Model predictive control of portable electronic devices under skin temperature constraints.

Thermal management is becoming a major challenge for electronics, and a better temperature control algorithm that could maximize the system performance will play a greater role in fully utilizing the existing cooling capacity. Unfortunately, the simplest look-up table method is still widely used as the temperature control algorithm in current portable electronic devices, especially laptops, resulting in a significant performance loss of devices. In this paper, a general temperature control framework for a commercial laptop that considers the skin temperature constraints is proposed based on the model predictive control algorithm. In specific, a high-accuracy compact thermal model is first generated through the model order reduction method and validated by abundant experimental data. Then the proposed MPC is numerically evaluated in three test scenarios, covering different workloads and performance indexes. The results show that the proposed MPC outperforms the baseline look-up table method by achieving about 10–20% higher performance index in different test scenarios. The open-loop optimal control method is also considered to estimate the optimality of the proposed MPC. Moreover, a parametric study is conducted to analyze the influence of different control parameters, indicating broad prospects for the future application of the proposed MPC algorithm. • A general temperature control framework for commercial laptops is proposed via MPC algorithm. • The constraints of skin temperature are considered. • A compact thermal model of the commercial laptop is generated and validated by abundant experimental data. • The proposed MPC could obtain about 10–20% higher performance index compared to the baseline method of the look-up table. • The proposed MPC could achieve more than 90% of the optimal value calculated from the open-loop optimal control algorithm. [ABSTRACT FROM AUTHOR]