In-process tap tool wear monitoring and prediction using a novel model based on deep learning
- PDF / 3,426,196 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 93 Downloads / 218 Views
ORIGINAL ARTICLE
In-process tap tool wear monitoring and prediction using a novel model based on deep learning Xingwei Xu 1,2 & Jianweng Wang 1,2 & Weiwei Ming 1,2 & Ming Chen 1,2 & Qinglong An 1,2 Received: 10 March 2020 / Accepted: 4 November 2020 # Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Tool wear monitoring and prediction are of great importance for machining precision and surface integrity. In order to ensure good quality of the machining components, tool wear should be monitored and predicted in time. However, the traditional methods greatly depend on feature selection and extraction, and accuracy and generalization are limited. In this paper, a novel model based on deep learning was proposed to monitor and predict tap tool wear. Firstly, 1D CNN was designed to extract features from the vibration data, and then the residual block with the dilated convolution was specially developed to accept the features from 1D CNN. After that, the fully connected neural networks were designed to predicate tool wear. Moreover, to verify the superiority and generalization of the proposed method, the taping experiment from a real engine cylinder head production line was conducted. During the tapping process, the vibration signal and tool wear were collected. And then the training experiment was conducted with the collected data. The prediction results with the proposed model were compared with those of the state-ofthe-art algorithms. The compared results showed that the proposed model was more robust and accurate for tool wear prediction. Keywords Tool wear monitoring . Tap wear prediction . 1D CNN . Dilated convolution . Deep learning
1 Introduction Tool wear monitoring and prediction play a crucial role in the machining process, especially for the machined part with a large added value. Because the surface integrity of the workpiece and production efficiency are greatly affected by tool wear during the machining process [1–3], accurate tool wear monitoring and prediction can make proper tool change at an early stage to reduce downtime and prolong the service life possible, and ultimately enhance safety and ensure product quality [4, 5]. Thus, it is necessary to monitor and predict tool wear to maintain high quality and save costs. Current tool wear monitoring and prediction methods usually contain four steps: data acquisition, data processing, feature engineering, and feature recognition. The objective of the signal acquisition is to collect the original signal by sensors * Weiwei Ming [email protected] 1
State Key Laboratory of Mechanical Systems and Vibration, Shanghai Jiao Tong University, Shanghai 200240, China
2
School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai 200240, China
(e.g., vibration, force, or power) [6–8]. Data processing and feature engineering aim to extracting the representative features from raw monitoring signals and filtering out disturbing noise and useless information. In feature recognition, the specific methods for tool wear monitoring and
Data Loading...
