Modeling of aluminum agglomeration at solid propellant burning surface by combining aluminum ignition with LDP-SC cluster analysis

Aluminum particles at burning surface may agglomerate during the combustion process of composite solid propellant, involving the aggregation and ignition of aluminum particles as well as the secondary coalescence of agglomerates. So far it is still a huge challenge to capture the agglomeration process and obtain the size distribution of agglomerates accurately. Modeling of aluminum agglomeration was carried out by constructing several submodels, including propellant combustion, cluster analysis of aluminum particles, aluminum ignition and coalescence based on the packed three-dimensional propellant microstructure unit. Dynamic combustion process of propellant microstructure unit was simulated firstly, which characterizes the temperature distribution above the burning surface. Next, aluminum clusters were grouped using a new local density peaks-spectral clustering (LDP-SC) algorithm for initial aluminum particles in microstructure unit. The ignition of aluminum particles in clusters exposed from burning surface was then examined by considering their distributions in the flame of propellant. These exposed aluminum particles in a cluster yield an agglomerate once an aluminum particle is successfully ignited. Secondary mergence of multiple agglomerates was also considered when the distance between two particles in different agglomerates was less than a critical separation distance. With these submodels, the agglomeration process of a typical propellant and properties of aluminum clusters and agglomerates were estimated. The predicted size distribution of agglomerates shows reasonable agreement with experimental data, indicating that the present model is capable of predicting aluminum particle agglomeration during composite propellant combustion.