Prediction of Machining Condition Using Time Series Imaging and Deep Learning in Slot Milling of Titanium Alloy

Hojati, Faramarz; Azarhoushang, Bahman; Daneshi, Amir; Khiabani, Rostam Hajyaghaee

doi:10.3390/jmmp6060145

Cited by 4 publications

(4 citation statements)

References 35 publications

(39 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Table 3 Accuracy of one-dimensional data input model Model Accuracy IHDGWO-SVM [19] 92.00% Random Forest [14] 93.68% PSO-SVM [18] 94.38% SSAEN [20] 93.04% ReliefF-PCA-SVM [21] 95.58% CNN-LSTM [9] 95.30%…”

Section: Fig7 Evaluation Indicator Valuesmentioning

confidence: 99%

“…In response to the limitations of traditional image encoding methods and neural networks in predicting tool wear states, which include inadequate generalization capabilities and a need for improved test accuracy, this study, drawing on the insights from previous research [11][12][13][14] , proposes a novel prediction approach for tool wear states that integrates the GAF and the MTF with AlexNet. This method is specifically designed for the classification and detection of tool wear.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Tool wear state prediction based on GAF-MTF-AlexNet

Yu,

Yang,

Liu

et al. 2024

Preprint

View full text Add to dashboard Cite

The manufacturing sector commonly employs traditional systems for predicting tool wear states, which rely on one-dimensional numerical inputs, in order to mitigate the issues of escalating costs and reduced efficiency that stem from excessive tool wear. However, these models face drawbacks such as high computational demands and limited generalization capabilities. To overcome these challenges, this study introduces a novel method for predicting tool wear states, which leverages the Gramian Angular Field-Markov Transition Field-AlexNet (GAF-MTF-AlexNet) architecture. Initially, the vibration signal's envelope spectrum is mapped into a two-dimensional image through the application of the GAF and the MTF transformations. Subsequently, the generated image is horizontally stitched to form a comprehensive dataset. The dataset is then fed into the customised AlexNet model for the purpose of prediction. Following meticulous research and optimisation, this method enhances signal visualisation and allows deep learning algorithms to extract signal features more effectively, resulting in a significant improvement in accuracy to 96.83%. This precision far surpasses that of conventional one-dimensional data input approaches, thus validating the efficacy of the measures implemented. The empirical results demonstrate that the method is proficient in predicting the transition of tools across diverse wear states. Consequently, this approach provides a robust technical foundation for the manufacturing industry to reduce production costs, improve operational efficiency, and facilitate predictive maintenance practices.

show abstract

Section: Fig7 Evaluation Indicator Valuesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Tool wear state prediction based on GAF-MTF-AlexNet

Yu,

Yang,

Liu

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…The authors suggest that incorporating temperature signals can be a reliable way to enhance wear predictions. Hojati et al [21] proposed a model to predict critical machining conditions concerning surface roughness and tool breakage during the milling of a titanium alloy (Ti6Al4V). After collecting the process signals through a Siemens SINUMERIK Edge Box computing device, the Gramian angular field (GAF) and a CNN method were applied.…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Approach for Predicting Critical Machining Conditions in Titanium Alloy Slot Milling Using Feature Selection and Binary Whale Optimization Algorithm

et al. 2023

Self Cite

View full text Add to dashboard Cite

Monitoring the machining process is crucial for providing cost-effective, high-quality production and preventing unwanted accidents. This study aims to predict critical machining conditions related to surface roughness and tool breakage in titanium alloy slot milling. The Siemens SINUMERIK EDGE (SE) Box system collects signals from the spindle and axes of a CNC machine tool. In this study, features were extracted from signals in time, frequency, and time–frequency domains. The t-test and the binary whale optimization algorithm (BWOA) were applied to choose the best features and train the support vector machine (SVM) model with validation and training data. The SVM hyperparameters were optimized simultaneously with feature selection, and the model was tested with test data. The proposed model accurately predicted critical machining conditions for unbalanced datasets. The classification model indicates an average recall, precision, and accuracy of 80%, 86%, and 95%, respectively, when predicting workpiece quality and tool breakage.

show abstract

“…The GAF allowed the transfer of CNC machine control system signals in images of the machined surface, which proceeded with the CNN algorithm. The trained classification CNN model resulted in recall, precision, and accuracy with 75%, 88%, and 94% values, respectively, for predicting workpiece surface quality and tool breakage [37] and tool life [38]. Currently, different types of tool wear can be predicted using artificial neuron networks and measured in real time.…”

Section: Introductionmentioning

confidence: 99%

Selected Aspects of Precision Machining on CNC Machine Tools

Peterka,

Kuruc,

Kolesnyk

et al. 2023

Machines

View full text Add to dashboard Cite

The work is new due to the type of process used—ultrasonic precision machining—to determine the possible effect of spindle heating (long-term machining) on the precision of the flat surface. It was carried out on a precise ultrasonic machining machine, and the material of workpiece was ceramic Al2O3. A flat surface was machined. Such an experiment has not been feasible until now. The experiment was divided into two days. On the first day, the machining time was 4 h. It is a long enough time to create a temperature-steady state. On the second day, with a cold tool and cold machine tool, we continued where we left off on the first day. This is how we monitored the accuracy of the dimensions of the workpiece on the plane surface. We have achieved the following: The average interface depth achieved values of 0.007089 mm and 0.003667 mm for cold and heated spindles, respectively. It means that when the spindle is not heated, the depth of the interface is higher by 93% (almost double the depth). The average standard deviation of the interface depth is 0.001683 mm and 0.000997 mm for cold and heated spindles, respectively. It means that when the spindle is not heated, the process is not as stable, and the standard deviation is higher by 69%.

show abstract

Prediction of Machining Condition Using Time Series Imaging and Deep Learning in Slot Milling of Titanium Alloy

Cited by 4 publications

References 35 publications

Tool wear state prediction based on GAF-MTF-AlexNet

Tool wear state prediction based on GAF-MTF-AlexNet

A Hybrid Approach for Predicting Critical Machining Conditions in Titanium Alloy Slot Milling Using Feature Selection and Binary Whale Optimization Algorithm

Selected Aspects of Precision Machining on CNC Machine Tools

Contact Info

Product

Resources

About