Ultimate compressive strength prediction for stiffened panels by counterpropagation neural networks (CPN)

Stiffened Panels are important strength members in ship and offshore struct ures. A new method based on counterpropagation neural networks (CPN) is pro posed in this paper to predict the ultimate compressive strength of stiffen ed panels. Compared with two-parametric polynomial, this method can take mo re parameters into account and make more use of experimental data. Numerica l study is carried out to verify the validation of this method. The new met hod may find wide application in practical design.