Multi-fidelity Co-Kriging surrogate model for ship hull form optimization

Abstract

For the simulation-based hull form optimization design, there are many methods to evaluate the hydrodynamic performance of the hull form. Although the high fidelity of the surrogate model can be guaranteed by evaluating a large number of new sample hulls based on viscous flow theory, the computational cost can be too high. Therefore, in order to release the burden of calculation, based on the traditional single-fidelity Kriging surrogate model, the multi-fidelity Co-Kriging surrogate model gives attention to both high accuracy and efficiency by combining the accuracy advantage of high-fidelity sample evaluation with the efficiency advantage of low-fidelity sample evaluation. This paper first introduces the construction process of the multi-fidelity Co-Kriging surrogate model, and then uses a series of numerical examples to illustrate the advantages of the multi-fidelity Co-Kriging surrogate model compared with the single-fidelity Kriging surrogate model in terms of fidelity and efficiency. Finally, a hull form optimization design for total drag of DTMB-5415 hull at the design speed is given in detail, where the viscous flow theory and potential flow theory are used for the hydrodynamic evaluations of the hull forms to obtain the high- and low-fidelity results respectively. Results show that the multi-fidelity Co-Kriging surrogate model can be established for hull form hydrodynamic performance optimization, which is superior to the single-fidelity Kriging surrogate model in accuracy, and the optimal hull obtained by Co-Kriging surrogate model has a better resistance optimization effect.