A method for the profile optimization of the primary barrier of a membrane-type LNG tank using surrogate modeling

Membrane-type cargo holds, known for their high capacity, are commonly used in LNG carriers and are composed of various insulating materials and primary and secondary barriers to provide good thermal insulation and structural strength. The primary barrier, directly in contact with the liquefied cargo, is designed with a corrugated profile to withstand cryogenic thermal loads, cargo pressure, and sloshing loads. In this study, a novel method for developing the optimal profile of the primary barrier in a membrane-type LNG tank is proposed. The profile is defined using a third-order Bezier curve to allow flexibility in design, and tensile tests for SUS304L, cargo pressure calculations based on IGC code, and sloshing loads derived from model tests are used for non-linear structural analysis. Global optimization of the primary barrier profile is performed using the genetic algorithm NSGA-II, applying surrogate models as objective functions. Various surrogate models, including traditional methods and DNN, are evaluated to identify the most suitable model for optimization. The proposed method was compared and analyzed with the primary barrier geometry of the MARKIII FLEX cargo containment system (CCS) of a 174K LNG carrier. The results show that the optimal primary barrier can minimize longitudinal stress, permanent deformation, and production costs.