Research on fault character recognition method of the reciprocating machinery based on SVM and local wave method

Based on the SVM principle and Local Wave Method, a new fault character recognition method for reciprocating machinery is presented. Vibration signal from reciprocating machinery analyzed by Local Wave Method, the parameters of the system character are obtained. The classification and recognition method for multi-faults based on SVM is applied and the whole diagnosis process is accomplished. Using this particular method, both the integrality of extracting information from the impacting signal of the reciprocating machinery and the recognition accuracy for the system status are confirmed, therefore the method is effective in practical application.