Go With the Flow

Here is a pro tip: If a mechanical engineer offers to explain computational fluid dynamics, ask for a demonstration that employs libations in a half-empty glass. You may, at least fleetingly, gain a (probably false) sense of comprehension.

Or so it seemed to me when Dr. Amir Isfahani, president and CEO of Flow Science (www.flow3d.com), used his partially consumed beer to explain the fundamental issue his company addresses through its FLOW-3D software, used by clients all over the world in industries that include aerospace, biotech, automotive, additive manufacturing and consumer products, to name a few.

Isfahani was the Kick-Ass Entrepreneurs of Santa Fe Storytime guest speaker on Sept. 5 at Back Road Pizza—the series highlights local entrepreneurs and their businesses' origin stories.

"I don't know if I can call myself an entrepreneur," Isfahani said by way of opening, before discussing his background, the company's history and the physics behind it all.

And while several scientists in the audience would have followed a higher-level explanation of computational fluid dynamics, the beer demo was definitely easier to grok for those of us who studied English in graduate school.

Isfahani completed his dissertation in theoretical and applied mechanics at the University of Illinois at Urbana-Champaign in 2011, focusing on flow, but in that case the microcirculation flow of red and white blood cells, the formation of which at such a tiny scale, he told me after the talk, "is not very well understood."

The 38-year-old became president of Flow Science approximately two years ago, but has worked at the company for more than a decade, starting out as a software developer before transitioning into research and development, sales and, ultimately, company leadership.

FLOW-3D allows users to create simulations for free surface flow. An example Isfahani shared with me after his talk included work helping design a "fish ladder" to address thwarted fish migration routes in rivers owing to a variety of factors, such as increased water temperatures and perched culverts. Flow Science, he says, has several clients in Japan who use the software for tsunami simulations and modeling.

Back to the beer. Essentially, Isfahani said, as he tipped his glass so the beer flowed to one side or the other of the glass, "if it has this much beer in it, and say we are going to move the glass likes this, the fluid starts moving like this." On a computer screen, he noted, you could try to design a different shaped glass and try different motions to see whether or not the liquid would spill. "This was a cheap example," he said. "The expensive version is you want to build a dam in the river and it's a multi-million-dollar project."

The company's FLOW-3D software began as just a few hundred lines of code written by then-Los Alamos National Laboratory scientist CW "Tony" Hirt as an appendix to a paper he and other scientists authored in the late 1970s. As Isfahani tells it, Hirt "said to the lab, 'what if I tried to commercialize this?' And they said, 'go for it.'" Hirt went for it, founding Flow Science in 1980. The original code, Isfahani says, is now "a few hundred thousand lines of code … and we protect it with our lives."

Over the decades, he says, Flow Science engineers have added and shaped the code in response to different problems and requests from its clients, "and in time we built this monstrosity that now serves 12 different major industries and does the best job worldwide compared to any other code."

That worldwide success means Isfahani travels approximately six months out of the year and works, he says, 24/7. The company is small: 36 people. "We are a tight-knit family," he said. Flow Science was recognized for its employee policies this summer with an award from Family Friendly New Mexico.

The company also faces challenges. On the physics side, calculating for turbulence remains inexact, and equations involving turbulence still can't be solved with zero error, he said (#physicsproblems). The code is written in Fortran, a classic programming language for physics, but not necessarily known by a new generation of programmers. Isfahani says the company talks about rewriting the code every 10 years, but ultimately decides it's "too scary" to risk accidentally breaking a code that has been working for decades.

I snuck in a question about my favorite existential threats—artificial intelligence and automation—which Isfahani acknowledged were areas the company needed to keep in mind "to make sure we are not left behind."

More pressingly, Santa Fe's infrastructure shortcomings make Flow Science's future here uncertain. Specifically, limited housing options for employees as well as concerns about the quality of the public schools makes it difficult at times to recruit top and younger talent with young families. The city is moving in the right direction, he said, but it's unclear if it's happening quickly enough. "If this is going to be the state of affairs," he told the audience, "… we won't be here in five years."

Drink up.