Multi-point transmission mechanism fault diagnosis method and system based on convolutional neural network

Abstract

The invention provides a multi-point transmission mechanism fault diagnosis method and system based on a convolutional neural network and aims to solve the technical problem that existing fault diagnosis efficiency and accuracy do not adapt to transmission mechanism development. The method comprises the steps that distribution positions of vibration sensors are optimized, and parallel vibration data corresponding to the vibration sensors is formed; and a convolutional layer, a pooling layer and a full-connection layer of an initial convolutional neural network are established, the initial convolutional neural network is adjusted, fault diagnosis precision of the initial convolutional neural network is verified, the fitting degree of the initial convolutional neural network is adjusted, anda fault diagnosis model is formed. Therefore, an automatic analysis process of numerous measuring point data of a complicated transmission mechanism is formed, noise information can be filtered out automatically and accurately, and fault characteristics can be separated; and as to a large amount of to-be-processed data, the purpose of processing the large-scale data can be achieved by adjusting the scale of the network, and therefore a complicated mapping relation is represented. Besides, the requirements on efficiency and reliability of fault early-warning in a full life cycle of the complicated transmission mechanism are met.

Description

technical field [0001] The invention relates to the technical field of fault diagnosis signal processing, in particular to a multi-measuring-point transmission mechanism fault diagnosis method and system based on a convolutional neural network. Background technique [0002] In the field of multi-measuring point transmission mechanism fault diagnosis, the existing fault diagnosis technology detects the location of the fault and analyzes the cause of the fault by monitoring the current operating state of the transmission mechanism to prevent catastrophic consequences. At present, the commonly used fault diagnosis technology process is as follows: first, the fault feature is extracted through signal processing technology; and then, the machine learning technology is used for fault diagnosis. There are four problems in this technology: 1. Due to the complexity of the transmission mechanism itself, it is necessary to set up multiple measuring points at different positions and mon...

AI Technical Summary

Problems solved by technology

[0003] In view of the above problems, the embodiment of the present invention provides a multi-measuring point transmission mechanism fault diagnosis method and system based on a convolutional neural network to solve the technical problem that the efficiency and accuracy of the existing multi-measuring point transmission mechanism fault diagnosis are not suitable for the development of the transmission mechanism

Images