Linear extrapolation method based on multiple equiproportional models for thermal performance prediction of ultra-large array

Abstract

Strong demand for developing the photothermal (PT) and electrothermal devices with ultra-large array is increasing. Thermal performance prediction is vital to optimize the key properties of the devices with ultra-large array. Finite element method (FEM) provides a powerful numerical approach for solving complex thermophysics issues. However, for calculating the performance of devices with ultra-large array, it is very memory-consuming and time-consuming to build an equal scale three-dimensional (3D) FEM model. For an ultra-large periodic array irradiated with a local heating source, the use of periodic boundary conditions could lead to considerable errors. To solve this problem, a linear extrapolation method based on multiple equiproportional models (LEM-MEM) is proposed in this paper. The proposed method builds several reduced-size FEM models to carry out simulation and extrapolation, which avoids dealing with the ultra-large arrays directly and greatly reduces the computation consumption. To verify the accuracy of LEM-MEM, a PT transducer with beyond 4000 × 4000 pixels is proposed, fabricated, tested and compared with the prediction results. Four different pixel patterns are designed and fabricated to test their steady thermal properties. The experimental results demonstrate that LEM-MEM has great predictability, and the maximum percentage error of average temperature is within 5.22% in four different pixel patterns. In addition, the measured response time of the proposed PT transducer is within 2 ms. The proposed LEM-MEM not only provides design guidance for optimizing PT transducers, but is also very useful for other thermal engineering problems in ultra-large array that requires facile and efficient prediction strategy.