When most people think of computer simulations, they imagine sleek graphics or Hollywood-style animations. But for Tayfun Tezduyar, the James F. Barbour Professor of Mechanical Engineering at Rice University, accuracy is everything.
“For engineering and science, you don’t just want something that looks realistic,” Tezduyar said. “You need the solution closest to the true solution. If you’re designing a parachute for astronauts or modeling blood flow through a heart valve, the difference between ‘close enough’ and ‘best solution’ could mean life or death.”
For more than three decades, Tezduyar has been developing and refining space-time computational flow analysis, a framework he introduced in 1990 for solving some of the toughest real-world problems in fluid dynamics. Much of this work has taken place at Rice since 1998 and in collaboration with Kenji Takizawa, professor of mechanical engineering at Waseda University, since 2007. The work is now drawing even more international attention. A new book Tezduyar co-authored with Takizawa, “Space-Time Computational Flow Analysis: A Chronological Catalog of Unconventional Methods and First-of-Its-Kind Solutions,” chronicles the breakthroughs — but the bigger story is how this research is reshaping industries ranging from medicine to aerospace.
Tezduyar said what makes their work unique is twofold: the scope of problems they can solve and the level of fidelity they achieve.
“Very few people in the world can solve this range of problems this accurately,” he said. “We take on problems that others have considered intractable, and we find a way to model them accurately, creating high-fidelity representations of the true solution.”
The problems the Tezduyar-Takizawa team has tackled span a remarkable range of real-world challenges. Their modeling helped NASA design landing parachutes for the Orion spacecraft, ensuring astronauts’ safe return to Earth. In medicine, space-time analysis has been used to simulate blood flow through heart valves with unprecedented accuracy, providing surgeons with better information for personalized surgical treatments for aorta and heart valve disorders. In transportation, tire manufacturers can use their simulations to improve tire performance and cooling, reducing the risk of tire damage. And in renewable energy, their models help predict how the turbulent wake of wind turbines affects nearby small aircraft, drones and even wildlife, offering critical insights for avoiding these wakes and for safer placement of turbine fields.
Traditional simulations use different methods for space and time representations of the flow patterns with the sophisticated representation methods developed in recent decades targeting the flow-pattern representation in space. Tezduyar’s view since 1990 has been uniting the two representations.
“In real life, flow patterns depend not just on location but on the instant in time,” he said. “You can’t underrepresent one and expect to get the best answer. Our method uniquely provides high-fidelity representation in both dimensions.”
Tezduyar explained that the complexity of a system’s geometry almost always produces equally complex flow patterns that change across both space and time. For the best solution, computer simulation methods must be as sophisticated in representing the flow patterns in time as they are in representing the flow patterns in space — a feat achieved by the space-time computational flow analysis.
The approach also allows the Tezduyar-Takizawa team to place a high density of computational “points” where they matter most — like around the contact between a tire and the road or where heart valve leaflets close to stop blood flow. While traditional methods require a choice between unrealistic gaps and reducing the density of the points, the simulations of Tezduyar and Takizawa are without any gaps or reduced accuracy in representing the flow patterns.
Tezduyar emphasized that the work is not just about equations but about responding to real-world challenges.
“Many of our projects began because someone came to us with a problem they wanted to be solved,” he said. “Whether it was NASA, the U.S. Army or a tire researcher, they needed answers that weren’t available with existing tools.”
From an idea first explored in the 1990s to decades’ worth of real-world solutions, Tezduyar stands as a true pioneer in his field.
“For many years after I first started, it was really just me and a few former students using these methods, but now there is an increased interest and effort in this class of computations, which makes this book timely,” Tezduyar said. “We’re not interested in simple problems or fast answers. We want challenging problems and the best accuracy, because in the real world, that’s what makes the difference.”