(625f) Thermo-Mechanical Modeling of Melt Casting of Explosives

Melt casting is one of the most popular and inexpensive approach as for explosives production. However, this method has a large number of limitations, such as component segregation, non-uniformity of the cast product, the introduction of porosity, separation between the mold and the casting, voids and cracks formation. A numerical model has been developed for the analysis of the fluid flow, heat transfer, and stress fields involved in the casting process. The explosive composition consisted meltable explosives including TNT and DNAN, and contained RDX which remained crystalline during casting. The model was built using commercially available FLUENT and ANSYS software, coupled dynamically. An enthalpy method was employed to model the solidification process of explosives. While the melt explosive material contains a large percentage of solid particles, it was considered as incompressible and Newtonian fluid and the variation of density with temperature is described by the Boussineq approximation. The coupled thermo-mechanical model was used for the prediction and suppression of defects during the casting of energetic materials in various molds. Successful numerical modeling was built upon a complete material property database. Modeling transport phenomena in melt casting and predicting mechanical characteristics of explosives uses properties such as thermal conductivity, specific heat, thermal expansion coefficient, liquid viscosity, and stress-strain relationships. The thermo-physical properties database of studied explosives was experimentally measured using Diamond TMA, Diamond DSC, Hot Disk apparatus and Brookfield viscometer. A comprehensive numerical investigation was conducted to improve the quality of cast explosives, by optimizing and controlling the processing conditions. The computational results showed that the thermal boundary conditions are critical not only for the solid front movement and void formation, but also in determining residual thermal stresses in the final product. Good agreement was observed between the model prediction of the size and location of separation and the experimental observation.