Bearing fault diagnosis in spindles using vibration and acoustic emission

Abstract

In modern automated manufacturing processes, machinery has become more flexible and automatic but also more susceptible to conditions in different parts like spindles. The main conditions affecting them are focused in the shaft and bearings. To successfully detect and identify each condition, methodologies based on new transforms and sensors are used in controlled environments, in order to recognize the conditions effects in vibration and sound. To detect machining conditions there are two important factors: (1) The feature extraction method and (2) the type of sensor used. In feature extraction methods reviewed, Cepstrum Pre Whitening (CPW) is remarkable useful for vibration. Its main feature is suppress the shaft speed waveform of the motor in rotational machine systems and help to detect bearing faults. The main sensor used is the accelerometers to acquire vibration, but recently acoustic emission (AE) acquired from transducers are studied to improve the diagnosis along with accelerometers. In this study, an experimental system was built to acquire vibration and AE signals from faulted bearings and a methodology based on CPW, tested for vibration signals, was applied for both type of signals to compare and enhance results on machining condition monitoring. A methodology proposed using CPW, envelope spectrum, trend removal, compression and RMS limit filters (the last two just for AE) was applied to 9 vibration and 9 AE signals taken from the experimental system with the purpose of diagnosing bearing faults in the inner race (IR), outer race(OR)androllingelement(RE)inlowfrequenciesforbothsignalsandhighfrequenciesinAE. For the 18 analyzed signals, in 5 the identification of fault components were easily made, in 12 signals the fault identification was possible; but there were peaks with similar amplitudes of the fault components and in 1 signal the identification of fault components was unsatisfactory because there was no peak that matches the bearing fault frequencies. The comparison between vibration and AE showed that in 6 from 9 tests, vibration have a better result diagnosing bearing faults than AE, specifically in the IR and RE, for the remaining 3 tests that correspond to OR, AE have a better result than vibration. Finally, the high frequencies in AE revealed that just RE faults had high frequency components in one of the three analyzed tests that can be related to remarkable faults in the ball.