Electrode Positioning And Plasma Temperature Diagnosis In Arc Stage Of Vacuum Switch

Research on visual diagnostic technology for vacuum switches is one of the most important tasks in the field of switchgear.Therefore,on the basis of previous studies,a visual detection technology based on the backpropagation gradient descent algorithm for locating the electrode position of the vacuum switch in arc firing stage is proposed,and deep studies of the vacuum switch arc plasma temperature diagnosis technology based on Radon inverse transformation and colorimetric temperature measurement method are put in this paper,with the arc images as the research object.Finally,Tensorflow,Open CV,C/C++,Python and other tools are used to program experiments to study visual diagnostic technology of electrode positioning and arc plasma temperature diagnosis in arc firing stage,in the Linux.First,two different CMOS high-speed cameras sampling systems for locating the electrode position and plasma temperature diagnosis are introduced respectively.Next,the noise sources and main noise types of CMOS image sensors are described,and the characteristics of common noise are discussed.Then the characteristics and research status of common filters are elaborated.Finally,the experimental models of gaussian noise and salt-and-pepper noise are established by C/C++ programming in the Linux,and the analysis of the filtering effects of mean filtering,median filtering,wavelet noise reduction,etc on common two noises are also realized with C/C++ programming.Which facilitates the extraction of arc image information in subsequent steps like electrode positioning analysis and plasma temperature diagnostic analysis.In order to locate the position of the electrode during the arc firing stage of vacuum switch,a hybrid algorithm that combines traditional algorithm and supervised backpropagation gradient descent algorithm used with small data sets of arc images is proposed in this paper for locating the electrode position.Firstly,the theoretical basis of target recognition and location of traditional integral value algorithm and supervised backpropagation gradient descent algorithm is elaborated,including the comparative theoretical analysis of common pooling methods in the model learning process.Finally,model training for locating the electrode position with small data sets of arc images is realized by programming combined with Tensor Flow,Opencv,C++,Pyhton and so on,in the Linux,and error analysis is performed on the predicted value of electrode positioning,which provides a new method for vacuum switching contact positioningBased on the previous research on colorimetric thermometry,further research on the method of back projection reconstruction algorithm for arc plasma temperature diagnosis is made in this paper.Firstly,the characteristics of interpolation methods and filters commonly used in the back projection reconstruction algorithm are discussed.Then quality analysis of images reconstructed respectively by different equal angles,different interpolation methods,direct back projection and R＿L back projection is realized by programming experiments with C/C++ in the Linux.Finally,quality analysis of arc plasma temperature reconstructed respectively by direct back projection and R＿L back projection is realized,and visualization results of the reconstructed arc plasma temperature fields are given too,which is of great significance for the study of vacuum switching performance.