An Effective and Portable Electronic Stethoscope for Fault Diagnosis by Analysing Machine Running Sound Directly

Machine sound is a typical kind of non-stationary signal which carries information regarding the operating conditions of the machine. In the past, human ears have been used for detecting any abnormality occurring in a machine as this method is simple and fast. However, the audible range of human hearing is broadband and has a low signal-to-noise ratio, making the method inefficient. The invention of the Fast Fourier Transform (FFT) for vibration-based machine fault diagnosis has helped to increase the efficiency of diagnosis. However, FFT fails to detect the transitory characteristics of signals that are fault-related. Moreover, the cost of a FFT based analyser is expensive and the accessibility of a wired transducer is limited. Therefore, we are proposing the use of the hearing method again ? not with a pair of human ears, but with an electronic stethoscope. We use Continuous Wavelet Transforms (CWT) to remove the noise from raw machine running sound signals and to detect nonstationary impulses generated from the impacts of defective components. The method of Trajectory Parallel Measure (TPM) is then used for fault detection and classification. From the results of the tests on a number of similar types of gas engines, the concept of the electronic stethoscope has been found to be feasible and promising. This electronic stethoscope uses CWT and TPM to diagnose faults by analysing the machine running sound directly.