Analysis of High Frequency Component of Ultrasound Signal for Fault Evaluation of Ball Bearing in Induction Machine

Abstract

Abstract Ball bearings fault detection with the aid of ultrasound technique is of great importance. Hence, this paper evaluates signal processing of ultrasounds sent by faults and addresses relevant challenges. In this regard, the waveforms of ultrasound signals were analyzed for different types of fault and considering specification of ultrasound channels to enhance the perception of ultrasound signal processing. For this purpose, different conditions have been considered such as waveform fading and the likelihood of frequency fading at lower signal-to-noise ratios (SNRs). Furthermore, the sensitivity of ultrasonic sensors to fault detection at higher frequencies and the probability of the modulation have been assessed, and modeling results have been presented. In addition to modeling, fault detection in a ball bearing of an electromotor has been studied through an industrial-empirical case study. Therefore, the bearing signals have been processed for fault-free, initial fault, and advanced fault conditions. The studied fault emerged in several months (without hand intervention) and were then intensified over time. According to results, the envelope technique is capable of extracting fault frequencies of ball bearings. Also, high-pass filters have been employed to demonstrate signal shape mode and impulse generation in detail. Moreover, it has been shown that the growth of initial faults can generate new frequencies and fault frequency harmonics other than the fault frequencies of ball bearings. Also, important tips on the ultrasound fault detection of electromotor were provided with a potential effect on the entire fault detection process. These tips came from the author’s experiences obtained by monitoring rotary machines. In which, they can serve as a beneficial tool for monitoring of such machines in industrial sites.