Multi-Objective Jellyfish Search Algorithm for Two-Level Stochastic Planning and Demand-Side Management of Renewable-Integrated Microgrids

Abstract

The optimal placement and power injection of photovoltaic (PV), wind, and battery energy storage system (BESS) units are critical for reliable microgrid operation under renewable-generation uncertainty. This study proposes a two-level planning-and-operation framework for a grid-connected microgrid. At the planning stage, PV and wind uncertainty is represented using correlated Latin Hypercube Sampling (LHS) with scenario reduction; candidate nodes for placing DG units are identified through Active Power Loss Sensitivity Factor (APLSF) screening and optimal DG power injection is obtained using a Multi-Objective Jellyfish Search Algorithm (MOJSA) that jointly minimizes power loss and pollutant emissions. At the operational stage, a rule-based energy management system (EMS) coordinated with demand-side elastic load shifting with an aim to minimize the 24 h operating cost. The two-level optimization framework is evaluated on 33-node and 118-node microgrid test systems across five demand-side management (DSM) participation levels. Relative to the no-DSM baseline, results show the operating cost reductions of up to 3.42% (33-node, 30% DSM) and 10.35% (118-node, 40% DSM), with peak-hour pollutant emission reductions of 52.43% and 21.556%, respectively.