Research: My Statement for Promotion

I am finally going up for promotion from Associate to Full Professor, at the request of my department. This process was an opportunity for reflection on my career, and I wrote about it in three statements that were required for the package of materials submitted for the approval process: teaching, research and service. Here is the first of three posts, with my research statement:


I love flow visualization, the observation of gases and liquids, especially the sinuous, spiraling flows of vortexes. This love has guided my research career. Initially, as a graduate student, post-doc and assistant professor, I hid this. My early experiences in academic culture emphasized that engineering researchers were expected to focus on issues important to industry and government, where research funding was abundant. Societal benefit was secondary, and an interest in the aesthetic aspects of fluid flows was considered useless, if not kooky.  So I focused on experimental work in combustion flows with military applications: rockets, ramjets and turbojet afterburners, and fundamental work on turbulent flame propagation that I hoped would contribute to improvements in air quality at some point, or be useful in industrial combustors (where there was little funding at the time). And I secretly enjoyed the flow visualizations that I was able to incorporate.

In 1997, around the time I received tenure, I learned that vortex rings had been observed in the human heart, and what that might mean for improving medical diagnoses and treatments. I was so excited to learn that the same kind of vortexes that can plague ramjets causing catastrophic failures can drive healthful flows in the human heart and can be used to diagnose certain cardiac illnesses. Since then, I’ve been fascinated with trying to get a better view of cardiac fluid dynamics, and a better understanding of the possible implications for, well, helping sick people. Who could resist? This has formed a consistent theme in my disciplinary work in fluid dynamics.

However, a few years later, in 2002, an unexpected opportunity changed my life, and made me split my research into a brand new branch. There was a program on campus that offered a seed funding for engineering faculty to collaborate with ‘the other side of campus’, i.e.  the humanities, to create interdisciplinary courses. Embracing the spirit of interdisciplinary work, I took this opportunity to marry engineering with the fine arts, and indulge my yearning for the aesthetics of flow visualization.  In collaboration with Alex Sweetman, a photography professor in the Department of Art and Art History, we offered a course titled “Flow Visualization: The Physics and Art of Fluid Flow.” I gave instruction in basic photography, optics, light-matter interactions, the physics of atmospheric clouds, and basics of flow visualization techniques. Alex gave the history of the development of photography from a technology to an art form. Students made aesthetic images of flowing gases and liquids. The assignments were quite open-ended, to accommodate backgrounds of both engineering and photography majors.

The response from students was astonishing. I got excellent FCQs for the first time, and students wrote to me about the impact the class had on them. “I’ll never ignore the sky again,” said one (who later became a rocket engineer at SpaceX). No student had ever written to me after taking my fluids, thermodynamics or measurement laboratory courses, although I thought I had brought similar energy, enthusiasm and expertise to them. What was going on?  I had to know, so I could elicit this response in all my courses, and to colleagues in their engineering courses. Thus began my work in engineering education research.

Since then, I have balanced two research directions: engineering education research (EER) and fluid physics, mostly cardiac fluid dynamics with a few related topics in addition. Although both draw on my visual approach to fluid physics, they have developed independently, so I will expand them separately below.

EER

I’ve been a member of the American Society of Engineering Education since early in my career. I had read in the Society’s magazine and journal how to develop curriculum in various engineering disciplines, and reports of innovative interventions around the country. Then in 2005 I was invited to present on the Flow Vis course at an American Association of Physics Teachers conference, and I got my first taste of what it looked like to apply the tools of quantitative scientific research to education. I was excited to see what kind of quantitative as well as qualitative measurements of teaching effectiveness and educational outcomes could be made, and conclusions drawn from them.

It turns out that doing research on how humans learn is crazy making. It takes years to do controlled experiments, sample sizes need to be large plus studies have to be rigorous enough to defuse the skepticism my colleagues have towards research outside their areas. As a classically trained mechanical engineer, I thought biological engineering research was difficult due to the variability in living organisms. I had no idea about how to address the challenges when human behavior was involved in education research.

Yet, I had to do this if I wanted to understand what was going on in my Flow Vis course, so that its benefits could be more broadly applied. In the process I found a thriving and welcoming research community here at CU Boulder in the form of the Disciplinary Based Education Research (DBER) and Physics Education Research (PER) groups. At the time, the College of Engineering was not particularly supportive of EER as a scholarly area, although recognition of this important area has improved somewhat since then.

After a number of attempts, I was able to secure funding from NSF for a mixed methods (quantitative and qualitative) study of the Flow Vis course. One of the first aspects we explored was how students reported ‘seeing fluids everywhere’ afterwards, demonstrating an ‘expansion of perception’. I found a collaborator in our Neuropsychology department, Prof. Tim Curran, who studies visual expertise. Radiologists, dog show judges, bird, plane and train spotters etc. demonstrate visual expertise in that they can immediately categorize to a granular level the subject of their expertise; they can immediately recognize that ‘that’s a red-breasted nuthatch’ without going through the mental process of ‘that’s a bird, a small one, it’s got a pointy beak and is reddish on the front, and it’s holding on to the feeder upside-down, so it must be a red-breasted nuthatch’.

We hypothesized that Flow Vis students gained a similar visual expertise and expansion of perception, allowing them to recognize fluid flows everywhere.  My PhD student, Kate Goodman, ran a visual expertise train-and-test study comparing ‘experts’, students who had been in Flow Vis or a fluids course with ‘novices’, non-engineering students. With training, both sets of students improved on recognizing turbulent vs laminar flows in Von Karman vortex streets and other types of flows. The experts were able to generalize from training using vortex street images only to other types of flows, while the novices were not able to generalize. This showed that visual expertise was indeed gained, and that the experts were able to apply prior knowledge, enhancing their learning. This result expanded our concept of learning, and is ripe for application in other disciplines.

The other portion of the project involved validating and analyzing a survey of Flow Vis students’ attitudes towards fluids perception and knowledge. The theoretical framework for this was the ‘transformative experience’ as defined by Pugh (2011) to consist of (a) motivated use, where students apply course learning outside the classroom, (b) expansion of perception (seeing fluids everywhere), and (c) experiential value, i.e. appreciating, even enjoying seeing and using fluid mechanics. This survey, administered before and after the course, showed that many students indeed had a transformative experience. The survey has been administered for most offerings of the course since then. The improvement in student attitudes has become more difficult to measure as the reputation of the course has grown and students enter the course with a positive affect. Answers to the Likert scale questions have become saturated at the top of the scale. The survey is currently being redesigned, and is also being expanded to other courses and content areas. However, analysis of the open-response questions which are most valuable is highly labor intensive. We’ve started a new collaboration with Katharina Kann, an assistant professor in Computer Science, to develop an artificial intelligence approach to analyzing open response answers; a prerequisite to scaling up the survey.

Pugh, Kevin J. “Transformative Experience: An Integrative Construct in the Spirit of Deweyan Pragmatism.” Educational Psychologist 46, no. 2 (April 2011): 107–21. https://doi.org/10.1080/00461520.2011.558817

 Summary of EER Contributions

Six archival publications, 51 conference papers and presentations, one PhD, two Master’s and more than 17 undergraduate researchers

Cardiac Fluid Dynamics.

Research on human subjects is an order of magnitude more difficult than typical well-controlled engineering fluid mechanics problems. IRB issues aside, the cardiac system is highly three-dimensional, asymmetric, and time dependent. It is dominated by fluid-structure interactions and the flow is transitional, neither completely laminar nor turbulent. This is my sweet spot, the kind of complex flow where the fluid physics can only be explained by data derived from visual analysis that then directs quantitative analysis. For example, in the early 2000s I began learning about a rapidly developing magnetic resonance imaging technology: 4-D phase contrast MRI. It is suited for measurement of phase-averaged (i.e. a ‘typical’ heart beat) large-scale flows throughout the heart, providing reasonably well resolved three dimensional flow fields with all three (xyz) velocity components. This allows analysis and visualization of vortical flows during diastole, when the heart is filling. I began working with clinicians at the National Jewish Health Center, Drs. Fenster and Schroeder, to develop a metric to track the progress of pulmonary hypertension (PHT). PHT is high blood pressure in the artery between the heart and lungs where making a pressure measurement is particularly difficult; our goal was to use vorticity (local fluid rotation or spin) calculated from the 4-D MRI-measured velocity fields. We have shown that vorticity in the right ventricle during diastole (filling) can indeed be used to track pulmonary hypertension.  Visual analysis of the vorticity shows that most of the vorticity is generated in the boundary layer that forms over the tricuspid valve flaps, which are themselves invisible to most imaging techniques. However, knowing this fact about the vorticity can allow a more sensitive, accurate tracking of the disease.  This project is still ongoing. Some of our results are easy to see in 3-D (on a stereoscopic monitor) but difficult to represent convincingly in two-dimensional media. Still the project to date is modestly successful; what is perhaps more important are the detailed visualizations of this fascinating flow (https://www.youtube.com/watch?v=zGU2rgWdDJ4 for example). We are exploring VR techniques, but these are not quite ready for prime time. I’m looking forward to better technology being available soon.

Since then I’ve been working with clinicians at the VA Hospital in Denver, looking at diabetic patients before and after an exercise regime. Men and women with diabetes don’t derive the same cardiac benefits from exercise as everybody else, and we are hoping to see why by examining the cardiac flows.

Another of the works that I’m proudest of was the first paper describing ‘echo PIV’. ‘Echo’ refers to ultrasound echo imaging, a safe, non-invasive, relatively inexpensive imaging technique that forms images from high frequency sound waves bounced off tissue and particles in the blood stream. We added an ultrasound contrast agent (lipid-coated microbubbles) to a fluid stream and made multiple rapid images of the microbubbles’ positions in time. This allowed us to calculate the velocity of the bubbles, and thus the fluid, everywhere throughout the image (a two-dimensional, two-velocity component technique). When coupled with the more detailed information about healthy vs diseased flow in general, gained from 4DMRI, this technique could prove to have more widespread clinical acceptance.   Prof Robin Shandas (CU Denver/Anschutz) had the original idea and access to the ultrasound equipment. I worked closely with my post-doc at the time, Hyoung Bum Kim (now a professor in the School of Mechanical and Aerospace Engineering, Gyeongsang National University, Korea) making detailed measurements to explore and validate this new measurement technique. Echo PIV is now an accepted method for measuring blood velocity in the human body and in other opaque flows where optical methods fail.

Summary of Cardiac Contributions

40 archival publications, 39 conference papers and presentations, five PhDs, six Master’s and more than 27 undergraduate researchers

COVID Aerosol Project

Most recently I have been working with Shelly Miller, Marina Vance and participants in the 

International Coalition Performing Arts Aerosol Study. Early in the pandemic superspreader events from singing threatened music education programs worldwide. If just singing emitted so much infectious aerosol, what about brass and woodwind instruments? The Coalition was formed of dozens of music education professional societies who came together to fund our research. We performed flow visualization on a variety of musical instruments and vocal performers, which informed the experimental designs for quantitative aerosol measurements associated with performances. Together we came up with a suite of mitigation recommendations: put “masks” on instruments and singers; increase room ventilation; and reduce time spent in rehearsal rooms. These measures were widely adopted, and I’m relieved to report that no schools that followed the recommendations found any COVID transmission traceable to the music program. This work generated quite a bit of media attention, which was stressful for us all: the work released in preliminary form had immediate life safety implications as well as protecting the employment of thousands of music teachers. I  had hoped that this project would become less important as the pandemic waned, allowing my current PhD student Abhishek Kumar to get on with a more academic study of instrument aerodynamics, but unfortunately this work remains distressingly relevant. Our first peer-reviewed publication has just been published at ACS Environmental AU:

Stockman, Tehya, Shengwei Zhu, Abhishek Kumar, Lingzhe Wang, Sameer Patel, James Weaver, Mark Spede, Donald Milton, Jean Hertzberg, Darin Toohey, Marina Vance, Jelena Srebric and Shelly Miller “Measurements and Simulations of Aerosol Released While Singing and Playing Wind Instruments.” ACS Environmental Au, August 27, 2021. https://doi.org/10.1021/acsenvironau.1c00007.

Other Fluid Physics Research

There are a number of other projects I’ve worked on, contributing both quantitative and qualitative flow visualization methods: analyzing airflows in hospital operating rooms, augmenting fuel sprays for small engines, evaluating performance of pumped thermal energy storage, and using synthetic jets to improve indoor air quality. All have allowed me to use, express, and demonstrate how my love of fluid physics and contributions in flow visualization motivates rigorous and useful science and engineering.

Summary of Other Contributions

13 archival publications, seven PhDs, 17 Masters and more than 14 undergraduate researchers.

Opportunities for Undergrads Interested in Research

Here in the Mechanical Engineering department at CU, there are many opportunities for undergraduate students to get research experience. Pretty much all of us professors have active research programs, and most of us welcome undergraduates to participate. If my projects don’t fit with your interests, don’t stop looking. On the department website you can find short descriptions of what each prof is interested in. They might also have an informative website, but don’t count on it. The best way to find out what a prof is doing is to ask them; make an appointment, say that you are interested in their work.

If you say you want to work with me, I try to find out what kind of experience you want so I can suggest projects to match. Some projects are hands-on design/build/test of a piece of laboratory apparatus for my fluids research or for the Flow Vis course. Some projects involve a bit of Matlab programming and/or data analysis from my research. A project might  be a literature survey on a topic of mutual interest, or might be interviewing other students and analyzing the results. Most projects are related to ongoing research, so you might be helping and be supervised by one of my graduate students. I’m also open to fluids-related ideas that you are passionate about. Whatever it is, I want a good match so you’ll be enthusiastic, self-motivated and dedicated.

Other things I look for in a research student:

  • Being a junior or a senior. This means that you have enough background in your discipline (whether it’s Mech Engin, some other engineering, filmmaking or whatever) to get started quickly. This is a guideline, not a hard and fast rule.
  • Having a partner or two lined up, with schedules similar enough that you can spend around 10 hours together per week, plus a short group meeting with me every week.
  • Being able to make a commitment to a total of 150 hours in a semester. Sometimes this can be spread out over more than a semester, and include part or all of summer. This means having a reasonable course load, and not a lot of other projects.
  • Production of a good final report. It will be due two weeks before the end of classes, so I have time to edit it and you have time for revisions.
  • I much prefer to work with CU students, with the hope that after I invest my time in you and get you trained up to be productive that you will want to stay on and work with me for more than one semester.

In return, you’ll get a taste of real research, including an experienced mentor (I’ve had over 150 undergrad researchers in my program), a great letter of reference for job applications, and maybe a research publication or two to put on your resume. You can also get either

  • 3 credit hours of Independent Study which will count as a technical elective in Mechanical Engineering. If this is in your plan, you’ll need to fill out the application form, get my signature, and get it to one of the ME undergrad advisors in time to register. Yes, it has to be typed, and we have to agree on the scope and methods.
  • OR
  • a bit of money. This is harder to set up, but I’ve been fairly successful helping CU students get funding from the Undergraduate Research Opportunity Program (UROP) here at CU (but watch out, the deadlines are waaaaay in advance). Depending on the project, there might be other pots of money around for funding.

I usually have 3 to 6 undergrads working with me at any given time. I’ll be posting about specific projects in the future, so if you are interested, check back here now and then.

Seeing into the heart

 Our visual systems are pretty impressive. Even when we are looking at images on a flat surface, our brains use all sorts of cues to create a 3D mental map. Shading on round objects helps, and of course the differences between what your right and left eyes see. If it’s moving, foreground objects will move faster than far field, and they’ll be larger (motion parallax). Shapes change as they rotate, and our brains interpret that easily. Occlusion, when an object passes behind another, is a big cue. Babies stop laughing at peek-a-boo when they figure that one out.

I’ve been spending a lot of time with my stereoscopic computer monitor, looking at flow in the right heart. It’s so complicated that even when I use lots of shading and occlusion, twisting and turning the representation, I still need the stereo cues to really see what’s going on. But it’s thrilling when I do.

So now I want to show it to everybody, but nobody else I know uses the 3D technology, even though it’s now dirt cheap (Thank you gamers!). So we’re back to motion parallax, shading and occlusion: it needs to move. The flow also needs to be simpler, so in this video, I’ve removed all the small velocities and vorticities, edited out almost all left heart and other extraneous flows, and tried to keep the colors from being overwhelming (although I love saturated colors).

Please enjoy this intro ride through the right heart. I hope to have more soon.

Not an accident!

Flow enters the right cardiac atrium from above and below, bringing wall generated vorticity with it.

Here’s the image I was trying to get before New Year’s. This shows flow entering the right atrium of a normal subject’s heart. Blood flow is shown by the white pencils. Vorticity (the amount of spin of a bit of fluid) is shown by the colored arrows. Only the strongest velocity and vorticity is shown here, to keep the image from getting too cluttered. The right atrium is shown in transparent white, and the right ventricle is the big triangular shape in yellow. Flow is entering the atrium from the top, through the superior vena cava (SVC). You can see vorticity ringing the flow, since it is being generated at the SVC surface. Flow up from the bottom, through the inferior vena cava (IVC), is more complicated. Venous (return) flow from the liver comes in and wraps around behind the main IVC flow from the lower abdomen. The two flows mix as they enter through the bottom of the atrium. This is useful, since you want all the important chemicals from your liver to mix with the rest of your blood before it gets pumped to your lungs and beyond. Too bad this type of imaging (4DMRI) can’t give more details about the mixing process.