作者: Giuseppe Quaremba , Vincenzo Niola , Antonio Donnarumma
DOI:
关键词:
摘要: In the present work a methodology useful for identification and classification of anomalies showed by an aleatory mechanical system was developed. order to perform study, more than 1000 tests, each with predefined characteristics goals, have been carried out means dynamical test-bed based on two circular-arc cam-follower mechanism. The acceleration follower applied torque were sampled electronically. signals obtained grouped into 4 main families and, family, groups according their features; signal processed applying Discrete Wavelet Transform (DWT). Therefore, identified through 10 energetic variables deriving from decomposition orthogonal components application DWT. Afterward, results classification, multivariate statistical analysis (i.e., discriminant analysis), compared ones fuzzy algorithm.