| 초록 |
Computational fluid dynamics or CFD simulation models consist of a massive number of balance and constraint equations in general. It brings the complexity of simulations and lots of computational burdens. In order to solve this problem, data-driven modeling approaches such as deep neural network techniques are applied to construct a surrogate model. The data are CFD simulation results of a LNG vaporizer. Compared to the original CFD model, the surrogate model has a more straightforward explicit formulation, and its computational loads decrease. Applications of the surrogate model are introduced, and its advantages and limitations are discussed in this presentation. |
