Articles techniques et présentations

The risk of thermal runaway highlights the critical need for efficient liquid cooling in lithium-ion batteries. This study employs COMSOL Multiphysics® to perform topology optimization of cooling channels within a battery cold plate. Using the density method, we couple turbulent flow and heat transfer models to enable simultaneous consideration of flow and thermal fields in the topology design. Four key parameters—Darcy penalization factor, projection slope, coolant volume fraction, and thermal-hydraulic preference weight—are parametrically modeled. Batch simulations are executed in COMSOL, and the exported data are used in MATLAB to independently construct a Kriging surrogate model and perform NSGA-II multi-objective optimization, balancing heat dissipation and flow resistance. Compared with full factorial design requiring 75,000 simulations, the proposed framework needs only 81 samples to build an accurate surrogate model, achieving approximately 925-fold improvement in efficiency. The optimized topology reduces maximum temperature by about 1.5 K relative to a Z-shaped parallel channel, and reduces pressure drop by 87.49% along with a 0.8 K temperature reduction compared with an S-shaped serpentine channel. This work demonstrates the engineering value of integrating accurate simulation with MATLAB's optimization capabilities in battery thermal management.