Astron. Astrophys. 346, 861-877 (1999)

5. Summary

For the first time a comprehensive numerical three -dimensional study is presented of wind-accretion with a density gradient using a high resolution hydrodynamic code. I vary the following parameters: Mach number of the relative flow (Mach =3 and 10), strength of the density gradient perpendicular to this flow (=3%, 20% and 100% over one accretion radius), and adiabatic index (=5/3, 4/3, and 1.01). The results are compared (a) among the models with different parameters, (b) to some previously published simulations of models with density and velocity gradients, and (c) also to the analytic estimates of the specific angular momentum.

1. All models exhibit active unstable phases, which are very similar to the models without gradients. Only the mildly supersonic case (=1.4) displayed a relatively steady flow (as compared to the faster flow models). The accretion rates of mass, linear and angular momentum fluctuate with time, although not as strongly as published previously for 2D models.

2. Depending on the model parameters, the average specific angular momentum accreted is roughly between zero and 70% of the analytical estimate, which assumes that all angular momentum within the accretion cylinder is actually accreted.

3. The mass accretion rates of all models with density gradients are equal, to within the fluctuation amplitudes, to the rates of the models without gradients (published previously), although the accretion rates might seem to decrease slightly when increasing the density gradient. The fluctuations of the mass accretion rate in all models hardly vary with gradient strength.

4. The overall qualitative flow dynamics as well as the mass accretion rates are very similar to what has been published on models with velocity gradients. Of course the accretion rate of angular momentum and its specific values (i.e per mass unit) are reduced if one compares equal gradients for both velocity and density, well in accordance with the analytic estimates. This reduction means that in the density gradient models the fluctuations due to the unstable accretion flow have a greater influence on the angular momentum accretion than for the velocity gradient models.

5. The models with small gradients (=0.03) display an initially quiet stable phase, in which the specific angular momentum of the matter accreted is within 10% of the analytic estimate. Thus for the quiescent phases the analytic values are appropriate. The average drops when the flow becomes unstable.

6. The model with very large density gradient ( over one accretion radius) was the only one for which the accreted angular momentum was always prograde with respect to the angular momentum available in the incoming flow. Here the amplitude of the perturbation due to the unstable flow is much smaller than the average angular momentum accreted. However the specific angular momentum of the incoming flow in this case is larger than the maximum given by the Kepler velocity times the radius of the accretor surface.

© European Southern Observatory (ESO) 1999

Online publication: June 17, 1999
helpdesk.link@springer.de