Stochastic, resilience-oriented optimal sizing of off-grid microgrids considering EV-charging demand response: An efficiency comparison of state-of-the-art metaheuristics

Abstract

The resilient microgrid (MG) capacity planning and optimisation problem is widely recognised as a non-deterministic polynomial time-hard (NP-hard) problem. Accordingly, metaheuristics – top-level algorithms inspired by various natural and physical processes – can be utilised to determine the near optimality in designing MGs. However, a comprehensive review of the mainstream literature has shown that the performance of several metaheuristics has not yet been evaluated. In response, this paper first systematically benchmarks the efficiencies of previously unexplored metaheuristics in MG sizing applications against the well-established metaheuristic in the literature, namely the particle swarm optimisation (PSO) algorithm. To this end, the metaheuristics are separately integrated into a novel MG sizing method, which is aware of the optimal demand response capacity procured from electric vehicle (EV)-charging loads. Two grid-independent, 100%-renewable MGs are modelled, which enable the reliable and robust supply of electrical loads in areas far removed from the grid. Furthermore, an advanced EV-charging demand response program is integrated into the overall method, whilst quantifying various sources of time-series data uncertainty and considering specific resilience constraints. The simulation results yielded from three real-world isolated community case studies in Aotearoa-New Zealand confirm the effectiveness of the proposed stochastic, resilience-oriented, EV-charging demand response-addressable MG sizing method. Importantly, the comprehensive statistics-based performance evaluations indicate that new metaheuristics have the potential to outperform the PSO by up to ∼6% in MG sizing applications. This indicates the potentially significant implications of using advanced metaheuristics for improving the economics – and, therefore, rolling out – capital-intensive grid-isolated 100%-renewable MGs.