Reaction mechanism and kinetic modelling of RDX/AP via a chemical reaction neural network

Energetic composite play a core role in the aerospace field. An in-depth understanding of their kinetic behavior is crucial for predicting their combustion and explosion characteristics. This work combines chemical reaction neural networks (CRNN) with thermogravimetric (TG) data. The CRNN models of single-component RDX and AP simulate the mass loss process of an RDX/AP composite, and a kinetic model applicable to energetic composite is established. The model predicted the mass loss values of RDX/AP at each stage, which were consistent with the experimental results. In the first stage of mass loss of RDX/AP, the primary process is the decomposition of RDX, with a small amount of AP catalytically accelerating the rapid thermal decomposition of RDX, resulting in rapid mass loss in the first stage. In the second mass loss stage, the substances generated from the thermal decomposition of RDX as catalytic materials significantly promoted the thermal decomposition of AP and accelerated the overall decomposition rate. The two components have a coupling effect during the thermal decomposition process. This model also well predicted the thermal weight loss process of RDX/AP at different ratios, indicating good generality. This model not only deepens the understanding of the coupling effects between RDX/AP components but also holds promise for extending to the study of other energetic composite, opening new research pathways.