Skip to content

Commit a696c39

Browse files
committed
Improve docu for contact force law
1 parent fb64b8a commit a696c39

File tree

1 file changed

+37
-8
lines changed

1 file changed

+37
-8
lines changed

docs/src/internal/ContactForceLaw.md

Lines changed: 37 additions & 8 deletions
Original file line numberDiff line numberDiff line change
@@ -148,15 +148,33 @@ In special cases (for example sphere rolling on a plane), the rotational coeffic
148148
can be interpreted as *rolling resistance coefficient*.
149149

150150
Coefficients ``c_{geo}, n_{geo}, \mu_{r,geo}`` depend on the geometries of the objects
151-
that are in contact. Only for spheres meaning values are provided based on Hertz' pressure,
152-
because currently the collision handling in Modia3D does no provide enough information for other
153-
geometries (``r_i`` is the radius of sphere ``i``):
151+
that are in contact. The coefficients are computed approximately based on the contact theory
152+
of Hertz [^5], [^6]: Here, it is assumed that each of the contacting surfaces can be described by a
153+
quadratic polynomial in two variables that is basically defined by its principal curvatures
154+
along two perpendicular directions at the point of contact. A characteristic feature is that
155+
the contact volume increases nonlinearly with the penetration depth, so ``n_{geo} > 1`` (provided
156+
the two contacting surfaces are not completely flat), and therefore the normal contact force
157+
changes nonlinearly with the penetration depth. In the general case, elliptical integrals
158+
have to be solved, as well as a nonlinear algebraic equation system to compute the normal
159+
contact force as function of the penetration depth and the principal curvatures at the contact point.
160+
An approximate *analytical* model is proposed in [^7].
161+
162+
In order that a numerical integration algorithm with step-size control
163+
works reasonably, the contact force needs to be continuous and continuously differentiable with
164+
respect to the penetration depth. This in turn means that the principal curvatures of the contacting
165+
surfaces should also be continuous and continuously differentiable, which is usually not the case
166+
(besides exceptional cases, such as a Sphere or an Ellipsoid).
167+
168+
Since the determination of the principal curvatures of shapes is in general
169+
complicated and the shapes have often areas with discontinuous curvatures, only a very rough approximation
170+
is used in Modia3D: *The contact area of a shape is approximated by a quadratic polynomial
171+
with constant mean principal curvature in all directions and on all points on the shape*.
172+
In other words, a sphere with constant sphere radius ``r_{contact}`` is associated with every shape that
173+
is used to compute coefficients ``c_{geo}, n_{geo}, \mu_{r,geo}``. A default value for ``r_{contact}``
174+
is determined based on the available data of the shape (see [shape data](https://modiasim.github.io/Modia3D.jl/stable/Components/Shapes.html)):
175+
176+
xxx
154177

155-
| Object 1 | Object 2 | ``c_{geo}`` | ``n_{geo}`` | ``\mu_{r,geo}`` |
156-
|:----------- |:--------- |:---------------------------------------- |:------------|:------------------- |
157-
| Sphere | Sphere | ``\frac{4}{3} \sqrt{1/(1/r_1+1/r_2)}`` | ``1.5`` | ``1/(1/r_1+1/r_2)`` |
158-
| Sphere | no Sphere | ``\frac{4}{3} \sqrt{r_1}`` | ``1.5`` | ``r_1`` |
159-
| no Sphere | no Sphere | ``1`` | ``1.0`` | ``1.0`` |
160178

161179

162180
## Regularized unit vectors
@@ -262,3 +280,14 @@ similar responses:
262280
[^4]: Andrea Neumayr, Martin Otter (2019):
263281
[Collision Handling with Elastic Response Calculation and Zero-Crossing Functions](https://doi.org/10.1145/3365984.3365986).
264282
Proceedings of the 9th International Workshop on Equation-Based Object-Oriented Modeling Languages and Tools. EOOLT’19. ACM, pp. 57–65.
283+
284+
[^5]: Hertz H. (1881):
285+
[Über die Berührung fester elastischer Körper](https://home.uni-leipzig.de/pwm/web/download/Hertz1881.pdf).
286+
Journal für die reine und angewandte Mathematik 92, S. 156-171.
287+
288+
[^6]: Johnson K.L. (1985):
289+
Contact Mechanics. Cambridge University Press.
290+
291+
[^7]: Antoine J-F., Visa C., and Sauvey C. (2006):
292+
[Approximate Analytical Model for Hertzian Elliptical Contact Problems](https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.1055.4455&rep=rep1&type=pdf).
293+
Transactions of the ASME, Vol. 128. pp. 660-664.

0 commit comments

Comments
 (0)