Machine Learning Algorithms and PV Forecast for Off-Grid Prosumers Energy Management

The actual context characterized by the high prices of the conventional power gives more and more credit to the Renewable Energy Sources (RES) to cover load requirements in large amounts. However, the volatility of RES (especially solar and wind) restricts their smooth integration into the residential consumption energy mix. One of the main challenges is to maximize the consumption of appliances from RES taking into account their availability. To fulfil this objective, first, a performant forecast is necessary to create the day-ahead schedule and optimize the operation of appliances. In this paper, we propose a framework to perform PV forecast with machine learning algorithms and various data sources for the energy management of the off-grid prosumers.