Adaptive Algorithm Based on Wavelet and SVM for Turbopump Fault Detection

Abstract

An adaptive fault detection algorithm based on wavelet and SVM (Support Vector Machine) is proposed for LRE(Liquid Rocket Engine) turbopump real-time fault detection. The algorithm firstly divides the historical signals into some segments by reasonable step length. Then for each segment it gets M-layer detail signals through Daubechies wavelet transform. Thirdly it divides every layer into K average segments and calculates there RMS values, gets M RMS sequences of detail signals. After that it constructs M-dimensional RMS vector as fault feature by extracting RMS values at the same position in every RMS sequence, and extracts all the fault feature vectors of historical signal to construct SVM training sample set and then obtains SVM classifier. At last the classifier will be real-time updated by a reasonable method in the testing process to improve the classification accuracy. To validate the algorithm, a track of the vibration acceleration signal of a certain type of turbopump was chosen as the test object. The test results showed that the algorithm met its demands of accuracy and real-time performance.