Deep-learning aided stochastic co-optimization of energy self-consistent system in highway service area

The synergy of electric vehicle (EV) charge demand and renewable generation in microgrids presents an economical and sustainable solution to managing the growing electricity demands in highway service areas (HSA). However, the stochastic nature of new energy power generation and electric vehicle charging demands in highway service areas poses challenges. This paper introduces an integrated optimization approach that enhances the expected comprehensive performance of HSA microgrids while minimizing capital costs by concurrently optimizing energy management strategies and system design parameters. A co-optimization approach is utilized to adjust system parameters, focusing on the anticipated best performance of various design options when faced with uncertain conditions. Furthermore, a deep-learning-aided model is incorporated to improve the efficiency of the optimization process. This method provides Pareto optimal solutions that represent various trade-offs. As the recommended reference design results, PS-2 achieves a reduction of 67.8% in capital cost an improvement of 61.2% in system comprehensive performance, compared with the empirical design.