Utilize este identificador para referenciar este registo:
https://hdl.handle.net/1822/74067
Título: | A comparison of anomaly detection methods for industrial screw tightening |
Autor(es): | Ribeiro, Diogo Aires Gonçalves Matos, Luís Miguel Cortez, Paulo Moreira, Guilherme Pilastri, André Luiz |
Palavras-chave: | Autoencoder Deep Learning Industry 4.0 Isolation Forest One-class classification Random Forest Unsupervised learning |
Data: | Set-2021 |
Editora: | Springer |
Revista: | Lecture Notes in Computer Science |
Citação: | Ribeiro, D., Matos, L. M., Cortez, P., Moreira, G., & Pilastri, A. (2021). A Comparison of Anomaly Detection Methods for Industrial Screw Tightening. In International Conference on Computational Science and Its Applications (pp. 485-500). Springer |
Resumo(s): | Within the context of Industry 4.0, quality assessment pro- cedures using data-driven techniques are becoming more critical due to the generation of massive amounts of production data. In this paper, we address the detection of abnormal screw tightening processes, which is a relevant industrial task. Since labeling is costly, requiring a manual effort, we focus on unsupervised approaches. In particular, we assume a low-dimensional input screw fastening approach that is based only on angle-torque pairs. Using such pairs, we explore three main unsuper- vised Machine Learning (ML) algorithms: Local Outlier Factor (LOF), Isolation Forest (iForest) and a deep learning Autoencoder (AE). For benchmarking purposes, we also explore a supervised Random Forest (RF) algorithm. Several computational experiments were held by us- ing recent industrial data with 2.8 million angle-torque pair records and a realistic and robust rolling window evaluation. Overall, high quality anomaly discrimination results were achieved by the iForest (99%) and AE (95% and 96%) unsupervised methods, which compared well against the supervised RF (99% and 91%). When compared with iForest, the AE requires less computation effort and provides faster anomaly detection response times. |
Tipo: | Artigo em ata de conferência |
URI: | https://hdl.handle.net/1822/74067 |
ISBN: | 978-3-030-86959-5 |
e-ISBN: | 978-3-030-86960-1 |
DOI: | 10.1007/978-3-030-86960-1_34 |
ISSN: | 0302-9743 |
Versão da editora: | https://link.springer.com/chapter/10.1007/978-3-030-86960-1_34 |
Arbitragem científica: | yes |
Acesso: | Acesso aberto |
Aparece nas coleções: |
Este trabalho está licenciado sob uma Licença Creative Commons