Bearing fault online diagnosis method and system based on depth separable convolution

Abstract

The invention discloses a bearing fault online diagnosis method based on depth separable convolution. The method comprises the following steps: inputting historical vibration data acquired by a data acquisition system into a depth separable convolution-based network for training to obtain a diagnosis model; converting and storing the diagnosis model in a TensorFlow.js layer format, and deploying the converted layer format model in a browser; and loading historical vibration data from a database in the browser, and inputting the historical vibration data into the converted layer format model for diagnosis to obtain the bearing state of the data. The invention discloses a diagnosis system, which comprises a model training module, a model conversion and deployment module and a fault identification step module. According to the invention, a fault diagnosis model is built through a deep separable convolution thought, and the model is deployed in a browser for online fault diagnosis; and according to the method, on the premise that the diagnosis accuracy is guaranteed, the size and the calculation amount of the model are greatly reduced, the operation speed of the model is greatly increased, and the operation cost is effectively controlled.

Description

technical field [0001] The invention relates to a method and system for online diagnosis of bearing faults based on depth separable convolution, belonging to the field of fault diagnosis of mechanical equipment. Background technique [0002] In order to detect hidden dangers in the industrial production process in time, it is necessary to monitor the health status of key components of the mechanical system in operation. Traditionally, the fault diagnosis of bearings is to use the acquisition equipment to collect and save the vibration data, and then use manual feature extraction for fault analysis, which is not only costly, but also highly dependent on prior knowledge, and the diagnosis results are greatly affected by human factors. In recent years, due to the rise of artificial intelligence, neural networks have also made a qualitative leap in the field of fault diagnosis. At present, the diagnostic accuracy and efficiency of intelligent diagnosis based on neural network h...

AI Technical Summary

Problems solved by technology

[0003] The present invention provides a bearing fault online diagnosis method and system based on depth-separable convolution. A fault diagnosis model is built through depth-separable convolution, and it is further used in a browser to realize online fault diagnosis, which can further solve the problem of existing diagnosis. The model parameters are large, occupying a lot of resources, and it is difficult to directly deploy large-scale networks on devices with poor computing power

Images