Application of machine vision method in tool wear monitoring

Peng, Ruitao; Liu, Jiachen; Fu, Xiuli; Liu, Cuiya; Zhao, Linfeng

doi:10.1007/s00170-021-07522-4

Cited by 39 publications

(10 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the practical process of image acquisition, capturing images of the tool's side is challenged by the speed of the tool's movement. In response to this challenge, this paper proposes an image selection method based on the SSIM [22]. The objective of this method is to effectively filter and select qualified side images of the tool, overcoming the obstacles posed by the tool's rapid motion during image capture.…”

Section: Tool Side Image Extractionmentioning

confidence: 99%

Tool wear monitoring based on scSE-ResNet-50-TSCNN model integrating machine vision and force signals

Nie,

Guo,

Lou

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

In the realm of mechanical machining, tool wear is an unavoidable phenomenon. Monitoring the condition of tool wear is crucial for enhancing machining quality and advancing automation in the manufacturing process. This paper investigates an innovative approach to tool wear monitoring that integrates machine vision with force signal analysis. It relies on a deep residual two-stream convolutional model optimized with the scSE(Concurrent Spatial and Channel Squeeze and Excitation) attention mechanism (scSE-ResNet-50-TSCNN). The force signals are converted into the corresponding wavelet scale images following wavelet threshold denoising and Continuous Wavelet Transform (CWT). Concurrently, the images undergo processing using Contrast Limited Adaptive Histogram Equalization (CLAHE) and the Structural Similarity Index Method (SSIM), allowing for the selection of the most suitable image inputs. The processed data are subsequently input into the developed scSE-ResNet-50-TSCNN model for precise identification of the tool wear state. To validate the model, the paper employed X850 carbon fibre reinforced polymer (CFRP) and Ti-6Al-4V titanium alloy as laminated experimental materials, conducting a series of tool wear tests while collecting pertinent machining data. The experimental results underscore the model's effectiveness, achieving an impressive recognition accuracy of 93.86%. When compared with alternative models, the proposed approach surpasses them in performance on the identical dataset, showcasing its efficient monitoring capabilities in contrast to single-stream networks or unoptimized networks. Consequently, it excels in monitoring tool wear status and promots crucial technical support for enhancing machining quality control and advancing the field of intelligent manufacturing.

show abstract

Section: Tool Side Image Extractionmentioning

confidence: 99%

Tool wear monitoring based on scSE-ResNet-50-TSCNN model integrating machine vision and force signals

Nie,

Guo,

Lou

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Flank wear is the most dominant failure mode, where the maximum flank wear width is usually used to define the failure criterion. Tool wear can be measured directly (i.e., measured from image data, collected in-process using high-speed cameras [14] or offline using microscopes), or indirectly (i.e., correlated to measured forces, vibrations, or acoustic emission data). Based on the data collected, tool wear characterisation and RUL prognostics have been studied extensively in the last several decades, with a range of modeling techniques explored.…”

Section: Overview Of Cutting Tool Phm Approachesmentioning

confidence: 99%

“…Physics-based degradation / lifetime models e.g., Modified Taylor (tool life with MPPs) ; Coromant [2]; Sipos (flank wear and MPPs) [15]; flank wear relationship to cutting force and MPPs [12] Highly accurate simulations can be used to predict MPPs' impact on cutting tool life; suitable when MPPs are constant Need to specify failure thresholds relative to failure time; costly to develop sufficient trials with varied experimental MPPs; difficult in estimation of unknown model parameters Statistical methods without covariates e.g., Bayesian models (i.e., GPR [16], PF [17,18], Wiener Process [19] ) with covariates e.g., Proportional Hazards models (PH) [20] Uncertainty bounds for prediction; modelling different degradation process distributions a-priori; deal with MPPs as weighted covariates; assume flexible parameter distributions on covariates; uncertainty bounds for predictions Choices of model structure can be difficult to develop when underlying process is unknown; feature engineering required on raw data; censored observations require special attentions in modelling process Machine Learning approaches e.g., Linear/Nonlinear Regression [21], SVM/R [22], Random Forest [23], Fuzzy methods [14] Capable of accurate predictions with smaller datasets Raw signals almost always subject to feature engineering or dimensionality reduction; no RUL probability distributions Deep Learning approaches e.g., Back-Propagation Neural Network (BPNN) [24]; CNN (wear classification [25]; wear prediction [26]; wear segmentation [27]); Wavelet Neural Network (WNN) [28]; LSTM (tool RUL prediction ); CNN-LSTM [29][30][31]; GRU [32]; Autoencoders [33][34] End-to-end trainable with feature learning mechanisms (i.e., no feature engineering); learning from heterogeneous data; generalising across huge datasets (millions of parameters); transfer learning; multi-task approaches; supervised/unsupervised training Model training resources are difficult for collection (require sufficient labelled data for supervised learning; computation); model overfitting on smaller datasets; regarded as black-box models, so inherently difficult to troubleshoot; negative transfer…”

Section: Modelling Process Strengths Weaknessesmentioning

confidence: 99%

See 1 more Smart Citation

Cutting tool prognostics enabled by hybrid CNN-LSTM with transfer learning

Marei

2021

Int J Adv Manuf Technol

View full text Add to dashboard Cite

show abstract

“…In addition, some other researchers believed that tool wear states would be reflected on the surface topography, so the surface roughness was used as the input data to establish the tool wear estimation models. For example, energy, entropy, inertial moment, and correlation were extracted from surface topography, and then, the tool wear states were estimated through these features [14,15]. Jain et al established a tool wear monitoring model with the current and previous surface roughness values as the input variables [16].…”

Section: Introductionmentioning

confidence: 99%

A Novel Multi-Task Learning Model with PSAE Network for Simultaneous Estimation of Surface Quality and Tool Wear in Milling of Nickel-Based Superalloy Haynes 230

Cheng

Jiao

Yan

et al. 2022

Sensors

View full text Add to dashboard Cite

For data-driven intelligent manufacturing, many important in-process parameters should be estimated simultaneously to control the machining precision of the parts. However, as two of the most important in-process parameters, there is a lack of multi-task learning (MTL) model for simultaneous estimation of surface roughness and tool wear. To address the problem, a new MTL model with shared layers and two task-specific layers was proposed. A novel parallel-stacked auto-encoder (PSAE) network based on stacked denoising auto-encoder (SDAE) and stacked contractive auto-encoder (SCAE) was designed as the shared layers to learn deep features from cutting force signals. To enhance the performance of the MTL model, the scaled exponential linear unit (SELU) was introduced as the activation function of SDAE. Moreover, a dynamic weight averaging (DWA) strategy was implemented to dynamically adjust the learning rate of different tasks. Then, the time-domain features were extracted from raw cutting signals and low-frequency reconstructed wavelet packet coefficients. Frequency-domain features were extracted from the power spectrum obtained by the Fourier transform. After that, all features were combined as the input vectors of the proposed MTL model. Finally, surface roughness and tool wear were simultaneously predicted by the trained MTL model. To verify the superiority and effectiveness of the proposed MTL model, nickel-based superalloy Haynes 230 was machined under different cutting parameter combinations and tool wear levels. Some other intelligent algorithms were also implemented to predict surface roughness and tool wear. The results showed that compared with the support vector regression (SVR), kernel extreme learning machine (KELM), MTL with SDAE (MTL_SDAE), MTL with SCAE (MTL_SCAE), and single-task learning with PSAE (STL_PSAE), the estimation accuracy of surface roughness was improved by 30.82%, 16.67%, 14.06%, 26.17%, and 16.67%, respectively. Meanwhile, the prediction accuracy of tool wear was improved by 46.74%, 39.57%, 41.51%, 38.68%, and 39.57%, respectively. For practical engineering application, the dimensional deviation and surface quality of the machined parts can be controlled through the established MTL model.

show abstract

Application of machine vision method in tool wear monitoring

Cited by 39 publications

References 34 publications

Tool wear monitoring based on scSE-ResNet-50-TSCNN model integrating machine vision and force signals

Tool wear monitoring based on scSE-ResNet-50-TSCNN model integrating machine vision and force signals

Cutting tool prognostics enabled by hybrid CNN-LSTM with transfer learning

A Novel Multi-Task Learning Model with PSAE Network for Simultaneous Estimation of Surface Quality and Tool Wear in Milling of Nickel-Based Superalloy Haynes 230

Contact Info

Product

Resources

About