Cnc Machine Tool Fault Signal Analysis And Feature Extraction

NC machine plays an important role in modernized production . If malfunction or accident happen in NC machine , it will bring the larger loss on factory. The reason is that the downtime will make the maintaining period longer , and the NC machine's price more expensive. The self-diagnosis in NC machine can diagnose the faults of joint parts with NC system . But the defect and forecast of faults in mechanism is scotoma . So on-line defect and early diagnosis can reduce fault probability.This paper brings forward the fault diagnosis and forecast system of NC machine ,and designs it's configuration and function . The paper diagnoses the cause made the fault occured , and draws a conclusion that the matter which caused mechanism's fault is mainly about the fault of gear and bearing . The paper creates the kinematics models of gear and bearing by the characteristic of impactive libration which been caused by flaws on gear and bearing . The paper uses Kalman filtering algorithm filtering noise and uses Laplace correlatied filtering algorithm extracting fault characteristic . The paper verified these two algorithm by simulation experiment. It diagnoses the NC machine's principal axis's libration , filters libration signal, and distills the fault's characteristic by Kalman filtering algorithm and Laplace correlatied filtering algorithm. The experiments show that these two algorithm are effective on signal filtering.