A rotor fault diagnosis method based on BP-Adaboost weighted by non-fuzzy solution coefficients

To ensure the long-term safe and reliable operation of mechanical equipment, it is necessary to carry out fault diagnosis key components in system. A rotor method based on non-fuzzy solution weighted BP-AdaBoost (NFSW-BP-AdaBoost) proposed this paper. First, vibration signal transformed into a high-dimensional feature set. Then, VC value used obtain sequence. Finally, NFSW-BP-AdaBoost diagnose type degree fault. The core article analyze fuzzy combined binary classifier applied multi-classification problems decision-making, put forward coefficient weigh conclusion multi-classifier, highlighting sensitivity solutions each multi-classifier. experimental analysis results show that recognition rate classification for types can reach more than 97%.