A Thompson Sampling Efficient Multi-Objective Optimization algorithm (TSEMO) for Lithium-ion Battery Liquid-cooled Thermal Management System

Efficient design of Battery Thermal Management Systems (BTMS) plays an important role in enhancing performance, life, and safety of Electric vehicles (EV). This paper aims at designing and optimizing cold plate based liquid cooling BTMS. Pitch sizes of channels, inlet velocity, and inlet temperature of outermost channel are taken as design parameters. Evaluating influence and optimization of design parameters by repeated CFD calculations are time consuming. To tackle this, effect of design parameters is studied by using surrogate modelling. Optimized design variables should ensure perfect balance between certain conflicting goals, namely cooling efficiency, BTMS power consumption (parasitic power) and size of the battery. So, the optimization problem is decoupled into hydrodynamic performance, thermodynamic performance, and mechanical structure performance. The optimal design involving multiple conflicting objectives in BTMS is solved by adopting the Thompson Sampling Efficient Multi-Objective Optimization (TSEMO) algorithm. The results obtained are as follows. The optimized average battery temperature after optimization decreased from 319.86 K to 319.2759 K by 0.18%. The standard deviation of battery temperature decreased from 5.3347 K to 5.2618 K by 1.37%. The system pressure drop decreased from 7.3211 Pa to 3.3838 Pa by 53.78%. The performance of the optimized battery cooling system has been significantly improved.