Optimal Demand Response Self-Scheduling Approach for a Grid-Tied Microgrid Based on Load Demand and PV Forecasts Considering Battery Degradation

In the present work, a day-ahead self-scheduling technique based on Demand Response (DR) program, load demand, and PV power forecast is proposed. The main target behind the developed method is to minimize the operation cost of the system. Thus, the exchange cost, the system's operation and maintenance costs, as well as the battery degradation cost, are considered in the objective function. First, simple Long-Short-Term-Memory (LSTM) and LSTM-based encoder-decoder models are proposed to predict the load demand and PV power output. The performance of the proposed models is compared with other forecasting models. To assess the effectiveness of the proposed DR approach, two other scenarios are constructed. Findings demonstrate that the operation cost of the MG is optimized more efficiently by implementing the DR program, in contrast to the other established scenarios.