E-mobilite uygulamaları için lityum-iyon bataryaların fizik tabanlı modellerinin karşılaştırmalı analizi

Abstract

The widespread adoption of electric vehicles (EV) has led to the development of lithium-ion batteries as a critical technological solution due to their high energy density and fast charging capacity. Continued improvement of these batteries will play a key role in achieving the goals of extending EV range, reducing charging times and improving overall performance. However, this technological development is hindered by several challenges. Battery models are generally classified into four main groups. Empirical models, such as Shepherd, Unnewehr universal, and Nernst, are based on experimental data and offer varying degrees of accuracy and computational complexity. Equivalent Circuit Models (ECM), simplify the representation of the internal electrical characteristics of a battery using circuit elements. They offer flexibility in designing structures suitable for specific applications, but may require more cell information and increase computer time. The data-driven battery modeling based solution utilizes machine learning methods to derive knowledge from large datasets of battery measurements. It excels at modeling complex nonlinear behaviors and provides high accuracy, but requires careful experimental setup and high-quality data. Physics-based models including the Single Particle Model (SPM), the Single Particle Model with Electrolyte (SPMe), and the Doyle-Fuller-Newman (DFN) model describe mathematically the basic physical and chemical phenomena in a battery. In-depth information about battery behavior can be provided and can be used to optimize battery performance. Beside this, as a key part of physics-based models, degradation models are vital in extending battery life and efficiency. This thesis addresses the physics-based simulation methods of lithium-ion batteries for electric cars, focusing in particular on SPM, SPMe and DFN models. Their performance has been extensively studied and compared with different chemical batteries. Their complexity and predictive capabilities are summarized and evaluated based on literature studies. The thesis begins with a comprehensive overview of all battery modeling methods. Next, physics based modeling methods for batteries with different chemical compositions are simulated and analyzed in detail. Optimizations are performed to obtain more efficient parameter values due to uncertainty in literature values. All the mathematical formulations needed to build an electric vehicle model are processed and the electric vehicle model is developed using simulation environments. The electric vehicle model is integrated with physics-based modeling methods for battery sizing and range calculation studies. A detailed simulation study is performed using vehicle models and battery modeling methods. The accuracy and precision of the simulation has been proved by comparing the obtained simulation results with the regional data presented earlier in the literature. The mechanisms that can influence cellular aging are widely studied and these changes are observed in cycles. Finally, this thesis presents a comprehensive analysis of battery modeling methods for electric vehicles. By comparing the performance of different chemical battery modeling methods and optimizing existing modeling techniques, it contributes to a more accurate and reliable prediction of electric vehicle battery performance. This in turn is a significant step towards extending battery life and increasing the durability of electric vehicles. The importance and contribution of the thesis is to provide an updated and comprehensive source of information on battery modeling methods. This is one of the few comprehensive studies that compares the performance of battery modeling methods with different chemical batteries. This helped identify the most appropriate modeling methods for batteries with different chemical compositions. It also facilitated more accurate prediction of battery performance by optimizing existing battery modeling techniques. The thesis provided new ideas and basic tools for the design and optimization of electric vehicles and promoted a more sustainable future with information related to extending the battery life of electric vehicles.