Grid load peak prediction

  • Sponsored by: sonnen GmbH
  • Scientific Lead: Dr. Stefan König and M.Sc. Eva Bammann
  • Project Lead: Dr. Ricardo Acevedo Cabra
  • Term: Winter semester 2018

sonnen's home storage system (the "sonnenBatterie") allows its owners to store their self-generated energy for later consumption. Typically, the sonnenBattery stores energy generated by the user's photovoltaic (PV) installation during the day and allows the owner to use this energy during evenings and nights. The sonnenBattery comes with an intelligent energy management system that adapts charging and discharging to the user’s needs and also enables sonnen to use the flexibility of home batteries for additional purposes like grid services.

Due to volatile renewable feed-in, power station outages and fluctuating consumption, the electricity grid is not always balanced. Grid operators are in charge of grid stability and security and must therefore match electricity generation and demand at all time. One of the grid operators’ means to stabilize the electricity network is to incentivize decentral generators (like home batteries) to feed electricity into the grid when network load is highest. However, the 15min with the highest peak load is only known at the end of the year.

Therefore, the goal of the present project is to forecast the 15min slot with the highest grid load of the year. Can we identify a number of 15min slots with highest probability to be the peak load slot? Given historical load and weather data, the goal is to generate an intelligent, re-useable forecast model that predicts future peak load slots of the electricity grid.

Results: The results of this project are explained in detail in the final documentation and presentation.