Simulation in Computer Graphics Elastic Solids. Matthias Teschner

Transcription

1 Simulation in Computer Graphics Elastic Solids Matthias Teschner

2 Outline Introduction Elastic forces Miscellaneous Collision handling Visualization University of Freiburg Computer Science Department 2

3 Motivation Elastic solids are modeled for particle sets, i.e. elements Forces at particles account for resistance to deformation, e.g., stretch, shear, bend, volume change Elastic solids in 1D, 2D, and 3D University of Freiburg Computer Science Department 3

4 Motivation Element in rest state / undeformed state No elastic forces Element in deformed state Elastic forces that accelerate particles towards the rest state of an element University of Freiburg Computer Science Department 4

5 Motivation Different types of deformation (degrees of freedom) can be derived from the incorporated particles Two particles: stretch, compression Three particles: area, shear More particles: volume, shear Rest state Deformed state with elastic forces Rest state Deformed states with elastic forces University of Freiburg Computer Science Department 5

6 Outline Introduction Elastic forces Miscellaneous Collision handling Visualization University of Freiburg Computer Science Department 6

7 Overview Define elements Compute deformation / strain Compute stress Compute elastic energy Compute elastic forces University of Freiburg Computer Science Department 7

8 Elements Particles form elements, e.g. Two particles form a line segment Four particles build a tetrahedron Element i Four particles and one tetrahedral element University of Freiburg Computer Science Department 8

9 Deformation / Strain Define a function that describes a deformation of element based on particle positions: Undeformed state: Deformed state: Example: relative stretching / compressing of two points with rest distance : Strain is dimensionless. University of Freiburg Computer Science Department 9

10 Deformation of Line Segments Five particles and four elements Exemplary deformation computations for the elements University of Freiburg Computer Science Department 10

11 Deformation of a Tetrahedron University of Freiburg Computer Science Department 11

12 Stress Deformation of an element causes element stress Internal pressure (force per area) Material stiffness Rest state Deformed state Stress due to deformation. Forces due to stress. University of Freiburg Computer Science Department 12

13 Elastic Energy Work that is performed to deform an element is stored as elastic energy: is the size (volume / area / length) of an element Quantifies the deformation of an element Undeformed state: Deformed state: University of Freiburg Computer Science Department 13

14 Elastic Forces Accelerate particles from positions with high elastic energy towards positions with low elastic energy Negative spatial gradient of the elastic energy Goal: Minimization of elastic energy, i.e. deformation For all particles of an element : University of Freiburg Computer Science Department 14

15 1D Distance Change Two particles form an element Strain University of Freiburg Computer Science Department 15

16 1D Distance Change Spatial derivatives of the strain University of Freiburg Computer Science Department 16

17 1D Distance Change Elastic force at particle : Elastic force at particle : University of Freiburg Computer Science Department 17

18 2D Area Change Three particles Edges: Strain: Forces: form a triangle University of Freiburg Computer Science Department 18

19 3D Volume Change Four particles Edges: Strain: Forces form a tetrahedron University of Freiburg Computer Science Department 19

20 Demo Tetrahedral mesh with combined constraints University of Freiburg Computer Science Department 20

21 Demo Volume forces can be used to mimic bending forces in triangulated surfaces Two adjacent triangles i, j can be combined to a virtual tetrahedron k. A volume change of k corresponds to an angle change between n i and n j. University of Freiburg Computer Science Department 21

22 Properties of Elastic Forces Preserve linear and angular momentum of an element or of a system of elements Distance forces change the distance between the two points, but not the linear and angular velocity of the spring Volume forces change the volume of the tetrahedron, but not its velocity. University of Freiburg Computer Science Department 22

23 Properties of Elastic Forces Sum up to zero for all elements Do not change linear momentum Do not cause torque Do not change angular momentum Also referred to as internal forces External forces can change linear and angular momentum of an element, e.g. gravitational force Forces sum up to zero, but change angular momentum of the element no elastic force University of Freiburg Computer Science Department 23

24 Outline Introduction Elastic forces Miscellaneous Collision handling Visualization University of Freiburg Computer Science Department 24

25 Damping Force proportional to a velocity Directed in the opposite direction of a velocity Models friction Improves the stability of a particle system in explicit integration schemes Typically omitted in implicit schemes with artificial damping University of Freiburg Computer Science Department 25

26 Damping Particle velocity Damping parameter Relative velocity Damping parameter Relative velocity projected onto direction x j x i Normalized direction University of Freiburg Computer Science Department 26

27 Damping Particle velocity External force Affects the global dynamics of a particle system, i.e. slows it down Relative velocity Internal force Reduces oscillations / noise Does not affect linear and angular momentum of a particle system University of Freiburg Computer Science Department 27

28 Particle Masses Should be proportional to the particle size Discretization should not affect the simulation E.g., Mass density of the material Particle size Elastic accelerations are accumulated at particles or University of Freiburg Computer Science Department 28

29 Time Step An element should not move farther than its size in one simulation step, e.g. its diameter : Time step limit: with can be interpreted as performance measure Time step size is only meaningful if related to the element size University of Freiburg Computer Science Department 29

30 Outline Introduction Elastic forces Miscellaneous Collision handling Visualization University of Freiburg Computer Science Department 30

31 Context Collision of a particle of an elastic solid with a plane Time t Time t+h University of Freiburg Computer Science Department 31

32 Plane Representation In 3D, a plane can be defined with a point on the plane and a normalized plane normal The plane is the set of points with For a point, the distance to the plane is University of Freiburg Computer Science Department 32

33 Concept If a collision is detected, i.e., a collision impulse is computed that prevents the interpenetration of the mass point and the plane We first consider the case of a particle-particle collision with being the normalized direction from to The response scheme is later adapted to the particle-plane case University of Freiburg Computer Science Department 33

34 Coordinate Systems Velocities before the collision response and velocities after the collision response are considered in the coordinate system defined by collision normal and two orthogonal normalized tangent axes and E.g. The velocity after the response is transformed back University of Freiburg Computer Science Department 34

35 Governing Equations Conservation of momentum Coefficient of restitution, elastic, inelastic Friction opposes sliding motion along and University of Freiburg Computer Science Department 35

36 Linear System University of Freiburg Computer Science Department 36

37 Solution University of Freiburg Computer Science Department 37

38 Particle Plane Plane has infinite mass and does not move: Columns 2, 3, 7, 8, 9 do not contribute to the solution To solve for the particle velocity response, rows 4, 5, 6 have to be considered Plane has infinite mass after collision University of Freiburg Computer Science Department 38

39 Implementation is difficult to handle and should be guaranteed is a useful simplification University of Freiburg Computer Science Department 39

40 Position Update The collision impulse updates the velocity However, the point is still in collision For low velocities, the position update in the following integration step may not be sufficient to resolve the collision Therefore, the position should be updated as well, e.g. which projects the point onto the plane The position update is not physically-motivated, it just resolves problems due to discrete time steps University of Freiburg Computer Science Department 40

41 Outline Introduction Elastic forces Miscellaneous Collision handling Visualization University of Freiburg Computer Science Department 41

42 Context Geometric combination of A low-resolution tetrahedral mesh for simulation and A high-resolution triangular mesh for visualization Supports simplified meshing for geometrically complex surface models Tetrahedral mesh Triangulated mesh University of Freiburg Computer Science Department 42

43 Illustration Two representations for simulation and visualization Tetrahedral elements with particles are simulated Triangulated elements with vertices are visualized University of Freiburg Computer Science Department 43

44 Barycentric Coordinates Surface vertex can be represented with the particles of a tetrahedron are Barycentric coordinates of with respect to University of Freiburg Computer Science Department 44

45 Properties is inside the convex combination of, i.e. inside the tetrahedron is on the surface of the tetrahedron is outside the tetrahedron University of Freiburg Computer Science Department 45

46 Computation leads to the following system Not solvable for degenerated tetrahedra is computed as University of Freiburg Computer Science Department 46

47 Implementation Preprocessing Determine the closest tetrahedron for surface points Compute Barycentric coordinates for surface points with respect to the corresponding tetrahedron Simulation step Compute surface-point positions from Barycentric coordinates and the positions of the particles of the corresponding tetrahedron Demo University of Freiburg Computer Science Department 47