Hybrid Optimization of Curvilinearly Stiffened Shells Using Parallel Processing

Abstract

With the advances being made in additive manufacturing, it is becoming increasingly possible to fabricate a broad class of complex-shaped designs for practical applications. This manufacturing capability has allowed structural designers to implement the use of bio-inspired curvilinear stiffeners for achieving better designs of stiffened plate and shell structures. Curvilinear stiffeners have proven to be useful over straight stiffeners in some applications for achieving better structural efficiency. A framework has been developed that employs a hybrid optimization technique of using particle swarm optimization for stiffener shape optimization and gradient-based optimization as implemented in MSC.NASTRAN SOL 200 for optimization of stiffener cross section and shell thickness. Parallel processing has been used to save extensive wall-clock time. The framework has been employed for optimally designing cylindrical shells/panels stiffened by arbitrarily placed stiffeners, the motivation being lighter-weight fuselage and space launch vehicle designs. Structural optimization results have been presented for the optimal design of stiffened shells with the objective of weight minimization for shells subjected to both buckling and stress constraints. Optimization studies show that both the stiffener placement and the stiffener geometric curvature influence the shell buckling load, which can help to decrease the structural weight by optimizing the cross-sectional dimensions for stiffeners and panel thickness. In this paper, cylindrical shells, stiffened by four stiffeners, placed arbitrarily, are studied under compression and shear load cases. It is seen that arbitrarily placed curvilinear stiffeners lead to a potential 13% weight saving as compared to the use of equally spaced straight stiffeners when designed for the compression load case.