A Study on Intraday Rolling Scheduling of Distribution Networks Based on Conditional Multivariate Normal Distribution Random Optimization

The paper introduces an intraday rolling scheduling model for power distribution networks, leveraging conditional multivariate normal distribution-based stochastic optimization. Initially, the forecast error data are binned, and their distributions are fitted. Subsequently, a set of temporally correlated wind power output scenarios is generated by sampling from the inverse transformation of the conditional multivariate normal distribution. These scenarios are then subjected to reduction via the synchronized backward reduction method and incorporated as deterministic inputs into the day-ahead rolling scheduling process of power distribution networks. Case study validates the proposed method, demonstrating its effectiveness in producing stochastic wind power output scenarios that more accurately capture wind power uncertainty, thereby leading to optimized scheduling outcomes that align more closely with actual conditions.