Bulletin of the American Physical Society
75th Annual Meeting of the Division of Fluid Dynamics
Volume 67, Number 19
Sunday–Tuesday, November 20–22, 2022; Indiana Convention Center, Indianapolis, Indiana.
Session G14: Energy: Storage and Combustion |
Hide Abstracts |
Chair: Jonathan Schuh, University of Illinois - Urbana Champaign Room: 141 |
Sunday, November 20, 2022 3:00PM - 3:13PM |
G14.00001: The effect of Rayleigh number on the temporal variation of the Nusselt number for buoyancy driven melting in a rectangular LHTES device. Kedar Prashant Shete, Esha Mujumdar, Steve M de Bruyn Kops, Dragoljub Kosanovic Buoyancy driven melting has been studied experimentally in a rectangular LHTES device of aspect ratio 2 at Rayleigh numbers between Ra=109 and Ra=1010 at Prandtl numbers 12 and 22. Based on the device geometry and operating Ra, it is hypothesized that if the flow is independent of the plate Reynolds number, three distinct regions should be present in the Nusselt number Nu as it varies with time, and should contain a local minimum and maximum. The highest Nu is expected in a region bounded by dimensionless time τa and τb. The melted liquid fraction η is expected to increase linearly with time for majority of the melting process, switching to an inverse exponential variation as it approaches unity. This is in contrast to melting at low Ra, where a monotonic decrease in Nu is expected, and the variation of η is expected to be inverse exponential throughout the process. The experimental results show that these local extrema occur at the same dimensionless time τ. The effect of plate Reynolds number on Nu is also explored to identify the minimum required Reynolds number. Time lapse images of the heat exchanger are used to calculate liquid fraction of the PCM. The identification and measurement of phases and calculation of η has been automated using unsupervised machine learning. Self-Organizing Maps (SOM) techniques show better performance than K-Means clustering due to their ability to identify number of clusters with limited user input. The SOM based analysis shows a linear variation of η with dimensionless time for majority of the melting process. The results indicate that although the LHTES device achieves the highest value of Nu between τa and τb, the rate of latent energy storage does not show such a maximum. |
Sunday, November 20, 2022 3:13PM - 3:26PM |
G14.00002: Utilizing water towers for pumped storage hydropower Jonathon K Schuh Pumped storage hydropower (PSH) stores electrical energy as gravitational potential energy. Water is pumped from a lower elevation reservoir to a higher one and later flows back to the lower reservoir through a turbine. For areas with naturally large elevation changes, PSH has been an effective way to store excess energy produced from renewable sources by pumping water into the higher reservoir during times of lower energy consumption (off-peak, lower associated energy cost) and discharging during times of higher energy consumption (peak, higher associated energy cost). However, areas that have relatively small elevation changes (such as the Midwest of the United States) require man-made height differences for PSH. Water towers could provide one method for obtaining the required height differences. Here, three different commercially available water tower designs (small, medium, and large) with varying pipe diameter and Pelton Wheel Turbine nozzle diameter are examined numerically to determine an optimal system configuration for energy storage. Maximum water level and Pelton Wheel blade angle are held constant across the three different water towers considered. The results suggest that each of the water towers considered has a maximum energy capacity that can be achieved with multiple combinations of pipe and nozzle diameter. Finally, historical data for differences in energy prices from the Midwest Indiana hub is used to estimate energy cost savings for the water tower PSH system. |
Sunday, November 20, 2022 3:26PM - 3:39PM |
G14.00003: Flame transitions and stabilization mechanisms in a freely falling fuel droplet encountering a co-flow Gautham Vadlamudi, Akhil Aravind, Saptarshi Basu The moving burning droplet is an interesting area of fundamental research that gives insight into spray combustion, where burning droplets encounter various flow velocities. Hence, the flame dynamics of a contactless burning droplet under free fall have been investigated in a drop tower facility, which is subjected to a co-flow. The droplet flame responds to the relative flow coherently but when it encounters a co-flow, the flame is no longer in equilibrium and tries to readjust till it equilibrates with the surrounding co-flow. In the initial readjustment period, the flame is not responding coherently with the co-flow, however, after it equilibrates the flame shape evolves based on the relative velocity. The droplet flame is observed to transition between different configurations like wake flame, reversed wake flame, and enveloped flame. In the wake, the flame is observed to stabilize based on triple-flame or bluff-body stabilization mechanism based on the relative flow velocity. The flame stabilization criteria and shape evolution are characterized based on the flow characteristics and transitions have been explained. Using a mathematical formulation based on the spring-mass system analogy, the evolution of droplet flame is predicted for different co-flow conditions imposed. |
Sunday, November 20, 2022 3:39PM - 3:52PM |
G14.00004: Application of an artificial neural network to sub-filter density function estimation for premixed combustion Hanying Yang, Tota Kobayashi, James C Massey, Yuki Minamoto, Nedunchezhian Swaminathan Machine learning (ML) is emerging as a well-suited and robust approach for modelling turbulent flames, but its powerful performance is limited by the scope of training cases. The present work aims to develop one ML approach, namely Artificial Neural Network (ANN), to resolve the lack of generalization while maintaining good accuracy. An ANN is trained using a comprehensive direct numerical simulation (DNS) dataset of Moderate or Intense Low-oxygen Dilution (MILD) combustion. A lot of premixed combustion cases, including planar flame, V-flame and swirl flame, with different fuel (hydrogen and methane) have been tested. It is observed the prediction of the Filtered Density Function (FDF) is satisfactory. This observation is further assessed by conditionally averaging FDF on the first and second moments of progress variable. Additional assessment is carried out through comparisons with the reaction rate using DNS data and the values obtained by using a presumed PDF method. It is detected that results from ANN and the presumed PDF approach are comparable, which illustrates the feasibility of replacing traditional look-up table with ANN in Large Eddy Simulation (LES) for cases distinct from the training cases. |
Sunday, November 20, 2022 3:52PM - 4:05PM |
G14.00005: Uncertainty Quantification of a Deep Learning Based Fuel Property Prediction Model Sahil Kommalapati, Pinaki Pal, Nursulu Kuzhagaliyeva, Abdullah AlRamadan, Balaji Mohan, Yuanjiang Pei, Mani Sarathy, Emre Cenker, Jihad Badra Deep learning surrogate models for predicting properties of chemical compounds and mixtures have recently been shown to be promising for enabling data-driven novel fuel design and optimization, with the aim of improving efficiency and lowering emissions from combustion engines. However, given the low interpretability of typical neural network models, uncertainty quantification (UQ) is critical to ensure the reliability of predictions as well as the training datasets, and for a principled quantification of noise and its various sources. In this study, UQ of a multi-task deep learning model that simultaneously predicts the research octane number (RON), Motor Octane Number (MON), and Yield Sooting Index (YSI) of pure components and multicomponent blends, is performed. Both epistemic and aleatoric uncertainties are incorporated by utilizing various implementations of Monte Carlo Dropout, Bayesian Neural Network (BNN), and Gaussian Negative Log Likelihood (GNLL) loss function. A comparative analysis exploring these approaches is carried out to achieve the best trade-off between accuracy and calibration of the surrogate model. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700