基于边缘计算的光伏阵列故障诊断系统:深度学习轻量化部署

The evolution of multi-regional interconnected power systems has been driven by the swift technological advancements of power-producing technologies and the increasing demand from power consumers. To meet the growing power demand, the power system relies on renewable energy sources, but their unpredictable nature influences the dynamic performance of the power system, especially frequency regulation. In response to this challenge, a robust cascade controller called FOPIDN (1+PIDN) has been recommended for the Load Frequency Control (LFC) system. This system encompasses two distinct areas that integrate a non-reheat thermal system with renewable energy sources such as solar photovoltaic and wind energy systems. Additionally, the study has introduced nonlinear constraints like Governor Dead Band (GDB) and Generation Rate Constraints (GRC) to enhance the realism of the analysis. The optimal parameters for the controller have been fine-tuned using the Grey Wolf optimization (GWO) technique by minimizing the ITAE performance index. The findings unequivocally show better performance in terms of performance indices, settling time, overshoots, and undershoots of the proposed controller over existing controllers such as PID, PIDN, FOPID, and FOPIDN. At last, a robustness study has been conducted to validate the efficiency of the proposed controller, assessing its performance across different load scenarios and under extreme variations in system parameters.