Study On Optimization Of 3D Electrode Of Sensor And Image Reconstruction Algorithm For An Oil-air Lubrication ECT System

Sensor design for oil-air lubrication ECT system is a primary step to monitor online parameters of oil-air two-phase flow in oil-air pipelines.The characteristic parameters,excitation detection mode and image reconstruction algorithm of the 3D electrode of the sensor for an oil-air lubrication ECT system are studied in-depth in this thesis,aiming at characteristics,such as small pipe diameter,thin oil film,less oil and more air flow and limited numbers of plate,of the oil-air lubrication ECT system.The main research work and conclusions are as follows,（1）An optimization model for a 3D structure of the ECT sensor is established,based on characteristics,mentioned above,of the oil-air lubrication system.And the capacitance values and the image correlation coefficients of the sensor are calculated taking the annular flow of oil-air in the pipeline as a typical flow pattern.Image accuracy coefficient（IAC）is defined,and relationship between the IAC and electrode’s structure parameters is studied.The initial value ranges before optimization of the 3D structural parameters,such as the electrode spacing,the single-layer electrode length and the electrode thickness,are obtained,after analyzing the capacitance value and IAC of the sensor in the full field condition.And then,the three structural parameters,mentioned above,of sensor are optimized taking the sensor performance evaluation function(_（80）1))as the target function,based on the method of the orthogonal experiment,employing the COMSOL software and the Design-Expert software.The research results show that,the image relative error is reduced by 39.69%,the image correlation coefficient is increased by 39.43%,and the image accuracy coefficient is increased by131.20%,comparing data from the sensor been optimized with that from the benchmark sensor for an oil-air lubrication ECT system.（2）Two kinds of excitation detection mode,including the horizontal and the longitudinal mode,of the sensor combining electrode are studied employing the MATLAB software and the COMSOL software,aiming at the sensor with optimized3D electrod.And then,the two modes are compared with the traditional single-electrode excitation detection mode from three aspects,such as value of capacitance,the distribution of sensitive field and quality of the reconstructing image.The research results show that,1)the error of the excitation detection mode for horizontal combining electrode is decreased by 15.61%,after comparing with the single electrode excitation detection mode;The error of excitation detection mode for the longitudinal combining electrode is decreased by 37.57%,after comparing with the single electrode excitation detection mode.2)Excitation detection mode of the longitudinal combining electrode is better than excitation detection mode of the horizontal combining electrode or single electrode excitation detection mode,based on 3D structure sensor for an oil-air lubrication ECT system.The effect of image reconstruction is better,and the distribution of sensitive field is more uniform.（3）The capacitance value of the sensor is calculated,aiming at the sensor of 3D structure ECT after the design of electrode optimization and excitation detection mode of longitudinal combining electrode in this thesis,based on the electromagnetic field theory.The mapping model of DBN is established,corresponding to the capacitance value and the image gray value;The DBN algorithm and the improving DBN algorithm are applied to the imaging process of pipeline section,of the two-phase flow of the 3D structure sensor for an oil-air lubrication ECT system.And the algorithm is evaluated,finally,taking image relative error（IME）and image correlation coefficient（CORR）as evaluation indicators.The research results show that,1)the DBN algorithm reduces the relative error of the image effectively,and the image correlation coefficient is improved effectively,compared with the traditional algorithm.2)The improving DBN algorithm improves significantly,such as the real-time performance and accuracy of the image reconstruction in an oil-air lubrication ECT system.