As the share of electric vehicles increases, electric vehicles are exposed to broader of driving conditions (e.g., extreme weather), which reduce the performance and driving ranges of electric vehicles below their nameplate rating. To ensure customer confidence and support steady growth in electric vehicle adoption rates, accurate estimation of battery state of charge and maintaining battery state of health through optimal charge/discharge decisions are critical. Recently, vehicle manufacturers have begun to employ machine learning techniques to improve state-of-charge management to better inform drivers about both short-term (state of charge) and long-term (state of health) performance of their vehicles. This comprehensive review article explores the intersection of machine learning and state of charge management in electric vehicles. Recognizing the critical importance of state of charge in optimizing electric vehicle performance, the article starts by evaluating traditional state of charge estimation methods. Subsequently, it delves into the transformative impact of machine learning techniques and associated algorithms on state of charge management. Through the lens of various case studies, this article demonstrates how machine learning-based state of charge estimation empowers electric vehicles to make informed and dynamic energy usage decisions, enhancing efficiency and extending battery life. The challenges of data availability, model interpretability, and real-time processing constraints are acknowledged as impediments to the widespread adoption of machine learning techniques. Despite these challenges, the future outlook for machine learning in state of charge management appears promising, with emerging trends such as deep learning and reinforcement learning poised to refine state of charge estimation accuracy. Moreover, this study sheds light on the transformative potential of machine learning in enhancing state of charge management efficiency and effectiveness for electric vehicles, offering critical insights.. Machine learning emerges as a game-changing force in state of charge management for electric vehicles, paving the way for intelligent and adaptive vehicles that are both environmentally friendly and efficient. This evolving field invites further research and development, making it a vital and exciting area within the automotive industry.
This article presents a novel approach to address the critical issues of stable rotation and energy efficiency in electric vehicles (EVs). To achieve these objectives, we propose a comprehensive control system that leverages the power of optimization through optimal predictive control methods. The central idea revolves around minimizing the predicted tracking error for future time steps by intelligently determining control inputs. In this innovative approach, we emphasize the dynamic adjustment of weight coefficients and optimization of wheel torque to strike a delicate balance between energy consumption and enhanced vehicle stability. The result is an adept controller that not only ensures vehicle stability but also significantly reduces energy consumption. Given the inherent limitations of electric motors, especially in terms of torque during vehicle operation, and the growing importance of energy conservation, our method tailors weight coefficients to generate optimal wheel torque. This ensures that the electric motors operate within their power range, thereby minimizing energy consumption and extending the overall efficiency of EVs. The combination of stable rotation and energy efficiency offered by this control system represents a promising step forward in the realm of electric vehicles, making them more sustainable and environmentally friendly while maintaining the high standards of performance and safety that consumers expect.
In the nonlinear dynamic equations of the electric vehicle, parameters such as the coefficient of friction between the tires and the road, the coefficient of traction, the resistance of the anchor, and so on have ambiguities. Designing a controller that is robust to the existence of these parametric ambiguities and also to external disturbances, while still satisfying the optimality criteria, is a challenging task. In practical applications, in addition to the problems mentioned above, the computing load of the control input should also be taken into account and a sensible interaction between the performance desired by the controller and the computing volume should be offered. In the present work, a robust, optimally stable fuzzy controller based on parallel distributed compensation is designed using the Takagi–Sugeno fuzzy model of the electric vehicle. The fuzzy model stabilizer feedback gains, the upper bound of uncertainties, the upper bound of disturbance effect, and the upper bound of the cost function are obtained completely offline by solving a minimization problem based on the linear matrix inequality. Therefore, the calculation volume of the control input is extremely small. This allows the proposed control to be put into practice. The good performance of the proposed controller is demonstrated in five-stage simulations.
14 June 2023 in International Journal of Emerging Electric Power Systems
Electric vehicles have many opportunities to participate in auxiliary operations as research interest in renewables and green energy sources grows. This paper discusses novel energy storage technologies for assessing the impact of electric vehicles and distributed flexible alternating current transmission systems in smart grids for improving various power system performance results using various optimization techniques with real and reactive power minimization as objective functions. The authors should strive to make this review article extremely beneficial to scholars, academics, and practitioners. As a result, this review article will also be beneficial to the social community.