Optimization design of helical carcass supported flexible cryogenic pipe based on BP-MOWOA

Abstract

In recent years, the demand for offshore liquefied natural gas (LNG) exploration and transportation has set higher performance standards for cryogenic pipes, particularly the helical carcass supported flexible cryogenic pipe (hc-FCP). Due to its multi-layered, spiral-wound, and non-bonded structure, optimizing the design of hc-FCP presents significant challenges. This paper proposes an optimization approach combining the backpropagation neural network and the multi-objective whale optimization algorithm (BP-MOWOA) to address the multi-objective optimization of hc-FCP structures. First, a high-precision finite element model of the hc-FCP was established and validated through experimental data. The model was used to generate a large dataset that captured the mechanical performance of hc-FCP under various operating conditions. Second, a backpropagation (BP) neural network was trained to predict the axial tensile and bending stiffness of the hc-FCP, forming the basis for the optimization process. Lastly, the BP-MOWOA was employed to optimize critical design parameters, such as the winding angles of the reinforcement layers, to achieve maximum axial tensile stiffness and reduced bending stiffness with minimal material consumption. The optimized design improved axial tensile stiffness by 24.35 %, reduced bending stiffness by 2.99 %, and lowered material consumption by 1.84 %. These results demonstrate the effectiveness of the BP-MOWOA in optimizing hc-FCP structures for enhanced performance and cost-efficiency, providing a flexible solution for engineering applications in varying operational conditions.