Pantograph Vibration Interference Signal Recognition Based on SVM Classification

In this article, a new pantograph vibration interference signal identification method based on SVM classification is proposed to deal with the abnormal signals caused by transient electromagnetic interference due to pantograph acceleration sensor's suffering from train over-voltage phase break. Firstly, the deviation position is determined by calculating the catenary impact index (CII) of pantograph vertical acceleration, then the acceleration attenuation coefficient of deviation at the said position is calculated, and finally the CII and acceleration attenuation coefficient is selected to form an identification feature vector, thus establishing an interference signal identification method based on support vector machine (SVM). The results show that such an identification structure has fast computing and high accuracy in judging whether the deviation is transient electromagnetic interference. Applying method to analyze the measured data of catenary inspection vehicles demonstrates its effectiveness.