Vous êtes sur la page 1sur 9

Universit de Lige

Aerospace and Mechanical Engineering Department


Mechanical Computer-Aided-Design

Prof. Mr. Eric Bchet


Find one of the intersections between two
parametric surfaces.




ALARCN SOTO Pablo

Academic year 2011-2012
1. Index

1.1. Contentindex

1. Index......................................................................................................................................2
1.1. Contentindex................................................................................................................2
1.2. Figuresindex.................................................................................................................2
2. Introduction..........................................................................................................................3
3. Creationofthealgorithm......................................................................................................3
3.1. Examinationofthecurves.............................................................................................4
3.2. Computationofthefirstpoint......................................................................................4
3.3. Marchingalongthecurve.............................................................................................5
4. Results...................................................................................................................................7
5. Bibliography..........................................................................................................................9

1.2. Figuresindex

Figuie 1 ueneial view of the inteisection.4
Figuie 2 Beteimination of the staiting point..S
Figuie S Inteisection cuive, view 1..7
Figuie 4 Inteisection cuive, view 2..8
Figuie S Inteisection cuive, view S..8








2. Introduction

Evaluate the inteisection of paiametiic suifaces is a common opeiation in computei
uesign anu soliu moueling; this task is commonly known as the pioblem of suiface to
suiface inteisection. In this woik we will focus on the inteisection between two
paiametiic suifaces uefineu as nbsplines suiface.
To compute the inteisection between two suifaces theie aie many uiffeient appioach,
among which we can highlight the uiviueanuconquei algoiithm (wheie the suifaces aie
subuiviueu into a gieat numbei of elements anu the inteisection of suifaces is
appioximateu by the inteisection paiis of these elements) anu the maiching algoiithm
(wheie one has to finu an staiting point on the inteisection cuive anu then walk along the
inteisection cuive to get the successive points). We will focus oui attention on the
maiching methous because they aie one of the simplest methous anu moieovei the
geometiical inteipietation is easy to unueistanu.
Taking the uefinition given by Patiikalakis
1
Naiching methous involve geneiation of
sequences of points of an inteisection cuive bianch by stepping fiom a given point on the
iequiieu cuive in a uiiection piesciibeu by the local uiffeiential geometiy, Bowevei,
such methous aie by themselves incomplete in that they iequiie staiting points (initial
conuitions) foi eveiy bianch of the solution. Consiueiing that infoimation we will
uevelop a simplifieu algoiithm following the piinciple of the maiching methous.

3. Creationofthealgorithm

To cieate the algoiithm we will follow in geneial teims the following steps.
- Examine the cuives to know in geneial teims the kinu of inteisection.
- 0btain a staiting point foi the inteisection cuive.
- Naich along the inteisection cuive until get the next inteisection points using as a
limit conuition the uistance between two points, one pei suiface.
- Continue with the iteiation until get the last inteisection point using as a limit
conuition the enu of the one the suifaces.

Surface to Surface Intersections. N. N. Patiikalakis, T. Naekawa, K. B. Ko anu B.


Nukunuan
3.1. Examinationofthecurves

Accoiuing to figuie 1 we can see that theie aie only one inteisection between both
suifaces, which is a smooth inteisection without singulaiities oi loops; foi that ieason we
can use a simplifieu veision of the maiching methou without pioblems.


Figure1Generalviewoftheintersection

3.2. Computationofthefirstpoint

Aftei examining the suifaces, we noticeu that the inteisection stait in the boiuei of one
suiface anu finish in the boiuei of the seconu one, helpeu foi figuie 2 we can see that the
inteisection stait when the value of v on suiface 1 is equal to u anu finish when the value
of u is equal to the maximum value on suiface 2.

















Figure2Determinationofthestartingpoint

3.3. Marchingalongthecurve

As soon we compute the fiist inteisection point, we have to maich along the cuive to
obtain the following inteisection points.
To compute the following inteisection point, we know that when two points (one pei
suiface) belong to the inteisection cuive, the cooiuinates of them take the same value anu
as they aie values in function of paiameteis u anu v foi each cuive, we can foimulate a
system of S equations with 4 unknowns. The auuitional equation is ielateu to the piocess
itself to walk ovei the inteisection cuive. 0n each iteiation, we compute an appioximation
point to the inteisection cuive, then using the NewtonRaphson methou we can obtain a
new appioximation point to the suiface inteisection (bettei appioximation than the
pievious one); the piocess is iepeateu until the eiioi is smallei than a given toleiance. The
piocess is uetaileu below
- Fiom suiface 1 we obtain the point P1(u
1
, :
1
), wheie u
1
onJ :
1
aie the
cooiuinates of P1 on the paiametiic space foi suiface 1; in the same way fiom
suiface 2 we obtain the point P2(u
2
, :
2
), wheie u
2
onJ :
2
aie the cooiuinates of
P2 on the paiametiic space foi suiface 2. Each point has S components, namely the
components ielateu to the Caitesian space (x,y,z).
- With these points, we uefine the functions to be minimizeu with the methou of
NewtonRaphson. These fiist S functions aie
F
0
: P1(u
1
, :
1
)
x
- P2(u
2
, :
2
)
x
= u
F
1
: P1(u
1
, :
1
)

- P2(u
2
, :
2
)

= u
F
2
: P1(u
1
, :
1
)
z
- P2(u
2
, :
2
)
z
= u
Beginningoftheintersection,
limitedbyoneborderofsurface1
Endoftheintersection,limited
byoneborderofsurface2
wheie the subsciipt uenotes the associateu cooiuinate on the Caitesian space.
- To solve the system, as we have 4 unknowns(u
1
, :
1
, u
2
, :
2
), we neeu an auuitional
ielation anu as we saiu this ielation is ielateu to the methou useu to walk ovei the
inteisection cuive. Foi that, we ueciueu to maich ovei the component :
1
on
suiface 1, namely in each iteiation the value of :
1
will be fixeu accoiuing to the
ielation
:
1
= u.uS - i
wheie i is a countei which stait fiom u anu inciease in a value equal to 1 aftei
finuing the inteisection point in each iteiation. The last equation is then
F
3
: :
1
- u.uS - i = u
- To simplify the iesolution of the system, we uefine a matiix F, which columns aie
the functions uefineu pieviously.
F = |F
0
F
1
F
2
F
3
]
- At the same time, we have to cieate a vectoi which contain the 4 unknowns,
namely
N = |u
1
:
1
u
2
:
2
]
t

- Auuitionally, to use the NewtonRaphson methou we neeu the associateu }acobian
matiix, which contain the paitial ueiivates associateu to each function.
(
(
(
(
(
(
(
(
(

c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
c
=
2
3
2
3
1
3
1
3
2
2
2
2
1
2
1
2
2
1
2
1
1
1
1
1
2
0
2
0
1
0
1
0
v
F
u
F
v
F
u
F
v
F
u
F
v
F
u
F
v
F
u
F
v
F
u
F
v
F
u
F
v
F
u
F
J
- Finally, the algoiithm to compute the next appioximation to the inteisection point
can be uefineu as
N
t+1
= N
t
+ [
-1
- F
Wheie N
t
is an appioximation point to the inteisection cuive anu N
t+1
is the next
appioximation (bettei appioximation) to the point onto the inteisection cuive. To
ueteiminate when a point is an inteisection point, in each iteiation we compute
the magnituue of the eiioi in the ueteimination of the cooiuinates of the
inteisection point accoiuing the uefinition of each suiface,
crror = F
0
- F
0
+F
1
- F
1
+ F
2
- F
2
+ F
3
- F
3

when the eiioi becomes smallei than a toleiance (in this case crror < 1u
-20
) the
point is an inteisection point.
- The algoiithm continue, now incieasing the value of the vaiiable i in 1 anu
iepeating the pieceuing steps until founu the last inteisection point, in this case
when the inteisection cuive ieach the enu of suiface 2.
4. Results

To save the iesults, we use S uiffeient vectois, namely
- pts: Which saves the values of the cooiuinate x, y anu z of each inteisection point
- paiam1: Which saves the values of the paiameteis u
1
onJ :
1
of the inteisection
point that belong to suiface 1.
- paiam2: Which saves the values of the paiameteis u
2
onJ :
2
of the inteisection
point that belong to suiface 2.
The numbei of computeu inteisection points is a function of the length of the chosen step
to walk ovei suiface 1 (namely the inciement on the value of the paiametei :
1
aftei each
iteiation). In simple woius, if we ueciease the length of the step, we will obtain moie
inteisection point, while in the othei hanu if we inciease the length of the step we will
obtain less inteisection points.
The inteisection cuive can be seen in figuies S , 4 anu S.

Figure3Intersectioncurve,view1


Figure4Intersectioncurve,view2


Figure5Intersectioncurve,view3
5. Bibliography

- An efficient suiface inteisection algoiithm baseu on loweiuimensional
foimulation. Shankai Kiishnan anu Binesh Nanocha.
- Suiface to Suiface Inteisections. N. N. Patiikalakis, T. Naekawa, K. B. Ko anu B.
Nukunuan.
- Tiacing suiface inteisections. C. L. Baja, C. N. Boffman, R. E. Lynch.
- A Naiching Nethou foi Computing Inteisection Cuives of Two Subuivision Solius.
XuPing Zhu, ShiNin Bu, ChiewLan Tai, anu Ralph Naitin.
- 0peiations on paiametiic cuives anu suifaces: Inteisection anu Tiimming. Lectuie
6, Piof. Zoiin, New Yoik 0niveisity.

Vous aimerez peut-être aussi