Full-Vector Signal Acquisition and Information Fusion for the Fault Prediction

Lei Chen,Yongxiang Cui,Wenping Lei,Jie Han,Zhenhong Guan

doi:10.1155/2016/5980802

Lei Chen, Yongxiang Cui + Show 3 more

Open Access

https://doi.org/10.1155/2016/5980802

Copy DOI

Abstract

Fault prediction is the key technology of the predictive maintenance. Currently, researches on fault prediction are mainly focused on the evaluation of the intensities of the failure and the remaining life of the machine. There is lack of methods on the prediction of fault locations and fault characters. To satisfy the requirement of the prediction of the fault characters, the data acquisition and fusion strategies were studied. Firstly, the traditional vibration measurement mechanism and its disadvantages were presented. Then, the full-vector data acquisition and fusion model were proposed. After that, the sampling procedure and information fusion algorithm were analyzed. At last, the fault prediction method based on full-vector spectrum was proposed. The methodology is that of Dr. Bently and Dr. Muszynska. On the basis of this methodology, the application study has been carried out. The uncertainty of the spectrum structure can be eliminated by the designed data acquisition and fusion method. The reliability of the diagnosis on fault character was improved. The study on full-vector data acquisition system laid the technical foundation for the prediction and diagnosis research of the fault characters.

Highlights

As the key technology of the predictive maintenance [1], fault prediction is utilized to predict the operating condition of the machine and give an early warning of the failure which may occur in the future
Including the Chinese Power Utility Industry, industry wide of a single transducer to assess dynamics has been replaced by use of two or more transducers and key-phasor as fully described in detail in the classic works by Dr Bently et al [18]
Since the fault characteristic is concerned with the spectrum structure, the characteristic and position of machine fault which may occur in the future can be determined by analysis of the predicted full-vector spectrum

Summary

Introduction

As the key technology of the predictive maintenance [1], fault prediction is utilized to predict the operating condition of the machine and give an early warning of the failure which may occur in the future. Researchers have conducted lots of work on the aspects of machine fault diagnosis [2,3,4,5,6], operation status prediction [7,8,9,10], and feature extraction and information fusion [11,12,13], and the application of fault prediction has been proved to be favorable to instruct production dispatch and equipment maintenance. To analyze the fault source and predict the fault character, failure type, and failure position of the equipment, one of the feasible means is to predict the trend of frequency spectrum. The data acquisition and signal fusion model for the fault character prediction will be analyzed.

Vibration Mechanism and Signal Monitoring

Full-Vector Data Acquisition and Information Fusion

Experiment and Discussion

Conclusions