-
Notifications
You must be signed in to change notification settings - Fork 1
/
Results.txt
72 lines (51 loc) · 2.94 KB
/
Results.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
PBF homework
Author: Rex Ying
This is my individual work.
-------------------
Demo
demo.gif shows the animation of water cube falling in a world space [0, 1] x [0, 1] x [0, 1].
In addition, another smaller cube is placed at [0, 0.4] x [0, 0.4] x [0, 0.4].
This is just to show that the code can handle any kind of mesh objects placed in the space, even
very complicated meshes (see Collision handling section).
I didn't draw the mesh in Mitsuba so both cubes are invisible...
But from the way the the water cube drops, you can roughly get where the smaller cube is.
-------------------
Intro
Run the program with Eclipse (Neon).
In this project, I implemented collision handling, finding particle neighbors, density constraint,
surface tension correction and vorticity confinement. In addition, I used Mitsuba to render
water cube falling with an invisible box placed in a corner.
-------------------
Collision handling:
I make use of the Proximity Query Package [1], which provides the C++ code for efficient collision
detection. Although in the project, I used simple meshes such as cubes, it is possible to do
collision detection for large meshes consisting of more than 100k faces, due to the efficient
collision detection algorithm of [1] using hierarchical bounding boxes.
To use it with the starter
code provided, I wrote a swig Java interface so that the PQP package can be called in Java.
*****Note
I compiled the library libPQP.so in Linux. To run the code, one needs to place the src/pqp/libPQP.so
under the java.library.path. (typically /usr/bin)
For mac OS, one needs to re-compile by running the Makefile in src/pqp.
After collision is detected, velocity (with damping) gets updated to the reflected direction.
--------------------
Particle neighbors:
I use the more efficient approach of constructing a hashmap which maps particle to a cube of side
len 0.1. This approach makes computing subsequent forces quicker.
This also has the advantage of being able to support large domains, comparing to using a fixed grid
of cells. See particlefluids.neighbors.CubeSpaceDivision class.
--------------------
Forces
All 3 required forces are implemented, see:
forces.Incompressibility
forces.ViscosityXSPH
forces.VorticityConfinement
Parameters are in particlefluids.Constants class.
--------------------
Rendering is done using Mitsuba.
--------------------
References
[1] OBBTree: A Hierarchical Structure for Rapid Interference Detection,
Stefan Gottschalk, Ming C. Lin, and Dinesh Manocha, Proc. ACM SIGGRAPH, pp. 171-180, Aug. 1996. Also available as tech. rep. TR96-013, Dept. Comput. Sci., Univ. N. Carolina Chapel Hill, 1996.
[2] Liu, G. Gui-Rong, and M. B. Liu. Smoothed particle hydrodynamics: a meshfree particle method. World Scientific, 2003.
[3] Matthias M?ller, David Charypar, Markus Gross, Particle-based fluid simulation for interactive applications, 2003 ACM SIGGRAPH / Eurographics Symposium on Computer Animation (SCA 2003), August 2003, pp. 154-159.