Semi-Autogenous Grinding Mill (SAG) Overload Forecasting Using Gram Penalized Matrices in a CNN

R. Hermosilla, H. Allende, C. Valle

2022

Abstract

In mining, detecting overload conditions is an opportunity to perform SAG mill to optimal operating conditions. With this in mind, several authors have prove using machine learning mechanisms to detect overloads. Our proposal establishes and tests a series of techniques to detect and forecast these events. Finally, we will look for an explanation of what the model considers for classification improving the phenomenon knowledge. Inspired by previous work and how operators classify overloads by analyzing behavior graphs of the most relevant variables, we proposed a framework that includes selection, encoding, and filtering improvement to finally discover the importance of the characteristics observed by our model using explanation techniques. Thus, using a group of novelty techniques, our advances exceed the results presented by other authors and by ourselves in previous publications, opening the door to a model based on attention in the future.

Paper Citation

in Harvard Style

Hermosilla R., Allende H. and Valle C. (2022). Semi-Autogenous Grinding Mill (SAG) Overload Forecasting Using Gram Penalized Matrices in a CNN. In Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC; ISBN 978-989-758-622-4, SciTePress, pages 391-396. DOI: 10.5220/0011946800003612

in Bibtex Style

@conference{isaic22,
author={R. Hermosilla and H. Allende and C. Valle},
title={Semi-Autogenous Grinding Mill (SAG) Overload Forecasting Using Gram Penalized Matrices in a CNN},
booktitle={Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC},
year={2022},
pages={391-396},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011946800003612},
isbn={978-989-758-622-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC
TI - Semi-Autogenous Grinding Mill (SAG) Overload Forecasting Using Gram Penalized Matrices in a CNN
SN - 978-989-758-622-4
AU - Hermosilla R.
AU - Allende H.
AU - Valle C.
PY - 2022
SP - 391
EP - 396
DO - 10.5220/0011946800003612
PB - SciTePress