Episode 37: SwRI’s 2021 R&D 100 Award Winners

Lisa Peña (LP): We're celebrating our 2021 R&D 100 Award winners. This prestigious science and technology competition recognizes the top 100 innovations of the year. SwRI picked up a record three awards this year for a revolutionary diesel exhaust fluid, game-changing connected vehicle technology, and a cutting-edge machine learning tool for chemicals. The winners next on this episode of Technology Today.

[MUSIC PLAYING]

We live with technology, science, engineering, and the results of innovative research every day. Now let's understand it better. You're listening to the Technology Today podcast presented by Southwest Research Institute. Transcripts and photos for this episode and all episodes are available at podcast.swri.org.

Hello, and welcome to Technology Today. I'm Lisa Peña. It's one of my favorite episodes where we talk to SwRI's newest R&D 100 Award winners. Every year, SwRI enters technologies into the competition known as the Oscars of Innovation, the prestigious R&D 100 Awards presented by R&D World. This year we have a record three winners, bringing SwRI's total to 50 R&D 100 Awards since 1971.

Our guests today are SwRI engineer Cary Henry who will tell us about the revolutionary catalyzed diesel exhaust fluid creating cleaner air for all. We'll also talk to SwRI research computer scientist Michael Hartnett back on the podcast to discuss Floodlight, a machine learning tool that analyzes chemicals we encounter every day.

But we'll start with SwRI engineer Sankar Rengarajan. His team developed connected and automated vehicle technology that lowers fuel consumption and emissions by optimizing power management, routing, and speed for passenger and fleet vehicles. Thanks for being here, Sankar, and congratulations on your team's R&D 100 Award.

Courtesy of SwRI

SwRI Engineer Sankar Rengarajan is part of the team that developed SwRI’s Eco-Mobility with Connected Powertrains, which delivered over 20% energy savings on a plug-in hybrid vehicle in real-world driving conditions by optimizing power management, routing and speed.

Sankar Rengarajan (SR): Thanks a lot, Lisa. It's great to be here.

LP: So tell us about your winning technology. How does the connected and automated vehicle technology work?

SR: Sure. So before we get into the finer aspects of it, at a broad level, let's understand, OK, what is connected and automated vehicle technology? So connectivity, as we normally associate, it's another folds these days. Like we are well-connected through regular 4G, LTE-style connectivity.

But in addition to that, in cars, there is another kind of communication, which is commonly referred to as short-range communication as V2X, which is Vehicle to Everything technology. So vehicles, think of it as a short-range WiFi communication that lets cars within a radius of, say, 300 to 400 meters can communicate with other vehicles, they can talk to infrastructure such as traffic lights, and also even pedestrians in some cases.

The second piece of the technology is automation. So there is the mobility as a service moment where there's a lot of investment and interest in the development of fully self-driving vehicles, which is normally referred as SAE Level 4 automated vehicle. But they are still very far off from production. So what currently exists in the market is what we regularly drive, which is just the regular, what is referred as SAE Level 2 automation, which is we have cruise control style and lane-keeping assist-style technology.

So be it the connected vehicle technology or the automated vehicle technology, one of the fundamental drawbacks has been that they have completely been decoupled from the powertrain technology. So that is where our technology comes into play where we operate at the intersection of connected vehicle technology, automated technology, and then how to improve powertrain efficiency by using the advanced sensing, as well as information that is available through these technologies.

So we develop three techs, as you pointed out. So the first one is called ecorouting. Essentially a framework that enables smarter travel between a given origin and destination taking into account various route information. And think of it as Google Maps just for, it's not just showing you travel times from point A to point B, it's also giving you, here is how much energy your vehicle is going to consume when you take this route.

The second one that was developed is, as we can tell, electrification is a very new aspect in powertrain technology. Hybrid vehicles, so when they have an electric machine as well as like a battery as well as an engine, now how can we use information about the route, traffic congestion, even simple information like grade or elevation, how can we use that information to more efficiently operate the electric machines or the engine?

And the third one is what's called as ecodriving, which is using information about other vehicles and, say, traffic lights, to essentially drive smarter, because the best form of efficiency is essentially reducing demand. So if that traffic light is going to turn from green to red in five seconds, there is no need to accelerate, so just removing that demand, and that saves a lot of energy. So that's kind of in a nutshell.

LP: Can you walk us through how an everyday driver would use the system? Would you turn it on? Is it already on when you get in the car? And then does it present like a map? Like you mentioned Google Maps, will it present right there on your screen and kind of give you the information as you go about?

SR: Yes, exactly. So right from the get-go, our team's motivation was definitely to keep it very practical and something an end user can readily appreciate that is integrated into their everyday lives. So think about how you would start a trip. So we get into a car, you open up your navigation system. In this case, let's say the example of Google Maps. So I enter my destination.

And in fact, our routing technology was built right on top of Google Maps technology. So effectively, we tapped into what is referred to as Application Programming Interfaces, APIs, of Google Maps. And then in addition to just the traffic information, now if you overlay on top of it a power train dynamics layer, so for example, if I'm driving my, let's say, an electric vehicle, now I can say that, all right, it is OK to use my battery a little more aggressively here because I know there is a downhill coming later part of my route, but I'm able to regenerate my battery and still keep essentially a good state of charge in my battery.

So using such information, it is able to plan what is the best optimal route, which is not just better in terms of travel times, but also is the best in terms of energy efficiency. And now, when you start the drive, now that is when the second piece of technology starts to come in. Now you have gone from a macro view of the route now to what we refer to as a mesoscopic scale information. So you're operating within information that is about 300 to 500 meters within your vehicle's radius. Now you're talking to other vehicles, you're talking to infrastructure like traffic lights.

And now there is a driver advisory system, or it could be integrated readily into your vehicle's cruise control technology. And what it's doing is it is essentially giving you, here an optimal speed to drive based on the conditions around you. And as pointed out previously in the example, it could be a vehicle that is going to get into your lane. So slow down to accommodate that other vehicle. Or it could be a traffic light that's going to change from green to red in five seconds. So if possible, increase your speed by five miles per hour. So that you can cross that light in green.

And the idea is that you essentially reduce power demand. And if you think about it from a heavy duty truck perspective, let's say the Class A big 18-wheelers that are driving in residential and urban neighborhoods delivering to, let's say, our grocery store, now the biggest fuel consumption or penalty in that truck is when it has to accelerate the 60,000 pounds of mass from 0 to 35 or 40 miles per hour. And here is a technology that is essentially either helping those trucks avoid completely stopping or come to a stop in the most energy-efficient manner. So this is how the technology integrates into a regular driver's trip.

LP: You are being recognized as one of the top innovations of 2021. How does that feel? What does this R&D 100 Award mean to your team?

SR: I don't recall the team and what we are as, even as an Institute. Like our code is to create clean mobility solutions. And then being recognized among all these other finalists, and we look at the organizations that we are around, there is Fortune 500 companies, National Research Labs, the top tier universities, it's a big motivation and a morale boost for our team.

It shows that technology that we have been working on, like we belong in that space. And it lends a lot of credibility. And like this is technology that is recognized by such a prestigious organization. And most importantly, I think it is the awareness that it generates.

Like nothing makes our day more than a city calling us and reading a press release saying that, hey, we saw this news release about you guys winning R&D 100 Award for this technology. We are interested in applying and seeing how this could benefit our city to improve our air quality because we are currently under non-attainable status. So really, truly using science and technology to improving mankind which aligns well with our institute's mission statement as well. I think all of this is well-recognized by our team winning such an award.

LP: Yes. Definitely aligned with our mission of research and development to benefit humanity. So great job to you and your team again. Congratulations on the award. Really thorough and great discussion today. Thank you for being here, Sankar, and for helping us understand the advantages of connected and automated vehicle technology, and really, the possibilities with interconnected vehicles on the road. It's exciting.

SR: Thank you, Lisa. Great being here.

LP: And we continue talking about technology, that's good news for drivers and for the environment. SwRI engineer Cary Henry joins us to discuss the game-changing catalyzed diesel exhaust fluid which reduces harmful emissions. Hi, Cary. Thanks for being here, and congratulations to your team on the award.
 

Courtesy of SwRI

SwRI Engineer Cary Henry, Catalyzed Diesel Exhaust Fluid principal investigator, says Cat-DEF™ is a revolutionary technology. The Cat-DEF system supports a deposit reduction potential of up to 98% during low-load operation, a 90% reduction in harmful NOX emissions and a 2% reduction in fuel consumption for heavy-duty diesel trucks.

Cary Henry (CH): Thank you, Lisa. Very excited to speak with you today.

LP: So let's start with a basic question. What is the function of diesel exhaust fluid?

CH: That's a great question. Diesel exhaust fluid is utilized in exhaust systems for diesel and other lean burn engines to reduce criteria pollutant emissions such as NO_X. NO_X emissions are regulated at the state, national, as well as international levels because it can directly impact human health, and it also has indirect implications with respect to the formation of ozone in the ambient environment.

LP: So how does a diesel exhaust fluid generally work? It's like a chemical reaction. When you put it in the vehicle what's the process to clean up those emissions?

CH: It's actually a combination of physical and chemical changes in the exhaust system. So if you think about a typical passenger car that's a gasoline vehicle that you operate, those cars utilize three-way catalyst technology. So they operate at stoichiometry air fuel ratios. And so they're always switching between lean and rich operation so that they can simultaneously reduce NO_X, hydrocarbon emissions, and carbon monoxide.

Diesel engines, on the other hand, operate lean over the entire drive cycle and they're designed to operate lean the entire time. One of the challenges when you operate lean is it's difficult to reduce emissions in a net lean environment. So an oxide of nitrogen is an oxidized nitrogen molecule, and so you have to reduce it to form nitrogen and water vapor. And so that's a challenge when you have a net lean, or net oxidizing environment.

And so what you do is you utilize diesel exhaust fluid where you spray it into the exhaust system in front of a catalyst. Ideally the diesel exhaust fluid, it carries urea, so diesel exhaust fluid is 32 and 1/2% urea with the remainder being simple water. And so when you spray the diesel exhaust fluid into the exhaust system, the ideal first step is you evaporate the water. You're left with a solid urea molecule.

And then that urea molecule undergoes thermal decomposition to form isocyanic acid. And then that isocyanic acid undergoes a hydrolysis reaction where it reacts with water to form ammonia. And then that ammonia is what is utilized to reduce NO_X in the diesel exhaust system.

LP: So you've described as SwRI's catalyzed diesel exhaust fluid as a revolutionary technology. What makes it truly game-changing, different from other fluids?

CH: The catalyzed diesel exhaust fluid has some additives that are incorporated into it to improve those physical and chemical reactions to create more of that desired ammonia to reduce NO_X. One of the challenges with conventional diesel exhaust fluid, when you spray it into a system at low temperatures, you may not have enough thermal energy in the exhaust to first evaporate all of the water, and so that evaporation process may take a long time. It would result in slow thermolysis of urea, so you can form some solid urea deposits.

And then you can also have issues where once you form that isocyanic acid from decomposition of the urea, it's kind of a chain branching reaction where you can form many different product species. And those product species can be desired species like ammonia, or undesired species that can form very durable and robust deposits in the system which can impede the performance of the exhaust system.
What we did with catalyzed diesel exhaust fluid is we incorporated a proprietary additive package that addressed some of the challenges with decomposition of diesel exhaust fluid. The first is a physical modifier that improves the evaporation of water, and the second is a chemical modifier that accelerates the desired reactions with respect to isocyanic acid. So the hydrolysis of the isocyanic acid to form ammonia in the system.

LP: So I know one of the highlights of this technology is that it targets problematic deposits. What are these deposits? How does that work?

CH: So these deposits are generated from the diesel exhaust fluid, and it's simply undesired reaction pathway is taking place over the preferred reaction pathways. And so they consist of urea or urea-derived products. The problem is is that they can impede flow in the exhaust system, they can mess up the uniformity or distribution of ammonia in the exhaust system which can prohibit the NO_X conversion performance that a manufacturer or a customer would expect of the exhaust system.

They also, because they are relatively stable in the exhaust, manufacturers have to develop regeneration strategies where they artificially increase the temperature of the exhaust to decompose these deposits. And of course, any time you artificially increase the temperature of the exhaust system, you're burning extra fuel. And so you're not only increasing the cost to the consumer, but you're also increasing the CO₂ and greenhouse gas emissions to the environment.

LP: So is the catalyzed diesel exhaust fluid already in use? Has it already hit the road?

CH: It has not hit commercial production. So we are in the final stages of development of the solution. So we have performed some optimization studies, some demonstration studies in a laboratory setting. The next step is we're starting to engage engine manufacturers, truck and vehicle manufacturers in order to conduct field trials and get it actually out there on vehicles, on the road with a target of entering commercial production within the next three years.

LP: What are your thoughts on winning an R&D 100 Award for this technology?

CH: Probably the most exciting part of it is the visibility that it gets in the industry. So with that, we have strong hopes that we can get this accepted into high-volume commercial production so that we can further improve emissions from mobile sources. The Cat-DEF technology is relatively unique in that it is able to simultaneously address criteria pollutant emissions in the form of NO_X, but it also can reduce greenhouse gas emissions because you're reducing the deposits that are formed in the exhaust system and reducing the amount of fuel required to remove said deposits.

And so it really is a unique technology, because typically in the automotive space and with respect to internal combustion engines, you're either improving greenhouse gas and fuel consumption while making increases or resulting in increases in criteria pollutant emissions or vice versa. And this particular technology has the ability to improve both greenhouse gas emissions as well as criteria pollutant emissions at the same time.

LP: All right. And we can all get behind cleaner air, of course. And that's why this is truly a technology that benefits all of us, benefits everyone. So thank you, Cary, for being here, and congratulations again to your team.

CH: Thank you, Lisa. Very much appreciated.

LP: And finally, our R&D 100 Award-winning technology, Floodlight. Floodlight, a cutting-edge software tool, illuminates patterns in complex chemical data, identifying chemicals present in products we use every day. We featured Floodlight in Episode 29. You can find that discussion at podcast.swri.org.

Computer scientist Michael Hartnett is back to talk about what's new with Floodlight and the team's big win. Michael, thanks for coming back, and congratulations.
 

Courtesy of SwRI

SwRI Computer Scientist Michael Hartnett co-created Floodlight™. Using machine learning, the software quickly recognizes and analyzes chemical signals. Floodlight developers are now expanding the software’s capabilities, building a suite of tools to simplify the process of sorting through complex chemical data.

Michael Hartnett (MH): Thanks, Lisa, it's great to be here.

LP: So Michael, for those listeners that haven't heard Episode 29 yet, please recap Floodlight for us. What is it? How do you describe it?

MH: Yeah. So Floodlight is a machine learning-driven decision support system. So it helps chemists perform high-throughput non-targeted analysis, which is, we can contrast that to traditional targeted analysis. I use the example of a groundwater survey. So someone might go out into the field and want to find contamination in the water supply that here in San Antonio comes from an aquifer and we all drink it.

So typically with a targeted analysis, that sample would be analyzed against a list of compounds that we're checking for the absence or presence of. And that's great. We can tell some information about known contamination. But what is missing there is all of the things not on that list. So a non-targeted analysis kind of opens that box and allows us to survey everything in the chemical sample.

Of course, with that comes a flood of data, and that was kind of the impetus of developing Floodlight, was to sift through that large amount of data and separate good signals from the instrument versus data artifacts. So Floodlight automatically removes those artifacts and just gives good data to the chemist for further processing.

LP: So it definitely clarifies what the chemist is working with, gives a better picture of all the compounds in a sample.

MH: Yeah. So it helps to remove a lot of the spurious signals that would trip somebody up when trying to do pattern analysis. If you put in bad data into downstream machine learning algorithms, you're going to pick up a lot of noise and background that's not important or even real data.

LP: So one of the highlights of this technology is its speed. How fast is it compared to what was done before, a chemist manually going through samples?

MH: So that manual process for a single sample would take hours. And it's very tedious work as well, I should add. Just looking at a few graphs some information over and over again, thousands of times, potentially. So that would take hours. Floodlight can process the data in under a minute.

LP: OK, huge, huge difference in speed there, saving the chemist tons of time.

MH: And boredom.

LP: And boredom, yeah. That's true, it's just, yeah, definitely a useful tool. So any new developments with the technology since we last spoke back in March?
MH: Yes. So an important part of machine learning, in order to stay relevant, is active learning. So what that means is periodically retraining the machine learning algorithm that's underlying Floodlight with new data so that it can keep up and generalize well to new chemical samples that it hasn't seen before.

So we're actively using that process to keep our machine learning algorithm up to date and most beneficial to avoid errors and problems from previous patterns it's learned that might not be accurate.

Another part is expanding this idea, this technological concept to other instruments as well. So we focused on two-dimensional gas chromatography paired with mass spec. We're also working to support liquid chromatography, Fourier transform infrared spectroscopy, autonomous spectroscopy, all kinds of new data to look at. Trying to transfer what we've learned developing Floodlight into new areas as well.

LP: All right. So what has been the response to Floodlight? Who is using it?

MH: The response has been pretty great. So internally, we've expanded our team up to six technical divisions out of the nine here at Southwest Research. We're pulling in domain experts from various fields using Floodlight with their workflows as well. And we're also working to integrate Floodlight into the workflows of other laboratories externally as well.

LP: All right. Do you expect that demand to just keep increasing as time goes on? What kind of feedback are you getting?

MH: Yeah. So non-targeted analysis is a rapidly growing field because of the power of all that new information going off of the target list. And exploring chemical samples in their full reality has been gaining popularity because it helps us to understand more about the world around us and the world within us as well.

So there's a number of fields, like exposomics, environmental chemistry that have begun to adopt this as an analytical technique. And that continues to grow rapidly. So we just expect everyone should be using non-targeted analysis in the future. One of the barriers to that is the problem that we're solving with our technologies here at Southwest Research. By removing the manual burden placed on the chemist to perform this technique, we're able to push the non-target analysis into the future and make it feasible to do across the board.

LP: So you said we are analyzing the world around us and the world within us. What do you mean by that?

MH: So we're surrounded by chemicals. Everything around us from the carpet to the curtains to the food we eat, everything is containing complex chemical mixtures, and some of that is consumed by us. And that has interactions in our own body. So it's exposomics is the study of the external chemical exposure and what internal effects that has on us as humans. So it's a fascinating field that has health effects, implications into how we target precision medicine and how we understand system biology as well.

LP: Yeah. Definitely an important topic for everyone. So the Floodlight team received a major recognition with a 2021 R&D 100 Award, and it's been recognized as one of the most significant innovations of the year, and for good reason, we just heard plenty of them. What does this honor mean to the team and for the future of Floodlight?

MH: It means a lot. Personally, the most enjoyment I get out of the work I do is bringing complex and challenging ideas into reality, and receiving this recognition is part of making the technology real. I think it will help us to grow and have our technology transferred to more industries and more collaborators.

As for the future of Floodlight, our vision is to develop a suite of software that we call Lighthouse. We hope that Floodlight is a key component of that software suite, and we want to support additional instruments to expand out the chemical space that we can see and analyze with our software suite.

We're also hoping to expand the application of our tools. So right now I mentioned that we're working with exposomics and the health effects of chemistry. We also want to look at environmental chemistry, like emissions from manufacturing plants. We're also looking at forensics applications as well. How we can keep ourselves safe. Just trying to get more application of these tools to better the world around us.

LP: All right. So much more ahead for this dynamic technology. Thanks for coming back to the podcast, Michael. Floodlight is always a fascinating topic. And congratulations again to you and your team on that R&D 100 Award.

MH: Thanks, Lisa.

And thank you to our listeners for learning along with us today and celebrating our 2021 R&D 100 Award winners. You can hear all of our Technology Today episodes and see photos and complete transcripts at podcast.swri.org. Remember to share our podcast and subscribe on your favorite podcast platform.

Want to see what else we're up to? Connect with Southwest Research Institute on Facebook, Instagram, Twitter, LinkedIn, and YouTube. Check out the Technology Today Magazine at technologytoday.swri.org. And now is a great time to become an SwRI problem solver. Visit our career page at SwRI.jobs.

Ian McKinney and Bryan Ortiz are the podcast audio engineers and editors. I am producer and host, Lisa Peña.

Thanks for listening.

[MUSIC PLAYING]