Ceres, Pallas encounters
The effect of close encounters between test particles and massive bodies
strongly affects the stability of swift_rmvs3 integrator. In the first
test run (2953-cp0) we used time step 36 days and RHSCALE = 7 (only for
Ceres and Pallas). This lead to approximately 1000 encounters
closer than (RHSCALE * Hill sphere radius) within 100 Myr. The jumps
in a thus occur very frequently and the overall amplitude
of semimajor axis might be upto 10-3 AU.
The time step and RHSCALE was decreased in the second run (2953-cp1):
dt = 2 days, RHSCALE = 1. The number of encounters within
the Hill sphere decreased downto 1 or 10, as we expected.
The most of a changes due to encounters is of order
10-4.
a(t) graphs for 2953-cp0 and
2953-cp1 runs (images 2953-cp0-a.gif, 2953-cp1-a.gif). |
The calculation using swift_bs (Burlisch-Stoer) integrator with the same
initial conditions. The time span is much shorter than in previous cases
(<= slower integrator). Jump in a are smaller then
a(t) graph for 2953-cp0bsf run
(image 2953-cp0bsf-a.gif). |
Miroslav Broz,
miroslav.broz@email.cz,
Nov 26th 1999