The Yuan requests your support! Our content will now be available free of charge for all registered subscribers, consistent with our mission to make AI a human commons accessible to all. We are therefore requesting donations from our readers so we may continue bringing you insightful reportage of this awesome technology that is sweeping the world. Donate now
AI aids fluid flow simulations
By Ricardo Vinuesa  |  Aug 23, 2022
AI aids fluid flow simulations
Image courtesy of and under license from
AI use is becoming ever more feasible in computational fluid dynamics, where it has already yielded promising results. Different computational means have their unique advantages and drawbacks, but it is not realistic to expect AI to fully replace these, as it is more likely to serve as a complement to them.

STOCKHOLM - Computational fluid dynamics (CFD) is the field focused on the use of computers to solve the governing equations of fluid flows. These governing equations are the so-called Navier-Stokes (NS) equations, which enable the determination of the flow velocity and pressure as a function of spatial location and time. These NS equations are partial differential equations (PDEs) which represent conservation of mass and momentum in the fluid flow under consideration.

These PDEs do not have general analytical solutions, and it is very challenging to solve them numerically, especially as the Reynolds number - the ratio of inertial and viscous effects in the flow - increases (i.e., as the conditions closer to those in industrial applications are approached). The NS equations can thus be integrated numerically via DNS (direct numerical simulation), an approach that is highly accurate but that also comes with high computational costs. They can also be solved with a certain degree of approximation via large-eddy simulation (LES), in which the largest scales are simulated, and the smaller ones are modeled (i.e., their behavior is assumed beforehand). LES has lower computational costs than DNS, and an even lower cost can be achieved through Reynolds-averaged Navier-Stokes (RANS) simulations. In the latter approach, Reynolds averaging is first used to divide the instantaneous flow velocities into mean and fluctuating components. Then, certain models are used to relate the so-called Reynolds stresses (basically the turbulent-fluctuation components) with other flow quantities. While RANS has the lowest computational cost and is widely used in industrial applications, there are problems in which its performance is limited. In certain applications involving flow control and/or optimization, e.g., it may be interesting to develop simplified versions of the original flow, usually by relying on projecting its temporal dynamics onto a different coordinate s

The content herein is subject to copyright by The Yuan. All rights reserved. The content of the services is owned or licensed to The Yuan. Such content from The Yuan may be shared and reprinted but must clearly identify The Yuan as its original source. Content from a third-party copyright holder identified in the copyright notice contained in such third party’s content appearing in The Yuan must likewise be clearly labeled as such.
Continue reading
Sign up now to read this story for free.
- or -
Continue with Linkedin Continue with Google
Share your thoughts.
The Yuan wants to hear your voice. We welcome your on-topic commentary, critique, and expertise. All comments are moderated for civility.