Enhancing protection in AC microgrids: An adaptive approach with ANN and ANFIS models.

• Adaptive protection scheme for AC microgrid (MG) is proposed based on ANN-ANFIS model. • Different short-circuit levels, dynamic distribution system parameters, and diverse network topologies in MG operations have been addressed. • Effect of signal noise on protection system accuracy has been investigated. The increased penetration of distributed generation (DG) in distribution networks is altering the conventional mode in which the power grid's functions. Therefore, microgrids (MG) are most efficient and feasible solution to integrate sustainable energy sources into power grids. Existing protection method uses numerical relays. Thus, it provides flexibility to upgrade to a more sophisticated artificial intelligence (AI) based protection strategy. The use of signal processing (SP) and AI techniques, such as artificial neural networks (ANN), could assist in decision-making process resulting in successful isolation of faults. In this paper, an intelligent adaptive protection method is proposed for AC-MGs. This study showcases the potential of ANNs and SP techniques, complemented by adaptive neuro-fuzzy inference systems (ANFIS), in MG protection. These approaches demonstrate the utility of proposed methodology in design, testing, and evaluation of protection strategies, ultimately bolstering system security and efficiency. The proposed approach encompasses fault detection, classification, and location based on signal analysis, assisted by ANN and ANFIS. The methodology accounts for dynamic behavior and different faults, even in two distinct operating scenarios— grid-connected and islanded modes. Through comprehensive validation, results affirm the efficacy of proposed methodology, encouraging deployment of AI-based protection strategies for AC-MGs leading to the foundation for more robust and intelligent protection system. [Display omitted] [ABSTRACT FROM AUTHOR]