No abstract

Results of Monte Carlo and molecular-dynamics simulations of Lennard-Jones systems are presented in order to compare various methods of computing interfacial properties of liquid-vapour systems. For the computation of the surface tension gamma a new method is developed, which makes use of the Bennett procedure for calculating free-energy differences. The method is compared with the conventional route to the surface tension via the virial expression. For the temperature derivative of the surface tension, gamma /dT, both a fluctuation equation and the Gibbs adsorption equation are employed. It is found that d gamma /dT is determined more accurately by the absorption equation (through the surface entropy). Results of simulations of binary Lennard-Jones mixtures are also presented. For the argon-krypton system, values of the adsorption of argon at the interface are determined from density profiles, and are compared with values predicted by the adsorption equation. Positive adsorption of argon manifests itself in krypton-rich mixtures as a significant 'bump' in the argon density profile near the interface.

A parameter study is performed for the case of two-dimensional sound propagation from a (source) city canyon to a nearby, identical (receiver) city canyon. Focus was on sound pressure levels, relative to the free field, in the shielded canyon. An accurate and efficient coupled FDTD-PE model was applied, exploiting symmetry of the source and receiver canyon. With the proposed calculation method, simulations were necessary in only half the sound propagation domain. The shielding in the receiver canyon in case of a coherent line source was compared to the shielding by an incoherent line source, by means of sound propagation calculations in a number of 2D cross-sections through source and receiver. It was found that the shielding is rather insensitive to the width-height ratio of the canyons. The presence of diffusely reflecting façades and balconies lead to an important increase in shielding compared to flat façades. Rigid façades yield significantly lower shielding compared to partly reflecting façades. Effects of a moving atmosphere were modeled in detail. Shielding decreases significantly in case of downwind sound propagation when comparing to sound propagation in a non-moving atmosphere. Refraction is the most important effect in the latter. In case of upwind sound propagation, turbulent scattering plays an important role and the shielding is similar to the one of a non-moving atmosphere for the parameters used in this paper. The combination of effects, as is shown by some examples, is in general not a simple addition of the separate effects.

No abstract

The numerical implementation of the Green's function parabolic equation ͑GFPE͒ method for atmospheric sound propagation is discussed. Four types of numerical errors are distinguished: ͑i͒ errors in the forward Fourier transform; ͑ii͒ errors in the inverse Fourier transform; ͑iii͒ errors in the refraction factor; and ͑iv͒ errors caused by the split-step approximation. The sizes of the errors depend on the choice of the numerical parameters, in particular the range step and the vertical grid spacing. It is shown that this dependence is related to the stationary phase point of the inverse Fourier integral. The errors of type ͑i͒ can be reduced by increasing the range step and/or decreasing the vertical grid spacing, but can be reduced much more efficiently by using an improved approximation for the forward Fourier integral. The errors of type ͑ii͒ can be reduced by using a numerical filter in the inverse Fourier integral. The errors of type ͑iii͒ can be reduced slightly by using an improved refraction factor. The errors of type ͑iv͒ can be reduced only by reducing the range step. The reduction of the four types of errors is illustrated for realistic test cases, by comparison with analytic solutions and results of the Crank-Nicholson PE ͑CNPE͒ method. Further, optimized values are presented for the parameters that determine the computational speed of the GFPE method. The computational speed difference between GFPE and CNPE is discussed in terms of numbers of floating point operations required by both methods.