A probabilistic analysis method for the impact of traction load and wind power on power grid voltage quality based on Shapley value incorporating spatial correlation and uncertainty

With the increasing integration of traction power supply systems (TPSS) and wind farms into power grids, existing assessment methods are difficult to apply for quantifying the interactive effects of multiple uncertainties on grid voltage quality. A probabilistic assessment method based on Shapley value is proposed to evaluate voltage quality impacts in regional power grids containing both railway traction systems and wind power. Firstly, a non-parametric probabilistic modeling method based on adaptive kernel density estimation (AKDE) is proposed for traction loads and wind power, with its accuracy being validated. Secondly, an improved Latin Hypercube Sampling (LHS) method incorporating equal probability transformation and Cholesky decomposition techniques is developed to enable efficient sampling of correlated random variables. Then, a three-phase probabilistic load flow (3PLF) model incorporating an asymmetric traction substation is established. The method for assessing the probabilistic impact of traction loads and wind power on regional power grid voltage quality is proposed by combining 3PLF with Shapley value theory. Finally, the case validation is conducted in a modified IEEE 14-bus three-phase system using measured traction load and wind power data. Results show that when the operating trains at the traction substation change from one HX D 1 to three HX D 2s, the probability of the three-phase voltage unbalance degree exceeding the limit at the traction load bus increases by 10 %. The research results provide a theoretical basis and decision-making tool for voltage quality management in regional power grids with high wind power penetration integrated with TPSS.