1. IE 590 Integrated Manufacturing Systems Lecture 5
Automated Extraction of Features

2. Concept of a ‘Feature’ Manufacturing is “feature” based Feature:
Structure, form or appearance
Makeup or appearance of the face or its parts
Prominent part or characteristics

3. Features ..Continued
A high level geometry which includes a set of connected geometries. Its meaning is dependent upon the application domain
Manufacturing is feature based

4. Manufacturing Features
For process selection
For fixturing
Round hole
Rot. Feature
Plane surface
Hole, profile
Free form
Surface
Cyl. Shell
Cyl. shell
Drilling
Turning
End Milling
Ball end
milling
Boring
Reaming
End Milling a slot

5. Extract and Decompose Features from a Geometric Model
Feature Recognition
Extract and Decompose Features from a Geometric Model
Controls ?
Identified
Features
Perform Feature
Recognition
Inputs ?
Mechanisms ?

6. Feature Extraction Importance
In CAD/CAM integration, automation of FE leads to automated process planning
this in turn bridges the gap between CAD/CAM
(contd.)

7. Feature Extraction (FE)
Steps involved in automating FE
Understand CAD Model (input)
Develop interface
Write translator
Formulate Feature Extraction rules
Implement software to complete FE
(contd.)

8. Related to this topic, we will cover
Methods to identify features
Data structures to support FE activities

9. Convex and Concave Edges (Nnaji 85)
Convex Edge
Edge which connects two faces whose angle is greater than 180°
Test: (f1 x f2). e3 < 0
Concave edge
Edge which connects two faces whose angle is less than 180°
Test: (f1 x f2). e3 > 0

10. Test for Convex Edge z Test for edge e1 rf x lf = i j k y
(f1 x f2)
= i
(f1 x f2).e1 = (i) . (-i)
= -1 (<0)
Therefore, e1 is convex y x f2 f1 f3 e1 e2

11. Test for Concave Edge z Test for edge e2 rf x lf = i j k y
(f3 x f1)
= - i
(f1 x f2).e1 = (-i) . (-i)
= 1 (> 0)
Therefore, e1 is concave y x f2 f1 f3 e1 e2

12. Multiple Convex Border
The feature, which is a through hole has two convex borders
Convex border1
Convex border2

13. A face containing a concave edge is called a concave face
Any concave face is added to a list called concave face group (CFG) if it shares a concave edge with one of the group’s members

14. Convex Border If a closed loop has edges which are all convex,
it is called a convex border
It is located by searching a closed loop of
convex edges from a face or a group
Convex border

15. Concave Feature
Convex borders which are part of the parent face form the borders of concave features
Parent face
Convex border
Concave
feature
Concave border

16. Convex Feature
Convex border
Convex feature
Parent face
Concave border

17. To be able to extract features, a robust data structure is needed
Winged Edge Model (WEM) is one such data structure
Baumgart initially proposed WEM
Braid further developed this concept to include faces with holes and introduced the concept of loops
In this course, we will use a modified WEM

18. Winged Edge model Starting right Ending left End Vertex (ev) Edge (sr)
Edge (el)
End Vertex (ev)
Left face (fl)
Right face (fr)
Start
Vertex (sv)
Ending right
Edge (er)
Starting left
Edge (sl)

19. Example of WEM data associated with each edge
Data structure for
Edge e1
sr = e2
er = e4
sl = e5
el = e7
sv = V4
ev = V1 V7 V2 e2 V6 e7 V1 e3 F1 F2 e6 e1 V3 V8 e4 V5 e5 V4

20. WEM (using loops) Faces can be represented using loops
Loop L1 = outside boundary of face F1
Loop L2 and L3 are boundaries representing holes in the face F1
Each loop  face F1
Face F1
Parent (L1)
Child (L2)
Child
(L3)

21. Types of loops P-loop or parent loop C-loop or child loop
A loop which forms the outside boundary of each face
Each face has at least one P-loop
There is a single P-loop for every face
Number of faces in solid = no. of P-loops
C-loop or child loop
A face can have more than one C-loop
(contd.)

22. A C-loop indicates the presence of a feature like protrusion or hole
In a C-loop, if there is a feature such as a hole, there will be a C-loop corresponding to the ring of edges (even though there is no face)
Child or C Loop