OPTIMUM DESIGN OF LAMINATED COMPOSITE STIFFENED PANELS

Abstract

Stiffened panels are generic structural elements in weight sensitive structural, aerospace and marine applications. These panels are becoming increasingly used in structural applications because of their high specific stiffness and specific strength. Laminated composites, which are non homogeneous and anisotropic/orthotropic, earlier limited to aerospace applications, are gradually being applied in structural applications. The use of laminated composite provides flexibility to tailor different properties of the structural elements to achieve strength and stiffness requirements. Stiffened panel is an appropriate structural form to utilize the advantages of laminated composites efficiently. Being a thin-walled structure the design of laminated composite stiffened panels is mainly governed by stability constraints. Mounting use of laminated composite stiffened panels in various structural applications such as long span bridge decks, ship hulls, etc., demands better understanding of the buckling behaviour. The stability of laminated composite stiffened panels is a current research area due to its complex buckling behaviour and rigorous design calculations. Additionally the design of laminated composite stiffened panels is more complex and prohibits human intuitions due to the involvement of large number of variables. Therefore, the use of artificial intelligence for analysis coupled with optimization procedures should be considered in improving and for obtaining economical designs of laminated composite stiffened panels. In the present work parametric studies are carried out to study the buckling behaviour of laminated composite stiffened panels subjected to in-plane shear, linearly varying in-plane compressive loading and combination of above loadings using FEA. Studies are carried out by changing the panel orthotropy ratio, stiffener depth, pitch length (number of stiffeners), smeared extensional stiffness ratio of stiffener to that of the plate and extensional stiffness to shear stiffness ratio of the plate. Based on the studies database is generated and few important parameters influencing the buckling behaviour are identified and guidelines for better stiffener proportioning are developed, which will be helpful for the designer. The FEA generated database is used to develop the computationally efficient analysis approaches in the form of neural networks. ANNs for predicting the buckling load are developed for in-plane shear and linearly varying in-plane compressive loading separately. It is observed that buckling load prediction by ANN takes fraction of second while the corresponding FE analysis takes five to twenty minutes. This proves the computation efficiency of developed artificial neural networks. Additionally the interaction equations are validated for laminated composite stiffened panels subjected to combined shear and linearly varying in-plane compressive loadings. The ANN based analysis tools are incorporated in GA based optimum design procedure. Numerical studies are carried out using this approach and the final results are reanalyzed using FEA to prove the adequacy of the developed analysis approaches. In nut shell the present work is carried out to understand the buckling behaviour of laminated composite stiffened panels with the help of exhaustive numerical studies for different states of in-plane stresses. Additionally, ANNs are trained to develop computationally efficient analysis procedures for GA based optimization studies. in