Product data exchange using STEP
C. Demartini, S; Rivoira, A. Valenzano
CENS and Dip. di Automatica e lnformatica, Politecnico
c.so Duca degli Abruzzi, 24-10129 Torino (ltaly)
tel. +39-11-5647016, fax. +39-11-5647099
Abstract
Product data technology and the related international standards have obtained
increasing interest both in the academic and industrial environments in the last
decade. Today several software tools are available to exchange and share product
data information according to the ISO 10303 specification. This paper presents a
case study for transforming the information model and data currently used in an
industrial organisation (legacy system) into a subset of the Configuration
Controlled Design (AP203) data scheme defined in STEP. This mapping is
important because it allows the user to follow a graceful migration path to the
STEP technology and gives herlhim the ability to exchange product information in
a standard form, while preserving data values already present in different formats
inside the legacy system.
Keywords
Product data technology (POT), STEP, Express language, product data modelling
1
INTRODUCTION
The exchange and sharing of product data between different enterprises or inside
the same organisation is becoming a primary need in a nurober of production
environments all over the world. In particular, the evolution of the markets into
new global seenarios is rapidly changing the relationships in the supplier-producercustomer chain, and new standard technologies are demanded to exchange product
information between partners at different stages of the production cycle, beginning
with the conception and design of a product right up to the moment when it has to
be disposed of.
G. Jacucci et al. (eds.), Globalization of Manufacturing in the Digital Communications
Era of the 21st Century
© Springer Science+Business Media New York 1998
258
The ISO 10303 standard collection (ISO 10303-1, 1995), better known as STEP,
is the most important standardisation effort in the last decade aimed at defining
rules and methods to manage product model data. Besides defining structures and
relationships in the application data models, ISO 10303 also includes the
specification of the formal methods used to describe the models themselves such
as, in particular, the EXPRESS data description language (ISO 10303-11, 1995)
that is extensively used to describe several parts of the standard documents.
STEP maintains a clear distinction between the conceptual model of data, its
physical implementations and the different views that applications have of the
.91-odel itself. This logical separation is enforced by the standard documents
organised in several autonomaus classes identified by a suitable numbering
scheme. In particular the conceptual model is modular in nature and is built up of
an (expandable) set of constructs called integrated resources that are used to build
the application views of the data.
Application views are called application protocols (APs) in ISO 10303
terminology and consist of formal models that specify how application
requirements, for the activities within the scope of the protocol, are satisfied by
means of the integrated resources. Application protocols are designed according to
the needs of specific fields: so, even though a relative small number of APs was
included in the initial ISO 10303 release, several new APs are currently being
developed so they can be added to the standard and involve, for instance, different
application areas in the automotive, ship building, avionics, electronic and
electrotechical industrial scenarios.
The physical views of the conceptual model are obtained by means of the socalled implementationforms (IFs). An IF specifies the mapping between a formal
description of the data model (that is a specification in EXPRESS language) and a
particular physical format for exchanging and/or sharing data items. In this paper
we are particularly interested in the "part 21" (ISO 10303-21, 1995)
implementation form, since it represents the most common way to exchange STEP
clear text data files between heterogeneaus systems.
Two different STEP implementations are able to exchange information if and
only if they adopt the same data model(s), or, in other words, if they comply with
the same application protocol(s). This requirement is more critical than reaching an
agreement on the data transfer syntax. In fact, very early on the existing legacy
data have to be converted and organised according to ISO 10303 standard models,
that specify uniquely the meaning of the various data entities and their
relationships.
Even though there are already good standard specifications for product data
modelling, the user is faced with the problern of how he/she can migrate to the
product data technology (PDT) architecture without wasting the significant
investments available in proprietary hardware and software and the existing
(valuable!) data already present in his/her own legacy systems.
This paper presents a case study on how this graceful migration can be achieved.
By modeHing the legacy information system with the formal techniques of the ISO
10303 standard, it is possible to map the entities that are present in the proprietary
data structure onto a coherent subset of an ISO 10303 application protocol. Once
259
the mapping has been identified, proprietary data can be converted according to the
standard model rules so that the user can benefit from all the advantages offered by
the STEP architecture.
SOURCB DATA MODBL
DESTINATION DATA MODEL
D
SOURCBDATA
MAPPING PROGRAM
DESTINATION DATA
Figure 1 Mapping of product data.
Transformation of product data is conveniently supported by mapping tools that
operate on EXPRESS representations of two different data models as shown in
Figure 1. The mapping specification describes how the entities in the source model
have to be transformed in order to obtain the entities in the destination model.
Data models are represented in standard EXPRESS language, while the mapping
specification is usually written using such languages as EXPRESS-M (Bailey,
1993) or EXPRESS-e (ISO PISA, 1994). These languages extend the data
definition language EXPRESS with constructs to model processes, events,
activities, transactions, build up execution models and so on. Then, the mapping
program resulting from the compilation of the mapping specification makes it
possible to read a STEP physical file of data which conforms to the source model
and transform it into a STEP physical file of data which conforms to the
destination model.
The paper is structured as follows: section 2 describes the model of the legacy
system developed using the same formal methods adopted in STEP. Section 3
deals with the subset of the application protocol AP203, which is relevant in order
to perform the mapping between the proprietary and the standard data models. For
the sake of simplicity we will present a simplified version of the model which is
sufficient to understand our work without boring the reader with a Iot of
unnecessary and, to some extent, confusing details.
260
2
LEGACY SYSTEM MODEL
A preliminary step in our project was the development of a model of the legacy
system according to the formalism and conventions adopted in the ISO 10303
standard. This aspect should not be underestimated since usually the structure of
data and procedures used in an industrial Organisation depends heavily on the
evolution of the different departments and plants of the organisation itself.
It should not be surprising, therefore, that information concerning a single
product be kept in a number of different places (archives), managed by different
(incompatible) systems and that often there is not even a complete and precise
description of the meaning of the data items, their location and their
consumers/producers !
A detailed and punctual analysis of the activities within the organisation, carried
out with the fundamental contribution of experts from different departments (i.e.
design, engineering, production, management departments) is then necessary to
develop a data model such as the one depicted in Figure 1. The reader is reminded
that, for the sake of simplicity, we will introduce in this paper only that subset of
the legacy system model, which is useful to illustrate the results of our work.
The (part of the) model in Figure 1 is specified in EXPRESS-G and, according to
that formalism, it consists of five entity objects and of their attributes and
relationships.
In particular, the key point here is that the central role in the whole product data
organisation is played by drawings and dash-number (part) elements, while the
standard models developed for STEP, such as the application protocols AP203
(ISO 10303-203, 1995) and AP214 (ISO 10303-214, 1997), are based on parts and
assemblies.
The mapping between the systems has then to take into account these different
points ofview, while trying to align the entities in the two models.
The abstract component entity in Figure 1 is used to describe product
components independently of their origin. A component, in fact, is either a
standard_part or a dash_number object. Standard parts are used to take into
account those elements, which are needed to assemble a product, but are not
produced inside the organisation. A standard part can be simply a screw, a bolt or a
more complex object (such as a pre-assembled mechanical subsystem) but in the
legacy system it can be considered as a basic element described by a textual
specification.
261
1,1(1,1)
1,1,4.......
-
.. .
セ@
Figure 2 EXPRESS-G specification of the legacy system model.
262
On the other hand, a dash_number is a product part designed and produced inside
the organisation, it is given a suitable identification number and is associated with
adrawing.
Both standard parts and dash numbers have common attributes such as a
description and a part. type that are taken into account in the definition of the
abstract component entity. Instead, a particular situation occurs when the
item_number attribute is considered; this kind of information, in fact, is not present
in this form in the legacy system and must be obtained by concatenating the
drawing_number and the dash number elements, so as to build up a unique
component identifier.
A complex product is obtained by assembling several parts together. The
connection between an assembled part and a component is modelled by the
structure entity, which represents an assembly relationship. A structure has two
main attributes to link an assembly to a component: while the former can never be
a standard part (standard parts are considered "elementary" components
independently of their complexity), the latter is necessarily a dash number element.
Note that the inverse is not necessarily true: a dash number can be either an
assembly or an elementary part.
In this way a product corresponds to a structured organisation of structures: if N
composite components are needed to build up a product, N-1 structure entities
must be used to specify the inter-component connection relationships. Structures
can be used "in parallel", that is, to specify assembly relationships among
components at the same Ievel, or hierarchically to connect an assembled subsystem
to a higher Ievel element.
A structure also contains an effectivity attribute used inside the target
Organisation to define a kind of Iimit to the validity of an assembly or of a change
to the usual structure of a product.
As mentioned above, a component is associated to a drawing. In the legacy
system model each drawing can in turn represent several dash-numbers: this is
outlined in Figure 1 by the inverse attribute represents that consists of a set of one
or more dash-numbers. Obviously each dash-number is associated to one and only
one drawing.
In addition each drawing has other important attributes: besides a unique number
and a textual description, in fact, this entity is assigned a security classification
field and a project code.
The legacy system model depicted in Figure 1 corresponds to a textual
EXPRESS specification automatically processed by the software tools used to
check the correctness of the model itself, or to verify that a physical file encoded
according to the ISO 10303 standard rules (that is a "part 21" file in our case) be
conformant to the STEP specifications.
Figure 2 shows a visual representation of some data items described according to
our Express legacy system model and obtained by means of the Karlsruhe
University ECCO toolkit (Staub, 1997).
263
l
。ウ」ゥエNセ、イキョァ@
111
IIIMIING:
llolg_-:·JS714125'
I
113
DIIAVINB:
ッセァ⦅NMイZGjUWQR@
projocLCado :'1134'
projact.,..COda:'D:M'
ait._cadl:41
IKU"'ity..cla•ificl\tan ...
ucuri\y_clanifh:at.hln ...
dncription:'ALA CENT Y ...
ウゥエ・N」ッ、セZMTQ@
dncrip'Lion:'ALA CENf Y .•.
Figure 3 Graphie representation of legacy data entities.
Bach box in the picture represents a different entity: in particular the entities
numbered 10, 15, 17 and 14 are abstract components represented by dash number
objects (entities 10, 15, 14) or by standard parts (entity 17).
The structure entities marked 12, 18 and 16 specify the assembly relationships
for components, thus, for instance, component 15 is the assembly for both
structures 12 and 18, but it also represents a constituent part for the assembly
described by structure 16. Finally drawing entities can be associated to a single
component (entity 11) or to multiple parts (entity 13).
The physical text file containing the same data instances of Figure 2 has
obviously a more concise form, suitable for data exchanges. The file structure, in
fact, conforms to the ISO 10303-21 specifications that are used to get a generic
data model written in EXPRESS to correspond with a clear text encoding of the
data item values. Instaoces are numbered progressively to distinguish them
uniquely, while their description contains the name of the entity type and the
ordered Iist of the attribute values.
3
CONFIGURATION CONTROLLED DESIGN MODEL SUBSET
The AP203 application reference model (ARM) describes the structure of a
product in terms of parts. A part entity is an element produced or employed in a
production process and the application protocol defines attributes to describe its
type (detail, assembly, customer supplied material etc.), name, version, whether
the part is a "standard" item or not and so on. Then, for our purpose, each part
instance can be made to correspond with a single drawing-dash number pair.
However, this mapping is to some extent complicated by the fact that each ARM
entity is built up by a network of integrated resources constrained to satisfy the
264
information requirements of the target application (in our case the "configuration
controlled design" application requirements). Figures 4 and 5 show the EXPRESSG representation of some of the integrated generic resources (ISO I 0303-41, 1995)
used in the AP203 application interpreted model (AlM) to describe the ARM part
entity. In practice, each part is mapped onto a product resource, while its attributes
correspond to attributes of either a product or product_related_product_category
element The latter is a subtype of the product_category entity which is used, for
instance, to describe classes of products such as mechanical parts, electrical parts
and so on and takes into account the part classification.
Each part (product) can be associated to several product_definitions used to
identify the part itself in different application contexts (i.e. physical design,
functional design, engineering etc.) that are grouped by means of a
product_definitionJormation entity.
Assernblies are obtained by means of product_definition:relationship entities that
are used for specifying associations between couples of components (parts). A
product definition can have one or more associated documents and this is enabled
by the product_definition_with_associated_documents subtype shown in Figure 5.
Documents are used in our mapping to transform the drawing entities in the
legacy system: in fact the drawing attributes and their meanings correspond quite
weil with the document attributes and can be translated by means of simple string
operations.
The effect of the transformation of the data instances is evident when Figure 3
and 6 are compared. As it is easy to verify, each assembly (structure) in Figure 3
corresponds to a product_definition_relationship entity in Figure 6, while each
component (dash_number) is mapped onto a different product element by means
of a product_definition_formation entity.
In the legacy model each structure is assigned its own effectivity (which is,
roughly speaking, a sort of range of validity for the associated assembly).
This information is maintained in the AP203 model where the effectivity entity is
linked to the product definition relationship. lt should be noted, however, that in
general the structure of the effectivity is more articulated and complex in the
AP203 schema.
4
CONCLUSIONS
This paper has presented a case study for transforming product data instances
stored inside a (proprietary) legacy system into equivalent information structured
according the STEP AP203 data model. This mapping is particularly important
when product data technologies based on the ISO 10303 international standard
have to be introduced in an organisation and a smooth migration path must be
ensured for such innovating approaches. The adoption of STEP, in fact, has a
strong influence on the architecture of the system software used for storing,
exchanging and sharing product data but also involves radical changes in the
procedures adopted to generate, organise and manage the data themselves.
265
In several situations the preservation of the valuable information already present
in a production environment is a mandatory issue and this can be achieved only by
carefully analysing the legacy data model and by getting it to correspond with one
or more application protocol data schemas .
...........
I
(
-
-
1
..................
,, ..
1......
I
..."..........
"......_...,.
(
----1·-----'
r
.. ,............
....
) (
I
:--"'"-"'
. ,,, ...........
I
----
I I ...__ I
)
----------
Figure 4 EXPRESS-G specification of the AP203 subset model (part 1).
266
u .........
...
1,1,1dtnttPIM
2.1,ta1
J,1,l4entlßllr
)
____)
•• _.. _.....
(...__ _
セ@
セ@
PiOiüöiJiilijJpo
)
....... :
Figure 5 EXPRESS-G specification of the AP203 subset model (part 2).
267
...
Figure 6 Graphie representation of data entities in the AP203 subset.
268
5
ACKNOWLEDGEMENTS
The authors are grateful toB. Chiavola, L. Garda, P. Verzino from Alenia for the
useful discussions and comments about the enterprise data models.
This work was partially supported by UNINFO and Camera di Commercio di
Torino.
6
REFERENCES
Bailey I. D., Mead M. (1993) EXPRESS-M- A Schema Mapping Language, Proc.
3ro EXPRESS User Group Conference, Berlin, 1993.
ISO 10303-1 (1995) Industrial Automation Systemsand Integration- Product Data
Representation and Exchange- Part. 1: Overview and Fundamental Principles,
International Standard 10303-1, 1995.
ISO 10303-11 (1995) Industrial Automation Systems and Integration - Product
Data Representation and Exchange - Part 11: Description Methods: The
EXPRESS Language Reference Manual, International Standard 10303-11,
1995.
ISO 10303-21 (1995) Industrial Automation Systems and Integration - Product
Data Representation and Exchange- Part 21: Implementation Forms: Clear
Text Encoding of the Exchange Structure, International Standard 10303-21,
1995.
ISO 10303-41 (1995) Industrial Automation Systems and Integration - Product
Data Representation and Exchange - Part. 41: Integrated Generic Resources:
Fundamentals of Product Description and Support, International Standard
10303-41,1995.
ISO 10303-203 (1995) Industrial Automation Systems and Integration - Product
Data Representation and Exchange - Part. 203: Application Protocols:
Configuration Controlled Design, International Standard 10303-203, 1995.
ISO 10303-214 (1997) Industrial Automation Systems and Integration - Product
Data Representation and Exchange - Part. 214: Application Protocols: Core
Data for Automotive Mechanical Design Processes, Draft International
Standard 10303-214, 1997.
ISO PISA (1994) Industrial Automation Systems and Integration - Product Data
representation and Exchange - PISA Information Modelling Language:
EXPRESS-C", TC184/SC4/WG5/WD N202, 1994.
Staub G., Maier M., Grabowsky H. (1997) The ECCO Toolkit User Reference
Manual, Karlsruhe University, Germany, 1997.
269
7
BIOGRAPHIES
Silvano Rivoira graduated in Electronic Engineering at the Politecnico di Torino,
he was associate professor of Computer Science at the University of Torino and
full professor at the University of Perugia. Since 1991 he has been full professor of
Computer Systems at the Politecnico di Torino, where he is director of the
Dipartimento di Automatica e lnformatica.
His research activity has concemed speech recognition, operating systems and
multiprocessor architectures. Currently his main interests are in product data
technology.
Claudio Demartini graduated in Electronic Engineering at the Politecnico di
Torino, he is currently full professor in the same institution, where he teaches
Computer Engineering and Database fundamentals. His interests include
communication protocols, industrial communication networks and database
principles. He is currently involved in international and national activities in the
area of industrial communications and product data technology.
Adriano Valenzano graduated in Electronic Engineering at the Polytechnic of
Turin. He is currently Director of Research at Centro Elaborazione Numerale dei
Segnali (CENS), Polytechnic of Turin where he is responsible for researches about
distributed systems, local area networks and communication protocols.
His current research interests are in the fields of computer networks and
communication protocols (in particular, industrial communications), formal
methods for protocol engineering and product data technologies.