Research On Intelligence Method For Improving Spatial Resolution Of Electrical Tomography Impedance Technique

Electrical impedance tomography(EIT)was first appeared in the 1980 s.This technology is based on electromagnetic field theory and has many advantages such as visualization,non-radiation,and low pollution,low cost and fast response.Now,it has attractive prospects in many fields including industrial measurement,medical monitoring and so on.However,due to the development level of measurement devices and the limitations of image reconstruction process,the spatial resolution of EIT is not very high.As far as we can see,EIT can only be used as an auxiliary measurement means,but not play its advantages to replace the mainstream measurement equipment in many engineering fields.Because of the inherent ill-posedness,morbidity and soft field effect of EIT technology,it is very difficult to obtain higher spatial resolution by traditional methods.The purpose of this paper is to further improve the spatial resolution of EIT technology by using the artificial intelligence and in-depth study of image reconstruction.The main research work of this paper is as follows:1)Convolutional neural network is used to train the sensitivity matrix parameters in EIT technology.It is concluded that the measured values of EIT system can be regarded as two-dimensional images to extract features from the principle of measurement of EIT system.So we can select the structure and parameters of corresponding convolution neural network.After that,we can select quantities of typical and representative input output samples to train the convolution neural network,and after training,we can use the test sample to validate the accuracy and generalization ability of convolutional neural network.2)Using fuzzy operators to improve the traditional EIT image reconstruction algorithm.First,we analyze the motivation and rationality of using fuzzy operators based on the undetermined and ill-conditioned nature of inverse problem of EIT.Then,we propose a method of selecting appropriate fuzzy operator according to the output optimization objective of EIT system.Finally,we compared the visual effect and error index of the output obtained by the algorithm before and after using fuzzy operators through simulation experiments.3)A new method for calculating phase separation holdup of two-phase flow using fast fuzzy clustering algorithm is proposed.First,we analyzed the limitations of the existing methods for solving the fractional phase holdup and compared the advantages and disadvantages of the fast fuzzy clustering algorithm and other clustering algorithms.Then,we can divide the reconstructed images of EIT into three parts using fast fuzzy clustering algorithm,and after that we can achieve the calculation of the fraction holdup.Finally,the algorithm is applied to the actual engineering data of the dredgers in the Tianjin municipal navigation bureau,and we compared the traditional algorithm and fast fuzzy clustering algorithm for solving the phase fraction and made the interactive validation.4)The application of sequential optimization algorithm in EIT technology is preliminarily studied.Firstly,we introduced the sequential optimization algorithm briefly and discussed that the sequential optimization algorithm without prior information cannot be used in EIT process.Then we explained the application ways of several prior information in EIT process.Finally,the possible application conditions of the sequential optimization algorithm are analyzed through simulation experiments.