Abstract:
Our group (Raghu Kancherla, Fahad Khan, Javier Gioia, and Subith Vasu) has been significantly contributing to make direct-fired supercritical CO2 power cycles a practical reality. Supercritical CO2 power cycles attracted global energy sector’s attention due to its remarkable promise of efficiency, economics, and environmental friendliness. However, the operation of cycle in supercritical state is unconventional to modern power sector. In supercritical state the working fluid is neither a liquid nor a gas, but it will have densities like liquid and properties like a gas. As per the current state-of-the-art, the pressure of the supercritical CO2 cycle combustor is 300 bar. It is around fifteen times higher pressure than a typical combustor in existing power plants. Any experimentation at these pressures is expensive, time consuming, and even dangerous. Thus, design and development of these combustors heavily rely on high-fidelity computational tools such a CFD.
Combustion CFD solvers are computationally intensive as the nature of the combustion ordinary differential equations are more stiff, coupled, and non-linear. On top of that, supercritical state adds more complex thermal, and transport property models. Therefore, supercritical CFD simulation can’t typically fit into the memory of a single machine and requires computation to be shared on multiple communicating nodes. The effective communication between the computing nodes and the availability of quick computing resources are two primary challenges for us to compete with other researchers in this area. Currently, we have utilized Azure HPC for these intense simulations. Our presentation discusses our experience with Azure HPC, our preliminary combustor development based on our high-fidelity CFD, and various challenges and lessons learnt in provisioning & deploying resources in UCF's Azure subscription while working with various technical teams.
Author Bios:
Dr. Raghu Kancherla obtained his Ph.D. in Mechanical Engineering from University of Central Florida and current working as Postdoctoral Scholar. He has his masters from Indian Institute of Technology, Madras, India. After his Master, he served Nissan Auto India as a CFD engineer for four years. His research mainly focuses on applied and fundamental combustion modeling and analysis.
Dr. Fahad Ahmad Khan is a Research Cyberinfrastructure Facilitator with the Graduate Research IT team at UCF. He facilitates researchers by advising them on their cyberinfrastructure needs which include advanced computing resources, data storage, networking, and security & privacy. He is an alumnus of UCF (PhD CE 2019) and UET Lahore (MSc EE 2014, BSc EE 2007). His current research interests include sensor networks, approximate computing (BDD minimization), and applied machine learning (biomedical imaging & natural language processing).
For further information:
RRoCCET21 is a conference that was held virtually by CloudBank from August 10th through 12th, 2021. Its intention is to inspire you to consider utilizing the cloud in your research, by way of sharing the success stories of others. We hope the proceedings, of which this case study is a part, give you an idea of what is possible and act as a “recipe book” for mapping powerful computational resources onto your own field of inquiry.
