Artificial Intelligence (AI)
AI applications have been increasingly extended to various fields while showing their superior ability. Our research group is pursuing utilizing recent AI techniques in cutting-edge turbomachinery and energy systems. For instance, high-fidelity computational approaches like Computational Fluid Dynamics (CFD) and Finite Element Analysis (FEA) require a substantial amount of computation time in rotordynamics analysis.
Here, the deep learning surrogate model can help to reduce the computation time, significantly. The representative surrogate models are Fourier Neural Operator (FNO) and Deep Operator Network (DeepONet), and these are pure data-driven methods. We have proposed various turbomachinery performance prediction models via neural operators based on physics data.
Also, we are interested in Physics-Informed Neural Network which can be a good choice if there is not enough data, and we have conducted the related research, as well. In addition, for the Smaller Modular Reactor (SMR), we have developed the optimal design methodology of the steam generator through machine learning and core start-up control strategies via reinforcement learning.
Research Interest in AI Area
Physics Informed Neural Network (PINN)
Fourier Neural Operator (FNO)
Deep Operator Network (DeepONet)
Reinforcement Learning (RL)