Rolling bearing fault prediction method based on continuous deep belief network

Abstract

The invention proposes a rolling bearing fault prediction method based on a continuous deep belief network. The method comprises the following steps: first, extracting the time-domain characteristic quantity of vibration signals of a rolling bearing; then, fusing the extracted time-domain characteristic information using a locally linear embedding method, so as to define a new comprehensive monitoring index for better quantitative evaluation of performance degradation of the bearing; training a continuous restricted Boltzmann machine step by step to construct a continuous deep belief network prediction model; and using a genetic algorithm to optimize the structure of the continuous deep belief network so as to further improve the prediction precision. The prediction method is reliable in result, has good real-time performance, is simple and feasible, and is suitable for rolling bearing fault prediction.

Description

technical field [0001] The invention belongs to the field of mechanical equipment health monitoring, in particular to a rolling bearing fault prediction method based on a continuous deep belief network. Background technique [0002] Rolling bearings are the most widely used mechanical parts in rotating machinery, and they are also one of the most vulnerable components. Rolling bearings may be damaged due to various reasons during operation, such as improper assembly, poor lubrication, moisture and foreign matter intrusion, corrosion and overload, etc. may cause premature damage to rolling bearings. Even if the installation, lubrication and maintenance are normal, after a period of operation, the rolling bearing will suffer from fatigue, peeling and wear and cannot work normally. Therefore, rolling bearing fault prediction is very important. It can not only ensure the safe operation of equipment, prevent major accidents, and improve economic benefits, but also provide a reli...

AI Technical Summary

Problems solved by technology

The neural network is easy to fall into local minimum, and its convergence speed is too slow; the support vector machine is more suitable for the learning of small samples but not for the learning of large samples, and the selection of its kernel function and parameters lacks corresponding theory

More importantly, these traditional models are essentially shallow machine learning models, which usually contain no more than one layer of nonlinear feature transformation. When given a limited number of samples and computing units, it is difficult for

Images