ANFIS (Adaptive Neuro-Fuzzy Inference System) based on Microgrid's Reliability and Availability

The applicability of machine learning algorithms used to solve microgrid optimization is investigated in this paper. This paper's main objective is to build a microgrid model to achieve maximum reliability and availability using renewable resources that cater to users' needs with different demands and supplies. The model generated from the adaptive neuro-fuzzy inference system (ANFIS) is used to get the optimum reliability and availability strategy to achieve the user expectations and needs of future microgrids. The ANFIS model is trained with different data sets from Markov modeling. The dataset is divided into three sections, 40% of the data is used to train the model, testing is performed with 40%, and the last 20% of the information is checked. Implementation results show that ANFIS models emulate Markov modeling methods and artificial neural networks model and enhance reliability and availability. Additionally, the ANFIS model is better than the Markov and artificial neural networks models regarding the impact of individual failure rate optimization from sub-models. Comparison is shown with Markov modeling, genetic algorithm, artificial neural networks, and fuzzy system. The optimization toolbox Matlab is used for ANFIS.