Deep Learning revealed statistics of the MgO particles dissolution rate in a CaO–Al2O3–SiO2–MgO slag

Abstract

Accelerated material development for refractory ceramics triggers enhanced possibilities in context to enhanced energy efficiency for industrial processes. Here, the gathering of comprehensive material data is essential. High temperature-confocal laser scanning microscopy (HT-CLSM) displays a highly suitable in-situ method to study the dissolution kinetics within the slag over time. However, a major challenge concerns the efficient and accurate processing of the large amount of collected image data. Here, the application of encoder-decoder convolutional network (U-Net) for the fully automated evaluation of the particle dissolution rate, overcoming manual evaluation drawbacks and providing accurate, fast and, sufficient statistical information is introduced. The developed U-Net allows an automated diameter evaluation of the MgO particles' dissolution in the silicate slag from 15 HT-CLSM experiments at three experimental temperatures 1450, 1500, and 1550°C. Moreover, the model can be applied to particle tracking and identification in various domains.