Strategies for modeling geometry in Knowledge Based Engineering systems
FöRFATTARE
Andersson, Petter
DATUM
2001-02-15
INSTITUTION
Maskinteknik / Datorstödd maskinkonstruktion
SAMMANFATTNING Volvo Aero is a company in the aerospace industry and is currently
developing new engineering design systems in this direction. One approach to
the problem is to use KBE (Knowledge Based Engineering), which is an acronym
for object oriented modelling system. Knowledge is built into the system
using “knowledge objects” and linked databases. This is a very powerful tool
with the ability to constrain geometry to abstract objects, (e.g. cost
objects, manufacturing objects and organisation objects) with rules and
databases. A challenge with this approach is that the system tends to expand
and become wide ranging, beyond any possibility to survey. The trick appears
to be how to keep it as simple as possible and to use KBE where it gains the
most. This is where the objective for this thesis appears, “To which level
of geometry detail is it worthwhile building models in a KBE-system?”
Another aim for this thesis has been to enlighten the difference between
traditional CAE (Computer Aided Engineering) and KBE.
In early phases of design, decisions are of more fundamental character and
have a large influence on the forthcoming product. Time is a severe
restraint and modelling is usually time consuming. Flexible modelling,
allowing design iterations in early phases, is a way to improve the decision
base significantly. This is where KBE shows its strength. Within minutes the
KBE system changes configuration possibilities constrained to rules and
related data, presenting the new output in the way a designer has programmed
the system. It can produce 3D geometry, product data sheets or/and
blueprints. Normally, such activities take days or weeks and may not be
carried out at all. The approach has been to study KBE and object oriented
modelling in general. A test case has been defined to illustrate the
difference between Traditional CAE and KBE. I-Deas has been used to
illustrate traditional CAE. The KBE system has been defined in AML and
explains the theory and thoughts around the conclusions.
Products that will benefit with a KBE system are:
In this work, geometric modeling in object oriented engineering designs
systems has been investigated. These new design systems present one approach
directly addressing the immediate needs for rapid design and evaluation
existing in today’s industry.
By analysing the product specification and identifying those design
activities that will benefit most from using a KBE system, a design cycle
can be formed that produce a product model based on these activities.
Creating a design cycle based on these design activities will generate a
generative model with the minimum level of geometry necessary for the model.
In general, design activities that are frequently repeated take advantage of
the KBE systems ability to iterate through a variety of configuration
possibilities. Tedious time consuming work such as computational
pre-processing can be automated by integrating computational modeling as a
part of the generative product model. The final product will then be modeled
in a traditional CAE tool. Examples of activities where a generative model
may be cumbersome to define are activities such as NC-code generation and
mpeg visualization. These activities have a high impact on the level of
complexity since the need of geometry precision is high. In addition such
activities normally have a relatively low impact on the decision-making
process.
ISSN 1402-1617 / ISRN LTU-EX--01/178--SE / NR 2001:178
