Replace Hohmann Transfer with Bi-Impulsive Transfers #1057
Conversation
The existing Hohmann Transfer planner only does circular-to-circular coplanar transfer planning. The new functionality does full Bi-Impulsive Transfer planning with non-coplanar, non-elliptical orbits, including hyperbolic source and target orbits. It does this with a first pass through a simulated annealing algorithm that does global optimization to try to find the lowest basin. It then hands that off to Levenburg-Marquardt to find the local minimum. Since the simulated annealing algorithm uses a random search, it may not find exactly the same answer every single time. The answers it finds should be nearby and nearly the same dV but typically off by a few orbital periods of either the source or destination or both (when the search space has many possible orbits -- it should pick one of the better ones, but not necessarily the absolute best). Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
NOTE: Bi-Impulsive transfers (two-burn) are not Bi-Elliptic transfers. Bi-Elliptic transfers are actually Tri-Impulsive (three burn). |
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
this also rolls back to the old LambertSolver, the GoodingSolver is clearly buggy. Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
Signed-off-by: Lamont Granquist <lamont@scriptkiddie.org>
|
This is what the new "Hohmann" transfer planner window now looks like. The drop down allows for scheduling the burn at a specific user-selected time using the familiar widget. Otherwise the time of the burn is optimized. The "intercept only" button is really the "mono-impulsive" option for doing flybys, impactors, launching nuclear missiles at things, where the second burn is of no consequence so is not included in the cost. About all I can think of that is missing is a way to specify a transfer window period to search (minimum time in the future and/or maximum time in the future) and/or the maximum transfer time period to search -- plus a bit of code cleanup in #1058. I think my UX brain threw an exception though and it is time to ship and playtest. |
|
What does "fractional target period offset" mean? |
The existing Hohmann Transfer planner only does circular-to-circular
coplanar transfer planning.
The new functionality does full Bi-Impulsive Transfer planning with
non-coplanar, non-elliptical orbits, including hyperbolic source
and target orbits.
It does this with a first pass through a simulated annealing algorithm
that does global optimization to try to find the lowest basin. It then
hands that off to Levenburg-Marquardt to find the local minimum.
Since the simulated annealing algorithm uses a random search, it may
not find exactly the same answer every single time. The answers it finds
should be nearby and nearly the same dV but typically off by a few
orbital periods of either the source or destination or both (when the
search space has many possible orbits -- it should pick one of the
better ones, but not necessarily the absolute best).
Signed-off-by: Lamont Granquist lamont@scriptkiddie.org