基于均匀设计和最小二乘支持向量机的电容层析成像传感器优化设计

提出一种采用均匀设计与最小二乘支持向量机相结合的电容层析成像传感器结构参数优化方法．该方法以敏感场均匀度为设计目标，采用均匀设计安排试验，试验因素包括ECT传感器的4个重要参数即：极板张角，绝缘管道材料相对介电常数，管道壁厚，以及屏蔽罩与电容极板间的间距．运用最小二乘支持向量机对试验结果进行回归分析，并用因素轮换法进行寻优计算，从而得出优化的传感器结构参数．结果表明：经过优化的电容成像系统具有较好的成像效果．该方法试验次数少，具有较强的实用性．

The Optimizing Design of Capacitance Tomography Sensors Based on Uniform Design and LS-SVM

Uniform design and Least Square Support Vector Machines(LS-SVM) were applied to optimize design parameters of capacitance tomography sensors in this paper.Experiments were performed with uniform design:the spatial sensitivity distribution was taken as the performance parameter,and four key parameters:the radian of electrodes,the permittivity of insulating pipe material,the thickness of pipe wall and the distance between earthed screen and electrodes were used as experimental factors.LS-SVM was used to establish the regression model based on the experimental results.Then,a set of optimum parameters was obtained from the regression model by the factor rotation method.Simulation results indicate that the presented method in this paper can find a set of optimal parameters for capacitance tomography sensors efficiently and effectively.