Limits...
Bearing Fault Diagnosis Based on Statistical Locally Linear Embedding.

Wang X, Zheng Y, Zhao Z, Wang J - Sensors (Basel) (2015)

Bottom Line: 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.Finally in the feature reduction space pattern classification and fault diagnosis by classifier are carried out easily and rapidly.The results indicate that the proposed approach obviously improves the classification performance of fault pattern recognition and outperforms the other traditional approaches.

View Article: PubMed Central - PubMed

Affiliation: College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing 210098, China. wangxiang@njit.edu.cn.

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.

No MeSH data available.


The four frequency-domain features value in the dataset: (a) Mean frequency; (b) Frequency center; (c) Root mean square frequency. (d) Root variance frequency (Note: sample data No.1–100, 101–200, 201–300, 301–400 represent normal, inner race fault, ball fault and outer race faults, respectively).
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4541876&req=5

sensors-15-16225-f007: The four frequency-domain features value in the dataset: (a) Mean frequency; (b) Frequency center; (c) Root mean square frequency. (d) Root variance frequency (Note: sample data No.1–100, 101–200, 201–300, 301–400 represent normal, inner race fault, ball fault and outer race faults, respectively).

Mentions: For the every obtained data set, we extract statistical 22 features following the time-domain, frequency domain and time-frequency domain for the next feature dimension reduction. Twelve time-domain and four frequency-domain statistical features could be calculated directly using the feature definition equations as shown in Table 1 and Table 2, and time-frequency domain features are extracted from the EMD energy. The calculated value of the six dimensional and six dimensionless time-domain statistical features are shown in Figure 5 and Figure 6, respectively, and the calculated value of the four frequency-domain statistical features are shown in Figure 7. The six time-frequency domain statistical features obtained from the first six IMFs energy are obtained by applying EMD method. The EMD result of a signal sample in the dataset is shown in Figure 8.


Bearing Fault Diagnosis Based on Statistical Locally Linear Embedding.

Wang X, Zheng Y, Zhao Z, Wang J - Sensors (Basel) (2015)

The four frequency-domain features value in the dataset: (a) Mean frequency; (b) Frequency center; (c) Root mean square frequency. (d) Root variance frequency (Note: sample data No.1–100, 101–200, 201–300, 301–400 represent normal, inner race fault, ball fault and outer race faults, respectively).
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4541876&req=5

sensors-15-16225-f007: The four frequency-domain features value in the dataset: (a) Mean frequency; (b) Frequency center; (c) Root mean square frequency. (d) Root variance frequency (Note: sample data No.1–100, 101–200, 201–300, 301–400 represent normal, inner race fault, ball fault and outer race faults, respectively).
Mentions: For the every obtained data set, we extract statistical 22 features following the time-domain, frequency domain and time-frequency domain for the next feature dimension reduction. Twelve time-domain and four frequency-domain statistical features could be calculated directly using the feature definition equations as shown in Table 1 and Table 2, and time-frequency domain features are extracted from the EMD energy. The calculated value of the six dimensional and six dimensionless time-domain statistical features are shown in Figure 5 and Figure 6, respectively, and the calculated value of the four frequency-domain statistical features are shown in Figure 7. The six time-frequency domain statistical features obtained from the first six IMFs energy are obtained by applying EMD method. The EMD result of a signal sample in the dataset is shown in Figure 8.

Bottom Line: 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.Finally in the feature reduction space pattern classification and fault diagnosis by classifier are carried out easily and rapidly.The results indicate that the proposed approach obviously improves the classification performance of fault pattern recognition and outperforms the other traditional approaches.

View Article: PubMed Central - PubMed

Affiliation: College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing 210098, China. wangxiang@njit.edu.cn.

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.

No MeSH data available.