Quantum-inspired Multi-scenario Optimization for Robust Battery Energy Storage System Control

Battery Energy Storage Systems (BESS) operate under varying environmental and load conditions, yet conventional optimization methods typically optimize control policies for a single or a limited set of operating conditions. As a result, these policies can exhibit degraded performance and accelerated capacity loss when actual temperature and load conditions differ. To address this, a scenario-based Quantum-Inspired Evolutionary Algorithm (QEA) framework is proposed for degradation-aware BESS control across diverse operating conditions. The Scenario Set Evaluation considers four ambient temperature ranges and three load profile types, with six ambient-temperature values uniformly sampled within each temperature range for each temperature–load combination, yielding 72 operating scenarios. Load profiles are modeled as superposition of Gaussian peaks with load-specific minimum depth-of-discharge (DoD) requirements, and penalties are imposed in the objective function when the achieved DoD falls below the required minimum.

The QEA optimizes binary-encoded decision variables, including C-rate, maximum and minimum State of Charge (SOC_max and SOC_min), and thermal set point (T_set) by combining a dimensionless battery-longevity score and fractional round-trip efficiency into a composite objective. The proposed framework achieves a mean composite objective value of 0.921 on a normalized 0 to 1 scale, exceeding the corresponding values of 0.871 and 0.883 obtained using Genetic Algorithm (GA) and Particle Swarm Optimization (PSO), respectively. The QEA also achieves 94.7% round-trip efficiency, while the model-based projected cycle life is 2589 cycles to 80% State of Health (SoH), corresponding to a 37.0% longer projected cycle life than GA and a 74.1% longer projected cycle life than rule-based control. The proposed framework therefore enables a single degradation-aware BESS control policy to balance battery longevity and efficiency across diverse evaluated temperature and load conditions.