Poster - Energy Management in Off-grid Homes by Forecasting.

In our work, we propose a smart home energy management system (HEMS) for off-grid homes, where electricity is produced locally by a photovoltaic (PV) system. Smart home appliances can be adjusted or switched on and off by the HEMS. We introduce a forecasting approach for energy production and consumption and develop a working system prototype based on MQTT and openHAB, managing three different home appliances and basic load. The prototype is used to calibrate component models used in an accompanying simulation study that leverages real weather data for forecasting. One major observation is that in particular in periods of bad weather conditions, the improvement when using the HEMS with forecasting is remarkable: In one experiment week, the system can maintain 100% power supply availability while there would have been 16.07% outage time when disabling the HEMS.