Integration of Second-Order Sensitivity Method and CoKriging Surrogate Model

The global exploring feature of the surrogate model makes it a useful intermedia for design optimization. accuracy is closely related with efficiency optima-search. cokriging approach described in present studies can significantly improve and cut down turnaround time spent on modeling process. Compared to universal Kriging method, method interpolates not only sampling data, but also their associated derivatives. However, derivatives, especially high order ones, are too computationally costly be easily affordable, forming bottleneck application derivative enhanced methods. Based sensitivity analysis Navier–Stokes equations, current study introduces low-cost compute high-order making derivatives practically achievable. For methodological illustration, second-order regression correlation models proposed. A model-based optimization tool was developed tested optimal an automotive engine cooling fan. This improves modern proposes novel direction large scale problems.