Bearing Fault Diagnosis Based on Statistical Locally Linear Embedding

Abstract

Fault diagnosis is essentially a kind of pattern recognition. The measured signal samples usually distribute on nonlinear low-dimensional manifolds embedded in the high-dimensional signal space, so how to implement feature extraction, dimensionality reduction and improve recognition performance is a crucial task. In this paper a novel machinery fault diagnosis approach based on a statistical locally linear embedding (S-LLE) algorithm which is an extension of LLE by exploiting the fault class label information is proposed. The fault diagnosis approach first extracts the intrinsic manifold features from the high-dimensional feature vectors which are obtained from vibration signals that feature extraction by time-domain, frequency-domain and empirical mode decomposition (EMD), and then translates the complex mode space into a salient low-dimensional feature space by the manifold learning algorithm S-LLE, which outperforms other feature reduction methods such as PCA, LDA and LLE. Finally in the feature reduction space pattern classification and fault diagnosis by classifier are carried out easily and rapidly. Rolling bearing fault signals are used to validate the proposed fault diagnosis approach. The results indicate that the proposed approach obviously improves the classification performance of fault pattern recognition and outperforms the other traditional approaches.

An automatic feature extraction method and its application in fault diagnosis

“…To testify the superiority of our method, five common dimensionality reduction tools combined with sparse filtering are adopted to process the gearbox dataset respectively. The five tools are: principal component analysis (PCA) [25], locality preserving projection (LPP) [26], Sammon mapping (SM) [27], linear discriminant analysis (LDA) [28], and stochastic proximity embedding (SPE) [29]. The classification results by the five methods are shown in Fig.…”

Fault Diagnosis from Raw Sensor Data Using Deep Neural Networks Considering Temporal Coherence

“…Table 10 shows classification accuracies of different methods, including Genetic Algorithm with Random Forest [22], Chi Square Features with different classifiers [23], Continuous Wavelet Transform with SVM, Discrete Wavelet Transform with ANN [24], Statistical Locally Linear Embedding with SVM [8] and the method proposed in this paper.…”

“…In this field, time domain feature methods [1,2,3], including Kernel Density Estimation (KDE), Root Mean Square (RMS), Crest factor, Crest-Crest Value and Kurtosis, frequency domain features [4] such as the frequency spectrum generated by Fourier transformation, time-frequency features obtained by Wavelet Packet Transform (WPT) [5] are usually extracted as the gauge of the next process. Other signal processing methods such as Empirical Mode Decomposition (EMD) [6], Intrinsic Mode Function (IMF), Discrete Wavelet Transform (DWT), Hilbert Huang Transform (HHT) [7,8], Wavelet Transform (WT) [9] and Principal Component analysis (PCA) [10] are also implemented for signal processing. These signal processing and feature extraction methods are followed by some classification algorithms including Support Vector Machine (SVM), Artificial Neural Networks (ANN), Wavelet Neural Networks (WNN) [11], dynamic neural networks and fuzzy inference.…”

“…Yu et al [7] used Window Marginal Spectrum Clustering (WMSC) to select features from the marginal spectrum vibration signals by HTT and adopted SVM to classify faults. Wang et al [8] used the statistical locally linear embedding algorithm to extract low dimensional features from high dimensional features which are extracted by time domain, frequency domain and EMD methods. Regression trees, K-nearest-neighbor classifier and SVM were adopted as the classifiers, respectively.…”

A New Deep Learning Model for Fault Diagnosis with Good Anti-Noise and Domain Adaptation Ability on Raw Vibration Signal

“…Therefore, it's necessary to find a way to extract useful features that represents the intrinsic information of the machine. Common signal processing techniques used to extract the representative features from the raw signal include time-domain statistical analysis [6], wavelet transformation [7], and Fourier spectral analysis [8]. Usually after feature extraction, a feature selection step will be implemented to get rid of useless and insensitive features, and reduce the dimension for the sake of computational efficiency.…”