Opening Velocity Detection For Vacuum Circuit Breaker Based On Image Sub-pixel

Opening velocity of vacuum circuit breakers has great influence on the vacuum arc form and performance.In recent years,with the infiltration in high-voltage and high-capacity areas of vacuum circuit breakers,a technical problem in urgent need to solve is how to choose operating mechanism with right velocity parameters.Operation mechanism has short running time,high acceleration,short stroke and other characteristics,it's difficult to diagnose the operating parameters.The traditional ways of opening velocity detection for vacuum circuit breakers are mainly grating sensors and external capacitance analysis.Application of these methods to measure the velocity can be realized,but the accuracy isn't high,the influence of electrode moving velocity on arc form transformation couldn't be analyzed either.In this paper,opening velocity detection of vacuum circuit breakers is taken as research target,and the arc edge is detected in image sequences of moving electrode by using digital image processing technology,recognizing electrode position and calculating the displacement of moving electrode.According to displacement of moving electrode and time interval between inter-frame images,moving electrode opening velocity can be calculated.Features of electrode moving sequence images collected by experiment are bright,dramatic transformation of arc form,and so much quantization noise.According to the feature of arc edge coinciding with electrode edge,arc edge is detected by using polynomial interpolating method.Then interference edge points are eliminated by using Hough transform and lines have been fitted by using least square for remaining edge points,finally the actual edge location of arc can be calculated.The electrode target detection problem is avoid,and transformed into edge detection problem.Finally,the instantaneous velocity is replaced by inter-frame average velocity v=s/t.Detection results show that the algorithm detects in high speed and can be feasible.In this paper,according to arc edge approximating ideal step edge model,gray moment method is adopted and 7×7 convolution mask is made to do convolution integral calculation with arc images.calculating normalized parameters of arc edge and obtaining sub-pixel edge points of it.Then interference edge points are eliminated by using Hough transform and finally linear fitting method is adopted to get real edge position of arc,obtaining displacement of moving electrode.Similarly,the instantaneous velocity is replaced by inter-frame average velocity to obtain opening velocity of vacuum circuit breakers.Test results show that this algorithm locate the arc edge with an error of 0.01?0.35 pixels.Meanwhile,the opening velocity curve fitted is smoother and the fitting error is smaller,which is more coincident with the true velocity curve.Compared with each other,the positioning accuracy of the sub-pixel edge detection method based on gray moment is higher than that of the polynomial interpolating method.Therefore the opening velocity curve fitted by the former method is more coincident with the true velocity curve.In conclusion,in this paper,the opening velocity of moving sequence images of electrode is deeply researched and analyzed and two classical sub-pixel edge detection methods are applied to detect opening velocity of electrode.Finally a method with higher precision is found out.Results demonstrate that the gray moment method has a certain degree of accuracy and stability,it provides a new method for detecting opening velocity and its detection accuracy is greatly improved.