Integration of CAPP with other CA systems
Ivan Kuric
1. INTEGRATION OF CAPP WITH OTHER CA
SYSTEMS
The flexibility and effects of a process planning system is increased with integrating
neighbouring systems (design, quality, manufacturing, scheduling and control).
1.1 Integration of process planning and scheduling
Process planning and scheduling have a close link. The process plan involves the time
order of manufacturing operations and information of workplace of process operation
realisation. The process plan is one of the significant input to the scheduling system.
There are several research projects concerned with the problem of a new approach in
industrial application. One of industrial problem is to originate a functional link between
process planning and scheduling.
The loading situation in a workshop is significantly determined by the selection of
resources performed by process planning. Traditionally process planning only considers
technological criteria and no logistical goals. To improve collaboration between process
planning and scheduling, it is needful to have a close relation between parameters of
process plan and the current situation in workshop.
Following are three different approaches for the integration of process planning and
workshop scheduling:
 dynamic process planning,
 Just-in-Time process planning,
 Non-linear process planning.
1.1.1 Dynamic Process Planning
The manufacturing of product is not realised according to forward planned process
operations. The process operation sequence considering this approach is not known for
the whole process at the beginning of part producing. This approach does not determine
the complete operation sequence and the corresponding resource allocation at the same
time.
After each of finished operation the actual workshop situation is recognised and the best
next operation and suitable resource is determined to continue manufacturing of this
piece. Some higher level planning has to be carried out to ensure that this process
planning approach does not frequently generate dead ends. Dynamic process planning
aims and supports full integration and concurrence between process planning and
scheduling activities.
Dynamic process planning provides flexibility for the process plan according the current
state of the workshop. The approach avoids all unnecessary planning effort on unused
alternatives. Disadvantages are that only local sub-optima can be achieved because only a
very limited time horizon is considered at the time of decision making.
1.1.2 Just-in-Time Process Planning
Just-In-Time is well known and an effective and popular method for product planning and
control. Just-In-Time process planning is started just before the first manufacturing step.
It is a very good alternative for re-using previous process plans or creating process plans
weeks before manufacturing. The new planning approach takes the actual workshop
situation into account for decision making about the resources used for manufacturing a
part.
This process planning approach is realised in the CAPP system PART, developed at the
University of Twente in the Netherlands. The PART system is a commercial software
system that is distributed by CONTROL DATA. The other commercial DTM-CAPP
system, distributed by Somatech, contains similar planning features.
Advantages of this approach are that a well balanced workshop load can be achieved and
it is not required to plan alternative routes in detail that are not used later. The result of
this process planning is a conventional linear process plan that can easily be exchanged
with existing MRP or workshop control systems.
Disadvantages of this approach are that a process planning session is started for a
complex part with many operations, the actual workload is hardly predictable. If the load
with is a mix of complex parts and simpler parts, it is not possible to achieve a planning
optimum. Also the re-use of process planning information from previous manufacturing is
not possible as the individual part may be manufactured each time in a different way.
1.1.3 Non-linear Process Planning
A basis for the new approach is linear process planning which also includes
manufacturing alternatives or possible changes in manufacturing sequence. Several
alternative routings or sequences of operations are combined in a net structure. Netted
process plans are called non-linear process plans. The required initial process planning
effort is high. The non-linear process plan gives full flexibility to optimally load resources
and also to re-allocate jobs in case of unforeseen disturbances.
Non linear process planning approach is used in the FLEXPLAN and COMPLAN
software system.
1.2 CAD/CAPP
1.2.1 System View on integration of CAD/CAPP systems
A process plan as a result of the CAPP system is created according to prepared data from
CAD system. As the process plan converts data of raw-stock into manufacturing
instruction, the CAD and CAPP systems have direct information relationship. It is
advisable to integrate CAD and CAD moduls in to one CAD/CAPP system with common
bases.
A data transfer between CAD and CAPP systems is one of the most complicated
problems in the information integration of CA systems. The ideal solution for CAD and
CAPP integration, is that both systems use the same internal representation of the data
model. As CAD systems are different, they use a different representation of the data
model.
There are two approaches for the creating and processing of CAD data for CAPP systems:
 the first approach is based on a general CAD part model and from the recognition of
separated part features. Design features are recognised from this model and saved as
manufacturing features. The recognising interface must be developed. Design
recognition is a very difficult and serious problem.
 the second approach comes from special created CAD systems, that are modelling the
design and manufacturing features of a part (Feature based modelling). This is
expanded modelling of the parts following predefined features. The feature has its own
geometry and must be associated with attributes of the part. The features are
geometrically independent. The designed part consists of design features (e.g. holes)
and manufacturing features (e.g. thread surfaces). Feature part model includes
geometrical, topological, technological and relational properties. There is specially
designed feature modeller. The feature modeller consists of design and manufacturing
features which are immediately recognizeable by the manufacturing planner and/or
computer system. The advantage is that all features on the part are manufacturable.
Feature based design is used in many generative CAPP systems. Feature based design
is the most recent and promising technique regarding the interface with process
planning. In the majority of cases, form futures are used and the manufacturing
features are defined by interpreting and combining the design form features from the
viewpoint of manufacture.
1.2.2 General CAD model of part
INTERNAL CAD DATABASE
AutoCAD
DXF, IGES
AutoLISP
TRANSFORMATION
CAD/CAPP MODUL
PART CODED MODEL
Geometrical, topological,
technological, relational properties
FEATURE
LIBRARY
FEATURE
CAD
Figure 1.A: Part coded model created in Cad system
During creating the model, data related to the existing engineering part are loaded in the
CAD system. All data are available in the internal CAD database. Some data are used for
local phases of the graphics system, other data are necessary during the whole cycle of
product creation. The information integration with other CA systems can be realised by a
part coded model.
Part coded model is generated by:
 the direct processing of the internal CAD database,
 the transformation of the standard digitized drawings (DXF, IGES),
 means of Feature base modeller.
Model can be transferred by a so-called „neutral“ format - DXF (Drawing Exchange
Format) and IGES (Initial Graphics Exchange Specification), that are used by most CA
systems working on the PC basis. These formats are exported by means of CAD systems
from the internal database.
This way only some data needed for processing in CAPP system are extracted. The
standard formats (e.g. DXF) include all drawing data needless for following processing.
The problem of the standard format transformation can be solved in two ways.
The first method is based on a developed graphic interface for extraction of the necessary
data. Another significant mean is the utilising of the implemented program language of
the graphic system. By means of the program language (e.g. AutoLisp) in a CAD system
it is possible to create from internal CAD database the own file structure including only
necessary data.
The preferable way is the program language utilisation of the graphic system. The
advantage consists in direct access to internal CAD database and access to all properties
of drawing entities. Generated data - the part coded model is in a file structure desirable
for processing in CAPP system.
1.2.3 Geometric reasoning
For planning process a detailed and precise model of part is necessary. Engineering
drawings are not only represented by geometry properties, they are also supplemented
with drafting symbols (hatch, roughness symbol, etc.) and texts. The overall shape
(geometry and surface properties) is very important for design of the process plan.
Transformation of a graphics model (digitised drawing) of part into a model of the part
which is useable in CAPP system is probably the most important and most difficult
interface task in process planning.
The overall shape of part must be identified from the lines and curves of the engineering
drawing. When the overall shape is found it is necessary to identify the properties of
separate surfaces. Transformation of a graphics design model of part into a manufacturing
model is possible to decompose on following research area:
 identification (recognition) of overall shape,
 decomposition of overall shape,
 identification of surface properties,
 transformation into manufacturing model.
This transformation task is called geometric reasoning, feature refinement or CAD
interface.
Normally a drawing is translated into process planning specific data by a human user. In
automated process planning the problem of transformation model is a complicated and
difficult task.
Transformation of a graphics model (digitised drawing) of a part into a product model,
which is advisable in CAPP system, is probably the most important and most difficult
interface task in process planning.
Feature based design is the most recent and promising technique regarding the interface
with process planning.
Data transformation or design inteface is executed in the following steps:
1. feature recognision - classifies and identifies the semantics of the feature,
2. model decomposition - separates the feature from a part model,
3. assigning of the geometrical, topological, technological and relational properties.
CAD
INTERNAL CAD
DATABASE
FEATURE
RECOGNITION
MODEL
DECOMPOITION
FEATURE
ATTRIBUTES
IDENTIFICATION
MODEL
TRANSFORMATION
PART CODED
MODEL
CAPP
Figure 1.B: Geometric reasoning – transformation of general CAD model
b.
c.
24
a.
20
Feature
Nr. 1
Feature
Nr. 2
Feature
Nr. 3
Des cription
Des cription
Des cription
DATA – Properties
FEATURES
Geometrical
Topological
Technological
Relational
Feature Nr.1
Feature Nr.2
...
Feature Nr.i
Figure 1.C: Transformation of Drawing Model into Part Coded Model
The modern approaches such as feature based design, geometric reasoning, CAD
interface, knowledge based system and artificial intelligence seem to be a natural
candidate for the application of transformation general CAD data to data suitable for
processing in CAPP systems.
2. TRENDS CONNECTED WITH CAPP
SYSTEMS
Machining process planning consists of machining operations which are required to
transform raw stock into a finished part. Conventional process planning is performed
manually and depends on the knowledge, competence and experience of the process
planner. Human reflection has the characteristic marks of reasoning, recognition of
significant and inessential events, abstraction, parallel or complex consideration,
intuition. There are also limitations of human reasoning – subjective approach, forgetting,
small computing power. Computer support aids removal of these disadvantages. Existing
programming tools for creating of computer systems limit computer aid. There are two
views on computer aid for engineering activities:
 hardware and
 software standpoint.
Computer hardware power determines computational power, data processing rate, data
transfer rate, possibility of image graphics, amount of saved data, retrieval speed.
Software limits the processing and processing methods. Previous software equipment and
programming tools were unable to perform advanced methods such as knowledge
processing, pattern recognition.
Activity – task – parameter
Human
Computer
Data processing
Parallel
Linear
Processing speed
Small
High
Data retrieval speed
Small
High
Reasoning
Complex
Individual
Computational power
Sufficient
Enormous
Abstractive reasoning
Yes
No
Forgetting
Yes
No
Recognition of significant and Yes
inessential events
No
Intuition
Yes
No
Subjective aspect
Yes
No
Learning, self-education
Yes
Partially
The first point at starting CAPP was help in report generation, storage and retrieval. It
was a computer assisted system. Today there is request to build an expert system
supplanting human planner activities and allowing to emulate the capabilities of an
experienced planner.
Current computer aided systems are based on modern respects of knowledge, heuristics,
feature modelling and artificial intelligence.
Generally it is possible to follow three main trends in CA system:
 integration,
 flexibility and
 intelligence.
The computer aided systems such as CAD, CAPP, CAM, CAQ, PPS with information and
material flux and relations constitute computer integrated manufacturing (CIM). CAPP is
an important link between CAD and CAM and often is the heart of an integrated system
in the engineering industry. Therefore the data integration and data transportability is a
very important property of CAPP systems.
Flexibility of CAPP system determines the ability to quickly fit and quickly respond to
client requirements.
Intelligence of software represents a higher level CAPP of adjustment to human
reasoning.
Intelligence of CAPP systems is above all in:
 knowledge and manipulation with them,
 modelling of product and process plan,
 methodology of process planning,
 data interface between CAD a CAPP.
2.1 Expert systems and artificial intelligence
The major activities of process planning are based on logic and knowledge. The process
planner makes decisions during various steps of planning.
The planner often applies heuristic solving methods, knowledge, experience and intuition.
Many years of practice and knowledge of researched area are often of the most
importance in the decision process.
An important tool in the development of CAPP systems are developments in artificial
intelligence and expert systems (ES). The artificial intelligence technique is used for
automated interpretation of the part and for process planning activities.
The CAPP expert system is a tool which has the capability to understand problem specific
knowledge of planning and use the domain knowledge to suggest an alternative path of
process planning action. A very important task in advanced CAPP systems is the
representation of knowledge. There are several known schemes and techniques
(production rules, predicate logic, semantic nets and frames). Production rules are one of
the most commonly used knowledge representation schemes.
Much of the CAPP software is developed on the basis of the object oriented
programming (OOP) method. OOP is an artificial intelligence paradigm which provides a
data structure for symbolic manipulation of conceptual information. Characteristic of
OOP is heredity of object properties. It is also an aspect of human reasoning.
The knowledge and expert systems present a new manner for the computer application.
They originated as a practical result of exercising the knowledge and experiences from
the artificial intelligence area. The principles of expert systems are based on human
experiences. The expert systems are based on the idea of acceptance of the knowledge
from experts including the heuristic method too.
Supposition for problem solved by the expert systems is resolution executable with
consideration. The expert systems simulates the solving procedures carried out by human.
The expert systems is used for problems solved with unknown deterministic resolving
methods. Acquisition of the expert systems is implementation heuristic process on
computer and use expert knowledge by non expert too. The knowledge assumed from
expert are encoded by advisable concept and saved in the knowledge base. The aim of
knowledge base systems and expert systems is to accomplish decision making on the
same as a human expert.
2.2 Learning, self-education CAPP system
One of the developing modern approaches is the implementation of human learning
capabilities in CAPP systems. The designed process planning system - hybrid system uses two approaches for planning - generative and variant. The basic approach is
generative. The developed generative-variant hybrid system is based on feature
modelling. For describing the part a GT code is applied. Individual process operations are
generated for each feature of the part. Selection and determination of process operations
is realised on basis of logical and heuristic approach. The variant approach of CAPP
system is used in the module of process operation selection. For concrete GT code is
recommended process operation from manufacturing knowledge base. The process
planner according to his experience can select some process operation from his
knowledge base. The selected process operation with the corresponding GT code is
subsequently stored in knowledge base. In this manner the CAPP system has learning
capability.
CAD
.....
Machine, Tool,
Fixture Base
Elementary Operation EO
generation
Manufacturable
Knowledge Base
Selection of Elementary
Operation EO
Advisable
Elementary
Operation EO
variant approach
Manufacturable
Knowledge Base
The Most Applied EO
Choice EO + GT Code
Learning
GT Code
Base of the Most
Applied Elementary
Operations
.....
PROCESS PLA N
Generative Variant
Hybrid System
Figure 2.A: Section of generative variant hybrid CAPP system
2.3 Trends in CAPP
The first computer application of process planning was to find the optimum cutting
conditions and report generation as well as documentation retrieval. After the variant
approach follows the generative one. Although the introduction of artificial intelligence
and expert systems improved the interest in planning problem and the capabilities of the
systems, the result are still far from desirable.
A majority of CAPP systems are still computer aided instead of computer automated. In
spite of this comment there are problems that are resolved in an automated way
(geometric reasoning, selection of machine equipment and optimisation of process
operation sequencing, etc).
From reviewing the literature, projects and research, the followings trends are emerged:
 feature based design and process planning,
 design and process planning interface,
 knowledge based process planning functions,
 product and process modelling,
 information technologies,
 concurrent/simultaneous engineering,
 future based process planning,
 geometric reasoning, CAD interface,
 utilising of artificial intelligence, expert systems
 neural networks,
 fuzzy logic,
 object oriented programming and genetic algorithms,
 holonic approach to CAPP,
 integration with other information and enterprise systems.
Research in concurrent and simultaneous engineering is the most promising current trend
towards an integrated development of products and processes.
There is a close relationship between process planning, modelling and tools for
concurrent engineering, such as design for manufacturing (DFM) and design for assembly
(DFA).
Further development is needed in tolerancing techniques and fixture planning methods. A
major success of the geometric modelling and process planning, with respect to feature
based modelling, is in CAD/CAM systems.
Process planning research is an extremely varied subject area. Currently there are some
well developed applications. New process planning research will be placed in the
concurrent engineering context and will be closely linked with the developments in
engineering design and production control.
The automatization process planning is not a simple matter. A single algorithm can not
model the complexity of the thinking process of an experienced human planner. The
modern approaches of feature based system, geometric reasoning, CAD interface,
knowledge based system and artificial intelligence seem to be natural candidates for the
application.
New approaches in CAPP are in the following area:
 information technology methods, e.g. neural network, OOP, genetic programming,
etc.,
 methodology of CAPP, e.g. expert planning methodology, parametrizing of process
plan, etc.