Coordinated Nonlinear Model Predictive Control for a distributed thermoelectric-based battery thermal management system

Abstract

Electric vehicles (EVs) rely on lithium-ion batteries that require sophisticated battery thermal management systems (BTMS) to maintain safety, performance, and longevity. A critical challenge is mitigating cell-to-cell temperature variations, which accelerate non-uniform ageing and degrade overall pack performance. Centralized thermal management controllers, while regulating average pack temperature, fail to address cell-to-cell temperature gradients. This paper proposes a novel Coordinated Nonlinear Model Predictive Control (CNLMPC) strategy for a BTMS with distributed cooling mechanism employing Thermoelectric Coolers (TECs). The CNLMPC architecture decouples the control problem into two layers: a primary NLMPC optimises average pack temperature and optimal energy use, while a secondary NLMPC dynamically adjusts the driving current of each individual TEC in real-time. The effectiveness of the proposed BTMS strategy is evaluated through simulations using a high-fidelity 180-cell electro-thermal model under aggressive WMTC and WLTC drive cycles. The results show the CNLMPC maintains the cell-to-cell temperature differential near the 5°C industry target. Compared to a non-adaptive fixed-gain controller, the CNLMPC achieves this thermal uniformity while reducing BTMS energy consumption by over 10%. This demonstrates an optimal and adaptive balance between thermal safety and energy efficiency.