Design of a high-accuracy energy management system for electric vehicles and V2G approaches considering battery aging

Abstract

Smart grid technology uses digital communication technologies to optimize the processes of energy production, distribution, and consumption, providing more flexible, reliable, and efficient energy management compared to traditional energy production and distribution systems. This system has the capability to monitor and manage energy demand in real-time, allowing for more reliable use of energy resources and faster detection of faults in the grid. The integration of renewable energy sources into smart grid systems is crucial for reducing the use of fossil fuels and meeting the increasing energy demand. When renewable sources such as solar, wind, hydroelectric, and geothermal energy are combined with smart grid technologies, they offer a sustainable and environmentally friendly approach to energy production and consumption. This integration increases energy efficiency while reducing the carbon footprint. With the increasing use of electric vehicles, the necessity of vehicle-to-grid (V2G) service technology is also rising. V2G technology transforms the batteries of electric vehicles into energy sources that allow bidirectional energy flow. This integration enables electric vehicles to become not only consumers but also energy producers, providing feedback to the grid to balance energy demand fluctuations. Thus, the integration of renewable energy sources is facilitated, energy storage solutions become more effective, and grid stability is enhanced. The V2G integration of renewable energy systems plays a critical role, especially in ensuring the continuity of intermittent and irregular sources like solar and wind energy. By storing energy from these sources, electric vehicle batteries can supply energy to the grid when energy production is insufficient. This optimizes the balance of energy production and consumption, providing both economic and environmental benefits. This study examines the frequency regulation service, one of the vehicle-to-grid services, in detail from the perspective of electric vehicle users. The study considers the battery aging and economic benefits of providing frequency regulation service when electric vehicles are connected to the grid. First, to perform a realistic analysis, a comprehensive model was developed by considering various forces and factors affecting electric vehicles. In this model, factors such as traction forces, road characteristics, environmental forces, auxiliary systems, driving profile, and gravitational forces were examined. The traction forces of the electric vehicle are related to the torque produced by the motor and the power losses during the transmission of this torque to the wheels. Road characteristics are defined by parameters like road gradient, surface type, and friction coefficient. Environmental forces include external factors like air resistance and wind. Auxiliary systems consist of in-vehicle electronic systems, air conditioning, and other energy-consuming components. The driving profile includes the vehicle's behaviors during acceleration, deceleration, stopping, and cruising. Gravitational forces determine the effect of the vehicle's weight on the road. Then, battery aging mechanisms and factors affecting battery aging were examined in detail, and the Arrhenius battery aging model was used for lithium iron phosphate battery cells. The aging model was considered for two separate conditions: cycle aging due to regular use and charge-discharge cycles, and calendar aging occurring over time under storage or preservation conditions. A battery pack was designed using four different lithium iron phosphate battery cells on the developed electric vehicle model. In this process, a selection algorithm was created by considering critical factors such as the volume, weight, cost, and aging of the battery pack. The algorithm aimed to make the most suitable selection by considering the different characteristics of each battery cell. For each battery cell, state of charge, state of health, and range data were analyzed. The state of charge indicates the current energy level of the battery, while the state of health evaluates the battery's lifespan and performance. Range data were examined to determine the distance the vehicle can travel on a single charge. Based on these analyses, the performance characteristics of each battery cell were compared, and the most suitable battery cell was selected. This comprehensive evaluation process aimed to offer an optimized solution for battery pack design in electric vehicles in terms of both performance and cost. Thus, the most suitable battery cell was selected for a long-lasting, cost-effective, and high-performance battery pack. To plan vehicle-to-grid services, Great Britain's grid frequency data was used. One year of frequency data was clustered using the K-means clustering algorithm. The K-means clustering algorithm is a type of algorithm used to partition data from complex and large datasets into similar clusters. The algorithm was run multiple times iteratively to determine the days closest to the centroid of each cluster. In this way, the days that best represented each cluster were identified. The variances of the selected days were compared, and the 10 most different days exhibiting distinct characteristics were identified. These selected days were used to derive 10 years of grid frequency data, considering the percentage sizes of the clusters. The number of days in the 10-year dataset was distributed according to the percentage sizes of the clusters. The derivation process was carried out based on the available one-year frequency data to realistically reflect long-term scenarios. Daily frequency data was divided into 30-minute periods, and optimization was performed with a binary genetic algorithm by assigning various weights in terms of revenue and battery aging of the electric vehicle. The binary genetic algorithm is a method where decision variables are represented in binary form and are evolutionarily optimized through genetic operators. This algorithm is population-based, with each individual evaluated as a solution proposal. New generations are formed using operators such as selection, crossover, and mutation, evolving towards the best solution. The scenario considers that the electric vehicle goes on a trip twice a day, once at 9:00 AM and once at 6:00 PM, each lasting 30 minutes. During each trip, the vehicle travels a total of 23 kilometers. The vehicle is charged every day between 11:00 PM and 11:59 PM and starts each day with an 80% state of charge. For the remaining time, the vehicle is assumed to be available for V2G frequency regulation service. The V2G service can operate bidirectionally when the state of charge is between 90% and 40%. If the state of charge is above 90%, the battery can only be discharged through the V2G service. If the state of charge falls below 40%, the battery can only be charged through the V2G service. The optimization algorithm decides whether to provide vehicle-to-grid service during different periods of the day according to various weights. During this process, an optimal strategy is developed in line with grid demands and the vehicle owner's income expectations. The study examines a total of six different scenarios. These scenarios compare the effects of different charging station powers, such as 11 kW and 3.7 kW, on revenue and battery aging. For each charging power, optimal results are obtained under conditions of minimizing battery aging, maximizing revenue, and 50% revenue, 50% battery aging weights. In conclusion, while the 11 kW charging station provides more revenue in the V2G service, the battery aging rate is also higher. This situation is related to the higher stress caused by the high charging power on the battery, which shortens the battery life in the long term. On the other hand, while 3.7 kW nominal power stations provide less revenue, the battery aging rate is also lower.