Fault pattern recognition of Power-Shift Steering Transmission based on support vector clustering

Fault pattern recognition is an important work in condition monitoring of Power-Shift Steering Transmission (PSST). Spectrometric oil analysis technology is a common and useful method to study the state of PSST. But, how to find the implicit information in data and classify the running state is a difficult work. In order to solve this problem, a support vector clustering(SVC) method is applied. The building process of this method is made. Four modes of PSST is made. And to get pattern information in data, three parameters of feature information are put forward. The influence of SVC model parameters for clustering regions is analyzed and optimal parameters are determined. On the basis of feature information extracting of spectrometric oil analysis data, fault pattern recognition is made with SVC model. The method has been proved that it has better accuracy in fault pattern recognition of PSST.