The Historically Black Colleges and Universities and Other Minority Institutions (HBCU-OMI) Program—sponsored by the Office of Fossil Energy and Carbon Management (FECM) and administered by the National Energy Technology Laboratory (NETL)—invests in developing a U.S. workforce that is diverse and highly skilled in science, technology, engineering, and math (STEM) through training and education grants. The HBCU-OMI Program is transformative for student researchers who will shape the future of the clean energy sector.
Over the next few months, FECM is highlighting students currently conducting research in STEM fields through the HBCU-OMI Program. Learn more about one of these students below!
Graduate Student (Ph.D.) in Computer Science
Florida International University (FIU)
Tell us about yourself and your experience in STEM.
I am a Computer Science Ph.D. candidate in the Knight Foundation School of Computing and Information Sciences department, FIU, advised by Dr. Shu-Ching Chen. As a member of the FECM/NETL HBCU-OMI program-funded project titled Development and Evaluation of a General Drag Model for Gas-Solid Flows via Physics-Informed Deep Machine Learning, I am also actively advised by Dr. Cheng-Xian Lin. My thesis aims to address several potential challenges in a data science project, including data availability and quality and finding a practical method to model strong relationships in the features extracted from unstructured data of multiple sources and modalities.
What made you start pursuing studies in a STEM field and your particular area of specialization?
I developed an early interest in the STEM fields in general, and mathematics has always been my favorite subject. Additionally, I was always fascinated by computers. From the moment I was taught how to use a computer at school, I decided to pursue studies in Computer Science. While I was pursuing my Master’s degree in Computer Science, I had the opportunity to start working on some exciting multi-disciplinary research projects related to infrastructure, sustainability, and technology adaptation. I was developing agent-based simulation models that provided various scenario-based analyses to assist in the resolution of complicated real-world problems. This experience instilled in me a strong interest in interdisciplinary research, which motivated me to pursue a Ph.D. in Computer Science. Currently, my primary focus is on applying sophisticated data science and machine learning to real-world challenges.
What research topics and/or technical areas have you worked on through this HBCU-OMI program?
Through this HBCU-OMI program, I worked mainly in the area of machine learning for computational fluid dynamics.1 More specifically, I have worked on the development of a general drag model using deep learning to predict particle drag coefficients from a variety of experimental scenarios using a wide range of features. The particle drag force coefficient is an important factor in modeling multiphase energy systems because it tells how much resistance a particle has in a fluid. In collaboration with my team members from a mechanical engineering background, we collected a large amount of particle drag data from more than thirty different published experiments that we could find in the literature. The data allowed us to investigate and develop a robust model trained to predict the particle drag from thousands of data samples.
What aspect of your work through this program are you most proud of?
Advancing the development of an accurate and general drag model is my most significant achievement thus far. With the help of my team members, I designed and developed a novel deep neural network that can adapt to different experiments and scenarios within wide parameter ranges while also producing very accurate results compared to other competing methods. My team and I believe that the new approach will be a reliable and low-cost solution for designing and operating industrial-scale design, supporting a wide range of single-particle experiments and features.
Have you published, or are you working on publishing any papers through your work in this program?
I have submitted and presented an abstract titled Development of a Deep Learning Model for Predicting the Drag Coefficients of Spherical and Non-Spherical Particles at the 2021 American Institute of Chemical Engineers (AIChE) Spring Meeting. More recently, I have led the writing of a new journal paper, currently in under-submission status, to the IEEE Transactions on Artificial Intelligence (TAI), a very priming venue of multidisciplinary studies in artificial intelligence. Meanwhile, we're making fantastic strides and achieving more fascinating findings, with further paper publications on the way.
How has your research through the HBCU-OMI program helped you to learn more about your field and career opportunities?
Through my current HBCU-OMI Program research, I learned more about research science positions outside of the traditional academic environment, such as in national labs and industry. I’m enthused about the many opportunities to work in a multidisciplinary research setting.
How do you feel your research helps to contribute to a sustainable, low-carbon energy future?
My research deals with the particles found in various technical and natural processes, including multiphase energy systems, essential for fluidized beds.2 Fluidized bed technologies have demonstrated exceptional heat transfer characteristics, allowing them to generate energy from a wide range of solid fuels while lowering toxic emissions and promoting environmental sustainability. Modeling the behavior of the particles inside the bed, on the other hand, requires a precise, robust drag model for a variety of particle shapes and flow conditions.
We developed a unique deep neural network3 architecture to predict particle drag trained on a large amount of experimental data to address the drawbacks of existing approaches. As a result, the advancement of this project will support the growth of energy, environmental, and economic benefits that bring forth the fluidized bed technology.