Power demand forecasting for demand-driven energy production with biogas plants
Celina Dittmer, Johannes Krümpel, Andreas Lemmer
Abstract
For the future energy system it becomes increasingly important that biogas plants produce electricity in a demand-oriented way to compensate electricity production from fluctuating sources like wind power and photovoltaics. Flexibilisation concepts provide a coordinated feeding management, which consider different gas production kinetics of used substrates to adjust the biogas production. To enable the generation of a prospective timetable, suitable forecast models for power demand were evaluated. The resulting 48-h forecasts of power demand of a “real-world laboratory” demonstrated that the four selected models achieve comparably good results with a mean absolute percentage error (MAPE) between 13 and 16%. Further evaluation showed that forecasts over longer periods of up to 14 days are advantageous as they are possible without compromising forecast quality.