AUTOMATED MODEL FITTING FOR METALLURGICAL AND OTHER INDUSTRIES

1 GAUDE-FUGAROLAS Daniel
Institution:
1 dgaude Prime Innovation SLU, c/. Alcalde Joan Batalla, 4, 08340 Vilassar de Mar (Spain), EU, dgaude@cantab.net
Conference:
26th International Conference on Metallurgy and Materials, Hotel Voronez I, Brno, Czech Republic, EU, May 24th - 26th 2017
Proceedings:
Proceedings 26th International Conference on Metallurgy and Materials
Pages:
1997-2003
ISBN:
978-80-87294-79-6
ISSN:
2694-9296
Published:
9th January 2018
Proceedings of the conference were published in Web of Science and Scopus.
Metrics:
437 views / 191 downloads
Abstract

One of the aspects defining the paradigm sometimes called Industry 4.0 is the availability of large amounts of data generated by modern processes. Nowadays, most process parameters can be collected automatically and therefore, the large volume of process data generated becomes close to impossible to analyse manually, or at least, to be able to extract from it all the useful information that it contains.Advanced mathematical and computational tools to analyse data and create models automatically have existed for some time, being labeled at one time or another as Expert Systems, Artificial Intelligence, Machine Learning, Data Mining, Statistical Pattern Recognition etc. However, only recently the combination of automated data collection and fast computers have allowed to extend the benefits of such techniques beyond some niche applications and industries.This work introduces a new project to develop a comprehensive set of such tools aimed to fit industrial needs. As starter, several advanced model fitting techniques are applied to a metallurgical example. As dataset, a publicly available database on creep rupture is used. Supervised learning algorithms (using gradient descent training, normal equation regression, artificial neural networks), when correctly applied, allow to fit models to any set of complex data and to make reliable predictions with them. Unsupervised learning methods on the other hand, may also be used to find structure in the data without any a priori knowledge of such underlying structure (anomaly detection, clustering).

Keywords: Automated Model Fitting, Regression, Artificial Neural Networks, Anomaly Detection, Cluster Analysis

© This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Scroll to Top