AI/ML Integration in CFD and Surrogate Modeling

Computational fluid dynamics (CFD) remains central to aerospace engineering, but its high computational cost limits rapid design and optimization. The integration of artificial intelligence and machine learning (AI/ML) into CFD workflows offers a powerful alternative through data-driven surrogate models that approximate high-fidelity simulations at lower cost. This chapter examines the foundations, architectures, and aerospace applications of AI/ML-enhanced CFD, including physics-informed neural networks, convolutional neural networks, Fourier neural operators, graph neural networks, generative adversarial networks, and deep reinforcement learning. Particular attention is given to turbulence modeling, aerodynamic shape optimization, flow reconstruction, and transfer learning. Challenges related to data scarcity, generalizability, and physical consistency are discussed alongside emerging benchmarks and future research directions. The chapter argues that AI/ML not only accelerates CFD workflows but also reshapes aerospace design methodologies.

Hali tarjima qilinmagan