Effective lightweight design of a rocket interstage ring through mixed-integer optimization

Abstract

The full lightweight potential of carbon fiber reinforced plastics in complex structures can only be exploited using an adequate combination of material and design. Potential combinations can be found, using an optimization routine integrated into the design process. This paper presents a detailed comparison of the optimization results for distinctive rocket interstage design concepts. The complex cylindrical grid structure has to withstand the mission loads while being as light as possible. For the efficient use of the superior material properties, carbon fiber reinforced plastics are combined with a foam core to form a sandwich configuration. To minimize the total mass, parametric finite element models are used within a constrained, mixed-integer optimization. The structure is modeled using a bottom up strategy. An optimization framework, incorporating the CAE software Ansys into the multi-purpose optimization package pyOpt, is presented. The augmented lagrangian particle swarm algorithm is used to find the configurations having minimal weight. As manufacturability is an important design requirement, a winding process is represented by constraints in the parameter set of the geometry. The stiffnesses of the structure, as well as strength criteria of the laminate are used as inequality constraints for the optimization. Further on, a lower bound of the first natural eigenfrequency using additional constraints, is investigated. Finally, the impact of the governing geometry, material parameters and hence their influences on the optimization problem are carefully reviewed. It is shown that appropriate design – material combinations can significantly reduce the structural mass.