Simulation Research on Reconstruction Algorithm of Complex Temperature Field Based on RBF Neural Network

Reconstruction algorithm is essential to reconstruct temperature field image in boiler temperature field survey. This article puts forward a new algorithm based on RBF neural network. The algorithm firstly does discrete cosine inverse transform to temperature distribution T(x,y) of ´typical section´ in the boiler to construct mapping relationship between low step item coefficients vector and average temperature vector of gas medium on sound wave flight path, and then implements the mapping using RBF neural network because it has strong functional approach ability. Through training the network with orthogonal least square method, the temperature field can be reconstructed. Simulation reconstruction experiments are done to different temperature distributions using this algorithm and the reconstruction results are compared with those using least square method. Simulation results show that the algorithm has high reconstruction precision, quick speed and good anti-interference ability