Theodore Young

Determination of detonation cell size and the role of transverse waves in two-dimensional detonations☆

Abstract Two-dimensional time-dependent numerical simulations have been performed to study the structure and propagation of self-sustained detonations. The simulations are first used to develop a systematic approach for determining the detonation cell size. This approach involves simulating systems with channel widths both larger and smaller than the transverse cell spacing. The cell size estimated using this approach is compared with experimental data. The simulations also provide insight into some aspects of the mechanism by which a two-dimensional, self-sustained detonation propagates. The evolution of the curvature of the transverse wave appears to be a crucial feature. It is shown that depending on the curvature of the transverse wave at the time of its reflection from either a neighboring transverse wave or a wall, flattened cells or pockets of unreacted gas can be formed.

Weak and strong ignition. I. Numerical simulations of shock tube experiments

Abstract Detailed one-dimensional calculations have been performed to simulate reflected shock tube experiments in the weak and strong ignition regime in hydrogenoxygenargon mixtures. It is found that the experiments and simulations agree well in the strong ignition case studied. In the weak ignition case, the simulations show the same qualitative behavior as the experiments. Here ignition starts at a distance away from the reflecting wall at a time much earlier than the calculated chemical induction time. This latter effect is shown to arise because of the sensitivity of the chemical induction time to fluctuations in the calculation. In the calculations these fluctuations arise because of numerical inaccuracies. In experiments, they can arise from a number of sources including nonuniformities in the incident shock wave leading to nonuniform reflection, thermal conduction to the walls, and interactions with boundary layers.

A study of detonation structure: The formation ofunreacted gas pockets

Schlieren photographs of detonations in low pressure H 2 -O 2 -Ar mixtures and numerical simulations of propagating detonations have revealed the presence of unreacted pockets of gas behind the shock front—Mach stem structure. These pockets are completely surrounded by burned gas, and they in turn burn more slowly, finally giving their energy to the system. Two-dimensional numerical simulations performed to study the development of unburned pockets show that they are peninsulas of cooler material cut off by the collision of two Mach stems or of a Mach stem and the wall. Their presence is observed when there is sufficient decoupling of the reaction zone from the incident shock. These unreacted gas pockets and the associated long induction distance are discussed in terms of possible mechanisms for cell re-initiation, cell generation, and the behavior of detonations near the detonation limits.

Application of time-dependent numerical methods to the description of reactive shocks

The coupling of the hydrodynamics and the chemical kinetics in a reactive mixture ignited by the passage of a shock is examined using a time-dependent numerical model. The fluid dynamic and chemical rate equations are integrated self-consistently on their own characteristic time-scales using the flux-corrected transport and selected asymptotic methods, respectively. Results are presented for a shock in an H 2 −O 2 mixture. The detailed calculations of the chemical and fluid structure of the shocked region clearly how the chemical energy released is partitioned between the fluid motion and the internal energy. A calculation is also shown for a reactive shock reflected from a rigid wall.

Large Scale Urban Contaminant Transport Simulations With Miles

Airborne contaminant transport in cities presents challenging new requirements for computational fluid dynamics. The unsteady flow involves very complex geometry and insufficiently characterized boundary conditions, and yet the challenging and timely nature of the overall problem demands that the turbulence be included efficiently with an absolute minimum of extra memory and computing time requirements. This paper describes the monotone integrated large eddy simulation methodology used in NRL’s FAST3D-CT (CT is contaminant transport) simulation model for urban CT and focuses on critical validation issues that need to be addressed to achieve practical predictability. Progress in validation studies benchmarking with flow data from wind-tunnel urban model simulations and actual urban field studies are reported. Despite inherent physical uncertainties and current model tradeoffs, it is clearly possible to achieve some degree of reliable prediction.

Numerical simulations of flow modification of supersonic rectangular jets

Numerical simulations have been performed to study the flowfield and near-field noise of a supersonic rectangular jet with two paddles inserted into the flow. The paddles cause a strong flapping motion to develop that enhances mixing of the jet with the surroundings. These simulations have been used to determine the flapping motions frequency, the mixing enhancement, the near-field noise, and the thrust loss associated with the paddles and to study the acoustic feedback mechanism that modulates the flapping motion. The flapping frequency has been estimated using the pitot pressure distributions at a sequence of times and Fourier analysis of the local pressure and z component of velocity. For paddles located x/h = 7.3 from the nozzle, where h is the narrow dimension of the nozzle, a frequency of 4700 Hz [St(h) - 0.136] with an amplitude of 157.5 dB at the nozzle lip has been predicted and is in agreement with experimental results. The pitot pressure drop, the mass, and the x-momentum fluxes along the flow direction have been used as a measure of jet mixing for jets with and without paddles inserted into the flow. In our numerical simulations, a control volume approach was used to estimate the thrust loss caused by the insertion of the paddles. The computational value of 13% is close to the experimental value of 14.4%, considering that the physical support for the paddles in the experiments is not included in the simulations. A special sequence of local pressure distribution plots, which highlight the acoustic waves, has been used to study the feedback mechanism that modulates the flapping motion.

Time-Dependent Simulation of Flames in Hydrogen-Oxygen-Nitrogen Mixtures

This paper describes a one-dimensional, time-dependent, Lagrangian model developed to study the initiation, propagation and quenching of laminar flames. The model incorporates a number of new approaches and algorithms which have now been tested by comparisons to less complex or analytic solutions and by comparisons to experimental data. These new elements include: ADINC [1] an implicit, Lagrangian method for solving the convective parts of the conservation equations; DFLUX [2,3] a variable accuracy algorithm for determining diffusion fluxes without having to invert matrices; SPLIT and MERGE, routines for dividing or merging computational cells as specified by external criteria; VSAIM, a vectorized version of the ordinary differential equation solver, CHEMEQ [4,5]; and a new method for treating an open boundary in an implicit, Lagrangian calculation. An asymptotic coupling method, used in conjunction with timestep splitting to couple the various processes, allows the use of entirely different algorithms for the physical processes represented by different mathematical forms.