Contents:

• Additive Process Chain

• Solid Modeling and conversion of solid model
into STL file
• STL file and related issues
• Integration of Reverse Engineering and Rapid
Prototyping
• Introduction to 3D scanner

1
There are a total of five steps in the chain and these are 3D modeling,
data conversion and transmission, checking and preparing, building
and post processing. Depending on the quality of the model and part
in Steps 3 and 5 respectively, the process may be iterated until a
satisfactory model or part is achieved.
2
Workflow of RP Processes:

3
 PCD /
Solid /
Surface
model
Finished
Product .STL file

Additive
Support Manufacturing
Removal Process Chain .STL file
and post verification
processing

Layer
3D
Slices
printed
and
job
Toolpath
Basic steps in Rapid Prototyping Process:
1. Creation of Solid model

2. Conversion of solid model into .stl file format

3. Conversion of .stl file into 2D layers using special rapid

prototyping program called as “Process Planner”. To
accommodate geometries with overhanging feature, the
parts are embedded in sacrificial supporting material
during buildup.

4. Generation of Rapid Prototype Model (Physical Model)

5. Removal of support material if required

6. Post processing if required

5
• It can be seen from above figure, that process starts with 3D
modeling of the product and then STL file is exported by
tessellating the geometric 3D model. In tessellation various
surfaces of a CAD model are piecewise approximated by a
series of triangles and co-ordinate of vertices of triangles
and their surface normals are listed. The number and size of
triangles are decided by facet deviation or chordal error.
These STL files are checked for defects like flip triangles,
missing facets, overlapping facets, dangling edges or faces
etc. and are repaired if found faulty.

• Defect free STL files are used as an input to various slicing

softwares. At this stage choice of part deposition orientation
is the most important factor as part building time, surface
quality, amount of support structures, cost etc. are
influenced. Once part deposition orientation is decided and
slice thickness is selected, tessellated model is sliced and
the generated data in standard data formats like SLC
(stereolithography contour) or CLI (common layer interface)
is stored.

6
• This information is used to move to step 2, i.e., generation of
physical model. The software that operates RP systems
generates laser-scanning paths (in processes like
Stereolithography, Selective Laser Sintering etc.) or material
deposition paths (in processes like Fused Deposition
Modeling). This step is different for different processes and
depends on the basic deposition principle used in RP
machine. Information computed here is used to deposit the
part layer-by-layer on RP system platform.

• The final step in the process chain is the post processing

task. At this stage, generally some manual operations are
necessary therefore skilled operator is required. In cleaning,
excess elements adhered with the part or support structures
are removed. Sometimes the surface of the model is finished
by sanding, polishing or painting for better surface finish or
aesthetic appearance. Prototype is then tested or verified and
suggested engineering changes are once again incorporated
7
during the solid modeling stage.
Solid Modeling & conversion of solid model
into STL file

8
Pre-processing:
• The first step in the RP&M process is virtually identical for all
of the various systems, and involves the generation of a three-
dimensional computer-aided design model of the object.

• A good, preferably solid CAD modeling or a total enclosed

surface – water tight – CAD model is a key component of
success RP processing.

• The CAD file is then translated into a triangulation tessellated

STL format, which is the standard of the RP&M field.

• Figure shows a typical example of STL model which is

composed of triangles and each triangle is described by a unit
normal vector direction and three points representing the
vertices of the triangle.

• Under preprocessing data conversion, validation, Part

Orientation, slicing and support generation are the important
steps.
9
Conversion of solid model into .STL file:
Figure shows a typical example of STL model which is composed of
triangles and each triangle is described by a unit normal vector
direction and three points representing the vertices of the triangle.

10
.STL File
 The STL file is derived from the word STereoLithography,
which was the first commercial AM process
 STL should stand for Stereolithography Tessellation
Language. STL files are generated from 3D CAD data within
the CAD system. The output is a boundary representation
that is approximated by a mesh of triangles.
 STL files can be output as either binary or ASCII (text)
format.
 The STL file is normally labeled with a “.STL” extension that
is case insensitive, although some AM systems may require
a different or more specific file definition. These files only
show approximations of the surface or solid entities and so
any information concerning the color, material, build layers,
or history is ignored during the conversion process.
 Furthermore, any points, lines, or curves used during the
construction of the surface or solid, and not explicitly used
in that solid or surface, will also be ignored. 11
.STL File
 An STL file consists of lists of triangular facets. Each
triangular facet is uniquely identified by a unit normal
vector and three vertices or corners.
 The unit normal vector is a line that is perpendicular to
the triangle and has a length equal to 1.0.
 This unit length could be in mm or inches and is stored
using 3 numbers. The STL file itself holds no dimensions,
so the AM machine operator must know whether the
dimensions are mm, inches, or some other unit.
 Since each vertex also has 3 numbers, there are a total of
12 numbers to describe each triangle.

12
 STL file generation:

Generation of Triangulation

Calculation of vectors

Calculation of surface normal of every triangle

Generation of STL file

13
 Rules for generation of triangulation:
To perform a triangulation following procedure can be used.

A connected, knot-to-knot surface triangulation is a finite

set of triangles satisfying the following conditions:

a) Each edge in the triangulation is shared by at

most two triangles (no bifurcations).

b) A vertex in the triangulation can be shared by any number of

triangles.

c) Each triangle has at least one point in common with another

triangle (connectivity; requires at least two triangles).

d) No triangle has an intersection with the interior of any other

triangle (no piercing / overlapping). Two triangles are called
neighbours if they share a common edge. 14
 Calculation of Vectors:
Let P1(3 2 1), P2(6 4 5) and P3(9 8 7) denotes three vertices of a
triangle. The two edges P2-P1 and P3-P1 denote two vectors lying in
the plane of triangle and intersecting at common vertex V1.

V1  i ( P2 x  P1x )  j ( P2 y  P1 y )  k ( P2 z  P1z )

i.e V1  i (6  3)  j ( 4  2)  k (5  1)

i.e V1  i (3)  j (2)  k ( 4)

Similarly


V2  i ( P3 x  P1x )  j ( P3 y  P1 y )  k ( P3 z  P1z )

i.e. V2  i (9  3)  j (8  2)  k (7  1)

i.e. V2  i (6)  j (6)  k (6)

The surface normal is given by vector product or cross product of

V1 and V2.
15
 Calculation of Surface Normal:
Surface normal = i j k
3 2 4
 
6 6 6 

i.e. S.N  i (12  24)  j (18  24)  k (18  12)

i.e. S.N  i ( 12)  j (6)  k (6)
 i ( 12)  j (6)  k (6)
i.e. unit S.N 
6 6

Generation of STL file:

facet normal 0.816548 0.408274 0.408274
outer loop
vertex 3.000 2.000 1.000
vertex 6.000 4.000 5.000
vertex 9.000 8.000 7.000
end loop
endfacet 16
17
Errors in .STL file

18
Errors in .STL file

19
Inverted Normals:

Intersecting Triangles:

20
 Noise Shells:
A shell is a collection of
triangles that form a single
independent object. Some
STL files may contain small
shells that are just noise.
These should be eliminated.

Nonmanifold meshes:

21
• In tessellation various surfaces of a CAD model are piecewise
approximated by a series of triangles and co-ordinate of vertices of
triangles and their surface normals are listed. The number and size of
triangles are decided by facet deviation or chordal error. These STL files
are checked for defects like flip triangles, missing facets, overlapping
facets, dangling edges or faces etc. and are repaired if found faulty.

• Defect free STL files are used as an input to various slicing softwares. At
this stage choice of part deposition orientation is the most important
factor as part building time, surface quality, amount of support structures,
cost etc. are influenced. Once part deposition orientation is decided and
slice thickness is selected, tessellated model is sliced and the generated
data in standard data formats like SLC (stereolithography contour) or CLI
(common layer interface) is stored.

22
Verification and fixing of STL file
• When one create 3D surface model using common surface
modeler, there is very little concern on the orientation of the
surface normal, as most of the tool path generation algorithms
can detect the material side correctly.

• However, when one generate STL data using these surface

models, many converters just use the normal data straight from
the NURBS surfaces – free form surfaces, thus the STL files
generated are not useable without repair.

• Triangles in an STL file must all mate with other triangle at the
vertex and must be properly oriented to indicate which side of
the triangle contains material. Many STL viewers like Magics RP
from Materialise read a STL file to analysis and to correct, the
connectivity and gap in the three-dimensional triangle matrix.
This process is totally automatic and simple to operate.

23
 Validation of generated STL files:

To validate the developed STL files three criteria are used:

[i] Euler’s Formula

[ii] Quality Factor

[iii] Deviation report between the Solid model and

generated STL file

24
 Validation of generated STL files:
[I] Euler’s Formula:
The Swiss mathematician Leonhard Euler proposed and proved a
very useful formula for polyhedra to check the validity of an object.

Polyhedra means the object having faces which are nonself –

intersecting and belong to closed orientable surfaces.

The Euler’s formula is: F - E + V = 2

where, F = number of faces
E = number of edges
V = number of vertices.
This law is applied to closed polyhedral objects only.

However, open polyhedral objects do not satisfy the above equation.

Open polyhedra means removing one of its faces and stretching it

out to lie flat. The simplified formula for open polyhedra is:
F - E + V = 1. To check whether the solid bodies or surfaces are
topologically valid Euler’s formula is used. 25
Euler’s Formula:
In the generation of STL file using point cloud data this Euler’s
formula needs to be satisfied after triangulation to check its is
topologically validity.

Depending upon the method of scanning the point cloud data is

obtained in two different formats i.e organized from continuous
contact scanning and scattered data from laser i.e non contact
scanning.

To perform triangulation for organized data, Matrix triangulation is

used and for scattered data, Delaunay triangulation method and
Matrix triangulation method using Row-Column technique is used.

Using the information from files required for the generation of

triangulation and output of triangulation, it is possible to calculate
the number of vertices, faces and edges.

26
[II] Calculation of Quality Factor of all the triangular elements
generated in a STL file:
Quality factor:

STL triangulation cannot be used directly in the finite element

method (FEM), mainly because it requires specific characteristics
in the geometrical description of the domain to be computed. In
FEM, the geometric and functional support is provided by the
elements (triangles or others), and they must have a specific
shape, i.e. the proper size and the proper quality factor for the
error estimator of the calculation to be as high as desired.

Elements with Qe larger than 0.7 are considered to be of a good

quality while elements with a value lower that 0.3 indicate poor-
quality elements. This quality factor is bounded between 0 (for all
kinds of degenerated triangles) and 1 (for equilateral triangles).

27
[II] Calculation of Quality Factor of all the triangular elements
generated in a STL file:
Quality factor:

In general quality factor can be looked upon as ratio of area of

triangle to the perimeter of triangle. Mathematically Quality factor
has been defined as:
QFACTOR = 4*1.7320*AREA/(E12+E22+E32)
where E1,E2,E3 denotes lengths of three edges of triangle.

A program name “ANAVAL” has been developed to estimate the

quality factor of all the triangles. Every individual triangle is
considered and its area and lengths of three edges are calculated.

28
Role of offset parameter in generating
tessellation:
In a process of tessellation there is some control over the size of
the triangles to be used in the model. Since STL uses planar
surfaces to approximate curved surfaces, then obviously the larger
the triangles, the looser that approximation becomes.

Most CAD systems do not directly limit the size of the triangles
since it is also obvious that the smaller the triangle, the larger the
resulting file for a given object.

An effective approach would be to minimize the offset between the

triangle and the surface that it is supposed to represent. A perfect
cube with perfectly sharp edges and points can be represented by
12 triangles, all with an offset of 0 between the STL file and the
original CAD model.

It is always a compromise between surface approximation and

excessively large file.
29
 An original CAD model
converted into an STL file using
different offset height (cusp)
values, showing how the model
accuracy will change according
to the triangle offset.

The exact value of the required offset would largely depend on the
accuracy of the AM process to be used. If the offset is smaller than
the basic resolution of the process, then making it smaller will have
no effect on the precision of the resulting model.
Since many AM processes operate around the 0.1 mm layer
resolution, then a triangle offset of 0.5 mm or slightly lower will be
acceptable for most AM technologies.
30
 Calculation of Each Slice Profile
 AM works by adding layers of material of a prescribed
thickness, starting at the bottom of the part and working
upwards, the part file description must therefore be
processed to extract the profile of each layer.

 Each layer can be considered a plane in a nominal XY

Cartesian frame. Incremental movement for each layer
can then be along the orthogonal Z axis.

 The XY plane, positioned along the Z axis, can be

considered as a cutting plane.

 Any triangle intersecting this plane can be considered to

contribute to the slice profile.

31
Algorithm for testing triangles & generating
line intersections

A vertex taken from an STL triangle

projected onto the y = 0 plane.
Since the height zi is known, we can
derive the intersection point xi. A
similar case can be done for yi in
the x = 0 plane

32
Factors affecting the quality of printed
parts:

 Part Orientation
 Support Generation and Editing
 Slicing (Layer thickness)
 Part Building
 Post processing

33
Part Orientation:
The part should be orientated with
minimum height in order to reduce
the number of layers.

For processes that need supports

structures, part orientation should
also be optimized such that it would
require minimal support hence
reduce to building time and the
building material.

The part will be strongest (Ultimate

strength) along the Y-Axis and the X-
Axis and less strong along the Z-Axis
which is illustrated in given figure.

34
Support Generation and Editing:
Support can reinforce delicate
part while building and in the
post processing stage.
However over support for
delicate part will increase the
difficulties of support removal
and even destroy the fine
details on the down face. Result of missing supports for overhanging
areas

35
Slicing (Layer thickness)
A STL model used for RP contains a collection of planar triangular surfaces.
These faces define an approximate boundary for the object. Horizontal layers
of equal thickness are produced while slicing to produce the outer boundary
of the part – slice curve. Then void and solid region of the slice curve can be
identified and proper fill pattern can be created for part filling. Typical layer
thickness of commonly RP system is ranged from 0.05mm to 0.15mm.

36
Part Building
The prototype can be built in the RP machine according to the toolpath and
control codes generated by the software of the RP system. To generate the
toolpath following parameters needs to be considered:

1. Outer boundary generation

2. Layer thickness
3. Shell thickness
4. Fill density and infill patterns
5. Support type
6. Printing speed [50 to 100 mm/sec.]

Outer boundary generation

37
Line Rectilinear Concentric Honeycomb
Pattern Pattern Pattern Pattern

Part Printed With Support

Material
Octagram
Hilbert Archimedean
Spiral Pattern
Pattern Pattern

Infill Pattern Support Types : Everywhere

Honeycomb Pattern and Line Pattern and Touching Baseplate
 Integration of Reverse Engineering & Rapid
Prototyping
• Both reverse engineering and rapid prototyping are
emerging technologies that can play a promising role in
reducing the product development time. The CAD model
developed by a reverse engineering process can be
converted to the physical prototype using a rapid
prototyping technique. It allows user to fabricate features
that are difficult or impossible to fabricate by machining
operation.
• In order to bridge rapid prototyping and reverse
engineering technologies, efficient point cloud
handling methods have to be developed first.
Secondly, an accurate geometry input format for RP
machines needs to be prepared.
39
Thank You !

40