A Four-Step Friendly Scheduling Framework for Hybrid Renewable Energy System

Abstract

In the context of the "dual carbon" targets, renewable energy has gained a good opportunity for large-scale development, but its uncertainty and volatility have brought severe challenges to the cooperation and coordination of "source, grid, load and storage" resources in hybrid renewable energy systems (HRES). This paper proposes a four-stage research framework for "friendly scheduling" HRES resources. It firstly focuses on low-carbon power dispatch modeling, incorporating economic and environmental factors. And then, it addresses the forecasting of wind and photovoltaic power generation by introducing a hybrid Convolutional Neural Network-Sparrow Search Algorithm-Long Short-Term Memory (CNN-SSA-LSTM) model that integrates CNN for spatial feature extraction and SSA-optimized LSTM parameters to achieve spatiotemporal feature fusion, significantly improving prediction accuracy. In the following, it undertakes a thorough analysis of wind and photovoltaic collaborative operation strategies, with the aim of resolving intermittency issues and optimizing system stability. Finally, it employs fuzzy chance-constrained programming and a Stepwise Carbon Trading mechanism to address HRES carbon constraints, enhancing system robustness and staged carbon management capabilities. Multi-scenario experiments show the effectiveness of these approaches. (1) The CNN-SSA-LSTM model reduces wind and solar forecasting MSE by 40% and 21%, respectively, compared to the LSTM model. (2) The incorporation of fuzzy programming and carbon trading enhances reserve capacity while improving utilization efficiency and economic-environmental benefits. (3) The integrated strategy ensures reliability and promotes sustainable power system development.

Keywords: HRES, Friendly scheduling, Source-Grid-Load-Storage Coordination, Fuzzy chance-constrained programming, Stepwise carbon trading