INTELLIGENT DISTRIBUTED MISMATCH CONTROL FOR EUROPEAN AEROSPACE INDUSTRY

boingcadgeMechanics

Nov 18, 2013 (3 years and 6 months ago)

167 views




INTELLIGENT DISTRIBU
TED MISMATCH CONTROL

FOR EUROPEAN
AEROSPACE INDUSTRY


V. Taratoukhine, Researcher, K. Bechkoum, Head of Department, M. Stacey, Senior Lecturer


Department of Computer & Information Sciences, De Montfort

University, Hammerwood Gate, Ke
nts Hill, Milton Keynes, MK7 6HP, U.K.



Abstract


The principal aim of this research is to support existing CAD/CAM systems by providing an automatic solution
for detecting design inconsistencies in a distributed
design environment especially for aerospace industry. The
proposed investigation is also aimed at supporting the designer in finding the modification that would eliminate
the mismatch while the overall design integrity is still preserved.


1

INTRODUCTION




One of the key challenges for Europe is to maintain and develop the European Aerospace sector as a world
competitive industry. The European Commission has fostered several collaborative research initiatives in
aeronautics yielding a number of successful

projects. In the Fifth Framework Programme of the EC the financial
support dedicated to the Aerospace industry alone is set to euro 700 million.

For many years the design and manufacture of major European aerospace products has been distributed across
the

continent; Airbus and EFA being typical examples. What makes co
-
operation amongst partners of the
European aerospace sector more challenging is the fact that the design process tends to be sequential and
requires centralised planning teams and a great dea
l of travel on the part of the distributed designers.

Distributed Design and Computer Supported Collaborative Work (CSCW) techniques promise to resolve most of
the difficulties above by replacing the paper and tape and physical meetings based methods by e
lectronic
communication and electronic meetings and provide a basis for Concurrent Engineering (CE) environment
[17][19][22].

According to [21] “Concurrent engineering is designing for assembly, availability, cost, customer satisfaction,
maintainability, m
anageability, manufacturability, operability, performance, quality, risk, safety, schedule, social
acceptability, and all other attributes of the product”. In this paper we focus on one aspect of the distributed
design process and Concurrent Engin
e
ering su
pport: design consistency checking during the assembly stage.

The design for assembly, as part of the CE process, is one of the most important issues, because the design of
large

complex products for automotive or aerospace industry is not possible without

using modern sophisticated
tools for full support of the virtual mock
-
up process. In this context, it is very important to analyse current
software for a virtual

mock up process. Some of the main tools are presented below:

dV/MockUp [9] is a family of t
ools for Interactive Product Simulation
-

the process by which design and
manufacturing companies can study the form, fit and function of their products.

Tecoplan Automatic Design Verification (ADV)[2] automatically uncovers all collisions and violations o
f
minimum distances during the early design phase. ADV based on Tecoplan’s formal model name Tecoplan’s
Space Management.

Another commercial systems such as CATIA [8] and I
-
DEAS [13] do provide facilities for assembly
mismatch control, but their approach
es are more focussed on the tolerances.

The commercial Virtual Mock
-
up systems mentioned above are very powerful tools. However, current virtual
mock
-
up software, in general, does not have facilities for detecting and resolving design mismatches. The
syst
ems can detect geometric inconsistencies based on syntax level, but have no capability of advising how to
change the design project in order to meet design requirements.




The following part of this paper describes some of the developments that have been mad
e to date in the field
using AI methods for Collaborative Design.

2 DISTRIBUTED ARTIFI
CIAL INTELLIGENCE IN

ENGINEERING DESIGN

Many of the recent developments in the field of “AI for Design” have been investigated and described by Akman
et al. [1], Brown an
d Grecu [7], Bento and Fejo [5], and Nyacinth and Nwana [12], Bechkoum and
Taratoukhine[4][20].

Subbu et al [18] present a Virtual Design Environment to support design
-
manufacturing
-
supplier planning
decisions in a distributed, heterogeneous environment.

The approach utilises evolutionary intelligent agents as
program entities, which generate and execute queries among distributed computing applications and databases.

Leiening and Blount [15] have described the implementation of knowledge
-
based engineeri
ng (KBE) for
aircraft wheel and brake industry. The paper investigates tools to increase productivity, explains the way a KBE
tool works, and describes possible KBE applications as design and diagnostic tools.


Distributed AI (particularly agent
-
based) is
increasingly being used to support distributed environments. A
recent description of the technology can be found in [7][5][11][12][16]. For example, Pan, et al [16] have
described an advanced CSCW technology which has a great impact on the communication a
nd cooperation
during the design process of products based on Multi
-
agent framework. Hale and Graid [11] have developed a
distributed intelligent system for aircraft design based on conception of a design integration framework. An
Intelligent Multi
-
discipl
inary Aircraft Generation Environment (IMAGE) is discribed, which uses state
-
of
-
the
-
art computing technologies.

The analysis of implementation of AI methods for collaborative design suggests that current design support
systems are highly specialised and u
se different approaches such as expert systems, multi
-
agent systems and
intelligent interfaces. Unfortunately, there are very few packages available in commercial or near to commercial
levels. Some systems do provide appropriate facilities for distributed
design however, the differences between
internal models of knowledge analysis and representations of these systems have the restricted these
implementation. What is also clear is that, current systems and models are not using any approaches for the
detecti
on of design inconsistencies, particularly at the assembly stage. The use of intelligent agents as
independent distributed knowledge entities promises to provide the missing link. In this context, the
investigation of methods and principles of organisation

of multi
-
agent system for mismatch design is
investigated. This multi
-
agent architecture will be at the heart of an intelligent distributed mismatch control
system (IDMCS) that aims at ensuring that the overall design is consistent and acceptable to all.


3 INTELLIGENT DIST
RIBUTED MISMATCH CON
TROL

A key ussue developed in the IDMC approach, is to allow the process to perform the design task described in the
scenario, involves the detection of non

compliances within the structural design. At this stag
e the parts which
are being designed concurrently by the distributed design team are brought together, integrated and any
mismatches have to be detected and resolved to ensure that the different components of the design are
compatible.

IDMC uses a concep
tion of distributed artificial intelligence
-
agents. In this case agents are represented as
“virtual designers” which have internal abilities to receive information, to identify design mismatches and to
prepare advice for the designer to find the best modif
ication to resolve the mismatch. Another key ability of
agents is their ability to adapt (to learn) using current information from designers. The general environment for
agent
-
based design is an Internet/Intranet as a more powerful environment for the orga
nisation of distributed
design is used.

In general, a mismatch control includes three stages: receiving design project information, identification
(classification) and generation of result. The vocabulary of indicators is using on the goals of identific
ation of
mismatches and include a classification tree and indicators (taxonomy of design mismatches). The taxonomy of
mismatches is described in next section.




4 TAXONOMY OF DESI
GN MISMATCHES

What is presented here is a part of the broad classification of

geometrical mismatches represented in [3]. An
example of taxonomy for Design For Assembly/Design For Manufacturability (DFA/DFM) is presented as
follows (Fig.1).



























Fig. 1. An example (part) of taxonomy for DFA/DFM.


This taxonomy

is especially oriented for implementation for mismatch detection during the integration phase
of an engineering design. It is particularly useful when dealing with a design assembly process. In this case, the
different (CAD) models are assumed to be error
-
free. The main concern is with ensuring that these models
(designed by geographically remote partners) do combine together to form an integrated design that is coherent
and acceptable to all.

Within the proposed taxonomy critical parameters are identifi
ed. It is the variation of these parameters that
causes a mismatch. For example, the bolted connection requires consistency between such parameters as thread
minor diameter, minor diameter and pitch (for bolt head type and assembly tools type). For bolt an
d nut


diameter, length and size, etc. Weld connections require consistency between types of materials and the material
thickness, as well as the geometric parameters of material.

For the mismatch detection process to be more successful, not only we need
to represent a wider variety of
mismatch types but we also need to represent geometric information as well as information about the material
which the parts are made of.

Taxonomy of DFA/DFM

Failed

assembly/

dissasembly

sequence plan

Dissasembly

mismatches

Sequence

mismatches

Assembly

mismatches

Forward

search:

impossible

assembly


Backward

search:

impossible

dissasembly


The time of

assembly

/

dissasembly

is too much

Non optimal

assembly/

dissasembly

sequence plan


Impossible

dissasembly

Unwanted

contacts


Bad

serviceability


Interaction

mismatches


Other

mismatches


No adequate

dissasembly

tools


Connection

mismatches


Impossible

connection


Interference

mismatches

Impossible

tools changes

operation

Time of

assembly is too

much

No tools for

automatic/manual

assembly

The number

of parts

Mating

direction

Difficulty

for

hand

operations

Symmetry

Assembly cost

is high

Directorability

Stability

Unwanted

contacts




The taxonomy is restricted by assembly mismatches. To define the taxonomy criteria of
classification should
be considered. In our case we have a main criterion
-

assembly process, and additional criteria as types of
connections and indicators (critical parameters
-

M
par
cr
).

5 A MULTI
-
AGENT FRAMEWORK

This section briefly reviews the agen
t framework and the components within. Then we describe the principles of
the agent network mechanisms as inter
-
agent relationships and cooperation.

Using a multi
-
agent approach the representation of the required knowledge will be distributed amongst seve
ral
“specialised” independent knowledge bases, or agents.

The architecture assumes that the design knowledge is encapsulated within the different members of the agent
community. The conceptual framework (CF) may be presented formally as follows:


CF = {A
P
1
, ... , AP
t
, ... , AP
n
},


where AP
t

is the t
th

Assembly Part, t = 1,2, … , n.


AP = {DA
1
, … , DA
i
, ... , DA
m
, CA
1
, ... , CA
j
, … ,CA
k
},


DA
i

is the i
th

Design Agent (D
-
agent), i = 1, 2, …, m,

CA
j

is the j
th

Control Agent (C
-
agent), j = 1, 2, …, k.


Each
DA
i

consists of: FB
-

facts base, which includes information about geometric characteristics of the part
and material type. KB
-

knowledge
-
base. K
-

corrector block
-

which updates the knowledge
-
base, as a result of
communication with other agents. I
-

in
ference engine. LI
-

local interface mechanism.

Each CA
j

consists of: MB
-

metaknowledge base, knowledge
-
base of control agent, inference engine,
corrector block, local and global interface mechanism (GI). Communication is a key ability of multi
-
agent
syst
ems. Agents exchange information using messages with syntax and semantics defined by the communication
protocol. The context of these messages can include declarative and procedural knowledge. The proposed
communication protocol (CP) for the above agents
is as follows:


CP={L
c1
, ... , L
cm
; L
i1
, ... ,L
im
; L
c
c
1
, ... , L
c
c
k
; L
i
c
1
, ... , L
i
c
k
},


where L
i

-

information language, which used by DA
i
, L
c

-

control language, used by CA
i

to update the fact
base of DA
i

with the view of resolving a mismatch. L
i
c

-

i
nformation language, which describes current
situation for control agents, L
c
c

-

control language, which updates meta
-

and knowledge
-
base of C
-
agents.

The conceptual framework is represented as a community of schedule and reactive agents.

A reactive age
nt is an entity that may be represented by an independent program that knows everything about
itself including its relationships with other agents. The principle of emergence states that intelligence in reactive
agents emerges from interaction of agents am
ong themselves and with their environment. The principle of
situatedness states that intelligence of reactive agents is situated in the world and not in any formal model of the
world build in the agent [6].

In our case DA
i
(D
-
agent) is an reactive agent, w
hich negotiate with other design agents using design’
schedule (assembly sequence) and scheduling preferences generated and supervised by the Control Agent (C
-
agent).

6 DISTRIBUTED KNOW
LEDGE BASE

This section presents aspects related to the organisation
of the distributed knowledge
-
base. This requires to be
organised as follows.




MB, KB of C
-
agents and KB of D
-
agents is represented using: (1) production rules for carrying out an
analysis of the design situation; (2) production rules for classification of
situations according to the necessity of
control actions; and (3) production rules for adaptation of distributed knowledge base.

Each rule M
i,
i = 1, ... m is characterised by a premise part, comprising the IF precondition statements, and
the consequent

part (THEN part), comprising the inferred outputs. FB of D
-
agent
-

includes frame facts
containing information about geometry and material type.

7 DYNAMICS OF MULT
I
-
AGENT FRAMEWORK. MIS
MATCH RESOLUTION PRO
CESS

It will be possible to analyse a commun
ication strategies in multi
-
agent network and to develop mismatch
detection and resolution scheme. The general structure of conceptual framework is shown in Figure 2.



































Fig. 2. The conceptual Framework. The Structure.



We have two main levels of vertical communications: (1) C
-
agents
-

D
-
agents (2) C
-
agents

Design Team
(DT) and two levels of horizontal communication as (1) between D
-
agents and (2) between C
-
agents.

The communication between C
-
agents and asso
ciated D
-
agents) is a process of elimination of incosistences
between assembly parts, when D
-
agents are unable to resolve it, using internal knowledge and/or horizontal
communication. This is client
-
server communication (under the supervision of the C
-
agen
t). C
-
agent receives the
D

D
-
agents level

C
-
agents level

Design Project

CA

CA

CA

M
1
i




new information from D
-
agents using syntax of L
i
. The result is external adaptation of knowledge
-
base of D
-
agents (using syntax of L
c
), according to C
-
agents meta
-
knowledge base information, if mismatches occur.

The communication
between C
-
agents and Design Team is a process of elimination of inconsistencies
between assembly parts, when C
-
agents are unable to resolve it, using internal knowledge and/or horizontal
communication. This is human
-
computer communication. The result of co
mmunication is external adaptation of
metaknowledge
-
base of C
-
agents.

Each D
-
agent operates as an independent entity and interacts asynchronously with other D
-
agents on peer
-
to
-
peer level. D
-
agent negotiate with other D
-
agents, using L
i
. D
-
agents are e
ndowed with the ability to negotiate
with one another to ensure that any mismatches are detected and that a solution is proposed. However, D
-
agents
cannot modify each other. “Modification” is the responsibility of the C
-
agent.

C
-
agent negotiates with other

C
-
agents, using L
i
c

and L
c
c
.

8 USING IDMCS IN A
EROSPACE DESIGN. AIR
BUS A340 WING ASSEMB
LY

Underlying this research was the hypothesis that current CAD/CAM systems do not support the sort mismatch
control process described here.

In cooperation with de
sign experts from British Aerospace Airbus and Cranfield University a set of practical
investigations of work practice and technology usage were designed. A wing box is the structural component of
an aircraft. Wing box model consists of several parts
-

st
ringers, skins, spars, ribs and connection details


boltlocks and rivets. In a large aircraft wing there can be over 50 ribs and 100 stringers [14]. The general
assembly process is represented in Figure 3.























Fig. 3. Wing box assembly

process




The IDMCS designed by the authors is a distributed knowledge
-
based design support system which detects
geometric and material irregularities at the assembly stage. The IDMCS provides the mismatch control for
assembly process using an initial se
t of data from British Airbus aircraft design sources. The development of
IDMCS is represented in Fig. 4.



Rear spar

Botton skin

Upper skin

Forward spar

Stringers

Win
g Box

Connection materials,
riveting and bolts

Wing box

Assembly

































Fig. 4. Development of IDMCS


When designing using IDMCS, the following steps are being performed: (1) analysis of as
sembly parts
-

assembly checks of stringers, skins, spars and etc., (2) analytically evaluation of assembly possibility
-

Collision
and Tolerance Analysis, (3) choosing the script (conditions) of virtual mock
-
up, and (4) progress analysis and
generation of

results.

The system analyses designer requirements to the design project given in the form of geometric information
and processes at the level of the distributed knowledge base.

The system is being designed using JAVA 1.2.1 in the Windows NT environmen
t using ZEUS agents building
toolkit [12] and PARASOLID CAD kernel.

9 CONCLUSION AND F
UTURE WORK

The general principles of Intelligent Distributed Mismatch Control are outlined. The taxonomy of distributed
design mismatches is defined. The Structural and

Dynamical Framework for a Multi
-
Agent System that handles
mismatches was developed. Organisation of IDMCS was outlined as well as the possibility of using IDMCS for
aerospace design, particularly Airbus A340 wing assembly process. The agent
-
based prototy
pe (based on
ZEUS), as an initial implementation of the framework, is briefly described.








IDMCS



Knowledge Engineering

Issues


Java

PARASOLID Interface




Java based Integration.



Agent Based Environment



PARASOLID
-
KID geometric
modeller



ZEUS Building To
ol
-
kit agents
definitions



Java
-

external programs



Distributed Design Environment




References


[1]

Akman, V, P.J. Hagen, and T. Tomiyama. A Fundamental and Theoretical Framework for an Intelligent
CAD System, Computer Aided Design Journal, Vol. 22,
pp. 352
-
367, (1990).

[2]

Automatic design Verification. AVD V3, Brochure, Tecoplan AG, Germany, (2000).

[3]

Bechkoum. K., Intelligent Electronic Mock
-
up for Concurrent Design // Expert Systems with Applications
Journal, Vol. 12, pp. 21
-
36, (1997).

[4]

Bechkoum, K.,

V. V. Taratoukhine. A Framework for Mismatch Control in a Distributed Design
Environment, Proc. Advances in Concurrent Engineering, Bath, 1
-
3 September, (1999).

[5]

Bento, J., B. Feijo. An Agent Based Paradigm for Building Intelligent CAD Systems, Artificial

Intelligence
in Engineering Journal, Vol. 11, pp. 231
-
244, (1997).

[6]

Brooks. R., Intelligence without Representation // Artificial Intelligence 47
(1
-
3): pp. 139
-
159, (1991).

[7]

Brown, D., D. Grecu.
Dimensions of Learning in Agent Based Design, 4th International Conference on AI
in Design
-

Workshop on Machine Learning in Design, Stanford, CA, June pp. 24
-
27, (1996).

[8]

CATIA, IBM: http://www.catia.ibm.com/catmain.html

[9]

dV/Mock/Up: http:// www.ptc.com/pr
oducts/

/division/mockup.htm

[10]

Fejo, B., R. Gomes, J. Bento et al. Distributed agents supporting design processes // Artificial Intelligence
in Design ’98, Kluwer Academic Publishers, pp. 557
-
577, (1998).

[11]

Hale, M., J. Craig. Preliminary Development of Ag
ent Technologies for a Design Integration Framework,
Proceedings of 5
th

Symposium on Multi
-
disciplinary Analysis and Optimisation, Panama City, FL, USA,
September 7
-
9, (1994).

[12]

Hyacinth, S., H. Nwana. ZEUS: An Advanced Tool
-
Kit for Engineering Distributed
Multi
-
Agent Systems,
Proceedings of PAAM'98, London, March, pp. 377
-
392, (1998).

[13]

I
-
DEAS:http://www.sdrc.com/nav/software
-
services/ideas/

[14]

Knowledge based engineering (KBE) at Airbus. British Aerospace Airbus
http://www.bae.co.uk/static/engistor.htm

[15]

Liening, A., Blount G. Influences of KBE on the aircraft brake industry, Aerospace engineering and
Aerospace Technology Vol. 70, Number 6, pp. 439
-
444,(1998).

[16]

Pan, Z., M. Zhang, J. Shi. Cooperation Manageme
nt for Collaborative CAD systems, Proceedings for CAD
and Graphics’97, Fifth International Conference on CAD and CG December 2
-
5, Shenzhen, China, pp.
622
-
627, (1997).

[17]

Sehdev, K., I. Fan, S. Cooper, G. Williams. Design for Manufacture in the Aerospace Ext
ended Enterprise,
World Class Design to Manufacture Vol. 2, Number 2, pp. 28
-
33, (1998).

[18]

Subbu, R., C. Hocaoglu and A. Sanderson. A Virtual Design Environment using Evolutionary Agents,
Proceedings of the 1998 IEEE International Conference on Robotics &
Automation, Belgium, pp. 247
-
253,
(1998).

[19]

Tarassov, V., L. Kashuba. Concurrent Engineering and AI Methogologies: Opening New Frontiers,
Concurrent Engineering Europe’97, Bulding Tomorrows’s Virtual Enterprise, April 16
-
18, Erlangen
-
Nuremberg, Germany,SCS I
nternational Publication, pp. 869
-
888, (1997).

[20]

Taratoukhine, V., K. Bechkoum. Towards a Consistent Distributed Design: A Multi
-
Agent Approach, Proc
Information Visualisation '99, London, IEEE Press,(1999).

[21]

Unan R., and E. Dean. Elements of Designing for C
ost, presented at
The AIAA 1992 Aerospace Design
Conference
, Irvine CA, 3
-
6 February, AIAA
-
92
-
1057, (1992).

[22]

Wallace, G., P. Sackett. Integrated design for low production volume, large, complex products // Integrated
Manufacturing Systems 7/3, 5
-
16, (1996)
.



Corresponding e
-
mail: vtaratoukhine@dmu.ac.uk