Review
Engineering, Manufacturing
Quade Butler, Youssef Ziada, David Stephenson, S. Andrew Gadsden
Summary: This review outlines the techniques and methods for feed drive condition monitoring, diagnostics, and prognostics in recent research. It also describes commercial and industry solutions to Industry 4.0 condition monitoring.
JOURNAL OF MANUFACTURING SCIENCE AND ENGINEERING-TRANSACTIONS OF THE ASME
(2022)
Article
Engineering, Manufacturing
Alexandra Schueller, Christopher Saldano
Summary: Tool condition monitoring (TCM) is an important research area because it can reduce manufacturing costs and improve process efficiency. Machine learning (ML) has advantages for TCM, but individual ML models lack generalizability and robustness to unbalanced datasets. Ensemble ML models have better performance in other fields, but their performance in TCM is not well understood. In this study, experiments were conducted using different cutting conditions and various performance metrics were compared to fill in these research gaps.
JOURNAL OF MANUFACTURING SCIENCE AND ENGINEERING-TRANSACTIONS OF THE ASME
(2023)
Review
Automation & Control Systems
Ruhong Jia, Caixu Yue, Qiang Liu, Wei Xia, Yiyuan Qin, Mingwei Zhao
Summary: Tool wear monitoring is necessary for the tool processing industry as the tool wear state is closely related to the quality of the workpiece and directly affects equipment performance. This research analyzes the relationship between tool wear and sensor signals to determine the required acquisition signal. Signal processing technology is used to preprocess the original signal, extract time and frequency domain features, and optimize the features using the extreme random tree (ET). The relevance vector machine (RVM) model proposed in this study effectively monitors tool wear.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2023)
Article
Engineering, Manufacturing
Alexandra Schueller, Christopher Saldana
Summary: Tool condition monitoring (TCM) is crucial for optimizing metal machining processes. This study addresses the challenge of TCM system generalizability and proposes ensemble machine learning techniques for TCM. Experiments were conducted to evaluate different models for predicting tool wear levels. The results show that cutting speed has a significant impact on model performance, while feed rate has little effect. The application of simulated noise data augmentation technique improves model generalizability. The extremely randomized trees ensemble machine learning model performs the best for this application.
JOURNAL OF MANUFACTURING PROCESSES
(2022)
Article
Engineering, Mechanical
Zhan Yie Chin, Pietro Borghesani, Wade A. Smith, Robert B. Randall, Zhongxiao Peng
Summary: Recently, the use of transmission error (TE) in gear diagnostics has become more feasible. TE is a parameter that describes the meshing error between two gears and can be accurately measured using shaft encoders. The measurement of TE has considerable benefits compared to vibration as well as traditional wear analysis, and it has been shown to be sensitive to the evolution of gear wear, making it a powerful tool for gear prognostics.
Review
Computer Science, Interdisciplinary Applications
Melvin Alexis Lara de Leon, Jakub Kolarik, Radek Byrtus, Jiri Koziorek, Petr Zmij, Radek Martinek
Summary: This article reviews and analyzes the approaches utilized for monitoring cutting tool conditions, focusing on the use of Machine Learning and statistical processes. It quantifies the typical signals used by researchers and scientists, such as vibration, cutting force, and temperature, to determine tool degradation and product quality. The article also presents statistical techniques used to cleanse collected data and extract relevant information for prediction and classification purposes.
ARCHIVES OF COMPUTATIONAL METHODS IN ENGINEERING
(2023)
Article
Engineering, Mechanical
Zhan Yie Chin, Wade A. Smith, Pietro Borghesani, Robert B. Randall, Zhongxiao Peng
Summary: Gear transmission error (TE) is a promising diagnostic tool as it is less affected by transfer path and can be conveniently measured using shaft encoders. Analysis of TE signals can provide information about gear wear depth and deviations from the perfect involute tooth profile, which are crucial in gear diagnostics.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2021)
Article
Engineering, Industrial
E. Traini, G. Bruno, F. Lombardi
Summary: This paper presents a data-driven framework for estimating tool wear status and predicting its remaining useful life using machine learning techniques. The framework includes data preprocessing, feature engineering, and development of prediction models. A case study in a milling process demonstrates the potential of the framework for tool condition monitoring.
INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH
(2021)
Article
Engineering, Manufacturing
Benjamin Stuhr, Rui Liu
Summary: In order to achieve smart manufacturing, tool condition monitoring (TCM) systems are used to detect the state of the tool and optimize tool conditions to prevent failures. However, many existing systems lack flexibility and require complex setup. To address these issues, a proposed algorithm utilizes repetitive machining operations in mass production settings and similarity analysis to achieve TCM. By comparing signals collected from the tool with known conditions, the algorithm can accurately predict the state of the tool. A case study confirmed the effectiveness and flexibility of the algorithm.
JOURNAL OF MANUFACTURING SCIENCE AND ENGINEERING-TRANSACTIONS OF THE ASME
(2022)
Article
Engineering, Mechanical
Bo Qin, Yongqing Wang, Kuo Liu, Shaowei Jiang, Qi Luo
Summary: This paper proposes an innovative approach that considers the tool wear law and the characteristic distribution of tool wear monitoring data to achieve accurate prediction of tool wear condition. Additionally, a data and mechanism-driven tool wear monitoring method is proposed to enable reliable adjustment of the monitoring results of the data-driven model without human intervention.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2023)
Article
Engineering, Mechanical
Yuanning Mao, Jun Tong, Zhan Yie Chin, Pietro Borghesani, Zhongxiao Peng
Summary: This study investigates the usage of vibration and transmission error (TE) to identify the presence of contaminants and measure gear wear severity caused by oil contamination. Tests were conducted on a spur gearbox rig using clean and contaminated lubricants with silica sand. The results show that TE can quantitively measure wear depth and vibration can qualitatively correlate with average wear depths.
Article
Engineering, Electrical & Electronic
Hao Guo, Xin Lin, Kunpeng Zhu
Summary: In this article, a pyramid LSTM auto-encoder is proposed for tool wear monitoring in high-performance CNC machining. The features are compressed layer by layer based on the frequency spectrum, simplifying the monitoring task and reducing model complexity. The efficiency of long-term signal processing is greatly improved by reducing the number of units. The introduction of auto-encoder further enhances the model's accuracy under complex working conditions through unsupervised learning.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2022)
Article
Automation & Control Systems
Siqi Wang, Shichao Yan, Yuwen Sun
Summary: This paper proposes a tool wear condition monitoring method based on nonnegativity-constrained autoencoder and grey wolf optimization algorithm, specifically for milling difficult-to-cut materials. Experimental results demonstrate significant improvements in modeling efficiency and prediction accuracy.
INTERNATIONAL JOURNAL OF ADVANCED MANUFACTURING TECHNOLOGY
(2023)
Review
Engineering, Mechanical
Ping Lu, Honor E. Powrie, Robert J. K. Wood, Terry J. Harvey, Nicholas R. Harris
Summary: This paper proposes an approach to enhance machinery condition monitoring through micro-sensing of tribological phenomena occurring between contacting surfaces, which can distinguish between benign and harmful wear scenarios. The wear process is influenced by lubricant condition, tribo-pair condition and operating condition, and requires a comprehensive monitoring strategy and targeted sensing technologies. Through a critical review of tribo-sensing technologies, potential monitoring technologies for early wear detection and related benefits are identified.
TRIBOLOGY INTERNATIONAL
(2021)
Article
Chemistry, Analytical
Yaping Zhang, Xiaozhi Qi, Tao Wang, Yuanhang He
Summary: Tool wear condition monitoring is crucial for mechanical processing automation, and this paper proposes a new deep learning model to accurately identify tool wear status. By transforming the force signal into a two-dimensional image using CWT, STFT, and GASF methods, the images are then analyzed using a CNN model. The results show that this method achieves an accuracy above 90%, outperforming other models like AlexNet and ResNet. The images generated by the CWT method exhibit the highest accuracy in tool wear state recognition due to its ability to extract local features and resist noise. These findings highlight the potential advantages of using a force signal transformed into a two-dimensional image and applying CNN models in tool wear state recognition, indicating promising applications in industrial production.
Article
Ethics
N. Jordan Jameson, Xin Song, Michael Pecht
SCIENCE AND ENGINEERING ETHICS
(2016)
Article
Automation & Control Systems
Noel Jordan Jameson, Michael H. Azarian, Michael Pecht
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2017)
Article
Mathematics, Applied
S. Sathananthan, N. Jameson, I. Lyatuu, L. H. Keel
STOCHASTIC ANALYSIS AND APPLICATIONS
(2013)
Article
Engineering, Multidisciplinary
David Mascarenas, Crystal Plont, Christina Brown, Martin Cowell, N. Jordan Jameson, Jessica Block, Stephanie Djidjev, Heidi Hahn, Charles Farrar
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2014)
Article
Engineering, Electrical & Electronic
N. Jordan Jameson, Michael H. Azarian, Michael Pecht
INTERNATIONAL JOURNAL OF ELECTRICAL POWER & ENERGY SYSTEMS
(2020)
Article
Chemistry, Analytical
Phillip Koshute, Nathan Hagan, N. Jordan Jameson
Summary: The study suggests using supervised machine learning classification models as a complementary approach to detect fentanyl analogs from mass spectra, achieving strong detection performance through extracting input features and identifying patterns within the values through techniques like random forests, neural networks, and logistic regression.
FORENSIC CHEMISTRY
(2022)
Proceedings Paper
Engineering, Manufacturing
Gregory W. Vogl, Brian C. Galfond, N. Jordan Jameson
PROCEEDINGS OF THE ASME 14TH INTERNATIONAL MANUFACTURING SCIENCE AND ENGINEERING CONFERENCE, 2019, VOL 1
(2019)
Article
Computer Science, Information Systems
Kai Wang, Haifeng Guo, Aidong Xu, Noel Jordan Jameson, Michael Pecht, Bingjun Yan
