Data-driven coordinated dispatch of source-grid-load-storage systems with renewable energy resources and storage integration

In response to the global imperative of green energy transition, this paper investigates data-driven coordinated dispatch strategies for source-grid-load-storage (S-G-L-S) systems integrating distributed energy resources (DERs) and energy storage (ES). Traditional centralized power systems suffer from inefficiency and inflexibility, motivating data-driven coordination of DERs and ES to enhance operational reliability and renewable energy utilization. This paper proposes a comprehensive S-G-L-S coordinated dispatch framework to address key challenges such as renewable intermittency, load uncertainty, and multi-objective optimization. The contributions of this study are threefold. First, a three-tier power data analysis framework is developed by integrating Gaussian Mixture Model (GMM)–based anomaly detection, seasonal-trend decomposition with linear interpolation for data cleansing, and a long short-term memory (LSTM) network for time-series power data forecasting. Second, an improved salp swarm algorithm (ISSA) is introduced, incorporating hierarchical evaluation, re-update mechanism for suboptimal followers, and dynamic leader rotation to enhance DER-ES coordinated dispatch. Moreover, a multi-objective extension of ISSA, ISSA-MO, is developed by integrating Pareto non-dominated sorting and constraint-handling preferences to effectively balance trade-offs among energy loss, voltage stability, and grid dependency. Experimental validation on IEEE 33-, 69- and 119-node systems demonstrates that ISSA reduces active power loss by up to 98.11% and minimizes daily energy loss, while ISSA-MO generates well-distributed Pareto fronts for multi-objective S-G-L-S dispatch. The results demonstrate improvements in economic efficiency, operational reliability, and environmental sustainability, providing valuable insights for the development of global low-carbon power systems.