Drying of refractory castables - How complex the modeling needs to be? A numerical and experimental study

Calcium aluminate cement (CAC)- bonded castables have numerous benefits when compared to monolithic refractories. Their main drawback is their long initial heating, when the physical and chemically-bonded water are released. If the mass flux of water to the ambient is lower than the vapor generation, the gas pressure can increase well above the material’s strength resulting in cracks and spalling of the ceramic lining. In order to avoid this, optimized heat up curves are needed, and one promising methodology by using numerical models. Most of these calculations are based on tools used for simulating Portland cement concrete structures on fire and they vary both in complexity and their basic assumptions. Thus, the main question that remains is how complex such models need to be in order to capture the fundamental aspects of this phenomenon. The present work aimed to solve this issue by implementing multiple numerical tools reported in the literature with distinct complexity levels and fundamental assumptions and by also conducting neutron tomography tests on a CAC-bonded castable. It was possible to see that the pressure values predicted by such methodologies are equivalent and the water content predicted is qualitatively similar within the models and with the experimental values, indicating that the simplest approach might be used for predicting the castables’ drying.