A novel intelligent method for bearing fault diagnosis based on affinity propagation clustering and adaptive feature selection

Wei, Zexian; Wang, Yanxue; He, Shuilong; Bao, Jianmin

doi:10.1016/j.knosys.2016.10.022

Cited by 185 publications

(80 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…6 Modifying the 0 value of M new as 3 do. 7 Perform the D-S fusion calculation once again as 4 does, and get the finally fusion results. , which is not consistent with normal acceptable target A.…”

Section: Of 21mentioning

confidence: 99%

“…Due to the unavoidable limitation in terms of accuracy and robustness in the feature extraction approaches of traditional shallow learning, Lu [6] investigated the deep learning method based on a convolutional neural network (CNN), the novel feature representation method for bearing data using supervised deep learning with the goal of identifying more robust and salient feature representations to reduce information loss, and two experiments had proved the efficiency of the proposed method. Wei [7] proposed a new adaptive features selection technique applied to the bearing fault diagnosis with affinity propagation clustering, and results demonstrated that the approach is able to reliably and accurately identify different fault categories and severities of bearings. For the studies shown above and others [8][9][10][11], they all aimed at the failure of one single point on the bearings; the current methods of feature automatic extraction and automatic classification didn't consider the failure conditions of multiple faults combined.…”

mentioning

confidence: 99%

See 1 more Smart Citation

A Novel Method for Intelligent Single Fault Detection of Bearings Using SAE and Improved D–S Evidence Theory

Zhang

Tang

2019

Entropy

View full text Add to dashboard Cite

In order to realize single fault detection (SFD) from the multi-fault coupling bearing data and further research on the multi-fault situation of bearings, this paper proposes a method based on features self-extraction of a Sparse Auto-Encoder (SAE) and results fusion of improved Dempster-Shafer evidence theory (D-S). Multi-fault signal compression features of bearings were extracted by SAE on multiple vibration sensors' data. Data sets were constructed by the extracted compression features to train the Support Vector Machine (SVM) according to the rule of single fault detection (R-SFD) this paper proposed. Fault detection results were obtained by the improved D-S evidence theory, which was implemented via correcting the 0 factor in the Basic Probability Assignment (BPA) and modifying the evidence weight by Pearson Correlation Coefficient (PCC). Extensive evaluations of the proposed method on the experiment platform datasets showed that the proposed method could realize single fault detection from multi-fault bearings. Fault detection accuracy increases as the output feature dimension of SAE increases; when the feature dimension reached 200, the average detection accuracy of the three sensors for bearing inner, outer, and ball faults achieved 87.36%, 87.86% and 84.46%, respectively. The three types' fault detection accuracy-reached to 99.12%, 99.33% and 98.46% by the improved Dempster-Shafer evidence theory (IDS) to fuse the sensors' results-is respectively 0.38%, 2.06% and 0.76% higher than the traditional D-S evidence theory. That indicated the effectiveness of improving the D-S evidence theory by evidence weight calculation of PCC.Entropy 2019, 21, 687 2 of 21 extracted features by Wavelet Packet Transformation, selected features through automatic encoder, and classified them by SVM. The proposed approach achieved very high accuracy and robustness to classify ball, inner race, and outer race faults even under very poor signal-to-noise ratio (SNR) conditions. Due to the unavoidable limitation in terms of accuracy and robustness in the feature extraction approaches of traditional shallow learning, Lu [6] investigated the deep learning method based on a convolutional neural network (CNN), the novel feature representation method for bearing data using supervised deep learning with the goal of identifying more robust and salient feature representations to reduce information loss, and two experiments had proved the efficiency of the proposed method. Wei [7] proposed a new adaptive features selection technique applied to the bearing fault diagnosis with affinity propagation clustering, and results demonstrated that the approach is able to reliably and accurately identify different fault categories and severities of bearings. For the studies shown above and others [8][9][10][11], they all aimed at the failure of one single point on the bearings; the current methods of feature automatic extraction and automatic classification didn't consider the failure conditions of multiple faults combined. Actually, current researche...

show abstract

Section: Of 21mentioning

confidence: 99%

mentioning

confidence: 99%

A Novel Method for Intelligent Single Fault Detection of Bearings Using SAE and Improved D–S Evidence Theory

Zhang

Tang

2019

Entropy

View full text Add to dashboard Cite

show abstract

“…El sistema carga el archivo de configuración con adquisición diaria y empieza a comparar el tiempo real hasta que coincide. La adquisición se realiza con una frecuencia de muestreo de 40KHz, que según el estado del arte [14] es apropiada para la detección de fallos de señales de vibración. Después de realizar la adquisición diaria, el sistema establece la comunicación con el servidor FTP.…”

Section: Figura 3 Módulo De Controlunclassified

“…Más específicamente, para monitorizar las señales de vibración de las máquinas giratorias se han desarrollado WSN, basado en Compuertas Programables en Campo (Field Programmable Gate Array FPGA) [6] o Sistemas Microelectromecanicos (Microelectromechanical Systems MEMs) [13], aunque con frecuencias de muestreo máximo de 20 kHz. Sin embargo lo anterior representa un problema, ya que muchas fallas que se quieren analizar al monitorear máquinas rotativas están presentes en frecuencias de hasta 12 kHz [14]. De acuerdo con el teorema de Nyquist, la frecuencia de muestreo para el monitoreo de máquinas rotativas debe ser de al menos 25 kHz.…”

Section: Introductionunclassified

Instrumentación para captura y transmisión de señales de vibración

Hernández

Echeverry²,

Mesa³

et al. 2018

Vis. Electron.

View full text Add to dashboard Cite

Las señales de vibración son usadas generalmente para detectar fallos en máquinas rotativas. En la actualidad existen diferentes metodologías para realizar análisis basado en dichas señales. Una metodología usada extensamente es el Mantenimiento Basado en Condición (CBM). CBM es un mantenimiento programado que recomienda acciones basadas en información recolectada. Actualmente, para la adquisición de señales de vibración se usan comúnmente Redes de Sensores Inalámbricos (WSNs por sus siglas en ingles). Los WSNs son redes formadas por una cierta cantidad de nodos, cada nodo está equipado con un sensor para identificar un fenómeno físico como la luz, presión, temperatura, etc. En este artículo, se propone un robusto sistema basado en WSNs para la adquisición, almacenamiento y transmisión de señales de vibración, el cual combina un mecanismo de acondicionamiento, una tarjeta central y un dispositivo para la transmisión inalámbrica. El sistema propuesto cumple todas las funciones anteriores de manera automática y precisa para dos señales de vibración y una señal de velocidad.

show abstract

“…A novel optimal feature selection approach was presented which can be a benefit for fault detection and diagnosis in smart buildings [8]. A novel intelligent fault diagnosis method was proposed for the rolling bearing based on the adaptive feature selection technique [9]. Kang et al suggested a hybrid feature selection method to reduce diagnostic performance deterioration in data-driven PHM system [10].…”

Section: Introductionmentioning

confidence: 99%

A Step-wise Feature Selection Scheme for a Prognostics and Health Management System in Autonomous Ferry Crossing Operation

Cheng

Ellefsen

et al. 2019

2019 IEEE International Conference on Mechatronics and Automation (ICMA)

View full text Add to dashboard Cite

Developing a reliable algorithm to detect faults automatically within critical components in autonomous ferries is essential for safe and cost-beneficial maritime operations. Autonomous ferries are equipped with hundreds of sensors. Thus, in order to support the algorithm, the input data should be subjected to a feature selection process. This paper introduces a novel step-wise feature selection scheme for prognostics and health management (PHM) system in autonomous ferries. The scheme mainly consists of two steps. The first step is the Pearson correlation analysis to reduce the redundant information among sensors. In order to study the importance of the selected features obtained by correlation analysis and removal of irrelevant features, the second step is sensitivity analysis (SA) based feature selection. The proposed scheme is evaluated on realoperational marine diesel engine data. In the experiments, both fault classification and fault detection demonstrate the feasibility of the proposed approach.

show abstract

A novel intelligent method for bearing fault diagnosis based on affinity propagation clustering and adaptive feature selection

Cited by 185 publications

References 23 publications

A Novel Method for Intelligent Single Fault Detection of Bearings Using SAE and Improved D–S Evidence Theory

A Novel Method for Intelligent Single Fault Detection of Bearings Using SAE and Improved D–S Evidence Theory

Instrumentación para captura y transmisión de señales de vibración

A Step-wise Feature Selection Scheme for a Prognostics and Health Management System in Autonomous Ferry Crossing Operation

Contact Info

Product

Resources

About