ENABLING FAST CHARGING AND DISCHARGING OF BATTERY SYSTEMS THROUGH ADVANCED THERMAL MANAGEMENT STRATEGIES

The increased reliance on renewable energy sources has made energy storage systems, such as batteries, commonplace. Battery-based devices usually require a supporting ecosystem to ensure proper functioning: specifically, they require frequent charging and discharging. One of the important aspects of a battery system is its thermal management system. Proper thermal control is critical for the efficient and safe operation of the devices, especially for fast charging and discharging applications. Such applications generate significant heating in various components of the systems, including the charging ports, charging cables, and the batteries themselves. This thesis addresses the challenges associated with cooling battery systems using two techniques: one leveraging phase change materials (PCMs) for passive thermal management and the other considering active thermal management using immersion cooling of battery cells and modules.

In the first portion of the thesis, I consider a PCM-based thermal management system (TMS) solution, originally intended for use within a direct current fast charging port (DCFC) during fast charging applications. More generally, this passive approach of using PCM-based thermal management can be applied to other aspects of the battery and electric vehicle systems, such as the power modules. At first, to accelerate the design of a PCMbased TMS, I developed a new figure of merit (FoM) to quickly, but quantitatively, compare the performance of different PCMs in potential design configurations. Compared to computationally expensive numerical simulations, this FoM can rank performance and enable design optimization in a fraction of the time. Overall, the proposed FoM can identify the best-performing geometrical configuration for a certain PCM or shortlist the best-performing PCM for a given setup of PCM-based TMS without detailed numerical simulations, making the design process more efficient. Furthermore, the newly developed FoM can accelerate the optimization of these TMS, such as composite PCMs (i.e., metal foams impregnated with PCM), by analytically predicting the optimum porosity in a given configuration. Subsequently, I focus on the degradation of PCM properties due to thermal cycling, which is a critical factor in selecting a PCM for a cooling solution. For this, we subject the PCMs to thermal cycling and then measure the phase change properties using differential scanning calorimetry. Based on the measured properties, we use a mix of numerical and analytical approaches to estimate the impact of thermal cycling on the performance of PCM-based TMS. To sum up, this portion of the thesis provides tools and insights that accelerate the design of a PCM-based TMS and make it more robust.

The second portion of this thesis focuses on immersion cooling of Lithium-Ion batteries (LIBs), where a dielectric fluid is in direct contact with the LIB cells. Immersion cooling is an effective cooling approach, but because of its complex nature, a thorough understanding of the underlying physics is required before it will see wide adoption into commercial systems. At first, I developed a fully-coupled numerical model that solves the detailed electrochemical submodel in conjunction with the thermal-fluid submodel to investigate the performance of a LIB subjected to forced immersion cooling. This multiphysics modeling approach is superior to previously developed models, which focused on thermal-fluid aspects and often neglected the coupling between temperature, cell potential, and heat generation. Batteries operated more efficiently at moderate temperature rise (∼ 15 K at 5C). Therefore, improved temperature control with immersion cooling leads to higher heat generation with increased capacity loss: a 3 K temperature rise corresponds to 10% loss, whereas a 42 K temperature rise results in 0.4% loss at 5C discharge. Building on this work, I develop a computationally efficient alternative to fully numerical models to analyze immersion cooling-based battery thermal management systems (BTMSs), considering the coupled electrochemical and thermo-fluid physics. The core strategy is to use simplified electrochemical submodels coupled with a lumped thermal submodel that uses a heat transfer coefficient (h) to account for the impact of fluid flow. Moreover, h is either based on the correlations from literature or estimated from a customized correlation trained from the numerical models. The significant reduction in computation cost [from hours or days for the fully-coupled numerical models to seconds for proposed models] makes the proposed approach more suitable for rapid analysis, optimization, and real-time implementation of the immersion-cooled BTMSs.

Further, we perform a comprehensive study of static immersion cooling, which spans from experimental measurements to data-driven models to provide an end-to-end analysis. We measure cell voltage and temperature for three different LIBs (two commercial and one customized with an internal thermocouple) and five cooling fluids (four dielectric fluids and air) across various discharge rates (0.25C to 5C). A general conclusion is that the higher cooling (or lower temperature rise) results in higher heat generation and lower discharge time (∝ energy supplied), which becomes more significant at higher discharge rates. Furthermore, the data enables the estimation of heat generation rates and the development of new convection correlations that incorporate the time and discharge rate dependence. Building on this, we develop two data-driven models for temperature prediction, where the first one is based on an energy conservation equation (error ≤ 0.99 K), and the last uses a modified electrochemical model (error ≤ 0.86 K). Based on the experimental results, we show the influence of fluids on LIB degradation, which suggests that extreme temperature rise (highest or lowest) results in higher irreversible capacity loss. In the end, we revisited the convection boundary condition used for validation purposes in the electrochemical-thermal models. Conventionally, h is chosen based on the values or correlations in literature or estimated scientifically by fitting the cooling curve of LIB (that is, temperature vs. time) after discharge into a lumped energy conservation model. However, we demonstrate that the conventional approach of assuming a constant h does not accurately capture convection cooling on the battery surface during discharging using two independent methods: experimental and numerical. To address this issue, we propose using a heat transfer correlation that depends on cell temperature and time. To demonstrate the effectiveness of the proposed approach, we show that it accurately predicts the heat generation rate in addition to cell voltage and temperature, unlike the conventional approach.

In summary, the last portion of the thesis provides novel techniques and insights that will be useful in the design, optimization, and real-time operation of battery thermal management systems.