HOLISTIC INVESTIGATIONS OF THERMAL MANAGEMENT SOLUTIONS FOR ELECTRIFIED TRANSPORT APPLICATIONS

Electrification of transport and refrigeration systems is a critical challenge that must be addressed as global warming necessitates decarbonization across all CO₂ emitting systems. These systems are often embedded parts of infrastructure that must work reliably and efficiently, necessitating careful design, simulation, and experimentation. The applications considered herein cover the optimization of battery electric vehicle lifetimes, electrification of transport refrigeration systems, and immersion cooling for batteries. To this end novel frameworks are developed for battery electric vehicle dynamic co-simulation as well as dynamic simulation of transport refrigeration systems. These novel simulation frameworks capture a holistic suite of multi-scale, multi-physics effects that govern these vehicles across dynamic drive cycles relevant to their design and optimization.

The dynamic co-simulation framework for battery electric vehicles captures localized battery aging effects across an ambient boundary condition and lifetime driving intensity framework. This framework relies upon battery data to form equivalent circuit and Arrhenius based aging models, of which there is a notable gap of models considering low-temperature aging and prismatic cell formats. The complete holistic aging framework is used to investigate the impacts of battery module construction, driving intensity, resting period, and fast charging on discretized battery aging and power fade across a theorized vehicle lifetime. Results indicate that battery cell selection is critical to reduce aging when designing battery electric vehicle (BEV) systems. An appropriately chosen cell can avoid worsened aging from high discharge rates because the thermal management system can appropriately control the heat generated inside of the battery pack during fast charging events. Discretized battery aging for each case is presented with an average capacity loss of 10.18% across its theorized lifetime. This demonstrates the utility of the designed framework, allowing the adaptation of reduced-order parameters and the production of BEV lifetime performance. The gaps identified in battery testing data, namely a lack of low-temperature aging studies of prismatic cell formats, motivates the design and construction of a dual-purpose testing facility that can perform combined battery aging and pseudo-thermal runaway testing under single-phase battery immersion cooling.

A dual-purpose experimental facility is developed to analyze immersion cooling of battery cells. This facility consists of a thermal bath, flow valves for control and charging of an immersion tank, and the immersion tank itself which suspends five prismatic battery cells with a variable gap size between each cell inside of a single-phase dielectric fluid. A copper heater block takes the place of the central cell in the group of five and acts as a heater to impose a pseudo-thermal runaway (PTR) heating boundary condition on the neighboring cells. During this PTR heat pulse, the thermal response of the neighboring cells is recorded and used to compare a dynamic numerical model. Once compared the numerical model is utilized to extend beyond PTR events to generate design guidelines for immersion cooled battery packs. Additionally, the ability of the test stand to hold a battery at an isothermal skin condition is proven at low temperatures and control of the battery management system established allowing for future construction of battery aging models at low temperatures and extended battery testing.

The dynamic BEV TMS framework is then extended to consider a commercial level BEV transportation refrigeration unit (TRU). This extended framework encompasses a trailer for last mile delivery of refrigerated goods to their destination for sale and consumption. This framework includes the refrigerated container, its thermal management system and control, the combined traction and refrigeration battery, and power electronics that move the vehicle. The battery is initially sized and then the framework is tested over a hot summer day and the thermal response of a baseline electrified system and with the addition of a photovoltaic panel. Results demonstrate that a battery size of 400 kWh is sufficient for electrification even over a hot summer day and energy harvesting of 6 kWh is achieved over the course of a single day of deliveries. This framework is then extended to consider the lifetime emissions of a theoretical BEV TRU considering its mission profile, localized climate, and nationally distributed grid emission intensities. It is found that the most sensitive parameter for BEV TRU emission lifetimes is the grid, and considering a non-homogenous grid distribution is critical for accurately capturing the impact of refrigerated transport.

This work completes a comprehensive framework for various BEVs which can account for thermal system response, system target temperatures, battery aging, and TRU performance and parasitic drain. A dual-purpose test stand is constructed to address open questions of immersion cooling safety and to provide low-temperature characterization of batteries in the context of immersion cooling. Immediate avenues for future work would be to extend the established experimental and simulation platforms for further parametric study and validation. This could include an experimental study on battery immersion cooling to characterize low temperature aging or integrate immersion cooling results into existing dynamic models. Additionally, the lifetime transportation refrigeration framework already used to examine lifetime performance of a BEV TRU across the United States, can be utilized to examine other countries, climates, or mission profiles and validated with original experimental data acquired from instrumentation and testing of a TRU. Additional pathways for future extensions of the work could be the integration of the existing immersion cooling test stand for a hardware in the loop platform or the testing of alternate immersion-cooled components such as server rack and for stationary battery applications. Longer-term future work for the BEV TRU space could be the application of neural networks as machine-learning-based surrogates for temperature prediction inside of the TRU volume where a direct co-simulation would be too computationally expensive.