The Japan News by The Yomiuri Shimbun

Home > Research > Knowledge Co-Creation - Profiles of researchers


Knowledge Co-Creation - Profiles of researchers

Using coupling analyses of structures and fluids to bring about engineering functionality for complex physical phenomena

Kenji Takizawa
Associate Professor, Faculty of Science and Engineering, Waseda University

Joint research with NASA

My areas of specialization are fluid and computational engineering. My work focuses on research into “fluid–structure interaction” which concerns the physical phenomena that result from strong relationships between fluids and structures. My interest in this field first arose thanks to a university professor I met while majoring in mechanical engineering and who specialized in numerical analysis of fluids. The feeling of understanding all phenomena in the natural world with computer simulations was a big attraction for me.

After completing my doctorate, I worked as a researcher at the National Maritime Research Institute (NMRI website) and addressed the topic of ship movements and ocean currents as a two-phase flow of structures and fluids. One well-known phenomenon that results from a two-phase flow is the “milk crown,” which is a crown-shaped spray that forms when a droplet is dropped into a liquid. This is a phenomenon of surface tension that occurs at the two-fluids interface between a gas and a liquid. In the case of ships and ocean waves, the question becomes even more complicated when you add changes in ship structure that occur as a result of ocean waves. This is what we refer to when we talk about fluid–structure interaction.

Fluid–structure interaction becomes more relevant when the structure is a large item or a soft item such as a living organism or fabric.
From the left: horizontal wind turbines; vertical wind turbines; ram air parachute; locust analysis

I was appointed as a researcher at Rice University, Texas, where I had the fortunate opportunity to conduct global, pioneering research under world authority Professor Tayfun E. Tezduyar. One of these projects involved carrying out parachute drop simulations in collaboration with NASA. Parachutes are used to drop spacecraft and measuring equipment etc. from the stratosphere to the earth, or to land on Mars. The goal of the research was to investigate the best parachute formation and when to use these formations in order to ensure landing is as stable as possible.

Parachutes are a perfect example of a structure that changes shape in response to air flow field. Their speed of descent is not fixed, and multiple parachutes result in contact and interference between parachutes. It is difficult to guess what the wind conditions will be when the parachute opens. The simulations were extremely complex and difficult and there was a great deal of complicated trial and error. We made our breakthrough when we hypothesized that the movement of multiple parachutes is symmetrical (in fact, it is asymmetrical), and then carried out simulations using simplified conditions.

A simulation of parachutes contacting with one another

We sought to identify conditions that would allow the phenomena to occur with no issues. We altered the conditions slightly each time, and built models based on realistic conditions. With these models we tried changing the parachute material, running cables around the parachutes, opening holes in the surface, and using simulations to limit the conditions prior to experiments. We then conducted experiments with real parachutes at NASA in order to design a steady parachute. Since returning to Japan, I have continued joint research with NASA as a member of Professor Tezduyar’s team. Unfortunately because I do not have United States citizenship it is difficult for me to observe NASA's experiments.

Fluid-structure interaction analysis of parachutes. Parachutes used by the Orion spacecraft.
(reproducing dimensions of the Apollo era (left); one of the test candidates (center); the final design (right))

Living organisms, a world with no blueprint

Another project I was involved in during my time in the United States was fluid analysis of blood flow in medical treatment. Rice University is located in Houston where one of the world’s big three medical hubs attracts numerous medical research institutions and hospitals. The project was a joint project between individuals in related medical fields. Fluid–structure interaction is effective when a structure changes shape easily, and is particularly suited for living organisms because they are soft and change shape. After learning techniques during my time on the project in America and then returning to Japan, I was invited to participate as a member of the CREST Project—“Alliance between Mathematics and Radiology” where I continue to research to this day.

We are analyzing large blood vessels such as the aortae and cerebral arteries. In recent years it has become clear that even if an aneurysm forms it is unlikely to rupture. Therefore, whenever possible, patients are not operated on but observed. However, there needs to be a method for deciding when operations are necessary. Ultimately, the goal is to be able to carry out analyses with individual patient data. Currently we are only at the stage of understanding basic mechanisms. We continue to hold discussions using numerous cases about overall trends in the phenomena of blood flow and we are continuing to work on a trial-and-error basis.

Living organisms do not have a blueprint like human-made objects, and understanding their structures poses a great challenge. The most difficult aspect is determining the degree of blood vessel contraction in medical imaging verses the contraction of healthy blood vessels. For example, it is possible to analyze the extent to which an old rubber hose has stretched in comparison to its normal state, but the blood vessels of a patient viewed through medical imaging are already stretched, making it necessary to estimate their original state. This is extremely difficult. If you cannot estimate the original state it is impossible to make accurate predictions.

Blood flow analysis graphics. Blood flow within the blood vessels (left); changes in blood flow as a result of changes wall shape (center); dramatic changes as a result of the opening and closing of valves, such as the heart valve (right)

If blood flow simulations become possible, we may be able to create operation plans based on objective evidence, rather than on current methods where doctors decide which blood vessels to reconnect and how to reconnect them based on their own experience. The experience of skilled doctors is extremely valuable, but this knowledge needs to be spread so it can become useful for larger numbers of patients. It is very difficult to observe fluid phenomena with your own eyes. Even skilled doctors cannot observe minute details of blood flow phenomena. When we show doctors the simulation images, they are often happy to see that the data backs up their own experience, saying things like “I knew the blood would be swirling in a vortex there and slowing down the flow!”

Tackling completely unknown phenomena

As computer technology advances, we can expect even greater possibilities for fluid analysis. In the case of the parachute analysis described above, we could start with a logical analysis of the phenomenon from an overall macro perspective and then move on to more local analysis such as examining micro changes in the surface area of the parachute before building a model that allows us to understand the details. In the future it is likely tools will become available that allow us to produce analytical results in real time. This will enable us to consider different suggestions during the planning phase such as enlarging a certain part or changing materials, etc. This requires immense computing resources, but possibilities will likely widen as computer technology continues to advance.

Multiscale Thermo-Fluid Analysis of a Tire
Graphic of Fluid Analysis of a Tire (Takashi Kuraishi, Kenji Takizawa, Shinichiro Tabata, Shohei Asada and Tayfun E. Tezduyar), which won an award at the 19th Japan Society for Computational Engineering and Science Conference Graphics Awards.

There are many faculty members at Waseda University with a strong background in automotive technology. We are taking full advantage of this environment to develop automotive-related fields. The research motto shared by all of these research themes is “identifying mechanisms for unknown phenomena.” The computing power of computers has improved dramatically in recent years and it is now common to carry out research that refines simulations of phenomena we already understand. However, I think the most rewarding aspect of computational engineering comes from taking on the challenge of unknown physical phenomena. The social function of big data has become a major topic recently. The basic idea is that, as long as trends of certain phenomena can be understood using the power of data volume, we can connect mechanisms while inside the “black box” and use these for the benefit of society. The role of my work is to shed light on this “black box” theoretically.

Analysis work requires collaboration and teamwork among different fields. It is impossible to conduct research without collaborating with those that manufacture things and those with practical capabilities in things such as demonstration testing in various fields such as space or medicine. A collaborative research team needs to be well organized. As numerical analysts we also need to possess a broad range of analytical techniques that allow us to respond to any situation. I learned what to look for in data from distinguished professors in the United States. Accumulated knowledge is effective in data analysis. This is why I always tell students in my laboratory to “examine each set of simulation data as carefully as possible and read as much as you can from the data.”

Computational Fluid–Structure Interaction: Methods and Applications was published in 2013. I co-wrote this book with Professor Tezduyar, who taught me so much during my time at Rice University, and with my research colleague Yuri Bazilevs (now a professor at University of California, San Diego). A Japanese translation of the book is currently being prepared for publication for use as a graduate school textbook.

Kenji Takizawa
Associate Professor, Faculty of Science and Engineering, Waseda University

In 2001, Professor Takizawa graduated from the Department of Mechano-Aerospace Engineering at Tokyo Institute of Technology. In 2005, he completed his doctoral degree at the Department of Energy Sciences, Interdisciplinary Graduate School of Engineering and received Ph.D. (Science). He worked as a Researcher at the National Maritime Research Institute, Japan and as a Research Associate and Research Scientist at Rice University, USA. In 2011, he was appointed Associate Professor at the Faculty of Science and Engineering, Waseda University. He is involved in a broad range of research fields, including fluid dynamics, structural mechanics and biomechanics, with a focus on computational mechanics. Main publications include: CG Simulations Using the CIP Method and Java [CIP-ho to Java ni yoru CG Shimyureshion] (coauthored by Takashi Yabe, Youichi Ogata and and Kenji Takizawa; Morikita Publishing, 2007); and Computational Fluid–Structure Interaction: Methods and Applications (Yuri Bazilevs, Kenji Takizawa and Tayfun E. Tezduyar; Wiley, 2013). Awards include the 2014 Waseda Research Award and the 2015 Young Scientists’ Prize, Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology, Japan.