Tool wear prediction based on multidomain feature fusion by attention-based depth-wise separable convolutional neural network in manufacturing

Abstract

Tool wear during machining has a great influence on the quality of machined surface and dimensional accuracy. Tool wear monitoring is extremely important to improve machining efficiency and workpiece quality. Multidomain features (time domain, frequency domain and time-frequency domain) can accurately characterise the degree of tool wear. However, manual feature fusion is time consuming and prevents the improvement of monitoring accuracy. A new tool wear prediction method based on multidomain feature fusion by attention-based depth-wise separable convolutional neural network is proposed to solve these problems. In this method, multidomain features of cutting force and vibration signals are extracted and recombined into feature tensors. The proposed hypercomplex position encoding and high dimensional self-attention mechanism are used to calculate the new representation of input feature tensor, which emphasizes the tool wear sensitive information and suppresses large area background noise. The designed depth-wise separable convolutional neural network is used to adaptively extract high-level features that can characterize tool wear from the new representation, and the tool wear is predicted automatically. The proposed method is verified on three sets of tool run-to-failure data sets of three-flute ball nose cemented carbide tool in machining centre. Experimental results show that the prediction accuracy of the proposed method is remarkably higher than other state-of-art methods. Therefore, the proposed tool wear prediction method is beneficial to improve the prediction accuracy and provide effective guidance for decision making in processing.

Deep learning-based CNC milling tool wear stage estimation with multi-signal analysis

“…In their study, Bazi et al [8] proposed a new method with CNN and bidirectional long short-term memory applications to estimate tool wear while cutting operations. To avoid the obstacles of time-consuming manual feature fusion, Li et al [35] conducted a new tool wear monitoring approach by using multi-domain feature fusion with depth-wise separable CNN. With this method, they obtained frequency and time domain features from shear force and vibration signals and combined feature tensors.…”

Tool Wear Prediction Model Using Multi-Channel 1D Convolutional Neural Network and Temporal Convolutional Network

“…Feature engineering is an indispensable upfront step to utilize ML techniques to effectively forecast and monitor tool wear. Li et al presented a feature transfer learning-based tool wear prediction approach [12], in which multi-domain features were extracted from cutting force and vibration signals and were integrated into feature tensors for tool wear prediction [13]. However, since feature engineering is usually a manual, iterative, and time-consuming process that needs expertise, this leads to longer projects and increased costs [14].…”

SoC-Based Early Failure Detection System Using Deep Learning for Tool Wear

“…Recently, deep learning approaches have automatically extracted features and provided accurate results in the tool wear monitoring fields [24], [25]. Most related studies have focused on special or one machining condition to establish a monitoring system for tool wear detection; however, there are machining variations in practical machining, such as different machining conditions and material variations, which affect the accuracy.…”