Flow Pattern Recognition And Three Dimensional Image Reconstruction Of Electrical Capacitance Tomography Based On Voids Model

Three dimensional image reconstruction of electrical capacitance tomography(3DIR-ECT)technology is one of the process tomography techniques,3DIR-ECT is used to constructe three-dimensional distributed images of the multi-phase flow medium in the measured region.It usually uses the capacitance measurement value as the projection data,the data is measured by a array of the capacitance sensor.Because it can detect a series of the key parameters of the multiphase flow,such as the distribution,the concentration and the mass flow in the measured area in real time,3DIR-ECT technology has gradually become the main research and development direction in the field of the multiphase flow detection.Because of the complexity of the multiphase flow state,and the lack of a lot of theoretical calculation data,which is needed for the iterative reconstruction algorithm,the accuracy of image reconstruction is low.Therefore,this paper used the finite element method,singular value decomposition method,electrostatic energy storage method and the other methods to analyze the principle of 3DIR-ECT technology.And combine with the parallel technology to establish 3DIR-ECT simulation program,achieve the automatic and continuous calculation of the theoretical data,reduce the amount of work required for repeated modeling and theoretical data calculation time.Aiming at the problem that the accuracy of image reconstruction is not high when the flow medium is in the bubble flow,wave flow and other flow patterns with the drastic change of flow state.In this paper,the sensitivity equation was used as the evaluation function,and the double chain quantum genetic algorithm was used to optimize the back propagation neural network to reconstruct the three-dimensional image of the measured region.The analysis theory data showed that the capacitance measurement data is related to the multiphase flow pattern,so this paper divided the capacitance measurement data according to the flow pattern,and selected the different image reconstruction methods for the different flow patterns.First,the calculation method of voids model was given by using the principle of the multi-phase flow energy conservation and 3DIR-ECT technology.This method used the average noise energy of the air pipe to reduce the interference of the environment to the measurement system,used the wavelet packet energy characteristic instead of the system energy to reduce the calculating time,and used the quantum algorithm to divide the flow status.Then the convolution neural network structure was improved by using the voids model,the theoretical data was transferred to the corresponding voids convolution neural network,the three-dimensional distribution image of the multiphase flow medium was reconstructed,the reconstruction accuracy was improved.And the time,which was reconstructed the three-dimensional image,was reduced.