A novel 3D thermal network model and multi-criteria optimisation of hybrid battery thermal management systems for electric vehicles

Thermal management of lithium-ion batteries under high-load conditions remains a critical bottleneck for the safe operation of electric vehicles. Current multi-criteria decision-making (MCDM) optimisation frameworks rely on computational fluid dynamics (CFD) or data-dependent surrogate models, severely limiting the range of design variables that can be explored. This paper proposes a novel, large-scale MCDM optimisation framework for a hybrid BTMS combining phase change material (PCM) and liquid cooling. A transient, 3D lumped thermal network model for a battery pack is developed and validated, enabling rapid and accurate evaluation of 78,000 distinct design configurations. The framework optimises seven design variables, including geometrical and flow control parameters, across three liquid cooling architectures and 26 PCMs, against four quantitative and eight qualitative, conflicting criteria. By coupling the TOPSIS ranking approach with six distinct weighting methods, the study reveals that the C₁ liquid cooling configuration, with a counter-current middle tube layer, shows better heat dissipation than the two other arrangements. Furthermore, organic PCM A28 emerges as the optimal thermal buffer across multiple weighting approaches under the current testing conditions. Spearman correlation analysis reveals strong rank alignment (>0.9) among the MEREC, AHP, and SWARA weighting methods. Ultimately, the proposed 3D model offers an agile and robust alternative for performing large-scale optimisation of hybrid BTMS.