SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Viewgraphs are presented on computation of turbulent combustion, governing equations, closure problem, PDF modeling of turbulent reactive flows, validation cases, current projects, and collaboration with industry and technology transfer.
An assessment of the plasticity-induced crack-closure concept is made, in light of some of the questions that have been raised on the validity of the concept, and the assumptions that have been made concerning crack-dp damage below the crack-opening stress. The impact of using other crack-tip parameters, such as the cyclic crack-tip displacement, to model crack-growth rate behavior was studied. Crack-growth simulations, using a crack-closure model, showed a close relation between traditional Delta K eff, and the cyclic crack-tip displacement (Delta eff) for an aluminum alloy and a steel. Evaluations of the cyclic hysteresis energy demonstrated that the cyclic plastic damage below the crack-opening stress was negligible in the Paris crack-growth regime. Some of the standard and newly proposed remote measurement methods to determine the 'effective' crack-tip driving parameter were evaluated on middle-crack tension specimens. A potential source of the Kmax effect on crack-growth rates was studied on an aluminum alloy. Results showed that the ratio of Kmax to Kc had a strong effect on crack-growth rates at high stress ratios and at low stress ratios for very high stress levels. The crack-closure concept and the traditional crack-growth rate equations were able to correlate and predict crack-growth rates under these extreme conditions.
Summary Two-phase flow is a common occurrence in pipes of oil and gas developments. Current predictive tools are based on the mechanistic two-fluid model, which requires the use of closure relations to predict integral flow parameters such as liquid holdup (or void fraction) and pressure gradient. However, these closure relations carry the highest uncertainties in the model. In particular, significant discrepancies have been found between experimental data and closure relations for the Taylor bubble velocity in slug flow, which has been determined to strongly affect the mechanistic model predictions (Lizarraga-García 2016). In this work, we study the behavior of Taylor bubbles in vertical and inclined pipes with upward and downward flow using a validated 3D computational fluid dynamics (CFD) approach with level set method implemented in a commercial code. A total of 56 cases are simulated, covering a wide range of fluid properties, pipe diameters, and inclination angles: Eo ∈ [10, 700]; Mo ∈ [1×10–6, 5×103]; ReSL ∈ [–40, 10]; θ ∈ [5°, 90°]. For bubbles in vertical upward flows, the simulated distribution parameter, C0, is successfully compared with an existing model. However, the C0 values of downward and inclined slug flows where the bubble becomes asymmetric are shown to be significantly different from their respective vertical upward flow values, and no current model exists for the fluids simulated here. The main contributions of this work are (1) the relatively large 3D numerical database generated for this type of flow, (2) the study of the asymmetric nature of inclined and some vertical downward slug flows, and (3) the analysis of its impact on the distribution parameter, C0.
The ability of the Abaqus progressive Virtual Crack Closure Technique (VCCT) to model delamination in composite structures was investigated for static, postbuckling, and fatigue loads. Preliminary evaluations were performed using simple Double Cantilever Beam (DCB) and Mixed-Mode Bending (MMB) specimens. The nodal release sequences that describe the propagation of the delamination front were investigated. The effect of using a sudden or a gradual nodal release was evaluated by considering meshes aligned with the crack front as well as misaligned meshes. Fatigue simulations were then performed using the Direct Cyclic Fatigue (DCF) algorithm. It was found that in specimens such as the DCB, which are characterized by a nearly linear response and a pure fracture mode, the algorithm correctly predicts the Paris Law rate of propagation. However, the Abaqus DCF algorithm does not consider different fatigue propagation laws in different fracture modes. Finally, skin/stiffener debonding was studied in an aircraft fuselage subcomponent in which debonding occurs deep into post-buckling deformation. VCCT was shown to be a robust tool for estimating the onset propagation. However, difficulties were found with the ability of the current implementation of the Abaqus progressive VCCT to predict delamination propagation within structures subjected to postbuckling deformations or fatigue loads.
A field study was conducted to ascertain the amount of protection that mesquite-dominated communities provide to the surface from wind erosion. The dynamics of the locally accelerated evolution of a mesquite/coppice dune landscape and the undetermined spatial dependence of potential erosion by wind from a shear stress partition model were investigated. Sediment transport and dust emission processes are governed by the amount of protection that can be provided by roughness elements. Although shear stress partition models exist that can describe this, their accuracy has only been tested against a limited dataset because instrumentation has previously been unable to provide the necessary measurements. This study combines the use of meteorological towers and surface shear stress measurements with Irwin sensors to measure the partition of shear stress in situ. The surface shear stress within preferentially aligned vegetation (within coppice dune development) exhibited highly skewed distributions, while a more homogenous surface stress was recorded at a site with less developed coppice dunes. Above the vegetation, the logarithmic velocity profile deduced roughness length (based on 10-min averages) exhibited a distinct correlation with compass direction for the site with vegetation preferentially aligned, while the site with more homogenously distributed vegetation showed very little variation in the roughness length. This distribution in roughness length within an area, defines a distribution of a resolved shear stress partitioning model based on these measurements, ultimately providing potential closure to a previously uncorrelated model parameter.
This paper develops a new approach for analysis of subgrid closures for turbulent combustion as modeled using direct quadrature (finite-rate chemistry) techniques. The approach, termed multiresolution analysis through mesh-sequenced realizations (MRA-MSR), conducts simultaneous, constrained large-eddy simulations on a set of hierarchically coarsened meshes. Furthermore, the availability of underlying fine-mesh (subgrid) data corresponding to coarse-mesh locations allows a clearer assessment of the effects of unresolved fluctuations on apparent reactivity. A key to MRA-MSR is the correlation of eddy structures at coarser mesh levels, which is facilitated by the transfer of filtered fine-mesh velocity information. A seven-mesh MRA-MSR hierarchy using three resolution levels is applied to one of the Sydney bluff-body stabilized methane–hydrogen flames. Analysis of the simultaneously evolved data at different resolution levels reveals several interesting trends. First, at high Damköhler numbers, there is clear evidence of attenuation of apparent reactivity due to the effects of unresolved fluctuations. Secondly, single-point, single-time filtered density functions of a normalized subgrid Damköhler number show a characteristic beta probability density function (PDF) form and display evidence of scale similarity. Interrogation of the MRA-MSR database also shows that the recently-proposed least-squares minimization (LSM) turbulence–chemistry interaction model can account for the observed diminishment in reactivity at high Damköhler numbers but cannot reduce scatter significantly. A new form of the LSM model, which makes use of the normalized subgrid Damköhler number beta PDF distribution, performs slightly better than the original model, illustrating the potential of MRA-MSR both in assessing existing closure concepts and in developing new ones.
The seismic slip rate along the Chile Trench estimated from the slip in the great 1960 earthquake and the recurrence history of major earthquakes has been interpreted as consistent with the subduction rate of the Nazca plate beneath South America. The convergence rate, estimated from global relative plate motion models, depends significantly on closure of the Nazca - Antarctica - South America circuit. NUVEL-1, a new plate motion model which incorporates recently determined spreading rates on the Chile Rise, shows that the average convergence rate over the last three million years is slower than previously estimated. If this time-averaged convergence rate provides an appropriate upper bound for the seismic slip rate, either the characteristic Chilean subduction earthquake is smaller than the 1960 event, the average recurrence interval is greater than observed in the last 400 years, or both. These observations bear out the nonuniformity of plate motions on various time scales, the variability in characteristic subduction zone earthquake size, and the limitations of recurrence time estimates.
Explore the source record for details and available documents.
The focus was on the use of databases obtained from direct numerical simulations of turbulent flows, for study of turbulence physics and modeling. Topics addressed included: stochastic decomposition/chaos/bifurcation; two-point closure (or k-space) modeling; scalar transport/reacting flows; Reynolds stress modeling; and structure of turbulent boundary layers.
A model for passive mixing of multiple scalars in homogeneous turbulence is developed. This model is based on the time-evolving mapping closure methodology. Calculations from the model are compared against heat conduction simulations (HCS) data. The agreement is excellent at the early stages of mixing and at the later stages it is fair.
The objective of the present research is to develop improved turbulence models for the computation of complex flows through turbomachinery passages, including the effects of streamline curvature, heat transfer and secondary flows. Advanced turbulence models are crucial for accurate prediction of rocket engine flows, due to existance of very large extra strain rates, such as strong streamline curvature. Numerical simulation of the turbulent flows in strongly curved ducts, including two 180-deg ducts, one 90-deg duct and a strongly concave curved turbulent boundary layer have been carried out with Reynolds stress models (RSM) and algebraic Reynolds stress models (ARSM). An improved near-wall pressure-strain correlation has been developed for capturing the anisotropy of turbulence in the concave region. A comparative study of two modes of transition in gas turbine, the by-pass transition and the separation-induced transition, has been carried out with several representative low-Reynolds number (LRN) k-epsilon models. Effects of blade surface pressure gradient, freestream turbulence and Reynolds number on the blade boundary layer development, and particularly the inception of transition are examined in detail. The present study indicates that the turbine blade transition, in the presence of high freestream turbulence, is predicted well with LRN k-epsilon models employed. The three-dimensional Navier-Stokes procedure developed by the present authors has been used to compute the three-dimensional viscous flow through the turbine nozzle passage of a single stage turbine. A low Reynolds number k-epsilon model and a zonal k-epsilon/ARSM (algebraic Reynolds stress model) are utilized for turbulence closure. An assessment of the performance of the turbulence models has been carried out. The two models are found to provide similar predictions for the mean flow parameters, although slight improvement in the prediction of some secondary flow quantities has been obtained by the ARSM model. It's found that the wake profiles inside the endwall boundary layers are predicted better than those near the mid-span.
The effects of several critical assumptions and parameters on the computation of strain energy release rates for delamination and debond configurations modeled with plate elements have been quantified. The method of calculation is based on the virtual crack closure technique (VCCT), and models of the upper and lower surface of the delamination or debond that use two-dimensional (2D) plate elements rather than three-dimensional (3D) solid elements. The major advantages of the plate element modeling technique are a smaller model size and simpler configurational modeling. Specific issues that are discussed include: constraint of translational degrees of freedom, rotational degrees of freedom or both in the neighborhood of the debond front, shear deformation assumptions; and continuity of material properties and section stiffness in the vicinity of the debond front. Where appropriate, the plate element analyses are compared with corresponding two-dimensional plane strain analyses.
The effects of several critical assumptions and parameters on the computation of strain energy release rates for delamination and debond configurations modeled with plate elements have been quantified. The method of calculation is based on the virtual crack closure technique (VCCT), and models that model the upper and lower surface of the delamination or debond with two-dimensional (2D) plate elements rather than three-dimensional (3D) solid elements. The major advantages of the plate element modeling technique are a smaller model size and simpler geometric modeling. Specific issues that are discussed include: constraint of translational degrees of freedom, rotational degrees of freedom or both in the neighborhood of the crack tip; element order and assumed shear deformation; and continuity of material properties and section stiffness in the vicinity of the debond front, Where appropriate, the plate element analyses are compared with corresponding two-dimensional plane strain analyses.
Accurate numerical simulations of turbulent free shear flows were performed in order to study the behavior of certain critical turbulence quantities which have been difficult to measure in the laboratory. The three-dimensional time-dependent nonlinear Navier-Stokes equations are solved by the direct numerical simulation technique without turbulence modeling or ad hoc closure assumptions. The pressure interaction terms were factored into diffusion and return-to-isotropy components according to two models. Simulations of the evolution of a turbulent wake with and without swirl present were performed at sufficiently high Reynolds numbers to insure that the flows were strongly nonlinear and turbulent, but low enough so that all scales of motion containing significant energy are adequately resolved. It is noted that the presence of swirl tends to enhance the amplitude of the rms pressure levels.
In this paper, we will discuss some recent model developments in two turbulence closure schemes. One is the Reynolds-stress algebraic equation model and the other is the Reynolds-stress transport equation model. Various model constraints required by the rapid distortion theory, the invariant theory and the realizability principle, etc. will be described in the model development.
The Z-pinch m = 0 instability as well as its stabilization by radially sheared axial flow is studied using the nonlinear ideal five-moment twofluid (5M2F) model with an extension of that model to include Braginskii heat and momentum transport. Using the ideal 5M2F model, linear growth rate results are compared with prior work using MHD and Hall MHD. At small normalized wavenumber, 1 < k a < 4 , where a is the effective pinch radius, 5M2F results agree with Hall MHD within 20% in scenarios without radially sheared axial flow. With the sheared flow and focusing on ka = 10/3, agreement with Hall MHD is excellent. In the limit of small ion inertial length, results also match with MHD. A comparison with PIC modeling of shear-free m = 0 stability focuses on a plasma scenario based on recent experimental results. In a scan of mode wavenumber, ideal 5M2F results are qualitatively similar to PIC: the growth rate rises to a peak at a moderate wavenumber and declines at a large wavenumber in contrast to MHD results, which show the saturation of the growth rate with the increasing wavenumber rather than a decline. The peak normalized 5M2F growth rate is γ τ A = 1.5 , where sA is the Alfven transit time across the pinch. The peak occurs at normalized wavenumber ka = 10. For comparison, PIC results have a peak growth of γ τ A = 0.77 at ka = 5. Including Braginskiibased closure of the 5M2F model does not qualitatively change the ideal results in this particular case. Nonlinear saturation is studied using the 5M2F model with the dissipative Braginskii-based closure in cases with pinch-edge sheared-flow speed equal to half the Alfven speed. Nonlinear mixing due to the sheared flow yields a quasi-steady state after modest losses of pinch ion inventory and pinch thermal energy, approximately 30% and 10%, respectively. 5M2F modeling captures the essential physics of m = 0 instability and offers a computationally tractable route to high-fidelity modeling of 3D Z-pinch behavior, including m = 1 instability.
Rayleigh-Taylor mixing in the presence of a third component with intermediate density is investigated through three-dimensional large-eddy simulation (LES) with a high-order compact finite-difference code. Two configurations are considered: (1) a symmetric configuration in which the Atwood number between the heavy and intermediate components matches the Atwood number between the intermediate and light components and (2) an asymmetric configuration in which the Atwood number between the heavy and intermediate components is an order of magnitude greater than the Atwood number between the intermediate and light components. Mass fraction covariances are extracted, and proposed Reynolds-averaged Navier-Stokes (RANS) closures for density-specific-volume and density-mass-fraction covariances are evaluated in an a priori fashion. Additionally, a multicomponent extension of the k - Φ - L - a - V RANS model [Morgan, Phys. Rev. E 104, 015107 (2021)] is presented which includes model equations for the upper-triangular elements of the mass fraction covariance matrix. This model, referred to as the k - Φ - L - a - C model, is compared against results from LES and against other RANS models. Profiles of average mass fraction, mass-fraction covariance, and density-specific-volume covariance obtained with the k - Φ - L - a - C model are found to agree well with LES data. Finally, the impact of three-component turbulent mixing on average reaction rate is investigated in both premixed and nonpremixed cases by heating the mixing layer and allowing it to undergo thermonuclear (TN) burn. A closure model for average reaction rate is proposed for use with the k - Φ - L - a - C model, and when this model is applied, improved agreement is obtained between LES and RANS in total TN neutron production.
This report shows that if a stochastic differential equation (Langevin equation) for velocity fluctuation vector is known, it is possible to derive the equations for scalar flux transport. Durbin and Speziale (1994) showed that the second moment of this stochastic differential equation gives an equation for the evolution of Reynolds stress tensor. Similarly, the stochastic equation will give an equation for scalar flux. Therefore, a coupling between these two is present. The basis for the present work is that there should be Langevin equations that can produce acceptable models for both the Reynolds stress tensor and the scalar flux vector. Having found this basic Langevin equation, the amount of work needed to model the second order closure problems is reduced; using the well developed models for Reynolds stress equations, it will be possible to derive corresponding models for scalar flux equation.