Duck Curve Aware Dynamic Pricing and Battery Scheduling Strategy Using Reinforcement Learning

The duck curve is becoming a global problem in energy technology due to the rapid increase solar power adoption and rise of prosumers. To address this issue, resource aggregator (RA) has emerged provide flexible solutions through aggregating prosumers demand response such as dynamic pricing. This paper proposes an optimal strategy for RA that dispatches pricing leverages battery system at both prosumer levels. proposed method based on model-free deep reinforcement learning (DRL) algorithm optimize each prosumer’s retail prices schedule usage RA’s station. An objective reward function used maximize profit, minimize cost, improvement curve. performance DRL-based was demonstrated by simulation experiments using actual wholesale price, demand, PV generation data. results show can improve standard deviation peak-to-average ratio net load up 57.1% 23%, respectively.