基于层组粒子群优化算法的城轨交通飞轮储能与逆变能馈协同优化配置OA

Efficient recovery and utilization of regenerative braking energy,which constitutes 30％～55％of total train traction energy consumption,is a key challenge in urban rail transit power supply system design.Traditional schemes suffer from low utilization rates of such energy,with a high proportion remaining unabsorbed.Existing technologies mainly rely on energy storage and inverter energy feedback,but they lack in-depth analysis of collaborative mechanisms in capacity matching and energy flow interaction.Particularly,in the mixed-variable optimization problem involving discrete equipment configuration parameters,such as the number of flywheel energy storage equipment(FESE)and inverter energy feedback equipment(IEFE)and continuous voltage thresholds,including charge-discharge voltage thresholds of FESE and grid-connection voltage thresholds of IEFE,achieving the comprehensive optimal system-level energy efficiency and economy remains a significant challenge. To address these limitations,this study proposes a collaborative optimization configuration method for FESE and IEFE.First,an AC-DC hybrid power flow analysis model for urban rail transit was established.This model integrates the steady-state characteristics of the AC power grid and the dynamic load properties of the DC traction network,enabling accurate simulation of energy interaction between AC and DC subsystems and overcoming the limitations of traditional models that focus solely on DC traction networks.Second,a two-layer framework for discrete-continuous mixed-variable optimization was constructed.The discrete variable decision layer focuses on optimizing the quantities of FESE and IEFE to determine the optimal equipment layout,while the continuous variable decision layer introduces a voltage optimization mechanism.This mechanism allows independent adjustment of FESE charge-discharge thresholds and IEFE grid-connection thresholds,breaking the rigid constraints of traditional static voltage preset strategies.Third,a layer-grouped particle swarm optimization(LGPSO)algorithm was developed.This algorithm constructs a comprehensive cost index using total cost and static payback period,and realizes collaborative optimization of multi-dimensional parameters through hierarchical decoupling and grouped evolution strategies. In optimization performance,compared with traditional differential evolution(DE)algorithm and particle swarm optimization(PSO)algorithm,LGPSO showed significant advantages:it reduced iteration counts by 18.75％and 13.33％respectively,while improving solution quality by 8.5％and 18.3％respectively,demonstrating high efficiency in handling high-dimensional mixed-variable problems.Typical case calculations verified the effectiveness of the proposed method.The collaborative optimization scheme(CO-OP)achieved a 43.2％lower static payback period than the FESE-only scheme and a 6.9％higher energy-saving rate than the IEFE-only scheme.In economic indicators,CO-OP reduced equipment investment compared to the FESE-only scheme,with daily total cost lower than both single-equipment schemes,and its regenerative braking energy recovery rate and energy-saving rate also showed comprehensive advantages. The collaborative optimization scheme integrates the technical strengths of both devices:inheriting IEFE's low investment characteristics while utilizing FESE's efficient dynamic buffering to suppress energy feedback issues in IEFE operation.Meanwhile,the voltage optimization mechanism further enhances energy recovery efficiency compared to static voltage strategies.These results indicate that LGPSO excels in high-dimensional optimization,and the collaborative mechanism significantly improves system comprehensive performance,providing theoretical support and practical paths for low-carbon transformation of urban rail transit power supply systems.