Bulletin of the American Physical Society
74th Annual Meeting of the APS Division of Fluid Dynamics
Volume 66, Number 17
Sunday–Tuesday, November 21–23, 2021; Phoenix Convention Center, Phoenix, Arizona
Session H09: Reacting Flows: LES |
Hide Abstracts |
Chair: Arash Nouri, University of Pittsburg Room: North 124 A |
Monday, November 22, 2021 8:00AM - 8:13AM |
H09.00001: Dynamically Dominant Subfilter-scale Structure for LES of Flame-Turbulence Interactions James G Brasseur, Yash Shah, Yuan Xuan A key weakness of large-eddy simulation (LES) of premixed turbulent combustion is its inability to predict key resolved-subfilter scale (RS-SFS) interactions that contribute to the full advective and chemical nonlinearities at first order levels. We explore the development of mathematical forms to approximate "dynamically dominant" (DD) SFS content of key species in key chemical nonlinearities for the key reactions within the species source terms, without the need for SFS grid support. Using DNS of 3D flame-turbulence interactions with large scale separation between integral/flame and flame/Kolmogorov scales, we identify the DD SFS content first in Fourier space by down-selecting from linear sums of nonlinear terms that contribute to RS-SFS interactions. From inverse-transformed structure in physical space, we find that the DD SFS species concentrations are localized to the lower curvature regions of the flame with two basic structural forms relative to the flame. The two relatively simple structural forms can be represented with 1D mathematical forms over discretized flame surfaces. We argue that these forms are likely generalizable to flame-turbulence interaction regimes more broadly. Key chemical nonlinearities miss over 40% of the SFS contributions to the source term that is reduced to less than 1% when the DD SFS contribution is included, suggesting a viable path to a new SFS modeling strategy for LES of premixed turbulent combustion. |
Monday, November 22, 2021 8:13AM - 8:26AM |
H09.00002: Subgrid-scale Models with Interpretable Machine Learning in LES of Transcritical Reacting Flows Wai Tong Chung, Aashwin Mishra, Matthias Ihme Many practical combustion systems operate under high pressures that surpass the thermodynamic critical limit of fuel-oxidizer mixtures. One challenge, which arises from the resulting real fluid behavior, includes the validity of existing subgrid-scale (SGS) models in large-eddy simulations of these systems. Data-driven methods can provide accurate closure in simulations of turbulent flames, but often lack interpretability, wherein they provide answers but no insight into their underlying rationale. The objective of this study is to investigate the accuracy of SGS models from conventional physics-driven approaches and an interpretable machine learning algorithm, i.e., the random forest, in a turbulent transcritical non-premixed flame. To this end, a priori analysis is performed on direct numerical simulation data of transcritical liquid oxygen/gaseous-methane (LOX/GCH4) inert and reacting flows. Results demonstrate that random forests can model SGS stresses as accurately as algebraic models, when trained on a sufficiently representative database. The random forest feature importance score is shown to provide insight that can be applied towards discovering SGS models through sparse regression. |
Monday, November 22, 2021 8:26AM - 8:39AM |
H09.00003: A dual grid strategy for subgrid reaction-rate closure using linear eddy mixing model for large-eddy simulation of turbulent combustion Reetesh Ranjan The reaction-rate closure for large-eddy simulation (LES) of turbulent combustion, referred to as RRLES, employs the linear eddy mixing (LEM) model at the subgrid for closure of the filtered reaction-rate term. The originally proposed RRLES strategy used a multilevel adaptive mesh refinement (AMR) framework to address some of the limitations of the well-established LEMLES strategy. In comparison to LEMLES, where the subgrid scalar field evolved using the LEM model is coupled with resolved level through a Lagrangian transport, RRLES only employs LEM to model the reaction-rate term. To enable the application of RRLES to complex geometries, a single grid-based strategy was developed. However, the multilevel technique tends to be more accurate for capturing the subgrid turbulence-chemistry interactions. In the present study, a local dual grid-based strategy is developed, which can potentially be used with different grid topologies, without the need for an AMR. The proposed approach is evaluated by simulating freely propagating methane/air turbulent premixed flames at different operating conditions. The effectiveness of the approach is assessed by comparing features of flame-turbulence interactions obtained from the single grid-based RRLES and reference direct numerical simulations. |
Monday, November 22, 2021 8:39AM - 8:52AM |
H09.00004: Large Eddy Simulation and Modal Analysis of the Dual Independent Swirl Combustor Nicholas Arnold-Medabalimi, Cheng Huang, Karthik Duraisamy Modern gas turbine systems face increased constraints from economic and environmental perspectives. These limits have pushed operations into instability-prone design points. To study the underlying physics and develop strategies to mitigate these issues, a laboratory-based gas turbine model combustor has been developed at the University of Michigan. This new combustor features a dual swirler setup with large, independently fed plenums with transverse fuel injection designed for studying active control strategies for gas turbine systems with an integrated control system. In this work, LES reacting flow simulations are conducted, and the simulations are validated against experimentally measured pressure signals, chemiluminescence imaging, and unsteady high-frequency PIV. The dynamics of coherent structures are studied using Koopman analysis. This work is part of an initiative to develop reduced models for use in the design of an active control strategy to suppress combustion instabilities. |
Monday, November 22, 2021 8:52AM - 9:05AM |
H09.00005: PeleLM-FDF Large Eddy Simulator of Turbulent Reactive Flows Aidyn Aitzhan, Shervin Sammak, Peyman Givi, Arash G Nouri After its success in several computational packages, the FDF is implemented into the PeleLM [1]. This is a finite volume code suitable for numerical simulation of low-Mach number, chemically reactive flows. The resulting PeleLM-FDF software is shown to provide an efficient and very reliable means of conducting LES of turbulent combustion. This is demonstrated by its implementation for the prediction of a CO/H2 jet flame modeled with skeletal kinetics. The high level of consistency and accuracy of the new simulator is assessed via comparison with DNS results of the same flame in capturing some of the detailed statistics, including joint PDFs and the conditional averages of the reacting field. |
Monday, November 22, 2021 9:05AM - 9:18AM |
H09.00006: Presumed subfilter PDF model for finite-rate oxidation of soot in turbulent reacting flows Hernando Maldonado Colman, Michael E Mueller Modeling soot evolution in turbulent reacting flows using Large Eddy Simulation is challenging due to the complex subfilter soot-turbulence-chemistry interactions. Soot particles form at fuel-rich mixtures and are subsequently oxidized as they are transported toward fuel-lean mixtures. In previous work, this phenomenology was explicitly encoded into a presumed subfilter PDF model for soot by confining soot to subfilter mixtures where growth rates exceed oxidation rates. However, this model implicitly assumed that oxidation is infinitely fast. In this work, a new presumed subfilter PDF model for soot is proposed that accounts for finite rate soot oxidation. The distribution of soot with respect to the flame structure (mixture fraction) is determined by comparing the local relative motion of diffusionless soot particles with respect to the flow with the local oxidation rate. When the oxidation rate is suppressed or the transport rate very fast, soot is allowed to penetrate further into fuel-lean mixtures. The new model is validated a priori against DNS databases of turbulent nonpremixed jet flames and then a posteriori against experimental measurements in a laboratory-scale turbulent jet flame. |
Monday, November 22, 2021 9:18AM - 9:31AM |
H09.00007: Spectral characteristics of the heat release rate in bluff body and swirl-stabilised flames Ankit D Kumar, James C Massey, Zhi X Chen, Nedunchezhian Swaminathan Direct combustion noise results from pressure fluctuations produced by the unsteady heat release rate (HRR). Predicting this broadband noise is challenging and its mitigation is important, as jet noise levels have declined significantly. The power spectral density of the direct noise —a measurable quantity in experiments, is related to the product of local spectral densities of a Green's function and unsteady HRR (Ψq), intregrated over flame volume appropriately. The quantity Ψq is challenging to measure but easy to obtain from large eddy simulation (LES) results. The behaviour of Ψq for a wide range of thermochemical and turbulence conditions is investigated. Three combustors operating at atmospheric conditions with CH4-air mixtures are studied: the dual-swirl burner developed by DLR, the PRECCINSTA single-swirl burner and a premixed bluff body burner. These cases are simulated using the LES-FlaRe (Flamelets revised for physical consistencies) model for sub-grid scale combustion. The dependence of Ψq on spatial location and flame configuration and its relation to volume-integrated HRR spectra are studied. It is observed that Ψq can be modelled using velocity and scalar spectra which will be discussed in the presentation. |
Monday, November 22, 2021 9:31AM - 9:44AM |
H09.00008: Multi-Component Reduced-Order Model Framework for Rocket Engines Cheng Huang, Karthik Duraisamy, Charles Merkle Combustion dynamics is characterized by the coupling between heat release, hydrodynamics and acoustics. In combustion engines, this complex coupling can lead to combustion instabilities that can cause devastating engine failures. Even with advances in modern computational capabilities, high-fidelity (e.g., Large Eddy) simulations of full-scale combustors remain out of reach. In this work, we develop a multi-component Reduced Order Model (ROM) framework to enable efficient prediction of combustion dynamics in rocket engines. Projection-based model reduction methods, leveraging the least-squares Petrov Galerkin methods, are adapted to obtain efficient and accurate ROMs to represent complex combustion dynamics near rocket injector elements. These ROMs are effectively built on small-domain (e.g., 2-3 elements) compared to the full-scale engine (> 100 elements) using offline LES simulations with excitation of essential dynamics at the training stage. Computational fluid dynamic (CFD) models are used in regions in which the dominant features are acoustically propagating waves (manifolds, injector posts and nozzle) and coupled to the aforementioned network of ROMs which represent the near-injector regions. The framework is demonstrated to predict combustion instability characteristics in a laboratory multi-element rocket combustor. |
Monday, November 22, 2021 9:44AM - 9:57AM |
H09.00009: Large eddy simulation of Darmstadt multi-regime burner with eddy dissipation concept and finite rate model Lorenzo Angelilli, Pietro Paolo Ciottoli, Francisco E Hernandez Perez, Riccardo Malpica Galassi, Mauro Valorani, Hong G Im The novel TU Darmstadt multi-regime burner constitutes a valid candidate to evaluate the performance of existing combustion models when premixed and non-premixed regimes co-exist simultaneously. The aim of this work is to investigate the premixed and non-premixed flame structure observed in such configuration via large eddy simulations while assessing the subgrid-scale chemistry model. The chemical source term is determined with the finite rate model based on the Arrhenius law combined with the eddy dissipation concept for the sub-grid closure. Two simulations are performed with OpenFOAM employing a reduced chemical kinetics mechanism and a detailed one. The numerical results are validated against the existing experimental data and the conditional flame structure to the mixture fraction and progress variable is discussed. Furthermore, the computational singular perturbation (CSP) and tangential stretching rate (TSR) analyses are also conducted to examine the non-adiabatic effects due to the cold bluff-body at the inflow and describe quantitatively how the diffusive and convective transport processes are connected to the the chemical evolution of this complex system. |
Monday, November 22, 2021 9:57AM - 10:10AM |
H09.00010: Linear response analysis and hydrodynamic transfer function of a reacting swirling jet Parth Patki, Benjamin L Emerson, Christopher M Douglas, Timothy C Lieuwen This presentation explores the linear hydrodynamic response of reacting swirling jets. The framework for this study is an analysis of the linearized reacting Navier-Stokes equations. The base flow for this analysis is a single modeled swirling jet. The homogeneous solution is calculated to explore the intrinsic instability characteristics, and then a linear response analysis algorithm is built to approximate the response to the additional non-linear terms of the inhomogeneous form. The numerical methods of this paper focus on discretization of the Navier-Stokes equations and applying boundary conditions for a specified geometry to obtain a weak variational form that can be solved using Taylor-Hood finite elements. The result of the analysis is a transfer function for the hydrodynamic amplification of external excitations. Together with this transfer function, the analysis elucidates the spatial structure of the flow’s externally excited coherent structures. |
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