Constructing Kriging Surrogate Model using Simple Alternative Sampling Approach based on Many Data Points

摘要

This paper presents a development of regression Kriging model with sequential sampling to reproduce the original data such as wind tunnel test data or flow field data from a computational simulation. In this sequential sampling method, maximum discrepancy location between the original data and Kriging model is used to construct Kriging model. Examples of force and moment coefficients measured from wind tunnel test and pressure data from CFD analysis show that proposed method can model the real responses with a reduced number of data and sufficient accuracy.