Session HV: Reacting Flows I

High-Order Quasi-Steady State Assumption for Chemistry Reduction

The quasi-steady state assumption (QSSA) is one of the most physically compelling concepts used for reducing large chemical kinetic mechanisms. However, QSSA is not helpful in chemistry regimes where experiential knowledge of the kinetic set behavior is lacking. This has lead to the development of more advanced kinetics reduction schemes which, while mathematically precise, are physically less insightful than QSSA. In this work, we develop a higher-order QSSA (HO-QSSA) formulation which is mathematically precise while preserving the physical clarity of the original QSSA. The talk will present examples of chemistry reduction and demonstrate the connection between HO-QSSA and other currently used reduction schemes such as Intrinsic Low Dimensional Manifold (ILDM).   

Dynamics in reactive bubbly flow

Multiphase flow in microfluidic channels encompasses a rich collection of phenomena of widespread interest in both fundamental and technological context. While studies on non reactive multiphase flow focus on the dynamics of bubble breakup, coalescence and stability, a reactive multiphase flow opens up a broader spectrum of dynamics, like nucleation, growth and detachment of bubbles as well as the secondary mixing in the slugs during these processes. Our interest lies in the flow in an electrochemical microfluidic fuel cell with liquid reactants reacting at catalyst walls producing gaseous products which choke the fuel cell efficiency due to uncontrolled bubbly flow. This challenge is an opportunity in itself provided the multiphase flow dynamics can be characterized to achieve a stable Taylor regime. Taylor regime allows for promisingly high efficiencies due to improved mass transfer of reactants to the concentration boundary layer of the electrodes achieved by the secondary flow in the liquid phase present between bubbles. Here, I will experimentally explore the different regimes of reactive bubbly flow in a microchannel. The phase diagram of the reactive multiphase flows would be used to identify the stable regime for efficient fuel cell operation. Further, I will study the mass transfer in the presence of multiphase flow to regimes of enhanced mass transfer, and compare it with numerical models.   

An ODT-Based Multiscale Radiative Transport Model in Participating (absorbing-emitting) gray media

A multiscale formulation for thermal radiation transport in participating (absorbing-emitting) gray media is developed. The model is based on a grid topology using the one-dimensional turbulence (ODT) model framework and the photon Monte-Carlo (PMC) method for radiative transport. The formulation is implemented within the context of large-eddy simulation (LES). The ODT solution for the evolution of temperature field are based on, (a) a deterministic implementation for diffusion, advection and reaction, and (b) a stochastic implementation for LES subgrid scale advective transport. Specific rules for ray tracing and the modeling of emission and absorption processes are designed to capture turbulence-radiation interactions and account for subgrid scale contribution due to residual temperature fluctuations. The model is implemented for turbulent premixed flame problem and compared with DNS. The results yield excellent agreement between the LES-ODT model for radiative transport and DNS predictions.   

Simulation of Compressible Reacting Flow using the Parallel Wavelet Adaptive Multiresolution Representation

The Wavelet Adaptive Multiresolution Representation (WAMR) provides spatial adaptivity which automatically supplies local grid resolution based on the demands of the solution. The WAMR method allows for a wide range of spatial scales to be captured while minimizing the number of degrees of freedom needed in the solution. Recently, the WAMR algorithm has been adapted for use on massively parallel computer architectures using an MPI-based domain decomposition approach. Results for parallel performance on test problems will be presented. Additionally, results from the application of the parallel WAMR algorithm to multidimensional compressible, reacting flows will be shown. Calculations include propagation of a detonation in a square channel and a combustible mixing layer using detailed chemistry models.   

An ODT-Based Flame-Embedding Approach for Turbulent Non-Premixed Combustion

A multiscale formulation is implemented to capture finite-rate chemistry in turbulent non-premixed flames. The formulation is based on a coupling of a large-eddy simulation (LES) solution for the transport of the filtered mixture fraction field and one-dimensional solutions embedded on the flame brush, which are tracked with the stoichiometric value of the filtered mixture fraction. The one-dimensional solutions are based on the one-dimensional turbulence (ODT) model. In the ODT solutions for the evolution of the reactive scalars, diffusion and reaction processes are implemented deterministically while LES subgrid scale advective transport have stochastic implementation. The ODT domains are allowed to be advected with the flow such that they remain attached to the flame brush. The formulation enables the implementation of effects that capture the fluctuations of reactive scalars within the LES grid, including finite-rate and non-equilibrium chemistry effects and radiation-turbulence interactions.   

A dynamic approach for nonequilibrium modeling of subfilter scalar dissipation rate in combustion LES

The filtered scalar dissipation rate is a fundamental parameter in combustion LES, appearing as an input parameter in all combustion models. Since subfilter dissipation is a small scale quantity, conventional dynamic modeling approaches are not valid. Typically, this quantity is obtained by assuming that the production of scalar variance at filtered scales is exactly balanced by dissipation, leading to an algebraic relation for the dissipation rate. However, this local equilibrium assumption is highly restrictive since it neglects spatial transport. Here, we propose a new modeling approach that overcomes this limitation. This nonequilibrium model uses a dynamic approach along with the scalar variance transport equation to determine the dissipation rate. \emph{A priori} studies using DNS are used to evaluate the accuracy of the method. In addition, a novel \emph{a posteriori} method is used to assess model performance in LES calculations.   

Quantitative Visualization of transverse annular jets

Transverse injection of fluid into an annular jet is a mechanism resulting in good mixing and is therefore utilized in engineering applications such as pintle rocket engines. Vigorous mixing occurs between the two jets. However, much of what we know about the flow behavior of such devices has been learned empirically with very limited studies exploring the fluid dynamics. The geometry under investigation is an axisymmetric radial jet of variable width impinging on a fixed annular jet. The main capability of the current facility is to reproduce start-up and quasi-steady flow conditions through the use of a fast acting valve which opens a pressurized air reservoir. The flow is then observed using laser interferometry giving quantitative measurements of the density fields that are compared with computations. The main parameters under investigation were the reservoir pressure and the area ratio between the axial and radial jet. Modeling of the fluid mechanics was performed at NASA-JPL.   

Ex vivo Characterization of Blast Wave Impact and Spinal Cord Tissue Deformation

Primary blast injury on central nervous system is responsible for many of the war related casualties and mortalities. An ex vivo model system is developed to introduce a blast wave, generated from a shock tube, directly to spinal cord tissue sample. A high-speed shadowgraph system is utilized to visualize the development of the blast wave and its interaction with tissue sample. Surface deformation of the tissue sample is also measured for the analysis of internal stress and possible injury occurred within the tissue sample. Understanding the temporal development of the blast-tissue interaction provides valuable input for modeling blast-induced neurotrauma. Tracking the sample surface deformation as a function of time provides realistic boundary conditions for numerical simulation of injury process.   

Experimental study on intensity of change in viscosity by a chemical reaction on a liquid flow

We have recently investigated a liquid flow involving viscosity change by chemical reactions. The liquid flow involves the displacement of the more viscous liquid by the less viscous one in a Hele-Shaw cell. So far, we reported the effects of Damkohler number (Da), which is defined as the ratio between a characteristic time of fluid motion and that of chemical reaction, on the flow. In reacting liquid flows involving viscosity changes, an intensity of change in viscosity, as well as Da, is supposed to be important. In the present study, we have experimentally investigated the effect of the intensity of decrease in viscosity by the chemical reaction on the liquid Hele-Shaw flow for the condition of infinite Da. We have found a threshold value of the intensity, beyond which the flow is dramatically changed by the reaction.   

Buoyancy-driven instabilities of acid-base fronts

Chemical reactions can produce buoyancy-driven motions in solutions by changing the local density in the gravity field. Starting from a stratification of one given miscible reactant solution on top of another miscible one, convective destabilization influenced by the reaction can emerge as a combination of several non-reactive hydrodynamic instabilities, including Rayleigh-Taylor and double diffusive instabilities. In the specific case of reactions between a strong acid and a strong base, we show that all the possible scenarios reduce to the composition of only two asymptotic situations because products generated at the reactive zone are always less dense and slower diffusing than the reactants. Experiments in a vertically orientated Hele-Shaw cell confirm the theoretical predictions and can be quantitatively compared to numerical simulations of a nonlinear reaction-diffusion-convection model.