Surrogate-based structural optimization design of large-scale rectangular pressure vessel using radial point interpolation method

Abstract

Large-scale rectangular pressure vessel possesses wide application due to volume efficiency and the convenience of transporting and installing. Despite there being increasing emphasis on weight reduction, the optimization design method for large-scale rectangular pressure vessel is still under development due to complicated geometry and engineering constraints that make conventional optimization approaches incapable. Therefore, a surrogate-based optimization design method is proposed in this work to resolve the aforementioned difficulties. The proposed technique adopts radial point interpolation method to construct surrogate model, which is validated with classical numerical examples and a rectangular pressure vessel, i.e., the test section of a large-scale water tunnel. Fair convergence and precision have been revealed. Then, the surrogate-based optimization framework is established on the basis of the surrogate model, involving optimization scheme, sampling range and constraints, sampling method, standard optimization model and method. Subsequently, the investigated rectangular pressure vessel is optimized using the developed method. The optimized structure maintains good machinability and reduces weight by 10.06% in comparison to the results of preliminary routine structural design, proving that the developed technique possesses good capacity for the optimization design of large-scale rectangular pressure vessels.