Search

Rolling element bearings are at the heart of most rotating machines and they bear the function of connectivity between the rotor and stator. It is important to distinguish the incipient fault before the bearing step into serious failure. The Slip Matrix (SM) construction method based on Singular Value Decomposition (SVD) is proposed in this paper. The SM based fault feature extraction and impulses intelligent detection methods are introduced as the key steps for rolling bearing fault diagnosis. The numerical simulation of rolling bearing fault signal is adopted which shows that the proposed method is good at fault impulses detection in strong background noise environment. The rolling element bearing accelerated life test is performed for the acquisition of experimental data of rolling bearing fault. With the rolling bearing running from normal state to failure, the initial fault signal part can be picked out from the whole life vibration data of the rolling bearing. The vibration signal is close to the nature fault signal which is acquired from a rolling bearing applied in industrial field. The analysis result shows that the proposed method has an excellent performance in rolling bearing fault detection. [ABSTRACT FROM AUTHOR]

• Novel CRB vibration model considering contact stiffness and natural frequency characteristics. • The outer raceway spalling fault with varying sizes is simulated in this paper, considering spalling defect extents and load distribution. • This paper showed that the effects of contact stiffness variation on natural frequency characteristics by the statistical distribution of fault feature signals. The modeling of early spalling defects in cylindrical roller bearings (CRB) is a challenging problem in fault diagnosis. Due to the absence of prominent "drop-in" excitation, accurately describing the signal characteristics of the early failure mechanism model for CRBs becomes difficult. In this paper, the failure of cylindrical roller bearings due to spalling caused by fatigue is a frequent occurrence. The previous defect models have been centered on sharp-edge impact between the roller and spalling defects. However, they have not provided an adequate depiction of the variation in contact stiffness caused by roller-fault interaction. In this paper, a novel CRB vibration model that takes into account contact stiffness and natural frequency characteristics is proposed. The model considers both the load redistribution and contact stiffness variation caused by spalling defects. This new model provides a spring-damper system based on contact stiffness between the roller and raceway, enabling an estimation of the CRB system's natural frequency characteristics. The outer raceway spalling fault with varying sizes is simulated in this paper, taking into account spalling defect extents and load distribution. The vibration response of the fault signal model is investigated. The simulation results show that the load and contact stiffness will change as the roller passes through the defect zone. Furthermore, the statistical distribution of fault feature signals demonstrates the effects of contact stiffness variation on natural frequency characteristics. The experiments given in this paper have verified the simulation results of the proposed signal model. [ABSTRACT FROM AUTHOR]

Vibration signals are widely used for bearing fault detection and diagnosis. When signals are acquired in the field, usually, the faulty periodic signal is weak and is concealed by noise. Various de-noising methods have been developed to extract the target signal from the raw signal. Stochastic resonance (SR) is a technique that changed the traditional denoising process, in which the weak periodic fault signal can be identified by adding an expression, the potential, to the raw signal and solving a differential equation problem. However, current SR methods have some deficiencies such us limited filtering performance, low frequency input signal and sequential search for optimum parameters. Consequently, in this study, we explore the application of SR based on the FitzHug-Nagumo (FHN) potential in rolling bearing vibration signals. Besides, we improve the search of the SR optimum parameters by the use of particle swarm optimization (PSO). The effectiveness of the proposed method is verifie d by using both simulated and real bearing data sets. [ABSTRACT FROM AUTHOR]

Gearboxes are at the heart of most rotating machines and they are considered as one of the main sources of vibration. As a key element in rotating machines, it is important to extract the gearbox vibration part from the mechanical system signal to assess the health state of the gearbox. In general, the gear meshing frequency part contains rich information which reflects its health state. In this paper, the authors find that in some cases when gearboxes work under heavy load, the meshing frequency part cannot be detected easily from the frequency spectrum. We think that the meshing frequency part may be modulated to the higher frequency band as the meshing impacts. To prove this point, a Multi-Input Single-Output (MISO) model is proposed to identify the local resonance excited by gear meshing impacts. In our method, a Meshing Impact Energy Distribution (MIED) graph is obtained through iteration to determine the demodulated frequency band. Experimental vibration data acquired form a healthy gearbox are illustrated to validate the performance of the proposed method. The forklift is taken as an example in the paper. Forklifts usually work in heavy load condition, and the gearbox vibration of a forklift is generally mixed with the ignition impacts of the diesel engine. The experiment result shows that for forklift machine system working in heavy load condition, its meshing frequency part of gearbox vibration signal is modulated into high frequency band. Further, the result shows that the proposed method is good at extracting the meshing frequency component when it is modulated into high frequency band especially when it is in heavy load condition. [ABSTRACT FROM AUTHOR]