiles from single-family houses are clustered and a PV/battery system is then fitted for each cluster profile. Battery dispatch algorithm is based on artificial intelligence models (AI) for probabilistic prognostication of electrical loads and solar PV generation for optimal battery dispatch.
The aim is also to demonstrate a battery dispatch algorithm based on an AI model at RISE’s Research Villa. This, to investigate how well the model can predict load and PV profiles and how well it can recover from wrongful predictions.
Swedish Energy Agancy
2019-01-03 – 2020-12-31
- Uppsala University
- Herrljunga Elektriska AB
- Trä- & möbelföretagen
Patrik Ollas (RISE)
Patrik Ollas (RISE), Joakim Munkhammar (UU)