Mechanical fault diagnosis method based on multi-sensor fusion

Abstract

The invention discloses a mechanical fault diagnosis method based on multi-sensor fusion. The method comprises the following steps: converting operation data, observed by sensors at different positions, of mechanical equipment into evidence information; calculating support intervals of focal elements in the evidences, and measuring conflict degrees among the evidences through interval distances so as to obtain support degrees of the evidences; quantifying the uncertainty degree of the evidence through the improved reliability entropy to serve as the information amount of the evidence; comprehensively considering the interval distance and the improved reliability entropy to determine the credibility of the evidence and obtain a weight factor; and carrying out weighted average on the obtained evidence by using a weight factor, and outputting a decision result of mechanical fault diagnosis. Compared with a traditional algorithm, the scheme can effectively measure the difference between evidences through the interval distance, quantifies the uncertainty of the evidences through the improved reliability entropy, and comprehensively considers the support degree and the information amount to jointly determine the weight factor of the evidences.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis, in particular to a mechanical fault diagnosis method based on multi-sensor fusion. Background technique [0002] At present, with the advancement of science and technology, the composition and structure of machinery and equipment are becoming more and more complex, and the accidents caused by the failure of machinery and equipment are gradually increasing. In order to avoid this situation, it is necessary to regularly monitor the health of machinery and equipment Inspection, by detecting the health of the machine and equipment operation, can effectively avoid production accidents caused by the failure of the machine and equipment itself. [0003] Multi-sensor information fusion technology can avoid the limitation of single sensor, so it is widely used in the field of fault diagnosis. Collect the data of the machine and equipment itself through multiple sensors in different positions, int...

AI Technical Summary

Problems solved by technology

For example, Murphy proposed a simple average method in the article "Combining belief functions when evidence conflicts" to eliminate the impact of conflicting evidence on fusion, but this method only performs a simple average of all evidence and does not consider the mutual influence between evidence; Influenced by Murphy, Yuan et al. used the Jousselme evidence distance proposed by Jousselme et al. in the article "A new distance between twobodies of evidence" and the Deng entropy defined by Deng in the article "Deng entropy" to jointly construct a weight factor to predict the fusion evidence. deal with

However, the Jousselme evidence distance cannot accurately measure the degree of conflict between two sets of completely conflicting evidence. At the same time, when the evidence is all composite focal elements, contradictory results will also appear when using Deng Entropy to quantify the degree of uncertainty of the evidence; Song Yafei et al proposed to use the cosine correlation coefficient to measure the degree of conflict between two sets of evidence in the paper "Evidence Conflict Measurement Method Based on Correlation Coefficient". degree of conflict, but when multiple focal elements appear, the cosine correlation coefficient cannot accurately measure the degree of conflict between evidence

To sum up, the methods used above have certain deficiencies when measuring the degree of conflict between evidence or quantifying the uncertainty of evidence, so there is still room for improvement when using the above method for mechanical fault diagnosis

Images