TY  - JOUR
AU  - Vasić, Milica
AU  - Jantunen, Heli
AU  - Mijatović, Nevenka
AU  - Nelo, Mikko
AU  - Munoz Velasco, Pedro
PY  - 2023
UR  - http://rims.institutims.rs/handle/123456789/516
AB  - Finding a solution to the problem of the large buildup of coal ashes is a vital necessity. Although the use of coal
ashes in fired clay bricks has been thoroughly investigated, there is insuffcient information on their industrial
utilization and researchers do not agree on whether or not this addition improves the quality of the fnal
products. Therefore, a database has gathered 20 years of research containing key factors related to the quality of
the bricks (i.e., chemical composition, fring temperature, soaking time, open porosity, water absorption and
compressive strength). Then, random forest regression and artifcial neural networks (ANN) modeling were used
to separately predict the parameters concerning the quality of the fnal products. The overall conclusions were
that the compressive strengths were the highest when using fly ashes and that class F ashes were highly suitable
to be used in the brick industry as a replacement material for brick clay. In addition, the ANN models showed
higher coeffcients of determination and an overall better fit to the experimental data. By changing the chemical
makeup of the initial materials and their proportions, the particle size of the ashes, the firing temperature and
soaking time, as well as the size of a product, the created models can be used to estimate the quality of the brick
containing coal ash. That is crucial because the inconsistent chemical composition of ash is generally the main
obstacle to its utilization. The local sensitivity analysis revealed the highest influence of the content of the alkali
oxides in the initial clay on the fired clay bricks due to their fluxing effect. In the case of ash-clay bricks, the
decisive factors were the type of furnace used, the ashes’ class, the Na2O content in raw clay, and the K2O
introduced with the ash. The F class ashes containing about 2–3% of K2O and <5% of CaO gave the highest
compressive strength in bricks fred at 1000–1100 ◦C.
Additional analyzes were made for 50% pond ash and 50% clay bricks to test the best-suited model and fill in
the knowledge gap. The results obtained in this study are important for supporting the decision in the selection of
materials and process parameter values that will increase the quality of the ash-clay-fired bricks.
PB  - Elsevier Ltd
T2  - Journal of Cleaner Production
T1  - Influence of coal ashes on fired clay brick quality: Random forest regression and artifcial neural networks modeling
VL  - 407
DO  - 10.1016/j.jclepro.2023.137153
ER  - 

@article{
author = "Vasić, Milica and Jantunen, Heli and Mijatović, Nevenka and Nelo, Mikko and Munoz Velasco, Pedro",
year = "2023",
abstract = "Finding a solution to the problem of the large buildup of coal ashes is a vital necessity. Although the use of coal
ashes in fired clay bricks has been thoroughly investigated, there is insuffcient information on their industrial
utilization and researchers do not agree on whether or not this addition improves the quality of the fnal
products. Therefore, a database has gathered 20 years of research containing key factors related to the quality of
the bricks (i.e., chemical composition, fring temperature, soaking time, open porosity, water absorption and
compressive strength). Then, random forest regression and artifcial neural networks (ANN) modeling were used
to separately predict the parameters concerning the quality of the fnal products. The overall conclusions were
that the compressive strengths were the highest when using fly ashes and that class F ashes were highly suitable
to be used in the brick industry as a replacement material for brick clay. In addition, the ANN models showed
higher coeffcients of determination and an overall better fit to the experimental data. By changing the chemical
makeup of the initial materials and their proportions, the particle size of the ashes, the firing temperature and
soaking time, as well as the size of a product, the created models can be used to estimate the quality of the brick
containing coal ash. That is crucial because the inconsistent chemical composition of ash is generally the main
obstacle to its utilization. The local sensitivity analysis revealed the highest influence of the content of the alkali
oxides in the initial clay on the fired clay bricks due to their fluxing effect. In the case of ash-clay bricks, the
decisive factors were the type of furnace used, the ashes’ class, the Na2O content in raw clay, and the K2O
introduced with the ash. The F class ashes containing about 2–3% of K2O and <5% of CaO gave the highest
compressive strength in bricks fred at 1000–1100 ◦C.
Additional analyzes were made for 50% pond ash and 50% clay bricks to test the best-suited model and fill in
the knowledge gap. The results obtained in this study are important for supporting the decision in the selection of
materials and process parameter values that will increase the quality of the ash-clay-fired bricks.",
publisher = "Elsevier Ltd",
journal = "Journal of Cleaner Production",
title = "Influence of coal ashes on fired clay brick quality: Random forest regression and artifcial neural networks modeling",
volume = "407",
doi = "10.1016/j.jclepro.2023.137153"
}

Vasić, M., Jantunen, H., Mijatović, N., Nelo, M.,& Munoz Velasco, P.. (2023). Influence of coal ashes on fired clay brick quality: Random forest regression and artifcial neural networks modeling. in Journal of Cleaner Production
Elsevier Ltd., 407.
https://doi.org/10.1016/j.jclepro.2023.137153

Vasić M, Jantunen H, Mijatović N, Nelo M, Munoz Velasco P. Influence of coal ashes on fired clay brick quality: Random forest regression and artifcial neural networks modeling. in Journal of Cleaner Production. 2023;407.
doi:10.1016/j.jclepro.2023.137153 .

Vasić, Milica, Jantunen, Heli, Mijatović, Nevenka, Nelo, Mikko, Munoz Velasco, Pedro, "Influence of coal ashes on fired clay brick quality: Random forest regression and artifcial neural networks modeling" in Journal of Cleaner Production, 407 (2023),
https://doi.org/10.1016/j.jclepro.2023.137153 . .