Cutting chatter identification method based on optimized VMD and CNN-ICBAM-MHSA-BiLSTM

By optimizing VMD parameters and constructing a CNN-ICBAM-MHSA-BiLSTM fusion network, the problems of parameter adjustment relying on manual intervention and insufficient feature capture in existing chatter recognition methods are solved, achieving high-precision and robust cutting chatter recognition.

CN122165242APending Publication Date: 2026-06-09SHENYANG UNIVERSITY OF TECHNOLOGY
0 Cites 0 Cited by

Patent Information

Application Number
CN202610380552.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-26
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing flutter recognition methods require manual adjustment of the number of modes and penalty factors, making it difficult to adapt to different processing conditions. Traditional deep learning models cannot effectively capture the local details and global context information of multi-sensor signals and have insufficient resistance to noise interference, resulting in a low recognition rate for minor flutter.

Method used

The arithmetic optimization algorithm AOA is used to optimize variational mode decomposition (VMD). Combined with convolutional neural network (CNN), improved convolutional block attention module (ICBAM), bidirectional long short-term memory network (BiLSTM), and multi-head self-attention mechanism (MHSA), a fusion network is constructed for signal processing and recognition.

Benefits of technology

It achieves adaptive parameter optimization and multi-scale feature capture, which improves the accuracy and recall of slight tremor recognition, enhances recognition precision and anti-noise interference capability, and achieves an overall recognition accuracy of 98.40%.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention provides a cutting chatter recognition method based on optimized VMD and CNN-ICBAM-MHSA-BiLSTM, relates to the technical field of machine tool cutting machining monitoring, and solves the technical problems that in existing cutting chatter recognition, VMD parameters depend on manual adjustment, and the slight chatter recognition rate is low. Comprising the steps of multi-sensor signal acquisition, AOA-VMD adaptive decomposition and IMFs screening reconstruction, multi-attention fusion network training and flutter condition identification. Wherein the arithmetic optimization algorithm AOA and the variational mode decomposition VMD are combined to realize parameter adaptive optimization; designing an improved convolution block attention mechanism ICBAM, and enhancing spatial feature extraction through a multi-scale convolution kernel; a CNN-ICBAM-MHSA-BiLSTM fusion network is constructed, and local features, time sequence dependence and key time step attention are integrated; experimental verification shows that three working conditions of stable machining, slight chatter and serious chatter can be accurately recognized, the overall recognition accuracy is larger than or equal to 98.40%, the slight chatter recognition accuracy is improved to be larger than or equal to 3.9%, the noise interference resistance is high, and reliable technical support is provided for chatter recognition of machine tool cutting machining.
Need to check novelty before this filing date? Find Prior Art