Simulation of the dynamics of dust-gas circumstellar discs is crucial in understanding the mechanisms of planet formation. The dynamics of small grains in the disc is stiffly coupled to the gas, while the dynamics of grown solids is decoupled. Moreover, in some parts of the disc the concentration of the dust is low (dust to gas mass ratio is about 0.01), while in other parts it can be much higher. These factors place high requirements on the numerical methods for disc simulations. In particular, when gas and dust are simulated with two different fluids, explicit methods require very small timestep (must be less than dust stopping time t stop during which the velocity of a solid particle is equalized with respect to the gas velocity) to obtain solution, while some implicit methods requires high temporal resolution to obtain acceptable accuracy. Moreover, recent studies underlined that for Smoothed particle hydrodynamics (SPH) when the gas and the dust are simulated with different sets of particles only high spatial resolution h < c s t stop guaranties suppression of numerical overdissipation due to gas and dust interaction.To address these problems, we developed a fast algorithm based on the ideas of (1) implicit integration of linear (Epstein) drag and (2) exact conservation of local linear momentum. We derived formulas for monodisperse dust-gas in two-fluid SPH and tested the new method on problems with known analytical solutions. We found that our method is a promising alternative for the previously developed two-fluid SPH scheme in case of stiff linear drag thanks to the fact that spatial resolution condition h < c s t stop is not required anymore for accurate results.

Circumstellar discs, from which planetary systems are formed, consist of gas, dust and solids. Simulations of self-consistent dynamics of gas, dust and solids in circumstellar discs is a challenging problem. In the paper we present fast algorithms for computing the drag force (momentum transfer) between solid phase and gas. These algorithms (a) are universal and applicable to dust and solids with any sizes smaller than the mean free path of gas molecules, (b) can be used to calculate the momentum transfer between dust and gas instead of one-way effect, as it is done in many models, (c) can perform simulations, without a loss in accuracy, with the time step determined by gas-dynamic parameters rather than by drag force, and (d) are compatible with the widely used parallel algorithms for solving 3D equations of gas dynamics, hydrodynamic equations for dust, and the collisionless Boltzmann equation for large bodies. Preliminary results of supercomputer simulation of the gas-dust disc dynamics within the developed approach are reported.

Simulations of gas-solid mixtures are used in many scientific and industrial applications. Two-Fluid Smoothed Particle Hydrodynamics (TFSPH) is an approach when gas and solids are simulated with different sets of particles interacting via drag force. Several methods are developed for computing drag force between gas and solid grains for TFSPH.Computationally challenging are simulations of gas-dust mixtures with intense intephase interaction, when velocity relaxation time t stop is much smaller than dynamical time of the problem. In explicit schemes the time step τ must be less than t stop , that leads to high computational costs. Moreover, it is known that for stiff problems both grid-based and particle methods may require unaffordably detailed resolution to capture the asymptotical bahaiviour of the solution. To address this problem we developed fast and robust method for computing stiff and mild drag force in gas solid-mixtures based on the ideas of Particle-in-Cell approach. In the paper we compare the results of new and previously developed methods on test problems.

We present a supercomputer numerical model for simulating 3D dynamics of dust-gas gravitating circumstellar disk. The model combines SPH method for solving gas dynamical equations, particle-in-cell (PIC) method for solving Vlasov equation, and convolution FFT-based grid method for solving Poisson equation. The approach is based on 3D domain decomposition, sorting of particles with regard to grid cell, and transferring necessary amount of particles between subdomains. The algorithm is aimed to be run on supercomputers with distributed memory and on hybrid supercomputers.

Simulation the dynamics of gas-solid mixtures is crucial in many applications: chemical reactor design, evolution of circumstellar discs, etc. Such mixtures where gas is a carrier phase and solid grains are dispersed phase can be simulated as interpenetrating continuum media. The characteristic parameter of this problem is relaxation time between gas and dust velocities. In many applications this parameter varies significantly during the simulation (from small to unity). Moreover, the drag force can be non-linearly dependent on solids to gas relative velocity. These factors place high requirements on the numerical methods for such problems. We propose a novel non-iterative algorithm for Smoothed Particle Hydrodynamics for computing gas-solid mixtures with exchange momentum between phases. Forces (except drag force) are approximated explicitly, while drag force is linearized and velocity relaxation time is approximated explicitly while relative velocity is approximated implicitly. The algorithm was tested on dynamical problems for dusty gas mixtures. It was shown that in the developed algorithm for stiff nonlinear drag we can use temporal and spatial resolution independent of drag parameters.

No abstract

No abstract

The paper is devoted to the first application of the SPH-IDIC method to simulation of three-dimensional dusty gas flow. The expansion of a sphere consisting of gas and monodisperse fine dust into vacuum is considered as a model problem. Computing was carried out using graphic accelerators based on nVidia CUDA technology. The numerical simulations show ability of the method to deal with three-dimensional dynamics. We found that in 3D case the method kept all the advantages previously demonstrated on one-dimensional problems: robustness, efficiency and simplicity in implementation.