ONTOLOGICAL APPROACH FOR THE CONFORMITY

CHECKING MODELING IN CONSTRUCTION

Catherine Faron-Zucker, Nhan Le Thanh

I3S, Université de Nice Sophia-Antipolis, CNRS, 930 route des Colles, BP 145, 06903 Sophia Antipolis, France

Anastasiya Yurchyshyna*, Alain Zarli

CSTB, Centre Scientifique et Technique du Bâtiment, 290 route des Lucioles, BP 209, 06904 Sophia Antipolis, France

* I3S, Université de Nice Sophia-Antipolis, CNRS

Keywords: Conformity checking, ontologies in construction, Semantic Web in Construction, graph validation,

knowledge extraction in construction.

Abstract: This paper presents an ontological method and a corresponding system C3R aimed to semi-automatically

check the conformity of a construction project represented by RDF graph against a set of construction

norms formalized as SPARQL queries. Our conformity-checking model has a two-level structure: the

reasoning relies on matching of these queries and graphs and on expert rules guiding the checking process

itself by optimal scheduling of matching procedures, according to semantic annotations of conformity

queries, which integrate the meta-knowledge on the checking process (formalized with the help of CSTB

experts). The reasoning results with a detailed (non)conformity report in terms of the Construction domain.

1 INTRODUCTION

The Construction industry is a major user of

increasingly complex technical regulations defining

the execution of construction projects (e.g. public

buildings). To reply the increasing demand of the

implementation of electronic regulation services,

multiple researches are recently conducted: e.g. the

ePOWER and the ISTforCE project, where CSTB

has participated (IST eGovernment projects, 2007),

aimed at the (semi)automation of the conformity

checking process of a construction project against a

set of technical norms. Our research work answers

this initiative and focuses on the development of the

ontology-enabled, construction-oriented conformity-

checking model.

Today, construction projects are usually

represented in the Industry Foundation Classes (IFC)

model, an object oriented file format that becomes a

standard for Building Information Modelling. The

IFC model captures information about all aspects of

a building throughout its lifecycle and its use is

compulsory for publicly aided building projects. The

IFC model allows an XML representation of a

construction project (ifcXML) that can be

automatically generated by architecture-oriented

CAD tools (e.g. AutoCAD). The IFC model (and the

ifcXML representation) fails, however, to represent

the whole semantic complexity of construction data

used for conformity checking (Yang, Zhang, 2006).

It is not particularly oriented towards the checking

problem either.

For this reason, we propose our knowledge

acquisition method (Faron-Zucker et al, 2008) to

model all the knowledge of the conformity-checking

process in construction. First, we develop an

ontology based on the classes of the IFC model,

dedicated to and oriented by the problem of

conformity checking. Second, in collaboration with

domain experts, we develop a base of semi-formal

representations of conformity queries and propose a

special query annotation of these queries to integrate

meta-knowledge on the checking process. Guided by

the conformity-checking ontology, we then extract a

construction project representation from its initial

ifcXML data that is oriented conformity checking.

Our conformity-checking model is based on the

matching of the norm representations with

representations of construction projects. Its

efficiency relies on two keystones: the ontological

492

Faron-Zucker C., Le Thanh N., Yurchyshyna A. and Zarli A. (2008).

ONTOLOGICAL APPROACH FOR THE CONFORMITY CHECKING MODELING IN CONSTRUCTION.

In Proceedings of the Tenth International Conference on Enter prise Information Systems - AIDSS, pages 492-495

DOI: 10.5220/0001700404920495

 SciTePress

representation of construction regulation knowledge

and the knowledge extraction process guided by the

acquired ontological knowledge. To control the

construction process itself, the model also integrates

meta-knowledge formalized with the help of CSTB

experts: we propose a special annotation of

construction queries and organize them according to

these annotations to schedule matching procedures.

Such semantic annotations are also used to generate

a conformity report and to explain the user the non-

conformity results of the validation process.

2 ACQUISITION OF USEFUL

REPRESENTATION OF

CONSTRUCTION PROJECT

We adopt the ontological approach and the semantic

web technologies (Berners-Lee, 2001) to acquire a

representation of a construction project oriented the

specific task of conformity checking. The success of

the checking process relies on the ability to develop

mechanisms of ontological reasoning: this

representation should be semantically richer than its

initial ifcXML description. Our research is based on

the works of (Bell, Bjorkhaus, 2006) aiming at the

development of a construction ontology

buildingSMART and the projects aiming at the

development of the IFC-to-OWL conversion tool

(Schevers, Drogemuller, 2005).

2.1 Acquisition of a Conformity Query

Base

The first phase of our knowledge acquisition method

aims to explicit formal representations of technical

norms. We use the CD REEF, the electronic

encyclopaedia of construction texts and regulations,

to extract a base of accessibility non-conformity

constraints, which we formalise as SPARQL queries

in terms of the IFC model: e.g. “The minimum width

of a door is 90 cm” is formalized by:

select ?door display xml where

{ ?door rdf:type ifc:IfcDoor

OPTIONAL {?door ifc:overallWidth ?w

FILTER ( xsd:integer(?w) >= 90)}

FILTER (! bound( ?w) )}

This is a manual process (the knowledge

extraction from texts is out of the scope of our

research) conducted in collaboration with CSTB

experts who help to explicit the domain knowledge.

However, these queries contain only conformity

constraints, but have no information useful for the

process of conformity checking: e.g. the information

of the regulation corpus from which the queries are

extracted, etc. To integrate this information into our

checking model, and thus to make it more

“intelligent”, we propose a special RDF annotation

of conformity queries, which contains all the

information related to the checking process not

represented by the query itself. It can be:

- Characteristics of the regulation text from

which a query was extracted: (i) regulation type (e.g.

circular); (ii) thematic (e.g. accessibility); (iii) title,

publication date, references; (iv) level of application

(e.g. national), (v) destination of a building (e.g.

private house); (vi) characteristics of extraction

process: article and paragraph from which a query

was extracted (e.g. 1

paragraph of Door article).

- Formalised expert knowledge: tacit « common

knowledge » on the process of conformity-checking

that is commonly applied by domain experts: (i)

knowledge on (sub)domain of the application of a

query (e.g. Stairs); (ii) knowledge on checking

practice (e.g. if a room is adapted, it is accessible)

- Application context of a query: the aspects of

query application for certain use cases. For example,

the requirements on the maximal height of stairs

handrail vary from 96 cm (for adults) to 76 cm (for

kids). In this case, it is important to know the

destination of a building (e.g. school).

Characteristics and possible values of the first

two groups are automatically extracted from the CD

REEF. The knowledge described by the last two

groups is defined partially and/or has to be explicitly

formalised by domain experts.

2.2 Acquisition of Conformity

Checking Ontology

The second phase is dedicated to the development of

a conformity-checking ontology based on the IFC

model. Guided by the goal of conformity checking,

this ontology includes only the primitive IFC

concepts (extracted from the ifcXML schema)

occurring in the acquired conformity queries. These

concepts are organized as hierarchies and described

in the OWL Lite ontology. If conformity queries

make use of some non-IFC concepts, we integrate

them into the ontology. The intervention of domain

experts is required in this case to define these

concepts with primitive IFC concepts. These

definitions are represented by RDF graphs (e.g.

GroundFloor is a subclass of IfcBuildingStorey

situated on the level of entering into a building).

ONTOLOGICAL APPROACH FOR THE CONFORMITY CHECKING MODELING IN CONSTRUCTION

493

Figure 1: GroundFloor definition by a rule of RDF graphs.

2.3 Acquisition of the RDF

Representation of a Project

The extraction of the RDF representation of a

construction project is conducted by a XSL

transformation of its initial ifcXML representation.

Such transformation is guided by the acquired

conformity-checking ontology. The acquired RDF is

then enriched with non-IFC concepts extracted from

conformity queries (e.g. a project representation is

enriched by GroundFloor concept defined by its

initial IFC-based data: IfcDoor, IfcStair, etc.).

3 CONFORMITY CHECKING

MODEL

Our conformity-checking model is based on the

matching of norm representations to construction

project representations. Such modelling corresponds

to the research on validation of knowledge bases

(Dibie-Barthélemy et al, 2004), constructed

according to the model of conceptual graphs (Sowa,

1984), (Baget, 2005). The results of matching and

the non-validation reasons are interpreted in terms of

conformity checking. To optimise expert reasoning,

we organise the base of conformity queries,

according to their semantic annotations, and define

the optimal scheduling of matching procedures.

3.1 Validation by Projection

The elementary reasoning mechanism of our

conformity-checking model is the matching of a

construction project representation with

representations of conformity queries. In practice,

we interest on the non-conformity condition and the

elements of the project causing its non-conformity

against a query: a construction project is conform to

a query, if there is no projection of the SPARQL

representation of this query into the RDF of the

project. If such projection is found for some

elements, these elements are non-conform. If the

RDF of the project does not contain enough

information “asked” by the query, the projection

cannot be established: no answer on the conformity.

3.2 Organisation of the Base of

Conformity Queries

The organisation of the base of annotated conformity

queries allows defining optimal scheduling of

matching procedures. First, we classify them

according to the regulation texts regrouping the

queries: (i) thematic (e.g. acoustics), (ii) regulation

type (e.g. NF norm); (iii) complex title, publication

date, etc.; (iv) level of application (e.g. European),

(v) destination of a building (e.g. public building).

This classification is generated automatically by

parsing the RDF semantic annotations of conformity

queries. Second, we classify conformity queries on

the basis of specialisation/generalisation relations,

which can be found in their graph patterns. For

example, queries concerning a building (e.g. public

building, three-floor house, school) are in the class

defined by a primitive concept IfcBuilding.

3.3 Formalisation of Expert Reasoning

By organizing the queries, we define the optimal

scheduling of matching procedures as a set of

explicit expert rules, which are defined by priorities

holding between whole classes of queries.

For conformity queries classified according to

regulation texts from which they were extracted, the

scheduling of these queries corresponds to the

order/hierarchy of their classes: e.g. queries

extracted from decrees are prior to circular ones.

For queries classified according to

specialisation/generalisation relations between their

graph patterns, queries representing more specialised

knowledge are treated in priority: e.g. an entrance

door query is prior to a door query (entrance door is

a specialisation of door): the non-conformity of a

project to the first one implies its non-conformity to

the second one.

For queries classified according to their semantic

annotations, priority is given to the queries with

most specific annotations: e.g. to check the

accessibility of a school, we should start by queries

applied to public building receiving sitting public

and continue by more general queries applied to

public building receiving public.

ICEIS 2008 - International Conference on Enterprise Information Systems

494

3.4 Analysis of the Results of

Conformity Checking Process

The results of the validation process (validation/non-

validation, elements causing the non-validation and

its possible reasons) are used to generate a

conformity report, which interprets them in the

terms of conformity checking: “what’s and why’s of

non-conformity”. It lists thus the failed queries: (i)

because of non-matching; (ii) which graph pattern is

more general in comparison to the ones previously

failed; (iii) which annotation is more general in

comparison to the annotation of a failed query.

Moreover, the other possible reason of validation

failure is the case when the representation of the

construction project does not contain enough

information for matching. For such incomplete

representations, it is useful to define the lacking

elements (the pattern sub graphs of the query which

can not be matched) for necessary modifications.

4 C3R SYSTEM

The conformity-checking model we propose is

called C3R that stands for Conformity Checking in

Construction with the help of Reasoning (Fig. 2).

Figure 2: C3R in a nutshell.

The main phases of the C3R approach are as

follows: (1) Acquisition of conformity queries; (2)

Semantic annotation of queries and hierarchical

organisation in query base; (3) Acquisition of

conformity-checking ontology; (4) Acquisition of a

useful annotation of a construction project; (5)

Application of expert rules; (6) Validation by

matching; (7) Generation of the conformity report.

For the checking operation itself, C3R relies on

the semantic search engine CORESE (Corby, Faron-

Zucker, 2007), which answers SPARQL queries

asked against an RDF/OWL Lite knowledge base.

5 CONCLUSIONS

We have presented an ontological approach for the

conformity-checking model in construction. Our

checking model is based on matching of an RDF

representation of a project to a SPARQL conformity

query. We developed a special semantic annotation

of conformity queries that integrates the meta-

knowledge on the checking process. The queries are

organized according to these annotations. This

allows formalising expert rules guiding the checking

process and interpreting the results of checking in

terms of conformity in construction. Ongoing works

focus on the incremental development of the C3R

prototype and its validation by construction experts.

REFERENCES

Baget J-F. RDF Entailment as a Graph Homomorphism, in

Proc. of the 4th international semantic web conference

(ISWC’2005), Galway (EI), LNCS 3729, Springer

Verlag, pp 82-96, 2005

Bell H., Bjorkhaug L., A buildingSMART Ontology. In

Proc. of the European Conference on Product and

Process Modeling (ECPPM-2006), Valencia, Spain,

2006, pp. 185-190

Berners-Lee T. (2001). Reflections on Web Architecture.

Conceptual Graphs and the Semantic Web. Available

at http://www.w3.org/DesignIssues/CG.html

Corby O., Faron-Zucker C., Implementation of SPARQL

Query Language based on Graph Homomorphism, in

Proc. of the 15

International Conference on

Conceptual Structures (ICCS’2007), Sheffield, UK,

LNCS, Springer Verlag

Faron-Zucker C., Yurchyshyna A., Le Thanh N. & Lima C

Une approche ontologique pour automatiser le

contrôle de conformité dans le domaine du bâtiment.

In Actes des 8èmes journées Extraction et Gestion des

Connaissances, EGC'2008, RNTI-E11, Cépaduès, p.

115-120. Sophia Antipolis, France (2008)

IST eGovernment projects, available at

http://ec.europa.eu/information_society/activities/egov

ernment_research/doc/pdf/wkshop05.05.04_rep.pdf

Schevers, H. & Drogemuller, R. Converting IFC data to

the web ontology language, in Proc. of the 1st

International Conference on Semantics, Knowledge

and Grid (SKG2005), Beijing, China

Sowa J.F. Conceptual Structures: Information Processing

in Mind and Machine, Addison-Wesley, 1984

Yang, Q., Zhang, Y. Semantic interoperability in building

design: Methods and tools, Computer-Aided Design

38(10): 1099-1112 (2006)

ONTOLOGICAL APPROACH FOR THE CONFORMITY CHECKING MODELING IN CONSTRUCTION

495