Optimization of micronizing zeolite grinding using artificial neural networks

Loading...
Thumbnail Image

Date

2024

Authors

Nikolić, Vladimir
Trumić, Milan
Tanikić, Dejan
Trumić, Maja

Journal Title

Journal ISSN

Volume Title

Publisher

University of Belgrade, Technical faculty in Bor

Source

Journal of Mining and Metallurgy, Section A

Volume

60

Issue

1

Abstract

The micronizing grinding of natural zeolite, of the clinoptilolite type, was investigated in a ring mill. The aim of the experiment was to determine the optimal grinding conditions to obtain a powder with appropriate physico-chemical and microstructural characteristics that would find its potential application as a binder and ion exchanger in structural composites. The analysis of specific size classes of zeolite e after micronization was performed by grinding kinetics. The research was carried out on previously prepared zeolite samples, on wider and narrower size classes (-3.35 + 0 mm;-3.35 + 2.36 mm; -2.36 + 1.18 mm; -1.18 + 0 mm) and different starting masses (50 g, 100 g, 200 g). Fine grinding was carried out at different time intervals (20 s, 45 s, 75 s, 120 s, 300 s, 900 s). A sieve analysis was performed on the grinding products, the content of the size class (-5 + 0) μm and the specific surface area of these products were determined. XRD analysis was performed on individual grinding products to take into account possible changes in the zeolite material itself. Based on the results obtained, an artificial neural network was developed and then compared with the experimental results. The artificial neural network models have achieved a satisfactory prediction accuracy (0.989 - 0.997) and can be considered accurate and very useful for the prediction of variable responses.

Description

Keywords

zeolite, micronizing grinding, specific surface, artificial neural networks.

Citation

DOI

10.5937/JMMA2401023N

Scopus

ISSN

1450-5959
2560-3159

ISBN

License

CC-BY-SA

Collections