| Issue |
MATEC Web Conf.
Volume 290, 2019
9th International Conference on Manufacturing Science and Education – MSE 2019 “Trends in New Industrial Revolution”
|
|
|---|---|---|
| Article Number | 02002 | |
| Number of page(s) | 10 | |
| Section | Management, Modelling and Monitoring of Manufacturing Processes | |
| DOI | https://doi.org/10.1051/matecconf/201929002002 | |
| Published online | 21 August 2019 | |
Methods to identify time series abnormalities and predicting issues caused by component failures
University Politehnica of Bucharest, Robots and Production System Department, 313 Splaiul Indepedentei, Bucharest, Romania
* Corresponding author: florinachiscop@gmail.com
Anomaly detection is a crucial analysis topic in the field of Industry 4.0 data mining as well as knowing what is the probability that a specific machine to go down due to a failure of a component in the next time interval. In this article, we used time series data collected from machines, from both classes - time series data which leads up to the failures of machines as well as data from healthy operational periods of the machine. We used telemetry data, error logs from still operational components, maintenance records comprising historical breakdowns and replacement component to build and compare several different models. The validation of the proposed methods was made by comparing the actual failures in the test data with the predicted component failures over the test data.
© The Authors, published by EDP Sciences, 2019
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.