Image Processing Of Intermediate Frequency Vacuum Arc And Investigations On The Combustion Rule Of The Anode

In order to balance the reliability and the cost of the project,the intermediate frequency is usually used as the switching frequency in the HVDC interruption.The change of the shape of the vacuum arc,especially the combustion state of the anode has a significant impact on the interruption of the circuit.The research on the basic characteristics of the arc shape and the changing rules of arc is not only helpful to the development of the arc control measures,but also to guide the design of the vacuum interrupter.Therefore,the anode of intermediate frequency vacuum arc is selected as the research object in the paper,and the digital image processing technology is used to pre-process and extract the feature of the arc image,also,the combustion law is analyzed.The paper provides the experimental data and image processing methods for the study of intermediate frequency vacuum arc.To obtain the intermediate frequency arc image,the experimental platform is built.It includes the main circuit,the vacuum generating and measuring device,and the high speed photographic arc acquisition system.The parameters are changed on the experimental platform,and the anode combustion images of vacuum arc are obtainedArc images often contain noise due to various factors.These noises can interfere with the subsequent analysis and even obscure the arc feature.In this paper,the BP neural network is applied as the noise detector to find the pulse noise of each frame,then the improved algorithm is used to filter out the noise points.The adopted algorithm is compared with the traditional median filtering algorithm,based on the objective evaluation index.The result shows that the adopted algorithm is superior to the traditional median filtering algorithm in the protection of image details.Based on the deficiency of the previous arc image processing and the gray distribution of the anode,a dual-threshold segmentation algorithm is adopted.The algorithm is compared with several traditional threshold segmentation methods,and the result shows that the dual-threshold segmentation algorithm is more suitable for the arc image,especially the anode combustion image segmentation.Finally,the technology of pseudo color image processing is introduced.The combustion process is divided according to the motion law of the arc in anode region by using the dual-threshold segmentation algorithm and the pseudo color image processing technology.The gray distribution of the anode with different frequency and amplitude is analyzed.The duration of anode high gray area is changed with the current frequency and the peak value.