A Rolling Horizon Scheduling Model for Accommodating High Penetration of Photovoltaic Power

This paper addresses the difficulty of the power system dispatch accommodating high penetration of photovoltaic power. Due to photovoltaic uncertainty, significant prediction error emerges, especially with long forecast horizon. The scheduling model proposed adopts a rolling horizon strategy, then the reserve capacity catering for the prediction errors of load and photovoltaic forecasts decreases. This reduces the operating costs of all the schedulable units. To better satisfy the power balance, several flexible units are also used to cooperate with the photovoltaic power. The prediction errors of load and photovoltaic output are described as fuzzy variables, and the scheduling model is solved using fuzzy chance-constrained programming method. The test on the 10-machine system demonstrates that the rolling horizon scheduling model is more practical compared to the day-ahead scheduling model.