J. Marine Sci. Appl. (2011) 10: 17-24 DOI: 10.1007/s11804-011-1036-7 Fault Detection and Diagnosis of a Gearbox in Marine Propulsion Systems Using Bispectrum Analysis and Artificial Neural Networks Zhixiong Li 1, 2 , Xinping Yan 1, 2* , Chengqing Yuan 1, 2 , Jiangbin Zhao 1, 2 and Zhongxiao Peng 3 1. Reliability Engineering Institute, School of Energy and Power Engineering, Wuhan University of Technology, Wuhan 430063, China 2. Key Lab. of Marine Power Eng. and Tech. (Ministry of Transport), Wuhan University of Technology, Wuhan 430063, China 3. School of Engineering and Physical Sciences, James Cook University, Townsville, Qld 4811, Australia Abstract: A marine propulsion system is a very complicated system composed of many mechanical components. As a result, the vibration signal of a gearbox in the system is strongly coupled with the vibration signatures of other components including a diesel engine and main shaft. It is therefore imperative to assess the coupling effect on diagnostic reliability in the process of gear fault diagnosis. For this reason, a fault detection and diagnosis method based on bispectrum analysis and artificial neural networks (ANNs) was proposed for the gearbox with consideration given to the impact of the other components in marine propulsion systems. To monitor the gear conditions, the bispectrum analysis was first employed to detect gear faults. The amplitude-frequency plots containing gear characteristic signals were then attained based on the bispectrum technique, which could be regarded as an index actualizing forepart gear faults diagnosis. Both the back propagation neural network (BPNN) and the radial-basis function neural network (RBFNN) were applied to identify the states of the gearbox. The numeric and experimental test results show the bispectral patterns of varying gear fault severities are different so that distinct fault features of the vibrant signal of a marine gearbox can be extracted effectively using the bispectrum, and the ANN classification method has achieved high detection accuracy. Hence, the proposed diagnostic techniques have the capability of diagnosing marine gear faults in the earlier phases, and thus have application importance. Keywords: marine propulsion system; fault diagnosis; vibration analysis; bispectrum; artificial neural networks Article ID: 1671-9433(2011)01-0017-08 1 Introduction 1 Gearboxes are extensively and widely used in various areas including aircraft, mining, manufacturing and agriculture, etc. The breakdowns of a gearbox mostly caused by the gear failures may result in significant economic losses (Fu et al., 2009). It is therefore crucial for engineers and researchers to monitor the gear conditions in a timely manner in order to prevent the malfunctions of the plants. In previous works (Yan et al., 2005; Yuan et al., 2005, 2007, 2008) we employed the oil monitoring techniques for condition monitoring of the marine propulsion systems. We found that there existed a strong coupling effect on different components in a marine propulsion system. However, the coupling characteristic has usually been underestimated and ignored in the gear fault diagnosis. As a consequence, the reliability and accuracy of the diagnosis results may suffer. Effective fault detection and diagnosis techniques need to be developed to deal with this case. Received date: 2010-03-26. Foundation item: Supported by the National Natural Sciences Foundation of China (No. 50975213 and No. 50705070), Doctoral Fund for the New Teachers of Ministry of Education of China (No. 20070497029) and the Program of Introducing Talents of Discipline to Universities (No. B08031). *Corresponding author Email: lzx___520@163.com © Harbin Engineering University and Springer-Verlag Berlin Heidelberg 2011 Among available techniques for gear fault diagnosis vibration analysis is becoming increasingly popular. Many signal processing techniques have been developed and applied for machine diagnosis in this area. They include the conventional techniques, such as the spectral analysis (Cheng et al., 2010), time domain averaging and time-series analysis (Zhou et al., 2007) as well as some new techniques, such as Wigner-Ville distributions (WVD) (Baydar and Ball, 2001), empirical mode decomposition (EMD) (Cheng et al., 2008), and wavelet transform (WT) (Lin and Zuo, 2003), etc. Among them, high-order statistical analysis, in particular, the bispectrum has a strong ability to process non-Gaussian signal and identify nonlinear system’s failure. Its excellent performance on fault detection is due to the fact that the third-order cumulant and bispectrum of zero-mean Gaussian noise are always equal to zero. Liu et al. (2008) used the bispectrum and 1(1/2)-dimension spectrum to diagnose a gearbox with various pitting faults. They concluded that bispectral analysis was very sensitive to the gear faults. Fu et al. (2009) applied the bispectrum technique to the gear wear fault detection and identification and Zhang (2006) to the gear crack fault diagnosis. Their studies showed that the bispectrum could be used as a diagnostic tool for gear faults detection. Hence, based on the existing studies, it is possible