| About this Abstract |
| Meeting |
MS&T23: Materials Science & Technology
|
| Symposium
|
Ceramics and Glasses Modeling by Simulations and Machine Learning
|
| Presentation Title |
Using Deep Learning to Develop a Smart and Sustainable Cement Manufacturing Process |
| Author(s) |
Aditya Kumar, Taihao Han, Jardel P. Gonçalves, Gaurav Sant, Narayanan Neithalath |
| On-Site Speaker (Planned) |
Aditya Kumar |
| Abstract Scope |
Cement manufacturing is widely recognized for its harmful impacts on the natural environment. The improvement in sustainability can be achieved by optimizing the cement manufacturing process, which encompasses refining the manufacturing parameters and the phases present in cement clinkers. In this study, a smart manufacturing process is developed to optimize manufacturing parameters and enhance the quality of cement clinkers. Manufacturers simply need to provide chemical compositions of their raw materials to the deep learning (DL) model, which in turn, optimizes the ratio of each component and calcination temperature, and predicts the phase composition of clinker. To train the DL model, thermodynamic simulations are employed to generate a database that encompasses a broad range of chemical compositions for raw materials and calcination temperatures. Utilizing the outcomes from the DL model, the optimal composition domains that produce high-quality clinker ((CaO)3(SiO2)>50%) at varying calcination temperatures are determined. |