This research aims to develop a multi-threading method for rapid tool wear detection by integrating image classification and object detection techniques to address the challenge of tool wear detection. The research proposes a two-stage method that leverages a fast image classification model (VGG-16) and a high-accuracy object detection model (YOLOv5)to enable efficient multi-threading detection of tool wear regions across a large number of the flank wear images. The experiment results reveal that, when the wear images account for less than 70% of the total, this method can achieve detection speeds exceeding that of YOLOv5 while maintaining comparable detection accuracy.
The objective of this study is to address the issue of data imbalance by augmenting the milling tool breakage dataset using Auxiliary Classifier Generative Adversarial Networks (ACGAN). The research team developed an ACGAN architecture capable of producing samples labeled with various states of tool breakage. To assess the fidelity of the ACGAN-generated data, this study employed evaluation metrics such as the Kullback-Leibler divergence, Euclidean distance, and the Pearson correlation coefficient, comparing the generated samples against actual samples. The findings indicate a high degree of similarity in data distribution between the synthetic and real samples, suggesting the effectiveness of the generated data for training purposes. This research introduces a cost-effective and efficient approach for data augmentation, significantly enhancing the capabilities of milling tool condition monitoring systems.
This study aims to achieve fast and accurate identification of tool breakage in multi-tooth milling cutters using methods such as thresholding, clustering, and neural networks. A multi-level thresholding strategy combining fixed thresholds and dynamic thresholds was designed to enable rapid and accurate response in tool breakage identification. The fuzzy c-means clustering algorithm (FCM) and one-dimensional convolutional Softmax classifier (1D-CNN Softmax) were employed to identify the tool breakage states, distinguishing between normal cutting, single-tooth breakage, and doubletooth breakage. Experimental results demonstrate that this method exhibits fast response and high accuracy in classifying the breakage states of the three-tooth milling cutter, achieving an accuracy rate of 98.6%. This research provides a rapid and accurate technique for tool breakage identification in the field of multi-tooth milling cutter tools.
Carbon fiber-reinforced polymer (CFRP) is a multiphase material consisting of fibers, interfaces, and matrix. Due to their excellent mechanical properties, they are widely used in the energy, military, and aerospace sectors. However, due to the anisotropic and non-homogeneous nature of the material, tool wear inevitably occurs during machining. In order to ensure the quality of material machining and to control tool costs, tool condition monitoring has become an integral part of machining. By monitoring the tool condition in machining, predictive maintenance can be achieved, and early warning of tool failure can be achieved, thus drastically reducing downtime and saving costs in terms of time and labor. On this basis, this paper proposes a novel physics-guided neural network approach for tool wear prediction. Firstly, the fusion of physical and data information is achieved through cross-physical data modeling. Second, a multi-channel 1D-CNN convolutional neural network is utilized to reduce the complexity of local feature extraction. In addition, a loss function considering physical subject factors is proposed to quantify the physical inconsistency. Experiments of the proposed model are carried out on carbon fiber reinforced ceramic matrix composites to validate the performance of the model in terms of MAE and RMSE.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.