MULTI-OBJECTIVE OPTIMIZATION DESIGN OF FLEXIBLE RISERS BASED ON BI-SCALE RESPONSE SURFACE MODELS

An optimized flexible riser design not only requests that the stress of local cross-section shouldn't exceed the allowable strength, but also can be compliant with the floater to improve the fatigue life. It should be particularly pointed out that the flexible riser is a typical multi-scale system, which consists of the local cross-sectional scale and the global configuration scale, which differentiates each other a lot from their geometrical scales. A bi-scale response surface model is established to perform the optimized design of flexible risers by considering the parameters of local cross-sections and global configurations simultaneously. The response surface model can be an effective surrogate model to integrate the local and global responses into one loop so that the computational efficiency can be increased significantly. In the bi-scale response model, design variables of a flexible riser are extracted and defined at both the local sectional scale and global configuration scale. Sensitivity analyses of the two objectives, ultimate tension and bending strength on the design variables are then deduced to establish the bi-scale optimization framework through the response surface methodology. Finally, the optimization framework is implemented on a flexible riser with lazy-wave configuration which is considered as a case study. The properties of the optimized flexible risers are compared with those without the optimization. It is found that the ultimate load bearing capacity and fatigue life of the optimized flexible riser are improved significantly. Moreover, the feasibility and effectiveness of the bi-scale optimization strategy are verified through numerical simulations, which indicates that the bi-scale response surface optimization methodology provides a new thought and approach to explore the design potential of flexible risers.