Main shaft bearing state evaluation method and device based on deep fusion network

Abstract

The invention discloses a main shaft bearing state evaluation method and device based on a deep fusion network, and the method specifically comprises the steps: converting a vibration signal, a current signal and a temperature signal of a main shaft bearing under different working conditions into a frequency spectrum signal through fast Fourier transform, taking a low-frequency component of the frequency spectrum signal, and processing the low-frequency component to obtain a training set and a test set; constructing an Inception branch network based on an Inception module and a convolutional neural network, constructing a central layer fusion network based on features extracted by each Inception branch network, and forming a deep fusion network by the branch networks and the central layerfusion network to realize state evaluation of the variable working condition spindle bearing. By fully mining the coupling characteristics of various sensing information, the defect that a single signal characteristic is difficult to comprehensively characterize the running state of the main shaft bearing is overcome, and the accuracy of main shaft bearing state evaluation is improved; coupling information of all sensing data is fully extracted from the perspective of multi-source information fusion, and the robustness to noise and different working conditions is enhanced.

Description

technical field [0001] The invention belongs to the field of spindle bearing state evaluation, and in particular relates to a method and device for evaluating the state of a spindle bearing based on a deep fusion network. Background technique [0002] Spindle bearing is a key component of CNC machine tools, and its performance has an important impact on the machining accuracy and reliability life of the machine tool. Since spindle bearings usually work in a high-pressure, high-load environment, the bearings are prone to various failures. According to statistics, about 40% of the failures in the spindle system are caused by rolling bearings. Once the spindle bearing fails, if it cannot be detected in time, it will not only affect the normal production cycle, but also may cause major mechanical accidents. Therefore, it is necessary to monitor the health and condition of the spindle bearings. [0003] The traditional bearing evaluation and diagnosis method usually evaluates ...

AI Technical Summary

Problems solved by technology

However, there is almost no research on the state evaluation of spindle bearings under variable working conditions by comprehensively utilizing multi-source information such as current, vibration, and temperature.

Images