Probabilistic Forecasting of Available Load Supply Capacity for Renewable-Energy-Based Power Systems

In order to accurately analyze the load supply capability of power systems with high penetration renewable energy generation, this paper proposes a probabilistic available (ALSC) forecasting method. Firstly, optimal input features are selected by calculating maximal information coefficient (MIC) between and target output. Based on this, stacking ensemble learning model is applied for prediction wind power, photovoltaic power. Secondly, distributions objects obtained based errors error statistics Finally, set as parameters system, then ALSC calculated using Latin hypercube sampling (LHS) repeated flow (RPF). simulate more realistic multiple slack buses introduced conduct two types imbalance allocations novel allocation principles during RPF calculation, which makes evaluation results reasonable accurate. The can provide reference capacity system in future, they also an early warning risk threshold overlimit. Case studies carried out modified IEEE 39-bus verify feasibility effectiveness proposed methods.