2019/07/27 by Jianlei Zhang, Zhang, Jianlei, Binil Starly +1
Engineering · #Advanced Machining and Optimization Techniques #Advanced machining processes and optimization #FOS: Electrical engineering #Industrial Vision Systems and Defect Detection #Signal Processing (eess.SP) #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1907.11848
openalex publication_date 2019/07/27 · openalex created_date 2022/07/28 · openalex updated_date 2026/07/28
Data-driven approaches to automated machine condition monitoring are gaining\npopularity due to advancements made in sensing technologies and computing\nalgorithms. This paper proposes the use of a deep learning model, based on Long\nShort-Term Memory (LSTM) architecture for a recurrent neural network (RNN)\nwhich captures long term dependencies for modeling sequential data. In the\ncontext of estimating cutting tool wear amounts, this LSTM based RNN approach\nutilizes a system transition and system observation function based on a\nminimally intrusive vibration sensor signal located near the workpiece\nfixtures. By applying an LSTM based RNN, the method helps to avoid building an\nanalytic model for specific tool wear machine degradation, overcoming the\nassumptions made by Hidden Markov Models, Kalman filter, and Particle filter\nbased approaches. The proposed approach is tested using experiments performed\non a milling machine. We have demonstrated one-step and two-step look ahead\ncutting tool state prediction using online indirect measurements obtained from\nvibration signals. Additionally, the study also estimates remaining useful life\n(RUL) of a machine cutting tool insert through generative RNN. The experimental\nresults show that our approach, applying the LSTM to model system observation\nand transition function is able to outperform the functions modeled with a\nsimple RNN.\n