Assessment Method for Wind Power Acceptance Capacity in Rural Power Grids Based on Improved Non-Parametric Estimation

This study addresses the integration capacity of distributed wind power in rural power grids and proposes an improved Kernel Density Estimation (KDE)-based wind power output forecasting model. The bandwidth matrix is optimized to enhance prediction accuracy, and the uncertainty of rural power loads is analyzed. Furthermore, a wind admissible power limit model is developed, and a wind power integration assessment model is established using chance-constrained programming. The model is solved using an optimized genetic algorithm combined with Monte Carlo simulation to improve computational efficiency. The findings provide technical support and decision-making guidance for the development and utilization of distributed wind power in rural areas.