FLUID MODES, DEEP LEARNING, AND CONSTRUCTAL LAW
DOI:
https://doi.org/10.59277/CLC.2024.39Keywords:
Reduced-order model, Proper orthogonal decomposition, Machine learning, Constructal lawAbstract
The fluid modes generated using a proper orthogonal decomposition (POD) method were predicted for a fluidized bed and a power generation turbine. The POD-based reduced-order models were solved using either a Galerkin projection or a deep learning strategy. In both cases, as the number of the fluid modes increased, the modes appeared to fragment/bifurcate, indicating that these modes follow the constructal law.
References
(1) Mahajan A.J., Dowell E.H., Bliss D.B., Eigenvalue Calculation Procedure for Euler/Navier-Stokes Solvers with Application to Flows over Airfoils. Journal of Computational Physics, 97, pp. 398–413 (1991).
(2) Holmes P, Lumley, J.L., Berkooz, G., Turbulence, Coherent Structures, Dynamical Systems and Symmetry, Cambridge University Press, 1996.
(3) Lindholm A., Wahlstrom, N., Lindsten, F., Schon, T.B., Machine Learning: A First Course for Engineers and Scientists, 2022.
(4) Cizmas, P.G.A., Palacios A., Proper Orthogonal Decomposition of Turbine Rotor-Stator Interaction. Journal of Propulsion and Power, 19, pp. 268–281 (2003).
(5) Yuan T., Cizmas P.G., O’Brien T.A., Reduced-Order Model for a Bubbling Fluidized Bed based on Proper Orthogonal Decomposition, Computers & Chemical Engineering, 30, pp. 243–259 (2005).
