Chapter 3

Problem with Explicit Equations of Straight Line:

y=mx+c

 Slope m is infinite for line parallel to y axis

 Near vertical lines have very large slopes
 Defining these values might be problemsome
o Very large number leads to numerous problems

Problems with Implicit Forms:

 They represent unbounded geometry

o For example 𝑎𝑥 + 𝑏𝑦 + 𝑐 = 0 denotes an infinite line
o But in CAD we need a line between two points
 Curves are multi-valued
o For example in 𝑎𝑥 2 + 𝑏𝑦 2 + 2𝑘𝑥𝑦 + 2𝑓𝑥 + 2𝑔𝑦 + 𝑑 = 0 for each value of x
there are 2 values of y
o Usually a unique point of the curve should be used while defining the curve
 Often CAD needs evaluation in an orderly sequence of points on a geometric entity
o Implicit equations do not offer a natural procedure for that
 The equation for the curve will depend upon the co-ordinate system used

Advantage of Alternate Representation:

 Faired shapes like airfoil are not amenable to representation by classical geometry
 Convenient for intersection curves

Solution to Problems:

 Description of geometric entities using a parametric form

 Interpolation of large numbers of linearly independent conditions using composite
entities that are formed piecewise from a number of segments

Parametric Representation of Geometry:

  Here, the relationship for x,y and z coordinates of points on a curve or surface or solid
are expressed in terms of one or more independent variables known as parameter
 For curve, single parameter is used (u)
 For surface, two parameters are used (u,v)
 For solid, three parameters are used (u,v,w)
 Form of relationship between coordinate positions and parameters can be quite
arbitrary

Parametric Cubic Polynomial Curves:

 Describes non-planar curves

 Avoids computational difficulties and unwanted undulations introduced by high-order
polynomial curves
 Very popular as a basis for computational geometry
 Polynomials expressed in Hermite basis are known as Ferguson or Coons Representation

Lagrange Interpolation:
The fitting of curve through points is known as Lagrange interpolation

 Named after mathematician Lagrange

Hermite Interpolation:
The definition of cubic curve through fitting two points and two slope conditions is known as
hermite interpolation
Bezier Curves:

 Was introduced by Pierre Bezier

 Used control polygon instead of points and tangent vectors for modeling
 Polygon is approximated by a polynomial curve
o Degree of the curve is one less than the number of polygon vertices/control
points/track points
 𝒑 = 𝒑𝟎 (𝟏 − 𝟑𝒖 + 𝟑𝒖𝟐 − 𝒖𝟑 ) + 𝒑𝟏 (𝟑𝒖 − 𝟔𝒖𝟐 + 𝟑𝒖𝟑 ) + 𝒑𝟐 (𝟑𝒖𝟐 − 𝟑𝒖𝟑 ) + 𝒑𝟑 𝒖𝟑

Advantage:

 Control points provide easier control over the shape of polynomial than the tangent
vectors
 Curves can be considered as a combination of blending functions representing the
influence each control point has on the curve
 Control polygon has more intersection with planes than interpolating curves
 Control polygon changes direction more frequently than interpolating curves
 As a result, Bezier curve will not show any unexpected behavior
 It can smooth put even rapidly varying point sets
 The curve lies within convex hull of defining control points
o Blending functions sum to one
 Curves show variation diminishing property in the intermediate part

Disadvantage:

 Approximates points without passing through them

 Only capable of being globally modified
o Moving one of the points in control polygon affects every position in the curve

Convex Hull:

The minimal convex region enclosing the control points is known as convex hull

Variation Diminishing Property:

  Smooths the control point

General Consideration for Multi-variable Curve Fitting:

Local Modification:

 Movement of one control point only affects the shape of the curve in the vicinity of the
point

Degree of Continuity:

 Characterized by C0…..Cn
 Nth derivative of its parametric form is continuous
 A polygon with discontinuous slopes and higher derivatives is C0 continuous
 In C1 continuity ,direction and magnitude of the tangent vector of a curve in parametric
space is continuous
 In C2 continuity, the first and second derivatives in parametric spaces are continuous
 Parametric continuity does not necessarily imply geometric continuity

Cubic Spline Curve:

Curve which is defined through boundary conditions continuity in first and second derivatives at
intermediate points is known as continuous second-derivative cubic spline curve.

Advantage:

 Not necessary to define slopes at intermediate knot points

Disadvantage:

 Only global modification possible

B-Spline Curve:

 Generalization of Beizer method

 Uses blending function to combine the influence of a series of control points in an
approximate curve
 For n+1 points p0,
𝑛

𝑝(𝑢) = ∑ 𝑁𝑖,𝑘 (𝑢)𝑝𝑖

𝑖=0

Here,

Ni,k = B-spline blending functions

 Overcomes limitation of Bezier curves-

 Local modification
 Constraint in number of control points

 If we define B-spline polynomial of order k for a set of k knot points, the blending
functions are identical to those for a Bezier curve
 With the decrease of order, local influence of each track point becomes more prominent
 Repeating points increase the influence of track points
o Pulls the curve towards the point and then makes it pass through the point
 Curves tangential to the line between first and last pair of points

Difference between Bezier Curve & B-Spline Curve:

 Bezier-
o Degree of polynomials denoted by number of track points within certain limits
o Blending function non-zero over entire interval
 B-spline-
 o Degree specified independent of number of track points
o Blending function non-zero for limited interval
 Each blending curve corresponds to particular point
 Moving the point will modify the curve only for that range of parameter
for which blending function is non-zero

Rational Curve:
The class of curves that is capable of exactly representing conic and more general quadratic
functions as well as various types of polynomials, is known as rational curve.

 Cublic spline, Bezier and B-spline curves are used for representing free-form
curves and data
 Engineering design involve both free-form and analytic geometry

Advantage:

 Allow representation of both analytic and free-form curves in single unified form
 Reduced database complexity
 Reduced number of procedures required for display and manipulation of geometric
entities

Non-Uniform Rational B-Spline (NUBRS) :

Rational B-spline function that allows a non-uniform knot vector is known as Non-uniform
rational B-spline or NUBRS.

 Capable of representing non-rational B-splines and Bezier curves in a single form

 It can also represent linear and quadric analytic curves
 Can also be used in approximating or interpolating mode
Rational Polynomial:

Functions in which one polynomial curve is divided by another is known as rational polynomial

 They make use of homogeneous coordinates

Techniques for Surface Modelling:

Surface Patch:

 Building block for surfaces

 May be termed bi-parametric
o U and v vary across the patch
 Parametric variables often lie in the range of 0 to 1

Isoparametric Curve:

Fixing the value of one of the parametric variables results in a curve on the patch defined in
terms of the other variable. This is known as isoparametric curve.

 Results in intersecting mesh of curves on the patch

Coons Patch:

 Sculptured surfaces often involve Interpolation across an intersecting mesh of curves

that in effect comprise a rectangular grid of patches, each bounded by four boundary
curves
  Coons patch is used for such interpolation
 Includes other patches which are blending of arbitrary boundaries

Here,

Linear Blending Functions:

𝒇(𝒕) = 𝟏 − 𝒕

𝒈(𝒕) = 𝒕

𝑷(𝑼, 𝑽) = 𝑪𝟎 (𝒖)𝒈(𝒗) + 𝑪𝟏 (𝒖)𝒈(𝒗) + 𝑫𝟎 𝒇(𝒖) + 𝑫𝟏 (𝒗)𝒈(𝒖) − 𝒑𝟎𝟎 𝒇(𝒖)𝒇(𝒗)

− 𝒑𝟎𝟏 𝒇(𝒖)𝒈(𝒗) − 𝒑𝟏𝟎 𝒇(𝒖)𝒈(𝒗) − 𝒑𝟏𝟏 𝒈(𝒖)𝒈(𝒗)

 Application of simply blending functions to the curve will give incorrect results at the
corners of the patch
o That is why function of corner points pij is introduced

Bicubic Patch:

 Important for surface descriptions defined in terms of point and tangent vector
information

Bezier Surfaces:

 Uses characteristic polygon for defining surfaces

o Polygon is named as characteristics mesh
 Points on Bezier Surfaces are expressed by-

𝒑(𝒖, 𝒗) = ∑ ∑
Here,

Pij =Vertices of characteristics polygon

Bi,m & Bi,n =Blending functions for defining curves

Similarity with Bezier Curve:

 Surfaces pass only through the corners of the characteristic polygon

 Edge curves are tangential to the edges of the characteristics polygon at the corner
points
 Surfaces are variation diminishing
 Surfaces have convex hull property

Limitation:

 Allows global modification only

 Constraining in case of achieving smooth transition between adjacent patches

B-Spline Surface:

 Overcomes limitations of Bezier curve

 Shares schemes with both Bezier & B-Spline Curve Scheme
 Approximates a characteristics polygon
 Passes through the corner points of the polygon
 Edges are tangential to the edges of the polygon
  Can also be defined through periodic blending functions of same shape

 Using closed boundary along with periodic blending function, the influence of control
points can be reduced
o Local modification becomes possible
 Expression-
𝒎 𝒏

𝒑(𝒖, 𝒗) = ∑ ∑ 𝑵𝒊,𝒌 (𝒖)𝑵𝒊,𝒋 (𝒗)𝒑𝒊𝒋

𝒊=𝟎 𝒋=𝟎

Techniques for Volume Modelling:

Boundary Model:
The model in which solid is defined from combination of geometric information about faces,
edges and vertices with topological data on how these are connected, is known as boundary
model.

 Major problem is which way to ensure that the models defined by the system will
always be topologically valid, even during interactive modification
 Done in 2 ways-
o Appropriate choice of data structure
o Ensuring that models conform to a set of mathematical rules controlling the
topology
 Also known as graph-based model
o Face, edge and vertex data is stored into nodes in a graph with pointers or
branches between the nodes to indicate connectivity
o Graphs are known as directed graphs
 Direction of the link between nodes is important
Features of Topological Consistency for a Convex Body without Holes:

 Faces should be bounded by a single ring or loop of edges

 Each edge should adjoin exactly two faces and have a vertex at each end
 At least 3 edges should meet at each vertices
 Euler’s rule should apply

Euler’s Rule:

 Named after Swiss Mathemetician Euler

Let,

V= Number of vertices

E= Number of Edges

F= Number of faces

Then-

V-E+F=2

Euler-Poincare Formula:

 For bodies with holes, protrusions from faces and re-entrant faces

Let,
H= Number of interior edge loops or holes in faces

P= Number of passages or through holes

B= Number of separate bodies

V-E+F-H+2P=2B

 Restricts the way the model can be manipulated during construction

Euler Operators:

 Defines legitimate operation on model

 Specify combination of edge, vertex and face that may be added or removed in a single
operation to maintain validity of the formula
Storage of Relationship between Model Entities and Maintenance of Valid Models:

 Modellers allowing non-manifold models cannot enforce strict validity conditions

 They use modeling’s basic representation as a topological representation of faces,edges
etc
 Their boundary representation solids form a subset when topological and geometric
conditions for validity are met

Constructive Solid Geometry:

 Graph is known as binary tree
o Nodes are connected by branches to root node
o Any node may have
 One parent node
 Two child node
o Root node has no parent
o Leaf nodes have no children
 Leaves are geometric primitives
 Internal nodes comprise of the Boolean set operations that construct the model

 Primitives may be defined by-

o Bounded Solids
o Half Spaces

Half-Space:
Half-space is the intersection of simpler primitives.

 These are surfaces that divide coordinate spaces into solid and space
Problems of CGS:

 Efficient calculation of intersection between elements of model is difficult to achieve

o For complex model, becomes too intensive for computer
o Intensity can be reduced by spatial division

Pure Primitive Instancing:

The modeling in which models are described by varying the dimensions of single primitives
recalled from library is known as pure primitive instancing.

 Simplest technique
 Applicable to shapes within part families which are geometrically and topologically
similar, but not dimensionally similar

Cell Decomposition:
Here, the model is described by the assembly of a number of small elemental shapes that are
joined together without intersecting

 Similar to CSG
 Only joining operation is conducted in stead of set-operation theory
 Not used widely
 Basis of Finite Element Analysis

Spatial Occupancy Enumeration:

  The model is divided into a number of small elements
 Involves identifying which of a regular grid of cubic elements are wholly or partly
occupied by the object being modelled
 Similar to cell decomposition
 Not used widely in geometric modeling
Associated Methods:
 Quadtree Sub-division
 Octree Sub-division
 Being increasingly applied
 Involve recursive or successive sub-division of a region into square and cubic shapes
 Representation is based on 4-ary and 8-ary tree structure
 When a shape is being approximated, the subdivision continues until each square or cube
is full of the shape, empty or until some predetermined resolution is reached.