Clustering Analysis on Internal Flow Field of a Spinning Solid Rocket Motor Nozzle

Clustering analysis was conducted on a solid rocket motor nozzle to investigate its internal flow field response under spinning condition. A Reynolds stress model and enhanced wall function were adopted to solve the Navier-Stokes equations with swirl flow. Wall temperature and tangential velocities distributions at three random locations of the divergent section, throat and convergent section were calculated and these data were clustered by a kind of hierarchical clustering method. Results show that rapid changes will occur when the spinning speed reaches some special points, such as 70rpm, 500rpm and 600rpm. Different physical quantities respond differently to both spinning speed and inner contour of the nozzle. Method established in this paper can be extended for other engineering fields such as analysis and prediction of fault diagnosis and failure behaviors. By combining data mining technology and traditional solid rocket motor design, design process can be improved.