Bulletin of the American Physical Society
77th Annual Meeting of the Division of Fluid Dynamics
Sunday–Tuesday, November 24–26, 2024; Salt Lake City, Utah
Session J11: Nonlinear Dynamics: Data-Driven Methods |
Hide Abstracts |
Chair: Krithika Manohar, University of Washington Room: 155 A |
Sunday, November 24, 2024 5:50PM - 6:03PM |
J11.00001: Data-driven modeling of multi-timescale systems by mixtures of neural ordinary differential equations Jake Buzhardt, Michael David Graham
|
Sunday, November 24, 2024 6:03PM - 6:16PM |
J11.00002: Active Learning of Model Discrepancy with Sequential Optimal Experimental Design Huchen Yang, Chuanqi Chen, Jinlong Wu Digital twins have been actively explored in many engineering applications, such as manufacturing and autonomous systems. However, model discrepancy is ubiquitous in most digital twin models. In recent years, data-driven modeling techniques have been demonstrated promising in characterizing the model discrepancy in existing models, while the training data for the learning of model discrepancy is often obtained empirically and an active approach of gathering informative data can potentially benefit the learning of model discrepancy. On the other hand, optimal experimental design (OED) provides a systematic approach to gather the most informative data, but its performance is often negatively impacted by the model discrepancy. In this talk, we build on sequential Bayesian OED and propose an efficient approach to iteratively learn the model discrepancy based on the data from the Bayesian OED. The results show that the proposed method is robust to escape local minimum and is efficient enough to handle high-dimensional model discrepancy, making use of data indicated by the sequential Bayesian OED. We also demonstrate that the proposed method is compatible with both classical solvers and modern auto-differentiable solvers. |
Sunday, November 24, 2024 6:16PM - 6:29PM |
J11.00003: Sensor Optimization of Nuclear Reactor Subsystems within a Digital Twin Network Niharika Karnik, Krithika Manohar, Mohammad G Abdo Nuclear power plants require continuous monitoring of systems, structures, and components for safe and efficient operations. Critical safety testing of new fuel compositions and power transient analysis on core temperatures are achieved through modeling and simulations, capturing dynamics associated with failure modes to create digital twins. Accurate reconstruction of temperature, pressure, and velocity fields from sensor measurements is essential for effective communication between physical experiments and models. Due to challenging conditions and spatial limitations, sensor placement in nuclear subsystems is highly constrained. This study develops a data-driven optimized constrained sensor placement algorithm to reconstruct the field of interest within a TRi-structural ISOtropic (TRISO) fuel irradiation experiment, a lumped parameter model of a nuclear fuel test rod, and a steam generator. The optimization process leverages reduced-order models of flow physics to achieve highly accurate full-field reconstructions of responses of interest, quantify noise-induced uncertainty, and identify physically feasible sensor locations. These precise sensor-based reconstructions lay the groundwork for digital twinning of subsystems, ultimately leading to a comprehensive digital twin aggregate of a nuclear power plant. |
Sunday, November 24, 2024 6:29PM - 6:42PM |
J11.00004: Actuator placement optimization by determinant-based greedy method in linearized Ginzburg-Landau model Masahito Watanabe, Yasuo Sasaki, Takayuki Nagata, Keigo Yamada, Taku Nonomura, Junshi Ito, Daisuke Tsubakino In this research a numerical algorithm is developed to optimize actuator placement in a linear system with impulsive forcing so that the quantity of state can be varied effectively in various situations. Further, the availability of the proposed method is evaluated by applying it to a linearized Ginzburg-Landau model, which is known as a simple model of fluid phenomena. We consider a multidimensional linear system, where some of the elements in the input is given by a delta function, while the others are set to zero. From the physical point of view this corresponds to a situation where impulsive forcing is added to the system from multiple actuators. We apply singular value decomposition to the matrix which connects the coefficient of input and terminal state. Then, actuator locations are selected by greedy method so that the determinant of a matrix associated with right singular vectors and singular values is maximized, where greedy method is a numerical method which selects elements one by one to gain the quasi‐optimum solution of combinational problems. In the numerical simulation of linearized Ginzburg-Landau model we show that the quantity of state can be varied more effectively by actuators placed by the proposed method than by those placed randomly or according to the trace instead of the determinant of the matrix mentioned above. |
Sunday, November 24, 2024 6:42PM - 6:55PM |
J11.00005: Nonlinear and nonperiodic optimal forcing analysis on the subsonic flow around an airfoil Nobutaka Taniguchi, Yuya Ohmichi, Kojiro Suzuki Optimal forcing analysis is a method for extracting the external forces to provide the largest disturbance growth on the base flow field. This provides valuable insights into the physical mechanisms and flow controls. In this study, we proposed an optimal forcing analysis for an unsteady base flow field with a non-periodic temporal distribution considering the nonlinear time-evolution of disturbance field. This extension of the optimal forcing analysis is critical for understanding shock-relating fluid phenomena because it enables the identification of triggers that induce shock waves in the nonlinear time-evolution. |
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. |
© 2025 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