In the rolling bearing fault diagnosis, the vibration signal of single sensor is usually nonstationary and noisy, which contains very little useful information, and impacts the accuracy of fault diagnosis. In order to solve the problem, this paper presents a novel fault diagnosis method using multivibration signals and deep belief network (DBN). By utilizing the DBN's learning ability, the proposed method can adaptively fuse multifeature data and identify various bearing faults. Firstly, multiple vibration signals are acquainted from various fault bearings. Secondly, some time-domain characteristics are extracted from original signals of each individual sensor. Finally, the features data of all sensors are put into the DBN and generate an appropriate classifier to complete fault diagnosis. In order to demonstrate the effectiveness of multivibration signals, experiments are carried out on the individual sensor with the same conditions and procedure. At the same time, the method is compared with SVM, KNN, and BPNN methods. The results show that the DBN-based method is able to not only adaptively fuse multisensor data, but also obtain higher identification accuracy than other methods.
A spiral-bevel gear is a basic transmission component and is widely used in mechanical equipment; thus, it is important to monitor and diagnose its running state to ensure safe operation of the entire equipment setup. The vibration signals of spiral-bevel gears are typically quite complicated, as they present both nonlinear and nonstationary characteristics and are interfered with by strong noise. The complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) method has been proven to be an effective method for analyzing this kind of signal. However, the fault feature information after CEEMDAN is not obvious and needs to be quantified. Permutation entropy can be used to quantify the randomness, complexity, and mutation of vibration time-series signals. This paper proposes to take the CEEMDAN-based permutation entropy as the sensitive feature for spiral-bevel gear fault identification. First, the raw vibration signal is decomposed by the CEEMDAN method to obtain a series of intrinsic modal functions (IMFs). The IMFs which included greater amounts fault information are selected as the optimal IMFs based on the correlation coefficient. Next, the permutation entropy values of the optimal IMFs are calculated. In order to obtain accurate permutation entropy values, the two key parameters, namely, embedding dimension and delay time, are optimized by using the overlapping parameter method. In order to assess the sensibility of the permutation entropy features, the support vector machine (SVM) is used as the classifier for fault mode identification, and the diagnostic accuracy can verify its sensibility. The permutation entropy of CEEMDAN-based/EEMD-based/EMD-based features, combined with SVM, is applied to identify three different fault modes of spiral-bevel gears. Their respective diagnostic accuracies are 100%, 88.33%, and 83.33%, which indicate that the CEEMDAN-based permutation entropy is the most sensitive feature for the fault identification of spiral-bevel gears.
Purpose The aim of this paper is to solve the problem of low accuracy of traditional fatigue crack growth (FCG) prediction methods. Design/methodology/approach The GMSVR model was proposed by combining the grey modeling (GM) and the support vector regression (SVR). Meanwhile, the GMSVR model parameter optimal selection method based on the artificial bee colony (ABC) algorithm was presented. The FCG prediction of 7075 aluminum alloy under different conditions were taken as the study objects, and the performance of the genetic algorithm, the particle swarm optimization algorithm, the n-fold cross validation and the ABC algorithm were compared and analyzed. Findings The results show that the speed of the ABC algorithm is the fastest and the accuracy of the ABC algorithm is the highest too. The prediction performances of the GM (1, 1) model, the SVR model and the GMSVR model were compared, the results show that the GMSVR model has the best prediction ability, it can improve the FCG prediction accuracy of 7075 aluminum alloy greatly. Originality/value A new prediction model is proposed for FCG combined the non-equidistant grey model and the SVR model. Aiming at the problem of the model parameters are difficult to select, the GMSVR model parameter optimization method based on the ABC algorithm was presented. the results show that the GMSVR model has better prediction ability, which increase the FCG prediction accuracy of 7075 aluminum alloy greatly.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.