A Two-stage Optimal Dispatching Method of Distribution Network Considering the High Proportion of Distributed Renewable Energy Penetration

The increased incorporation of distributed renewable energy (DRE) into distribution networks has presented new obstacles to the planning and operation of these systems due to its intermittency and volatility. This research examines the integration of DRE by modeling the output characteristics of distributed energy and the load fluctuation characteristics of the distribution network with deep learning technology. Specifically, the Long Short-Term Memory (LSTM) model is adopted to create a dynamic output prediction model, which is used to analyze the impact mechanism of the dual volatility of “source and load” on distribution network planning with the incorporation of DRE. Thus, a dual-phase optimization approach for day-ahead and intra-day planning is suggested to improve distribution network decisions from both safety and economic perspectives. The aim is to improve the system’s capacity for renewable energy utilization, financial benefits, and operational reliability. Additionally, a real-time optimization method and a model validation framework are developed.