A deep learning-based microgrid market modeling with planning assumptions*

Microgrids (MGs) can be considered as one of the best solutions for the distribution grid's resiliency and reliability. The research study suggests a hybrid stochastic-robust optimization method for determining the optimum schedule of a MG within usual and resilient operation situations. A new optimization method is used to model the effects of main electrical network cost uncertainty on the optimum planning of MGs. A stochastic optimization process is modeled with other dominant uncertainties, such as wind turbine power, photovoltaic power, and reactive/ active loads, through creating appropriate case studies for each. MG's resilient operation is improved further by implementing curtailable and shiftable demand response programs.The suggested method enhances the performance of MGs under uncertainties when the network is in its normal and resilient mode. In addition, the paper discusses different approaches for RO planning and evaluates the effectiveness of the suggested structure in a large-scale MG trial system.