Sizing of Hybrid Islanded Microgrids using a Heuristic approximation of the Gradient Descent Method for discrete functions

Abstract

Hybrid microgrids can handle fuel scarcity, reduce harmful emissions, increase flexibility, efficiency, and reliability. Nevertheless, these benefits are strongly related to the sizing and the energy management strategy (EMS) of the microgrid. In this regard, the sizing and the EMS acquire high importance for the planning of stand-alone microgrids. The present paper aims to design a sizing methodology that has a nested simulation model for power and energy balances. The sizing uses a heuristic approximation of the gradient descent method for discrete functions. The simulation model can evaluate the operation of the microgrid, even with EMSs using optimal criteria. A study case evaluation shows the effectiveness of the proposed algorithm compared to a traditional heuristic technique as Particle Swarm Optimization and an exhaustive search, which prove the feasibility of the algorithm to be used in stand-alone microgrid planning.