An MINLP model for optimal sandwich pipe design based on structural and thermal criteria

Sandwich pipe (SP) can be a promising solution in ultra-deepwater applications. Over a decade of research has validated its commendable structural strength, superior thermal insulation, and minimal submerged weight. Due to the vast choices of the material and thickness for each layer of SPs, the design process should be tailored for a specific scenario. As an attempt to fully exploit the advantages of SPs, this work is dedicated to developing an optimization model. First, a cost model for SPs, including material cost, fabrication cost, and welding cost, was developed. Then, a thermal analysis model for single-phase flow in a multilayer pipeline was built. Along with the existing prediction equations of the collapse pressure of SPs, a mixed-integer nonlinear programming (MINLP) model was proposed for the optimal design of SPs. A case study representing a practical working condition was selected. The cost and configuration of the optimal SP were compared with those obtained for a single-wall pipe (SW) and a traditionally designed SP. The results indicated that the optimization model could decrease the cost of SP yield by existing design procedures by 32.7%-43.5%. With the verified optimization model, a parametric study that combines 540 practical subsea working conditions was carried out. Based on the results, insightful understandings on how to improve and design an SP were concluded. Furthermore, similar to the established optimization model, an optimization model of SWs was also developed. Finally, the cost ratios of the optimal SP and the optimal SW under different working conditions were calculated. The results show that for large-diameter pipelines, replacing the SW with the SP in deepwater applications could reduce the pipeline cost up to 75%.