关键词: 风光储微电网, 协同调度, 储能配置, 改进差分进化算法

Abstract: To address the limitations of traditional differential evolution (DE) algorithms—specifically their tendency to fall into local optima and weak physical interpretability—in the optimal configuration of multi-park microgrid wind-solar-storage systems, this paper proposes an optimization framework integrating an improved DE algorithm with physical mechanism analysis. First, a wind-solar-storage configuration model is constructed with the objective of minimizing daily power supply costs, incorporating constraints on energy storage charge/discharge efficiency and state of charge (SOC). Second, a triply adaptive improved DE algorithm is designed: it employs a dual-phase linear decay mechanism to adjust scaling factors and crossover probabilities, integrates an elite historical experience reuse strategy to accelerate convergence, and introduces dual-mode oscillatory perturbation to enhance population diversity. Subsequently, intrinsic relationships between energy storage configuration and wind-solar-load curves are analyzed based on the physical essence of source-load matching. Case studies demonstrate that: 1) The improved DE algorithm performs better than traditional DE, particle swarm optimization (PSO), and genetic algorithms (GA), reducing joint-operation costs to 15,424.06 RMB; 2) Joint operation decreases total power supply costs by 6.11% (15,439.59 RMB vs. 16,444.78 RMB) compared to independent operation, saving 30.77% in total storage power (448.10 kW) and 50.00% in capacity (1,469.14 kWh), while eliminating curtailed wind/solar power; 3) A universal physical law is revealed: storage power approximates maximum single-period curtailment, while capacity is determined by total maximum continuous curtailment. Applying this law to Park B yields a physical estimation (626 kWh) costing 5,065.43 RMB—lower than the optimized algorithm result (5,066.22 RMB). This study provides a high-precision, strongly interpretable solution for wind-solar-storage system optimization through algorithmic refinement and physical mechanism exploration.

Key words: Wind-Solar-Storage Microgrid, Coordinated Scheduling, Energy Storage Configuration, Improved DE