Study on three-dimensional effect correction of surface heat flux estimation based on neural network

Based on the existing sequential function method for one-dimensional and two-dimensional surface heat flux identification, this paper presents a combination of neural network and sequential function method considering the real-time difficulty of three-dimensional identification. In this paper, a neural network is designed to correct the three-dimensional effects based on the results of one-dimensional identification, which can get accurate real-time results of peak heat flux. Moreover, PSO algorithm is introduced to optimize the initial weights and thresholds of the neural network in order to obtain a better model. It can be seen from the test results that the method presented for the peak heat flux is accurate with errors less than 4%. The method avoids time complexity of three-dimensional identification and has good noise immunity and stability at the same time.