Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach
Stamping processes remain crucial in manufacturing processes; therefore, diagnosing the condition of stamping tools is critical. One of the challenges in diagnosing stamping tool conditions is that traditionally, the tools need to be visually checked, and the production processes thus need to be hal...
Published in: | Applied Sciences |
---|---|
Main Authors: | , , |
Format: | Text |
Language: | English |
Published: |
Multidisciplinary Digital Publishing Institute
2021
|
Subjects: | |
Online Access: | https://doi.org/10.3390/app11156959 |
id |
ftmdpi:oai:mdpi.com:/2076-3417/11/15/6959/ |
---|---|
record_format |
openpolar |
spelling |
ftmdpi:oai:mdpi.com:/2076-3417/11/15/6959/ 2023-08-20T04:06:09+02:00 Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach Zaky Dzulfikri Pin-Wei Su Chih-Yung Huang agris 2021-07-28 application/pdf https://doi.org/10.3390/app11156959 EN eng Multidisciplinary Digital Publishing Institute Mechanical Engineering https://dx.doi.org/10.3390/app11156959 https://creativecommons.org/licenses/by/4.0/ Applied Sciences; Volume 11; Issue 15; Pages: 6959 deep metric learning (DML) stamping tool triplet network siamese network Text 2021 ftmdpi https://doi.org/10.3390/app11156959 2023-08-01T02:18:13Z Stamping processes remain crucial in manufacturing processes; therefore, diagnosing the condition of stamping tools is critical. One of the challenges in diagnosing stamping tool conditions is that traditionally, the tools need to be visually checked, and the production processes thus need to be halted. With the development of Industry 4.0, intelligent monitoring systems have been developed by using accelerometers and algorithms to diagnose the wear classification of stamping tools. Although several deep learning models such as the convolutional neural network (CNN), auto encoder (AE), and recurrent neural network (RNN) models have demonstrated promising results for classifying complex signals including accelerometer signals, the practicality of those methods are restricted due to the flexibility of adding new classes and low accuracy when faced to low numbers of samples per class. In this study, we applied deep metric learning (DML) methods to overcome these problems. DML involves extracting meaningful features using feature extraction modules to map inputs into embedding features. We compared the probability method, the contrastive method, and a triplet network to determine which method was most suitable for our case. The experimental results revealed that, compared with other models, a triplet network can be more effectively trained with limited training data. The triplet network demonstrated the best test results of the compared methods in the noised test data. Finally, when tested using unseen class, the triplet network and the probability method demonstrated similar results. Text DML MDPI Open Access Publishing Applied Sciences 11 15 6959 |
institution |
Open Polar |
collection |
MDPI Open Access Publishing |
op_collection_id |
ftmdpi |
language |
English |
topic |
deep metric learning (DML) stamping tool triplet network siamese network |
spellingShingle |
deep metric learning (DML) stamping tool triplet network siamese network Zaky Dzulfikri Pin-Wei Su Chih-Yung Huang Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
topic_facet |
deep metric learning (DML) stamping tool triplet network siamese network |
description |
Stamping processes remain crucial in manufacturing processes; therefore, diagnosing the condition of stamping tools is critical. One of the challenges in diagnosing stamping tool conditions is that traditionally, the tools need to be visually checked, and the production processes thus need to be halted. With the development of Industry 4.0, intelligent monitoring systems have been developed by using accelerometers and algorithms to diagnose the wear classification of stamping tools. Although several deep learning models such as the convolutional neural network (CNN), auto encoder (AE), and recurrent neural network (RNN) models have demonstrated promising results for classifying complex signals including accelerometer signals, the practicality of those methods are restricted due to the flexibility of adding new classes and low accuracy when faced to low numbers of samples per class. In this study, we applied deep metric learning (DML) methods to overcome these problems. DML involves extracting meaningful features using feature extraction modules to map inputs into embedding features. We compared the probability method, the contrastive method, and a triplet network to determine which method was most suitable for our case. The experimental results revealed that, compared with other models, a triplet network can be more effectively trained with limited training data. The triplet network demonstrated the best test results of the compared methods in the noised test data. Finally, when tested using unseen class, the triplet network and the probability method demonstrated similar results. |
format |
Text |
author |
Zaky Dzulfikri Pin-Wei Su Chih-Yung Huang |
author_facet |
Zaky Dzulfikri Pin-Wei Su Chih-Yung Huang |
author_sort |
Zaky Dzulfikri |
title |
Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
title_short |
Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
title_full |
Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
title_fullStr |
Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
title_full_unstemmed |
Stamping Tool Conditions Diagnosis: A Deep Metric Learning Approach |
title_sort |
stamping tool conditions diagnosis: a deep metric learning approach |
publisher |
Multidisciplinary Digital Publishing Institute |
publishDate |
2021 |
url |
https://doi.org/10.3390/app11156959 |
op_coverage |
agris |
genre |
DML |
genre_facet |
DML |
op_source |
Applied Sciences; Volume 11; Issue 15; Pages: 6959 |
op_relation |
Mechanical Engineering https://dx.doi.org/10.3390/app11156959 |
op_rights |
https://creativecommons.org/licenses/by/4.0/ |
op_doi |
https://doi.org/10.3390/app11156959 |
container_title |
Applied Sciences |
container_volume |
11 |
container_issue |
15 |
container_start_page |
6959 |
_version_ |
1774717082307395584 |