Optimal Configuration of Inverter-Based Regenerative Feedback and Flywheel Storage in Urban Rail Power Systems via LGPSO

The coordinated optimal configuration of inverterbased regenerative feedback systems (IEFS) and flywheel energy storage systems (FESS) in urban rail transit power supply faces significant challenges due to high-dimensional mixed-variable optimization. The strong coupling between discrete equipment deployment and continuous voltage threshold parameters leads to curse of dimensionality. Traditional optimization methods struggle to achieve global optima efficiently owing to high computational complexity and large solution space dimensionality. To address this, this paper proposes a Layer-Grouped Particle Swarm Optimization (LGPSO) algorithm through three innovative mechanisms: hierarchical decoupling, dynamic resource allocation and differentiated co-evolution. Case studies on a sixstation metro system demonstrate that the optimized configuration achieves a 10.05% annual operational cost reduction, 93.20% decreased braking resistor energy dissipation, and a 2.32 -year payback period, validating the algorithm’s efficacy in mixedvariable optimization. This method provides a computationally efficient and engineering-feasible decision-making framework for intelligent power system configuration in urban rail transit, extendable to other high-dimensional power system optimization scenarios.