Adaptive Coordinated Optimization and Dispatch of a PV–BESS Under Highly Variable Weather Conditions

Abstract

To address significant photovoltaic (PV) power fluctuations, large forecast errors, and differences in the operating states of battery energy storage system (BESS) units under highly variable weather conditions, this study proposes an adaptive coordinated optimization and dispatch method for a PV–BESS. Rolling-horizon optimization is first used to generate the total BESS control command. Elastic-net regularization and an abrupt-change protection mechanism are then introduced to decompose the command into a low-frequency smooth component, a high-frequency dynamic component, and a sparse anomaly component. Based on the decomposition results, state-aware power allocation is performed considering the state of charge, state of health, warning state, and fault status of each BESS unit. The updated available power and energy limits are fed back to the rolling optimization model to form a closed-loop coordinated dispatch process. The case study results show that the correlation coefficient between the extracted high-frequency component and the reference dynamic component reaches 0.983. Under the adopted battery degradation model and equivalent operating conditions, the time until the first BESS unit reaches the end-of-life threshold is extended by 13.20% relative to proportional allocation. Under abnormal operating conditions, the proposed method enables warning-state derating, fault isolation, and regulation-task transfer, while dynamic operating-limit feedback reduces unserved command energy by 94.8%. The proposed method improves tracking of the grid power schedule while coordinating the requirements for preserving rapid power variations, protecting BESS-unit lifetime, ensuring command executability, and maintaining safe operation under abnormal conditions.