Sintering fan faults diagnosis based on wavelet packet analysis and fuzzy recognition

According to the nine fault signal collected from 4th sintering fan in Shaogang Steel Group, energy analysis of the reconstructed signal by wavelet packet analysis was used. Extract feature vectors and a typical fault table that reflect running of fan. The results of analysis indicate that detecting signal with db10 wavelet six layers wavelet packet decomposition can obtain fan fault vector, applying the feature vector composition method and fuzzy recognition can diagnose the fan faults; 6-layer high-frequency decomposition can reflect the nature of the fan failure and 0.5, 0ne~five frequency multiplication composition was in 1~7 frequency range, it provides comprehensive information for fan fault diagnosis. The actual diagnosis result shows that using the feature vector typical characteristic fault, imbalance fault reaches to 0.951 through the fuzzy pattern recognition, showing that there exists fan´s imbalance fault and this analysis method is more accurate fault diagnosis.