Study on centrifugal gas compressor bearing fault character obtaining based on data mining

According to the characteristics of frequent fault and difficult diagnosis of rotating parts,selecting more interrelated vibration parameters as original character vectors,how to collect the characters from fault database and analyze rotating parts' running conditions by applying fuzzy clustering way is demonstrated in this paper.The theory of establishing model,specific algorithm and steps are given as well.Applied example showed that the method is correct and the result of fault diagnosis has also been proved to be reliable and accurate.