This Week in CFD

We begin 2021 Q4 with a lot of reading. No internet sound bites this week: V&V, quantum computing, physics and math, machine learning, and lots more. For students, do not leave this blog post until you’ve looked into the scholarships (for those getting started) and the jobs (for those now finishing) and the free software (for those in the middle). Shown here is a tiny version of this week’s MUST SEE image and an award-winning image at that. Created in Tecplot, it shows what goes on inside a thermocapillary liquid bridge.

Meshing

  • “Hex and Tet are what you call frenemies, they are friends but are secretly jealous of each other and green with envy.” And so begins a comparison of hex and tet finite elements. Two nits to pick.
    • A stated pro of hex elements is that “The mesh looks athletically [sic] pleasing to the eye.” This reminds me of a favorite aphorism: “Your opinion, while interesting, is irrelevant.” I love meshes more than the average guy but we have to stop glamourizing their appearance. Just like we have to stop body-shaming geometry models.
    • The choice needn’t be either or. But sometimes our friends on the solver side force the mesh to conform to their choices.
    • I also am curious as to why the author chose to anthropomorphize hex and tet and females.
    • They surely meant “aesthetically.”

CFD for…

Reading

  • I can get kinda preachy about verification and validation. NAFEMS offers this article on an Introduction to V&V with comments from an expert panel. Users of commercial off-the-shelf (COTS) software aren’t immune: “it’s appropriate for the organization that buys the license to do their own code verification testing.” The paper Is the Problem with the Mesh, the Turbulence Model, or the Solver? provides the interesting insight that not only is the blame evenly distributed the user is a contributor to the problem too.
  • Would you pay US$270,000 for this wind tunnel? The Aero-Optim targets cycling but there could be other uses. Maybe you could combine it with this pressure rake with built-in, pressure-sensitive LED lighting/
  • Research Rabbit claims to be like Spotify but for technical papers. [Given that I think Spotify is destroying the music industry by paying artists micro cents per play, I’m not certain that’s a good analogy.] Good thing their logo is a mesh.
  • The PLM Atlas Directory is “the most comprehensive directory, classification taxonomy, relationships map, and history of the global ecosystem of Product Lifecycle Management (PLM) related technologies, solution markets, software products, service providers, consultants, user industries, professional organizations, events, and resources.”
  • This Twitter thread shares must-read physics and math books that are freely available for download as PDFs. [I have read precisely one: Carlo Rovelli’s Seven Brief Lessons on Physics.]
  • This one I’ll have to read: Machine Learning Based Optimal Mesh Generation in CFD.
It’s important to keep in mind that Cadence had CFD before acquiring Numeca and Pointwise. Here’s a video introduction to the CFD capabilities in the Celsius thermal solver for a printed circuit board. An easy way to think about it is that Cadence is expanding our computational software expertise from EDA to what I’ll call mechanical CAE. Other CAE providers are doing the same thing but in the other direction: expanding from MCAE to EDA.

More Reading

  • Do I really want to add Quantum Computation of Fluid Dynamics to my reading list? Maybe one of you can read it and provide a Reader’s Digest version in the comments. [What percentage of our readership knows what Reader’s Digest is and what the implications are?]
  • Same question [the first one, not the rhetorical second one] about Machine Learning for Fluid Mechanics from the Annual Review of Fluid Mechanics.
  • This unsolicited advice targets astronomers but is probably applicable to CFD people too.
  • CFD SEMI is building “a new generation hardware accelerator that allows us to democratize LES computations” based on DG methods, ASICs with high-bandwidth RAM, and linear scaling of thousands of these ASICs.
  • Rescale offers the ebook [e-book? E-Book? Ebook?] HPC for R&D. (registration required)
I’ve been playing with CFD Insights, a dashboard from Astrid Walles’ CFDsolutions, for “design space exploration.”

Software

Taken without context from a tweet by Devender Singh (@DevSDhami).

Events

Jobs and Scholarships

Full Circle

Let’s close how we started. Here’s the full-size version of the image that received the Best Picture in Physical Science award. Frank Muldoon, and his colleague, Hendrick Kuhlman, at the Vienna University of Technology, produced this visualization of their research results using Tecplot. Read the article at Tecplot for the details on what this image illustrates. Just beautiful.

Award-Winning Image of Invariant Streamtubes of a Hydrothermal Wave in a Thermocapillary Liquid Bridge. Image from Tecplot.com. See link above.
This entry was posted in Applications, Events, Hardware, News, Software and tagged , , , , , , , , , , , , , , , , , , , , , , , , , , , . Bookmark the permalink.

4 Responses to This Week in CFD

  1. Jan says:

    Hi, seems the link to the ‘Machine Learning Based Optimal Mesh Generation in CFD’ pdf is wrong. Besides thanks for these, as always very interesting selection of topics!

  2. Gabriel says:

    Hi,
    The link for “Machine Learning Based Optimal Mesh Generation in CFD” seems to be broken. I believe the correct link would be: https://arxiv.org/abs/2102.12923

Leave a Reply