Feature strengthening method and system for on-line monitoring of state of thin-walled workpiece milling cutter
A technology for milling tools and thin-walled parts, which is applied in the field of condition monitoring, can solve the problems that the monitoring model is easily affected by environmental noise, the correlation between signal features and recognition targets is weak, and the calculation efficiency is low, so as to achieve reduced complexity and strong adaptability , the effect of high recognition accuracy
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0038] This embodiment provides a feature enhancement method for online monitoring of the state of a milling tool for thin-walled parts, including:
[0039] Calculate the signal characteristics of the sensing signal of each pass before strengthening, and obtain the tool wear amount after each pass;
[0040] Calculate the correlation between each sensing feature vector and the tool wear value according to the signal feature and the tool wear amount, and obtain the feature correlation value of the entire cutting process;
[0041] Calculate the feature weight coefficient for each feature correlation value to obtain the feature weight coefficient vector of the entire cutting process;
[0042] Based on the feature matrix and the weight coefficient matrix, the enhanced feature component matrix is obtained, the enhanced feature component matrix is summed and the average value is calculated to obtain the enhanced feature vector, so as to improve the correlation between the feature...
Embodiment 2
[0110] In order to test the validity of the first embodiment, a feature enhancement method for online monitoring of the thin-walled milling tool state of the first embodiment is verified based on the "PHM2010 tool wear data set" published by the International Association for Fault Diagnosis and Health Management.
[0111] Among them, each tool has a total of 315 sets of cutting signals, that is, m=315, and each set of signals gives a tool wear value (average wear width of the flank), and the average value and variance of each set of signals are extracted according to step S2. , standard deviation, root mean square, skewness, and kurtosis, a total of 6 signal characteristics, that is, n=6.
[0112] According to formula (18), the correlation vector ε of each signal feature and tool wear value can be obtained, Figure 4 The correlation calculation results of six signal features and the predicted target (i.e. tool wear value) are shown, from Figure 4 It can be seen that in this ...
Embodiment 3
[0116] This embodiment provides a feature enhancement system for online monitoring of the state of a milling tool for thin-walled parts, including:
[0117] The feature acquisition module is configured to: calculate the signal features of the sensing signal of each tool pass before strengthening, and obtain the tool wear amount after each pass;
[0118] The feature correlation value acquisition module is configured to: calculate the correlation between each sensing feature vector and the tool wear value according to the signal feature and the tool wear amount, and obtain the feature correlation value of the entire cutting process;
[0119] The feature weight coefficient vector acquisition module is configured to: calculate the feature weight coefficient for each feature correlation value, and obtain the feature weight coefficient vector of the entire cutting process;
[0120] The enhanced feature vector acquisition module is configured to: obtain the enhanced feature component...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


