Adaptive Neuro-fuzzy Inference System-Based Data-Driven Model for Optimal Recharging of Electric Vehicles and Cost Prediction in Energy Hubs

Abstract

This study presents a hybrid neuro-fuzzy prognostic framework that develops an energy allocation method for current urban power infrastructure. This is accomplished by combining a nature-inspired intelligent optimization method with the learning capabilities, interpretability, and reasoning powers of neural networks, known as an adaptive neuro-fuzzy interface system. To align the problem’s representation with real-world situations, this study includes different charging demand tactics enforced by electric vehicles (EVs) along with the limitations of the power-generating systems at energy facilities to achieve optimal power generation. The implementation of the one-day pricing structure entails a complex optimization process aimed at reducing costs and minimizing the release of detrimental emissions. The proposed model was constructed using an adaptive neuro-fuzzy approach, with the training data derived from this optimization problem as the central component. The proposed approach effectively handled the diverse energy demand patterns demonstrated by EVs, encompassing the Electric Power Research Institute recommendations, stochastic behavior, and both peak and off-peak charging. This cooperation occurs systematically. The Crow Search Algorithm-Adaptive Neural Fuzzy Inference System achieved a reduction in operational expenditures in terms of percentages, such as 2.66%, 3.39%, 3.94%, and 2.63% for all recharging scenarios. These values are lower than those of other advanced approaches. The hybrid strategy that has been established has several benefits, such as the efficient management of various scenarios related to the demand for charging energy for EVs and the development of a predictive cost scheme. This scheme can assist policymakers in formulating cost-effective energy policies and budget plans for future EV loads. Moreover, they offer the advantage of self-reliance, allowing EV owners to charge their cars economically in many accessible situations. Another important component is the ability of urban planners to reduce greenhouse gas emissions from power generation, thereby promoting the development of an ecologically sustainable charging infrastructure. Finally, a benchmark test system was employed to evaluate the efficacy of the proposed approach across different energy consumption patterns.