Rotary machine fault diagnosis method based on deep clustering

Abstract

The invention discloses a rotary machine fault diagnosis method based on deep clustering. The method comprises the steps of collecting a label-free mechanical vibration signal; preprocessing the label-free mechanical vibration signal to obtain a signal data set; constructing an auto-encoder structure, and carrying out the pre-training of the auto-encoder structure; the auto-encoder structure beingexpressed by pre-training initial features of learning data; setting hyper-parameters of a manifold learning method and using the manifold learning method to search for a manifold which can be more clustered to re-learn the stacked auto-encoder embedding; and based on the updated embedding, completing clustering by applying a shallow clustering algorithm, and performing result evaluation. The method has the beneficial effects that the shallow clustering algorithm is adopted, the calculation complexity of the model is reduced, the required time consumption is reduced, and meanwhile, the precision of a diagnosis result is ensured, so that the method can be better applied to an actual industrial scene.

Description

technical field [0001] The invention relates to the technical field of mechanical fault diagnosis, in particular to a method for diagnosing rotating mechanical faults based on deep clustering. Background technique [0002] In recent years, mechanical fault diagnosis technology has made great progress. There are many types of data-driven fault diagnosis methods in practical applications, including artificial intelligence, information fusion, multivariate statistical analysis, rough sets and signal processing. Due to the rapid development of artificial intelligence technology in recent years, the mechanical fault diagnosis technology driven by intelligent data has also been greatly developed. In this process, a large amount of model training is required, and the effect of model training will affect the final diagnosis result. accuracy. However, most of the existing diagnostic models are trained in a supervised manner, that is, the training samples need to be labeled. [0003...

AI Technical Summary

Problems solved by technology

However, in practical industrial applications, it is usually difficult, expensive, and sometimes impossible to collect enough labeled data, and these difficulties also greatly hinder the application of intelligent diagnostic methods.

Images