The Order Statistics Correlation Coefficient and PPMCC Fuse Non-Dimension in Fault Diagnosis of Rotating Petrochemical Unit

In this paper, the advantages of dimensionless indices and two types of correlation coefficients are combined and two methods are proposed to enhance the efficiency and accuracy of fault diagnosis in the petrochemical rotating machinery. The order statistic correlation coefficient and Pearson's...

Full description

Saved in:
Bibliographic Details
Published in:IEEE sensors journal 2018-06, Vol.18 (11), p.4704-4714
Main Authors: Xiong, Jianbin, Liang, Qiong, Wan, Jiafu, Zhang, Qinghua, Chen, Xuehua, Ma, Rubao
Format: Article
Language:English
Subjects:
Accuracy
Correlation
correlation coefficient
Correlation coefficients
correlation measure
Data integration
dimensionless index
Dimensionless numbers
Fault diagnosis
Indexes
Machinery
Pearson’s product moment correlation coefficient (PPMCC)
Petrochemicals
Preprocessing
Rotating machinery
Sensors
The order statistics correlation coefficient (OSCC)
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, the advantages of dimensionless indices and two types of correlation coefficients are combined and two methods are proposed to enhance the efficiency and accuracy of fault diagnosis in the petrochemical rotating machinery. The order statistic correlation coefficient and Pearson's correlation coefficient are used to calculate the correlation coefficients of dimensionless indices, which are given by dimensionless algorithms after preprocessing the raw data. Different fault types are recognized by comparing the correlation coefficient and each dimensionless indicator. The numerical results revealed that the proposed method has the highest accuracy of 80% while the average of 50%, and an overall accuracy improvement of 10.89% compared with the conventional method. The results clearly indicate that the accuracy of the proposed fault diagnosis method is superior compared with its counterpart.
ISSN:1530-437X
1558-1748
DOI:10.1109/JSEN.2018.2820170