A fault diagnosis method based on k-means clustering and comprehensive correlation

Abstract

The invention discloses a method for diagnosing mechanical faults based on K-means clustering and comprehensive correlation, comprising the following steps: first, acquiring the original observation information of sensors for each fault type; using K-means clustering to find data of each fault type The centroid of the group; calculate the difference between the mechanical equipment operating data collected by the sensor and the centroid of each fault type data cluster JS Divergence and gray correlation degree build comprehensive correlation degree; normalize comprehensive correlation degree and convert it into basic probability assignment function; finally, use Dempster combination rule to fuse basic probability assignment functions one by one, and output the final output for mechanical equipment The decision result of the fault type. The fault diagnosis method based on K-means clustering and comprehensive correlation proposed by the solution of the present invention can effectively diagnose the fault of mechanical equipment, and has important theoretical significance and application value.

Description

technical field [0001] The invention relates to the technical field of fault detection, in particular to a fault diagnosis method based on K-means clustering and comprehensive correlation. Background technique [0002] At present, with the rapid development of science and technology, machinery and equipment tend to be large-scale and automated, and intelligent machinery and equipment can solve many problems that cannot be solved by manpower. However, the intelligentization of mechanical equipment also means that the internal structure of mechanical equipment is becoming more and more complex, and accidents caused by mechanical equipment failure are also gradually increasing, which will bring serious losses to the factory. By effectively diagnosing the working characteristics of mechanical equipment, the production accidents can be reduced to a great extent, and the economic losses of the factory can be reduced. [0003] The multi-sensor information fusion technology compreh...

AI Technical Summary

Problems solved by technology

In the article "Intuitionistic Fuzzy Recognition Decision-Making Method Based on Gray Relation and Evidence Theory", Wang Hong et al proposed to use the gray relation degree to construct the basic probability assignment function, but if some abnormal data deviates from the normal data, the gray relation degree will be is greatly reduced, affecting the effectiveness of the basic probability assignment function

In the article "Basic Probability Assignment Generation Method and Application Based on Interval Numbers", Kang Bingyi and others proposed to use the interval number similarity to construct the basic probability assignment function, but also did not consider the impact of abnormal data that deviates from normal data on the basic probability assignment function. The impact of the generation, so the existing methods can not meet the actual needs

Images