An improved neural network algorithm for efficient scheduling of distributed energy resources in microgrid

The promotion of environmental sustainability and pollution reduction has been facilitated by microgrids based on renewable energy sources (RESs). In microgrids that have multiple energy resources and storage systems, optimal scheduling is critical to maximizing power sharing between components and ensuring efficiency, reliability, and economical operation. This research aims to find the optimal generation strategy of different distributed energy resources (DERs) and energy storage system (ESS) based on an improved neural network algorithm (INNA) to minimize the generation cost of the microgrid. INNA integrates a simple quadratic interpolation (SQI) to enhance search ability. This study investigates three cases of a typical grid-connected microgrid over 24 hours. The results clearly indicate that the generation cost is minimal when the microgrid actively participates in electricity trading compared to the case where the grid operates passively. Moreover, INNA obtains better results than other methods for all cases, which shows its high performance in finding the optimal generation strategy of the microgrid.