Temperature Field Reconstruction Algorithm Based On Radial Basis Function In Acoustic Tomography Research

Temperature field measurement technology can be divided into contact andnon-contact. Contact method must contact with the measured medium directly. So itrestricts the applications in the corrosive and high temperature. As a non-contacttemperature measurement technology, Acoustic temperature measurement technology usesthe data of sound wave time, backstepping the temperatures distribution of measuringregion. It has many advantages, such as non-contact, non-interference the measuringtemperature field, wide scope of application, measurement wide area and high accuracy.Temperature field measurement technology has been widely applied to various areas, suchas the grain storage temperature distribution monitoring, the deep-sea hydrothermaltemperature distribution testing, and the industrial boiler temperature field monitoring.This paper introduces the principle of the acoustic temperature tomography,temperature field typical reconstruction and interpolation algorithm, and reconstructionevaluating standards and influencing factors. On the basis of radial basis function,temperature field reconstruction algorithm of gauss radial basis function is proposed.Compare with the typical least squares, it without requiring pixel number less than acousticflight path. So the reconstruction algorithm is better than least squares for complextemperature field. Thirty-two and twenty sound transceivers are both used. Four typical3Dtemperature fields are reconstructed, and they are one-peak symmetry three-dimensionaltemperature field model, one-peak bias temperature field model, two-peak symmetrytemperature field model, four-peak symmetry temperature field model. The reconstructionresults and error analysis are also given. The results show that the algorithm has highreconstruction accuracy and strong anti-jamming capability. In addition, the influences ofthe parameters on reconstruction accuracy are analyzed. The paper provides the affect ofthe number and location of the transceivers, the effective path selection, the number ofpixels, the shape parameters of algorithm, the peak position of the one-peak temperature field model, and filtering parameter of the regularization on the reconstruction accuracy.Eventually, Kriging method is introduced to interpolation and extrapolation ofthree-dimensional temperature field reconstruction. Compare with accuracy of the commoncubic spline interpolation. As it turned out, the interpolation and extrapolation accuracy ofKriging method was higher than the cubic spline interpolation. The paper provides aneffective method to solve detail description and extrapolation of temperature field. SoKriging model can be introduced in Acoustic Computer Tomography.The three-dimensional temperature field reconstruction software is designed andintegrated based on Matlab. It can achieve the three-dimensional temperature fieldreconstruction, interpolation and extrapolation. And it is very convenient, shortcut,accuracy, and supporting the three-dimensional slice display.