Radial compression pressure estimation of carcass layers in unbonded flexible pipes based on neural networks

Abstract

Radial pressure resistance is a key factor in the design of unbonded flexible pipe carcass layers because it directly affects the safe installation and use of marine flexible pipes. It is challenging to evaluate the performance of radial compression because the carcass layer, which is a large-angle helical winding structure in innermost flexible pipes with a special-shaped cross section, experiences numerous contact frictions under radial pressure. In this study, a radial pressure estimation approach based on neural networks is proposed to address the aforementioned issues. This study explores the nonlinear relationship between the limited structural responses (circumferential strain) and the overall load (radial compression pressure) from the radial compression experimental data and estimates the radial compression pressure of carcass layers. The input variables are the structural strains, and the output variable is the radial pressure. Furthermore, some examples are used to test the generalization of the method, including robustness and understanding, considering the geometric characteristics of the carcass layer. The results demonstrate that the method can accurately predict the radial pressure applied to the carcass layer and provide a theoretical framework for the design of unbonded flexible pipes in deep-sea environments.