Rarely do I get a feeling of déjà vu in my professional life, but in March of this year something I was shown gave me a flashback to the 1980s. Some people will remember those halcyon days of two decades ago when commercial CFD was still in its infancy. Way back then the big commercial CFD vendors (Fluent, CHAM, Fidap, Flomerics, STAR-CD, and AEA) would regularly bring out CFD Comparison Charts comparing their CFD software with all their major competitors, usually in the form of long technical feature lists. It was like a Cold War nuclear arms race within CFD – one vendor would claim to have a Reynolds Stress Turbulence Model, another vendor would claim to have a Renormalization Group k-e model, and so on, until codes were offering 20 or more turbulence models! It became counter-productive as the different codes’ unique selling points and specialist capabilities got lost amongst all the other statements.
In those days CFD was a gentlemanly business, so it was all relatively amicable and honourable, if a bit ‘techie’. Such charts seemed to die out in the early 1990s and so I thought their story had ended. Not so…
On the 19th March 2010 Virginia-based Blue Ridge Numerics (a privately held company but supported by Washington DC based venture capital company Global Environment Fund Management Corporation) published the “CFD Comparison Chart” below, via a download feature inside an e-Newsletter from Desktop Engineering Magazine, purporting to be a comparison between their CFdesign CFD code, our FloEFD Concurrent CFD software, ANSYS Inc’s CFX and FLUENT software packages, and SolidWorks Flow Simulation. This “CFD Comparison Chart” claimed to be a guide for helping engineers decide on what CFD code to choose for their needs – and the check list clearly shows CFdesign coming out very well in every category.
When our Product Management team scanned the document we were dismayed to find 27 (yes 27!) false or misleading statements in it relating to our FloEFD software and SolidWorks’ Flow Simulation software, all of which are in Blue Ridge Numerics’ favour – and that does not include our count of the numerous erroneous statements we reckoned were made against ANSYS Inc’s product line from what we know of their capabilities!
With so many of the statements made in the “Chart” about FloEFD being wrong, it clearly went way beyond reasonable technical advertising, fair comment, and arguably even free speech. It was either very badly researched or very defamatory towards Mentor Graphics (and we suspect towards ANSYS and SolidWorks from what we know about their product lines). I was really surprised that a company like Blue Ridge Numerics, who have been in business for some 20+ years could be so badly informed about other products in their market. To their credit, when we notified them, Desktop Engineering apologised to us, sent out an apology to all of its readership and promised that such a chart would not go out again.
As an example of one very glaring mistake, look closely in the “Chart” above and see that the Fluid Flow Capabilities section of the “CFD Comparison Chart” claims that FloEFD cannot do transonic and supersonic (i.e. compressible) flows. A quick glimpse on our website reveals the following example:
Besides the ability to simulate airflow over the entire range of engineering-relevant Reynolds numbers from laminar, incompressible, compressible and transitional to subsonic and supersonic, FloEFD offers a wide range of additional physical models for characterization of the aerodynamic behavior of an aerospace object. It can even predict hypersonic flows, not mentioned in the “CFD Comparison Chart”!
At this time of writing, a search on Google for “Mentor Graphics” + “supersonic” has this Design World article third in the list. The article clearly shows one of the 20+ validation examples that we deliver along with FloEFD, being a classic illustration of supersonic flow in a 2D convergent-divergent duct. The agreement with theory is excellent, and the case also shows the solution-adaptive mesh capability in FloEFD that is essential for precise shock capture in such engineering situations!
Next time I’ll point out some more mistakes, and then in future blogs let you know what happened when we pointed these out to Blue Ridge Numerics.
Dr. J, Hampton Court