Multiwavelet transform and its applications in mechanical fault diagnosis - A review

Mechanical fault diagnosis is important to reduce unscheduled machine downtime and avoid catastrophic accidents. It is significant to extract incipient fault and compound fault features as early as possible, which is a complex and challenging task that requests advanced analytical methods with high reliability, high accuracy and high efficiency. Compound fault features are mutually coupled in dynamic signals from the complex system. Weak features of incipient faults are always submersed in background noises. Multiwavelet transform is a remarkable development of wavelet transform, which uses vector scaling functions and wavelet functions. Multiwavelets possess the property of orthogonality, symmetry, compact support and high vanishing moments simultaneously. These advantages promote the development of multiwavelets and their applications in mechanical fault diagnosis in the past decades. This paper attempts to summarize the recent development of multiwavelet transform and its applications in mechanical fault diagnosis. First, the history of wavelets and multiwavelets is introduced. Second, the necessity and the overview of preprocessing methods for multiwavelets are summarized. Third, the advantages of multiwavelets and improvements of different generation multiwavelets are addressed. Fourth, different algorithms of these multiwavelet transforms and their flow charts are presented. Fifth, engineering applications of multiwavelets in mechanical fault diagnosis are investigated. This review also describes a simulation experiment and three application examples which provide a better understanding of different generation multiwavelets for compound fault detection. Finally, existent problems and prospects of further researches are discussed. It is expected that this review will construct an image of the contributions of different generation multiwavelets and link the current frontiers with engineering applications for readers interested in this field.