Damage localization in hydraulic turbine blades using kernel-independent component analysis and support vector machines

Wang, X H; Mao, Hanling; Zhu, Canhong; Huang, Zhenfeng

doi:10.1243/09544062jmes1296

Cited by 18 publications

(17 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thus, the support vector machine (SVM) was used. Wang et al 16 used SVM to locate crack of turbine runners. The result shows that the recognition rate in the crack region is good with small samples.…”

Section: Aementioning

confidence: 99%

“…[16], the different methods and different number of samples were used to solve the damage localization in turbine blades. The motivation of this research is to verify the validity of crack location based on KICA and WNN and to find out which is more suitable for the large-size complex structure with a large input dimensions and output patterns the large number of samples were used in this study.…”

Section: Aementioning

confidence: 99%

See 1 more Smart Citation

Crack Localization in Hydraulic Turbine Blades Based on Kernel Independent Component Analysis and Wavelet Neural Network

Wang¹,

Mao²,

Hu³

et al. 2013

IJCIS

View full text Add to dashboard Cite

Hydraulic turbine runner has a complex structure, and traditional location methods can't meet its requirement. This paper describes a source location of cracks in turbine blades by combining kernel independent component analysis (KICA) with wavelet neural network (WNN). The research shows that the location accuracy of WNN combined with KICA feature extraction is the best comparing with the results of WNN and back propagation neural network (BPNN). The method decreases the dimension of input parameters and improves the accuracy of location as well.

show abstract

Section: Aementioning

confidence: 99%

Section: Aementioning

confidence: 99%

Crack Localization in Hydraulic Turbine Blades Based on Kernel Independent Component Analysis and Wavelet Neural Network

Wang¹,

Mao²,

Hu³

et al. 2013

IJCIS

View full text Add to dashboard Cite

show abstract

“…The work by Wang et al. in [14] and Samanta et al. in [4] concluded that the prediction performance using the essential AE parameters (i.e.…”

Section: Three‐dimensional Location Experiments Of Ae Sourcesmentioning

confidence: 99%

“…For example, Wang et al. [14] classified major crack regions in a hydraulic turbine blade into three patterns and recognised regions of AE sources by SVM. However up to now, in these situations, the damage localisation does not represent the real positioning (i.e.…”

Section: Introductionmentioning

confidence: 99%

Pattern Recognition Method for the Source Location of Acoustic Emission Generated During the Damage of Hydraulic Concrete

Jianjie³

et al. 2012

Strain

View full text Add to dashboard Cite

To overcome the disadvantages of current acoustic emission (AE) source location methods, such as classical approaches based on times of arrival and artificial neural networks based on AE signal features, support vector machines (SVM)-based models have been employed to recognise AE source regions in structures. However, in some circumstances, it seems that a more accurate positioning of AE sources is needed. This study concerns the spatial three-dimensional (3D) positioning (i.e. coordinates) for damages in hydraulic concrete structures using the least squares SVM (LS-SVM) regression with AE signal features. The data of artificial discrete AE sources were acquired from simulated AE events on a hydraulic concrete specimen. Various combinations of signal features were chosen to adequately excavate effective information and to obtain the multi-output LS-SVM regression model of the best performance. The training and testing results show that the proposed model can realise the accurate spatial 3D positioning of damages in hydraulic concrete structures in laboratory situations and reduce human factors (e.g. judgment of AE propagation velocity, etc.) in the AE source location process. Meanwhile, the work remaining in taking this idea to a practical implementation was discussed.

show abstract

“…Moreover, the limited human resource of expert personnel restricts a widespread application of the fault diagnosis techniques. As such, in order to help inexperienced operators to quickly make an objective and accurate decision on machinery running conditions, a wide variety of automated classification paradigms have been introduced for fault classification such as expert systems (ESs), fuzzylogic inference [25,26], NNs [6,[27][28][29][30][31][32][33][34][35], and SVMs [36][37][38][39][40]. ESs may be the earliest attempt made to automate fault diagnosis, which documents the expertise of human experts into a computer system and emulates human reasoning process.…”

Section: Introductionmentioning

confidence: 99%

Fault diagnosis based on optimized node entropy using lifting wavelet packet transform and genetic algorithms

Zhang

Xiong

Liu

et al. 2010

Proceedings of the Institution of Mechanical Engineers, Part I:

View full text Add to dashboard Cite

The current paper presents a novel scheme for bearing fault diagnosis based on lifting wavelet packet transform (LWPT), sample entropy (SampEn), support vector machines (SVMs), and genetic algorithms (GAs). In the proposed scheme, bearing vibration signals were first decomposed into different frequency sub-bands through a four-level LWPT, resulting in a total of 31 node signal components throughout all layers of the LWPT decomposition tree. The SampEns of all 31 components were then calculated as an original feature pool to characterize the complexity of the bearing vibration signals within the corresponding frequency bands. For selecting the most informative features thus reducing the number of features, a GA was applied to simultaneously pick out the salient features and optimize the parameters of SVMs so as to avoid the curse of dimensionality, alleviate computational burden, and improve the subsequent classification. Experiments were conducted on an induction motor with respect to various bearing faults and a range of fault severities. As an example here, three salient features were selected from the original 31 features, with the dimension of features reduced dramatically. The selected three features were then presented into the optimized SVM to identify various bearing conditions. The scheme was compared with the widely used WPT-Energy method and with a neural network classifier with respect to feature extractions and fault classifications, respectively. The results are in favour of the proposed scheme in terms of the feature dimension, number of support vectors, robustness to the data partition, and classification rate.

show abstract

Damage localization in hydraulic turbine blades using kernel-independent component analysis and support vector machines

Cited by 18 publications

References 9 publications

Crack Localization in Hydraulic Turbine Blades Based on Kernel Independent Component Analysis and Wavelet Neural Network

Crack Localization in Hydraulic Turbine Blades Based on Kernel Independent Component Analysis and Wavelet Neural Network

Pattern Recognition Method for the Source Location of Acoustic Emission Generated During the Damage of Hydraulic Concrete

Fault diagnosis based on optimized node entropy using lifting wavelet packet transform and genetic algorithms

Contact Info

Product

Resources

About