Intelligent LFC for multi-microgrid system integrating ASIC and BESS

High renewable penetration in multi‑microgrids reduces effective inertia and introduces fast, stochastic power imbalance, degrading load‑frequency control. This work proposes an adaptive artificial neural network-based proportional-integral-derivative secondary controller coordinated with an application‑specific integrated circuit and battery energy storage system in the power system having a conventional generator, PV, and wind systems. The application‑specific integrated circuit load has been utilized in a cryptocurrency mining system, allowing for the dynamic adjustment of power consumption based on generation levels. This approach stabilizes the system by using surplus power during periods of high generation and reducing demand during periods of low generation. The proportional-integral-derivative controller gains are pre-trained offline using a genetic algorithm sweeping across diverse operating scenarios and encoded into a feed-forward artificial neural network for online adaptation. Across all scenarios, the proposed controller consistently outperforms traditional methods such as genetic algorithm-based proportional-integral-derivative, particle swarm optimization-based proportional-integral-derivative, and gray wolf optimization-based proportional-integral-derivative controllers, achieving substantial reductions in error metrics such as Integral Square Error, Integral Absolute Error, and Integral Time Absolute Error by up to 28.95%, 16.85%, and 11.65% respectively. It has also achieved a lower rate-of-change of frequency, a 15% rise time, a 13% peak amplitude, a 7% settling time, and control effort for both loads, and the proposed scheme eventually enhances the system stability.