One of the main objectives of HugYourEngine is to shine a spotlight on individuals who are helping make internal combustion engines cleaner and more efficient. This spotlight is on Gina Magnotti.
Dr. Gina Magnotti is a Research Scientist at Argonne National Laboratory. She joined Argonne in 2017 as a Postdoctoral Appointee after receiving her PhD from Georgia Institute of Technology. In her research, she develops physics-based and data-driven models for multiphase flows, with a particular focus on fuel injection applications. Gina’s current research interests include fuel spray atomization, cavitation-induced erosion, machine learning, and emulation. She is an active member of the Institute for Liquid Atomization and Spray Systems (ILASS), American Society of Mechanical Engineers (ASME), and Society of Automotive Engineers (SAE).
Kelly Senecal (KS): When and why did you get started in combustion research?
Gina Magnotti (GM): As a mechanical engineering undergrad, I loved my thermal fluid science and controls classes and wanted to work on interesting problems that would allow me to use my knowledge from all of these subjects. In my Thermo II class, I learned that engines are pretty much an engineer’s playground! The complex physics that underlie its operation can be broken down into its fundamental components by using conservation laws and engineering principles. For senior design, I convinced my friends to work with a master’s student who was modifying a single-cylinder diesel engine to operate in homogeneous charge compression ignition (HCCI) mode. This experience solidified my decision to pursue a research career in this area. I then had the opportunity to go to grad school at Georgia Tech and focus on fuel injection and spray physics. Thus began my path to becoming a [self-proclaimed] sprays nerd!
KS: Describe your lab facilities.
GM: As a member of the Multi-Physics Computation group at Argonne, I get to work with some great people and resources. For example, I work with Dr. Riccardo Scarcelli, who has been previously highlighted on HugYourEngine – you should definitely check out his story and work on modeling advanced ignition systems! As a computational researcher, my work focuses on developing predictive simulation tools for fuel injection systems. Fuel injection and spray formation are complex problems to unpack, so we utilize a broad suite of experimental and computational facilities to interrogate the underlying physics. For example, we work with experimentalists at Argonne’s Advanced Photon Source, who utilize X-ray diagnostics to characterize the turbulent multiphase flow development inside of injectors. When you realize that diesel injector orifices are on the same length scale as the diameter of a human hair, it’s incredible to see that fine scale details in the injector geometry and flow structures can have large impacts on engine performance and emissions.
Additionally, we are fortunate at Argonne to have access to powerful computing resources that allow us to run detailed and large-scale simulations of fuel injectors and engines. The picture below is of me right in front of Mira in 2017, a 10 petaflop machine which at the time was one of the fastest supercomputers in the world. In 2021, Aurora is slated to come online and will be Argonne’s first exascale computer. To put all of that into perspective, Aurora will be able to perform calculations as fast as the human brain, and 100 times faster than Mira. This kind of computing power allows us to ask different scientific questions and affords us the opportunity to simultaneously perform simulations, data science, and learning.
KS: What are you working on now?
GM: One of the projects I am working on is funded by DOE’s Vehicle Technologies Office (VTO) to develop predictive simulation tools for compression ignition engines. As a part of that effort, I am leading the development of a predictive model that can represent damage inside of an injector due to cavitation-induced erosion. This is a critical problem facing the heavy duty engine community because cavitation-induced erosion compromises reliable fuel delivery from the injector over time. To date, industry has not had a simulation tool that can link injector design, scheduling, and operation to erosion risk. We’ve developed the cavitation-induced erosion risk assessment (CIERA) tool that links multiphase flow predictions with the progress towards material failure. So far, we’ve seen promising results in applying CIERA to injector simulations and we’re working with our experimental colleagues to further validate and fine tune our model using X-ray images from injector endurance tests.
Our group routinely uses internal nozzle flow simulations as a valuable tool to understand the injector performance and the initial spray development. However, these simulations are computationally expensive to run. To help expedite the time-to-solution of multiphase flow simulations, I am working with some colleagues at Argonne to explore the use of machine learning tools. Using Argonne’s clusters and supercomputer Theta, we are generating an injector database from a large number of CFD simulations. This database contains important information on how design parameters, fuel properties, and operating conditions impact injector performance. We are using machine learning frameworks like autoencoders and deep neural networks to develop reduced order models and emulators that can predict injection conditions for new design points.
KS: What’s your favorite type of flame?
GM: Diffusion flames! I have a love for diesel engines, and conventional diesel combustion is characterized by mixing-controlled turbulent diffusion flames. I also enjoy making and using candles, and they are a classic example of a laminar diffusion flame.
KS: What’s your favorite fuel?
GM: Diesel fuel has a lot going for it. By and large, it is the fuel powering diesel engines that have the durability, fuel economy, and torque needed to move nearly 70% of our nation’s goods, and provide reliable power output for the marine, rail, mining, construction, and agriculture sectors. But breaking the soot-NOx tradeoff remains a challenge. The promise of and challenges associated with conventional diesel combustion is what motivates much of my work. We have the computational models to simulate conventional diesel combustion, but predictive simulations are lacking for exploring new combustion regimes and injection strategies and their ultimate impact on engine performance and emissions.
KS: What advice would you give students thinking about going into combustion research?
GM: If combustion excites you, I would say to go for it! Combustion research is such a broad field that covers a wide range of applications - there really is something for everyone. If I could give one piece of advice, it would be to get involved in your professional societies as soon as possible. Attending conferences is a great way to stay up to date with the latest scientific and technological advances, understand where the field is headed, and meet members of the research community. This can also open up opportunities for you to eventually serve as a reviewer and session chair, which is a great way to give back to the community.
KS: Is the IC Engine dead?
GM: As you like to say, “the future is eclectic,” and I think there is a lot of truth to that. The use of electric, hybrid, and conventional vehicles (and the energy sources required to power them) to transport people and goods can be structured as an optimization problem. There is no single transport solution that is optimal under all conditions and constraints. While electric and hybrid vehicles may be better suited under urban driving conditions that are characterized by short distances, frequent stop and go drive cycles, and availability of charging stations, IC engines may be better suited for long-haul driving scenarios or in regions where investment in charging infrastructure has not yet happened. Based on the recent and projected progress in electric, hybrid and conventional vehicle technologies, it only makes sense for IC engines to remain a part of the transportation solution.
KS: How is your work helping improve fuel efficiency or reduce emissions?
GM: The development of predictive computational design tools is key to realizing improvements in fuel efficiency and reduction in emissions. As a computational researcher, my job is to distill the key physics into a computationally efficient model that can be used in engine simulations. By making these tools available to industry, the latest advances in simulating fuel injection and spray development can be used to accelerate the time-to-design of next generation engines and control strategies.