Neuro-Fuzzy and Networks-Based Data Driven Model for Multi-Charging Scenarios of Plug-in-Electric Vehicles

Abstract

In recent times, significant progress has been achieved in the domain of intelligent and eco-friendly transportation. Electric mobility emerges as a viable and effective solution, offering cost-efficient means of transportation. However, the rise in fuel expenses, climate change, and unregulated charging of electric vehicles (EVs) have necessitated a transformative shift in the operation of smart grids. The exponential rise in EV charging requirements has the potential to adversely affect power grids, resulting in peak demand, grid overload, energy hub emissions, and possible infrastructure overburden. This study presents a predictive cost model that utilizes a hybrid search and rescue (SAR) and adaptive neuro-fuzzy interface system (ANFIS), denoted as the SAR-ANFIS approach. The model is designed to effectively model complex dynamic energy emission dispatch scenarios that are integrated with transient charging loads. These scenarios include peak, off-peak, electric power research institute (EPRI), and stochastic scenarios. The proposed methodology aims to minimize the cost of charging scenarios while providing policymakers with a tool to create financial budgets for forthcoming electric vehicle loads. This is achieved through the use of a fuzzy model that has the ability to predict costs. The ANFIS exhibits robust predictive capabilities owing to its aptitude for acquiring and representing intricate non-linear associations between input and output variables. The incorporation of ANFIS into a SAR algorithm results in improved predictive capacity through the optimization of the learning process, enhancement of model accuracy, and facilitation of efficient parameter tuning. The integration of ANFIS and SAR algorithms enhances predictive accuracy and robustness by leveraging the former's adaptive and learning capabilities and the latter's global search and optimization capabilities. The model considers various charging strategies and dispatch constraints within an energy hub. The attainment of the 24-hour pricing scheme is achieved by solving a minimum-cost optimization problem, which serves as the initial training data for the development of the proposed model using an adaptive neuro-fuzzy approach. The proposed approach effectively coordinates the various charging behaviours of electric vehicles, including those identified by the EPRI, as well as stochastic, peak, and off-peak charging, at the system level. The proposed methodology offers several advantages, including the facilitation of coordination among various charging scenarios for EVs and the creation of a cost prediction model that can assist policymakers in devising budgetary plans for future EV loads. One of the benefits of this technology is its capacity for autonomy, which enables vehicle owners to charge their electric vehicles in a cost-effective manner, regardless of the specific scenario they find themselves in. Furthermore, the suggested plan has the potential to mitigate the release of greenhouse gases from the power generation sector, thereby facilitating the establishment of a viable charging network that is environmentally sustainable. The present study examines the efficacy of the SAR-ANFIS approach and model on a standardized test system across a range of load profiles.