Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts

A signal fusion and data dimensionality reduction technology, applied in computer parts, machine learning, instruments, etc., can solve data complexity, high data compression rate, inefficient multi-sensor signal fusion strategy, and large computational load in the training process, etc. problem, to achieve the effect of improving sensitivity, reducing the number of training samples, and reducing data complexity

Active Publication Date: 2022-04-22
SHANDONG UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies in the prior art, the purpose of the present invention is to provide a multi-sensor signal fusion monitoring method and system for dimensionality reduction of thin-walled milling data, which solves the problem of multi-sensor signal feature data dimensionality reduction in the existing multi-sensor signal feature data dimensionality reduction method. The problem of low efficiency of sensory signal fusion strategy, low data compression efficiency, and large amount of calculation in the training process can reduce data complexity, high data compression rate, high calculation efficiency, and effectively integrate multiple sensors without affecting the accuracy of machine learning training. signal characteristics

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts
  • A multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts
  • A multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0043] This embodiment provides a multi-sensor signal fusion monitoring method for dimensionality reduction of thin-walled milling data, such as figure 1 shown, including:

[0044] Obtain the original signal feature matrix composed of multi-sensor signal wavelet packet energy;

[0045] Calculate the natural frequency of the workpiece and tool according to the frequency response function, calculate the spindle rotation frequency and cutter tooth passing frequency under the corresponding milling parameters, and calculate the frequency band where all wavelet packets are located;

[0046] Determine whether the wavelet packet frequency band includes the natural frequency of the workpiece, the tool natural frequency, the spindle rotation frequency and the tooth passing frequency, and take values ​​for different wavelet packet frequency bands to obtain the eigenvector of the fusion multi-sensor signal;

[0047] According to the eigenvector of the fusion multi-sensor signal, the orig...

Embodiment 2

[0079] This embodiment provides a multi-sensor signal fusion monitoring method for dimensionality reduction of milling data of thin-walled parts, including the following steps:

[0080] S1. Input the original signal characteristic matrix composed of wavelet packet energy of each sensing signal (acceleration signal, sound signal, milling force signal in this embodiment). There are 9 groups of experimental samples in the original data set, and each group of experiments collects three kinds of signals: acceleration, sound, and milling force, and each signal is decomposed by three layers of wavelet packets, thus 2 3 wavelet packet energy, that is, the original data set has a total of 27*8 signal features, and the scatter distribution of the original data set is as follows figure 2 shown.

[0081] S2. It is easy to obtain the frequency response function and natural frequency of the workpiece and the cutting tool according to the existing public technology. In this embodiment, the...

Embodiment 3

[0103] This embodiment provides a multi-sensor signal fusion monitoring data dimensionality reduction system for thin-walled parts milling, including:

[0104] The original signal feature matrix acquisition module is configured to: acquire an original signal feature matrix composed of multi-sensor signal wavelet packet energy;

[0105] The data calculation module is configured to: calculate the natural frequency of the workpiece and the tool according to the frequency response function, calculate the spindle rotation frequency and the cutter tooth passing frequency under the corresponding milling parameters, and calculate the frequency band where all wavelet packets are located;

[0106] The fusion eigenvector acquisition module is configured to: judge whether the frequency band of the wavelet packet contains the natural frequency of the workpiece, the natural frequency of the tool, the rotation frequency of the spindle and the passing frequency of the cutter tooth, and take va...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts. The technical solution is to: obtain the original signal characteristic matrix; calculate the natural frequency of the workpiece and the tool according to the frequency response function, and calculate the corresponding milling parameters The main shaft rotation frequency and the tooth passing frequency are calculated, and the frequency bands where all wavelet packets are located are calculated; whether the wavelet packet frequency band contains the natural frequency of the workpiece, the tool natural frequency, the main shaft rotation frequency and the tooth passing frequency, and the frequency bands of different wavelet packets are respectively Take the value to get the eigenvector of the fusion multi-sensor signal; according to the eigenvector of the fusion multi-sensor signal, the original signal feature matrix is ​​converted into the fusion feature matrix; according to the requirement of the contribution degree in the milling state identification, the dimension-reduced feature vector matrix is ​​obtained , and output data dimensionality reduction results. The invention can reduce the complexity of data, and can effectively fuse the features of multi-sensing signals without affecting the training accuracy of machine learning.

Description

technical field [0001] The invention relates to the technical field of machining process detection, in particular to a multi-sensing signal fusion monitoring method and system for dimensionality reduction of milling data of thin-walled parts. Background technique [0002] Thin-walled parts such as aero-engine fans and low-pressure compressor rotor blades are mostly made of difficult-to-machine materials. Due to their difficulty in machining and poor local rigidity at the edges, problems such as machining chatter, tool wear, and damage during cutting seriously affect production efficiency. and workpiece service fatigue performance. The data-driven milling process monitoring technology can realize real-time online monitoring of the machining process through the real-time interaction of sensor data and self-learning of historical processing data by means of signal processing. However, different sensing signals in the processing process have different sensitivities to milling s...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/62G06N20/00
CPCG06N20/00G06V10/52G06F2218/08G06F2218/12G06F18/213G06F18/253
Inventor 宋清华王润琼刘战强马海峰
Owner SHANDONG UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products