Prediction of the Rapid Hardening Property of Calcium Aluminate Cement Based on Mineralogical Composition by Neural Network
Öz
Quantitative X-ray diffractometry using a Rietveld-based computational method was carried out for a series of Calcium Aluminate Cement (CAC) samples. This indicated that the CA content ranged between 37.7% to 47.7% while Brownmillerite (C4AF) amount varies between 11.0% to 23.6%. Magnetite was found in all the samples, ranging from 0.7% to 3.9% while Gehlenite amount varies between 0.5% and 6.5%. The Spinel amount was between 0.1% to 1.3% with an average of 0.5%. The amorphous content of CAC is ranged between 12.0% and 32%. The Mayenite and amorphous content could be a good indicator of the Rapid Hardening (RH) property of CAC. Samples with the high Mayenite content showed less RH properties, whereas RH increased as the content of amorphous material increased. The RH properties of CAC based on its mineralogical composition was predicted through various neural network techniques. The R2-value of the models was 0.39 for Linear Regression analysis model (LR), 0.56 for feed forward neural network (ANN) and 0.78 for Generalized Regression Neural Network (GRNN) approaches. The best prediction approach for RH value of the CAC with an Al2O3 content of 40% was GRNN that can be applied to predict RH.
Destekleyen Kurum
Çukurova Üniversitesi
Proje Numarası
FYL-2018-9911
Teşekkür
Financial support from the Cukurova University Scientific Research Project Unit (Under Project No: (FYL-2018-9911) is gratefully acknowledged. The support of the Cimsa Cement Plant, Mersin, Turkey is also acknowledged.
Kaynakça
- Pöllmann H. 2012. Calcium aluminate cements - raw materials, differences, properties and hydration. Reviews in Mineralogy & Geochemistry, vol. 74, no 1, pp 1-82.
- Pöllmann H. 2001. Mineralogy and crystal chemistry of calcium aluminate cement. Calcium Aluminate Cements. Proceedings of International Conference, Edinburgh, Mangabhai R J and Glasser F P (Eds). London, IOM Communications, pp 79-119.
- Bensted J. 2002. Calcium aluminate cements, Structure and Performance of Cements, 2nd ed., (Eds.Bensted J, Barnes P), London.
- Ukrainczyk N. Šipušić J. Dabić P. Matusinović T. Microcalorimetric Study on Calcium Aluminate Cement Hydration, 2008. 13. International conference on Materials, Processes, Friction and Wear - MATRIB'08, Vela Luka, Croatia, pp 382-388.
- H.M. Rietveld, 1969. A profile refinement method for nuclear and magnetic structures, J. Appl. Cryst. 2, pp 65-71.
- Svozil, D., Kvasnicka, V., & Pospichal, 1997. J. Introduction to multi-layer feed-forward neural networks. Chemometrics and intelligent laboratory systems, 39(1), pp 43-62.
- Hagan, M.T., and M. Menhaj, 1999. Training feed-forward networks with the Marquardt algorithm, IEEE Transactions on Neural Networks, Vol. 5, No. 6, pp 989–993.
Prediction of the Rapid Hardening Property of Calcium Aluminate Cement Based on Mineralogical Composition by Neural Network
Öz
Quantitative X-ray diffractometry using a Rietveld-based computational method was carried out for a series of Calcium Aluminate Cement (CAC) samples. This indicated that the CA content ranged between 37.7% to 47.7% while Brownmillerite (C4AF) amount varies between 11.0% to 23.6%. Magnetite was found in all the samples, ranging from 0.7% to 3.9% while Gehlenite amount varies between 0.5% and 6.5%. The Spinel amount was between 0.1% to 1.3% with an average of 0.5%. The amorphous content of CAC is ranged between 12.0% and 32%. The Mayenite and amorphous content could be a good indicator of the Rapid Hardening (RH) property of CAC. Samples with the high Mayenite content showed less RH properties, whereas RH increased as the content of amorphous material increased. The RH properties of CAC based on its mineralogical composition was predicted through various neural network techniques. The R2-value of the models was 0.39 for Linear Regression analysis model (LR), 0.56 for feed forward neural network (ANN) and 0.78 for Generalized Regression Neural Network (GRNN) approaches. The best prediction approach for RH value of the CAC with an Al2O3 content of 40% was GRNN that can be applied to predict RH.
Anahtar Kelimeler
CAC GRNN neural network rapid hardening Rietveld CAC, GRNN, neural network, rapid hardening, Rietveld
Proje Numarası
FYL-2018-9911
Kaynakça
- Pöllmann H. 2012. Calcium aluminate cements - raw materials, differences, properties and hydration. Reviews in Mineralogy & Geochemistry, vol. 74, no 1, pp 1-82.
- Pöllmann H. 2001. Mineralogy and crystal chemistry of calcium aluminate cement. Calcium Aluminate Cements. Proceedings of International Conference, Edinburgh, Mangabhai R J and Glasser F P (Eds). London, IOM Communications, pp 79-119.
- Bensted J. 2002. Calcium aluminate cements, Structure and Performance of Cements, 2nd ed., (Eds.Bensted J, Barnes P), London.
- Ukrainczyk N. Šipušić J. Dabić P. Matusinović T. Microcalorimetric Study on Calcium Aluminate Cement Hydration, 2008. 13. International conference on Materials, Processes, Friction and Wear - MATRIB'08, Vela Luka, Croatia, pp 382-388.
- H.M. Rietveld, 1969. A profile refinement method for nuclear and magnetic structures, J. Appl. Cryst. 2, pp 65-71.
- Svozil, D., Kvasnicka, V., & Pospichal, 1997. J. Introduction to multi-layer feed-forward neural networks. Chemometrics and intelligent laboratory systems, 39(1), pp 43-62.
- Hagan, M.T., and M. Menhaj, 1999. Training feed-forward networks with the Marquardt algorithm, IEEE Transactions on Neural Networks, Vol. 5, No. 6, pp 989–993.
Ayrıntılar
| Birincil Dil | İngilizce |
|---|---|
| Konular | Maden Mühendisliği |
| Bölüm | Araştırma Makaleleri |
| Yazarlar | |
| Proje Numarası | FYL-2018-9911 |
| Yayımlanma Tarihi | 10 Haziran 2022 |
| Yayımlandığı Sayı | Yıl 2022 Cilt: 55 Sayı: 1 |
