改进典型场景下的风光储牵引供电系统源储协同优化配置OA

The integration of distributed renewable energy sources into traction power supply systems represents a key step toward realizing low-carbon and resilient electrified railways.To enhance both the economic efficiency and supply reliability of such systems,this study investigates the coordinated planning and operation of a combined wind-photovoltaic-storage traction power supply system(CWPS-TPSS).The main objective is to determine the optimal configuration of renewable and storage capacities under uncertain operating conditions and to quantitatively assess the bidirectional reliability between the CWPS-TPSS and the external grid. To capture the stochastic characteristics of renewable energy generation,a data-driven scenario generation and reduction framework is developed.Specifically,a Wasserstein generative adversarial network with gradient penalty(WGAN-GP)is employed to learn the complex temporal correlations and non-Gaussian features of wind and photovoltaic(PV)power outputs,generating diverse and physically consistent scenarios that reflect the variability of renewable resources.Subsequently,an improved hierarchical clustering algorithm based on probabilistic distance metrics is designed to perform scenario reduction,ensuring that the reduced set maintains statistical representativeness while significantly decreasing computational complexity during the optimization process.This hybrid approach overcomes the limitations of conventional sampling methods that often rely on limited historical data and linear assumptions. On this basis,a bi-level multi-objective coordinated capacity optimization model is constructed.The upper-level model addresses long-term capacity planning by minimizing the daily equivalent annualized investment cost,optimizing the capacities of wind turbines,PV arrays,energy storage systems(ESS),and traction substations subject to technical,reliability,and power balance constraints.The lower-level model describes short-term operational dispatch under varying renewable and traction load conditions.It minimizes the system's operational cost while accounting for reliability indices that characterize the dynamic interaction between the CWPS-TPSS and the main grid.Two reliability indicators are defined:(1)The source-load fluctuation rate,quantifying the temporal mismatch between renewable generation and traction demand.(2)The average locomotive load loss probability during islanded operation,measuring the risk of supply interruption under grid faults.The coordination between the two levels is achieved through iterative information exchange until convergence in both investment and operational decisions is attained. The proposed framework is applied to a representative electrified railway corridor equipped with wind,PV,and ESS units.Simulation results show that the integrated optimization method yields a rational allocation of source and storage capacities that effectively balances cost,reliability,and renewable utilization.Compared with traditional single-objective or independent subsystem designs,the coordinated bi-level strategy achieves smoother power interactions with the external grid and a lower reliance on grid-side supply support.Under the obtained optimal configuration,the total system cost is significantly reduced while the reliability indicators satisfy operational constraints.Sensitivity analyses are further conducted to evaluate the effects of renewable resource availability,storage cost,and reliability weighting factors on the capacity allocation results.The findings indicate that increasing storage capacity enhances the system's ability to accommodate renewable fluctuations and reduce locomotive load loss,although with diminishing marginal benefits beyond a certain threshold. The study demonstrates that combining advanced scenario modeling techniques with hierarchical multi-objective optimization provides an effective pathway for the planning and operation of renewable-integrated traction power systems.The modeling approach and quantitative indicators developed herein can be generalized to other railway electrification projects and distributed energy systems that require simultaneous consideration of investment economy and power-supply reliability.Future work will focus on extending the framework to include carbon trading mechanisms,real-time dispatch under uncertainty,and resilience evaluation under extreme operating conditions.