Abstract
High-speed train will produce pantograph arc during the driving process, which is harmful to pantograph-catenary system. In order to reduce pantograph-catenary system damage, a method based on Gray Wolf algorithm to optimize the binary Support Vector Machine classifier to identify pantograph arc is proposed. In this article, 5 groups of pantograph current experiments under different conditions are carried out, and the current data in the pantograph-catenary system under different conditions are measured. The current data state obtained from the pantograph experiments is divided into normal current state and arc current state. Select the mean value, variance, standard deviation, mean value of the first-order difference, and mean value of the second-order difference of the current data as the characteristic value of the pantograph current, and calculate the contribution rate of each characteristic value at the same time, then the current eigenvalue data with a high contribution rate is used as a training sample for learning and recognition through the classifier optimized by Gray Wolf algorithm. The experimental results show that the Gray Wolf optimization algorithm can quickly and accurately identify the pantograph arc, and the classification model obtained is more accurate than the commonly used optimization algorithms such as genetic algorithm and particle swarm. In addition, an engineering implementation of on-line identification of pantograph arc based on industrial computer is proposed.
Highlights
With the rapid development of science and technology, the speed of railway locomotives has been improved
Some scholars began to study the process of pantograph arc generation and to model the pantograph arc
Karakose et al [7], a scholar who classifies and recognizes electric arcs by extracting the pantograph current signals, proposed an arc detection method based on signal processing and S transformation
Summary
With the rapid development of science and technology, the speed of railway locomotives has been improved. Karakose proposed a method to analyze arc faults through actual video capture and image processing of the contact point between pantograph and catenary [4]. Karakose et al [7], a scholar who classifies and recognizes electric arcs by extracting the pantograph current signals, proposed an arc detection method based on signal processing and S transformation. As shown in the area circled, this module can set parameters, adjust the sliding speed, contact current and pressure of the device to simulate the moving state of the train and generate two states of normal current receiving and fault current receiving. The five eigenvalues of the mean, variance, standard deviation, first-order difference mean, and second-order difference mean are extracted from the current receiving waveform for each 100 sampling points to form a vector group matrix in the form of a vector.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
