Applications
- Argonne National Labs shares this video of a 50-million cell, high fidelity CFD simulation of a diesel engine. You can read more about it here.
- The Maritime Research Institute of the Netherlands has a new 4,000 core cluster on which customers can run their CFD code ReFRESCO.

This Ohio State Univ. designed car set a land speed record of slightly over 212 mph. Image from International Science Grid This Week. Click image for article.
Hardware
- How would you like to have 36 cores on your desktop? You can with the new Intel Xeon E5 v3 if you have two.
- The Call for Papers is open for Nvidia’s 6th GPU Technology Conference.
- What CAD/CAE/PLM providers made the list of the top 500 cloud apps? Autodesk was ranked highest at 120.
- If you like your CAD free and in the cloud, try the beta of Honeycomb3D. It’s designed for 3D printing.
Visualization
- Two nice collections of resources from Visualizing Data:
- 273 data visualization resources
- best of the visualization web for July 2014
- The ParaView User’s Guide is getting reworked for the 4.2 release.
- Tecplot RS 2014 R1 is now available for visualization of oil and gas reservoir simulations.

This oil saturation plot is an example of one new feature available in Tecplot RS 2014 R1. Image from Tecplot.
Things to Read
- Siemens PLM opens the discussion on CAD/CAE integration by saying it’s the key to effective simulation driven design, especially early in the design process. Do you agree? Watch for follow-up posts on this subject including the people, process, and technology issues. [This article is the first time I’ve ever encountered the word “centisecond.”]
- It’s nice to see Prof. ZJ Wang blogging again about high order CFD.
Annoying Things
- The SolidSmack CAD blog surveyed CAD jockeys to find out what they find most annoying. It should not surprise anyone that the most annoying thing in CAD was “Getting 3D data from one CAD app to another.” You might be surprised by #2.
- The worldwide CFD market is forecast to grow at 12.85 percent by the year 2018. This is great news because it exceeds the projected growth rate of the entire simulation and analysis segment (7.5%) which in turn exceeds the overall PLM market (5%). [What’s annoying is that the growth rate is presented to two decimal places which makes me automatically reject that number as contrived.]
Deals
- How good of a deal is an undergraduate engineering education? Of all majors surveyed, engineering has the best ROI at 21% versus an average return of about 15%.
- Boeing’s use of products from MSC Software will extend through four decades with the recent agreement to a 5-year extension of the licenses.
- FS Dynamics will be using CD-adapco‘s software for another 3 years according to a recently signed deal.
Let Your Problems Melt Away
We have each experienced times when we’re so frustrated with our work that we’d like to toss it all in the trash or burn it.
Artist Roger Hiorns has elevated such destruction to fine art by atomizing an entire jet engine. He started by melting the engine and spraying the resulting liquid through a nozzle where it mixed with water. This caused the liquid engine to condense into droplets that became a fine powder. Therefore, you can say that fluid dynamics was involved in the process.
The result, shown below, was a candidate for the Turner Prize in 2009.
It’s kind of a cross between those “Will it Blend” videos and those pictures of disassembled objects with all their components neatly arrayed.
UPDATE: Image removed at request of owner.
I’m certain that my friends at General Electric, Pratt & Whitney, Rolls Royce, and the other engine manufacturers will enjoy this. In fact, this would be a good piece to install in the lobby of their corporate headquarters.
In regards to the higher order mesh discussion, I’m under the impression that higher order solvers are not used to gain engineering accuracy on a coarser grid, but instead to gain more (i.e. scientific) accuracy on a given grid. For example for tracking vorticity and LES/DES turbulence models. I’d be curious to see how a better representation of a nonlinear surface actually affects the solution.
Hi Martin. I see it (imperfectly) as better accuracy on a grid of fixed size (cell size and cell count) and similar accuracy on a smaller grid (larger cells and fewer of them). Since I am not an expert, all I can point to is the workshop coming up prior to AIAA SciTech 2015: https://www.grc.nasa.gov/hiocfd/