1
0
Fork 0
fgdata/Aircraft/c172p/dev/fdm/gerris
Stuart Buchanan 60d7073794 Initial checking of c172p-detailed.
From https://github.com/Juanvvc/c172p-detailed.git

Commit df6b2c5f4f75d000c5f5ef9c6cc4446784348f20
2015-06-16 22:20:54 +01:00
..
buoyancy3D.gfs Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
create_experiment.sh Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
floats.blend Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
floats.gts.base Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
README Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
run_experiment.sh Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00
summarize_experiment_jsbsim.sh Initial checking of c172p-detailed. 2015-06-16 22:20:54 +01:00

Setup for computing buoyancy forces and moments of a 3d shape using the
Gerris Flow Solver, http://gfs.sourceforge.net/wiki/index.php/Main_Page

1. The hull of the vessel needs to be described by a 3d model.
   The hull model must be a manifold
   It is preferable if the origin of the 3d model is the same as the
   origin of the FDM structural frame.

2. Convert the hull 3d model to GTS format for use with Gerris.
   Example with the model originally in blender:
   1. Export the model in STL format.
   2. Convert the STL model to GTS format with the stl2gts tool (from the gts
      library). For unknown reasons the normals have to be reverted in this
      step.
      % gts2stl --revert < floats.stl > floats.gts.base
   3. Check the model with gtscheck
      % gtscheck -v < floats.gts.base
      If there are any problems, such as incompatible faces or non-manifold
      edges, the model is very likely not to work in the latter stages.
      Make a cleaned up version for the buoyancy computation.

   The model.gts.base will be the original for the various cases/configurations
   described below.

3. Edit the reference length and solid model scaling in the buoyancy3D.gfs
   file.

4. Prepare the cases/configurations of interest. A case/configuration is one
   position and orientation of the 3d model with respect to the water surface.
   Each case becomes a separate gerris3D run.
   The create_experiment.sh script can be used to create a set of cases.
   The script needs to be edited to set the location of the hydrodynamic
   reference point, HRP, (which should be close to the CG), the desired set
   of data points in roll and pitch, and finally the distance between the HRP
   and the water surface. The last parameter currently has to be manually
   entered in the script file for each set of cases.
   Example:
      (Uncomment the distance line "HAGL=1.2192 # 4ft")
      % ./create_experiment.sh 20150529 p4ft
      (Uncomment the distance line "HAGL=HAGL=1.524  # 5ft")
      % ./create_experiment.sh 20150529 p5ft
      ...
   This creates a base directory 20150529 with a multitude of p4ft_* and
   p5ft_* etc. subdirectories, each one being a case.

5. Run the cases.
      % ./run_experiment.sh <base dir> <name base>
   Example:
      % ./run_experiment.sh 20150529 p4ft
   You can also leave out the second argument to run all cases in 20150529.
   Do consider running a suitable number of cases in parallel
   ./run_experiment.sh calls to speed up this step. E.g. as many as you have
   CPU cores if you don't run out of memory.

6. Summarize the case results into tables that can be copied into a JSBSim
   hydrodynamics configuration.
      % ./summarize_experiment.sh <base dir>/<name base>
      Will summarize the cases (roll and pitch) for one distance between the
      HRP and the water surface.
   This script must be edited to set the roll and pitch range and the
   scaling factors used.
   The resulting "coefficients" are scaled for imperial units and expressed
   as F/(Rho*G) and M/(Rho*G).

   Example:
      % ./summarize_experiment.sh 20150529/p4ft

      See Systems/c172p-hydrodynamics.xml to see where the tables for buoyancy
      lift force and pitch and roll moment "coefficients" go and how the
      "coefficients" are transformed back into forces and moments.
      (Search for 'breakPoint="4.0"'.)