Soft Sensor for Online Cement Fineness Predicting in Ball Mills

DOI

10.2991/aer.k.201221.069How to use a DOI?

Keywords

Cement fineness, artificial neural network, ball mill, product quality

Abstract

The cement fineness is a determining factor in product quality. Estimating this variable in real-time can be extremely useful to maintain the desired characteristics of the product during the cement grinding process, which will also allow a significant increase in the system energy efficiency. This paper describes the design and implementation of a soft sensor based on a backpropagation neural network model to predict the cement fineness online in a ball mill. The input variables of these models were selected by studying the cement grinding process, applying Spearman’s rank correlation, and the mutual information (MI) algorithm. The fineness results of laboratory tests were collected to obtain the output variable and for training the models. The procedure of extracting, analyzing, treating, and cleaning raw data received from the factory and the intensified hyperparameter adjustment of the predicting model provided excellent soft sensor performance. The developed system was tested in a cement grinding process and demonstrated the ability to provide information about the variables previously obtained only through offline laboratory tests.

Copyright

© 2020, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Karina Andreatta
AU  - Filipe Apóstolo
AU  - Reginaldo Nunes
PY  - 2020
DA  - 2020/12/22
TI  - Soft Sensor for Online Cement Fineness Predicting in Ball Mills
BT  - Proceedings of the International Seminar of Science and Applied Technology (ISSAT 2020)
PB  - Atlantis Press
SP  - 422
EP  - 428
SN  - 2352-5401
UR  - https://doi.org/10.2991/aer.k.201221.069
DO  - 10.2991/aer.k.201221.069
ID  - Andreatta2020
ER  -