Stefan Schlamp

Higher moments of the velocity distribution function in dense-gas shocks

Abstract Large-scale molecular dynamics simulations of a M s xa0=xa04.3 shock in dense argon ( ρ xa0=xa0532xa0kg/m 3 , T xa0=xa0300xa0K) and a M s xa0=xa03.6 shock in dense nitrogen ( ρ xa0=xa0371xa0kg/m 3 , T xa0=xa0300xa0K) have been performed. Results for moments (up to order 10) of the velocity distribution function are shown. The excess even moments of the shock-normal velocity component (i.e., in the direction of shock propagation) are positive for most parts of the shock wave, but become negative towards the hot side of the shock before reverting back to zero. The even excess moments of the shock-parallel velocities and the odd moments of the shock-normal velocity do not change signs within the shock. The magnitude of the excess moments increases with the order of the moment, i.e., the higher moments correspond less and less to those of a Maxwell–Boltzmann distribution.

Molecular alignment in a shock wave

Molecular dynamics simulations of dense nitrogen show that nonspherical molecules have a weak tendency to align their molecular axis such that it lies parallel to the plane of a shock wave front. As a consequence, there is also an even weaker tendency for the molecular rotation axis to align perpendicular to the shock front. The underlying mechanism is discussed and it is argued that this phenomenon can only be observed for dense fluids and only when considering realistic molecular interactions. A single relevant nondimensional parameter is proposed.

Sound wave channelling in near-critical sulfur hexafluoride (SF6)

Strong density and speed of sound gradients exist in fluids near their liquid-vapor critical point under gravity. The speed of sound has an increasingly sharp minimum and acoustic waves are channelled within a layer of fluid. Geometrical acoustic calculations are presented for different isothermal fluid columns of sulfur hexafluoride (SF6) under gravity using a semiempirical crossover equation of state. More than 40% of the emitted acoustic energy is channelled within a 20 mm high duct at 1 mK above the critical temperature. It is shown how, by changes in temperature, frequency, and gravitational strength, the governing length scales (wavelength, radius of ray curvature, and correlation length of the critical density fluctuations) can be varied. Near-critical fluids allow table-top sound channel experiments.

Low-coherence self-referencing velocimetry.

Low-coherence self-referencing velocimetry optically measures the relative velocity between a point in a particle-laden fluid and a (potentially moving) reference surface. Low-coherence light scattered off the particles and off the reference surface is coupled into an interferometer with variable optical delay in one arm and an acousto-optical modulator in the second arm. The measurement location is set relative to the reference surface. Its location can be scanned along a line by adjusting the optical delay in the interferometer. The spatial resolution is typically tens of micrometers. Only one low-coherence light beam is required for each component of the velocity vector. Proof-of-principle measurements in Taylor-Couette flow are presented.

Higher moments of the velocity distribution function across a shock wave

Large‐scale molecular dynamics simulations of a Ms =4 .3 shock in dense argon ( = 532 kg/m 3 , T = 300 K) and a Ms =3 .6 shock in dense nitrogen ( = 371 kg/m 3 , T = 300 K) have been performed. Results for moments (up to order 10) of the velocity distribution function are shown. The excess even moments of the shock‐normal velocity component (i.e., in the direction of shock propagation) are positive for most parts of the shock wave, but become negative towards the cold side of the shock before reverting back to zero. The even excess moments of the shock‐parallel velocities and the odd moments of the shock‐normal velocity do not change signs within the shock. The magnitude of the excess moments increases with the order of the moment, i.e., the higher moments correspond less and less to those of a Maxwell‐Boltzmann distribution function.

Experimental Considerations for Laser-Induced Thermal Acoustics (LITA) in Compressible Turbulent Flows

The accuracy of LITA measurements is limited by the signal–to–noise ratio as well as by the number and the depth of Doppler and Brillouin oscillations in the signal. The influence of the controllable experimental parameters on those is discussed. For highly diffusive fluids, e.g. highly turbulent flows, minimizing the laser spot size and the laser beam crossing angle maximizes the signal amplitude as well as the signal modulation. An improved, partially fiber–coupled experimental setup is presented, which reduces beam steering effects when using heterodyne detection for simultaneous speed of sound and flow velocity (Mach number) measurements.

Error surface topology in the data analysis of laser-induced thermal acoustics signals

Laser-induced thermal acoustics (LITA) promises remote, instantaneous and non-intrusive point-measurements of the speed of sound (temperature), thermal diffusivity (density), flow velocity (Mach number) and species concentration simultaneously in harsh environments. The data analysis relies on a nonlinear fit of an analytical model to the acquired data. The measured quantities are parameters in the model. Computational cost and convergence behaviour depend on the dimensionality of the parameter space, the initial guesses for the parameters and on whether the data analysis is performed in the time or frequency domain. The topology of the four-dimensional error surface is discussed and a characteristic allowable distance of the initial guesses from the global minimum is defined and quantified for typical configurations. Noise has no significant influence on the convergence neighbourhood or the computational cost. If improved initial guesses (10% maximum error) for the speed of sound and the flow velocity are obtained by data preprocessing, convergence of the fitting algorithm is ensured.