Ph. D. Project
Semi-automatic formalisation of a knowledge repository supporting the digital twin for building renovation: application to the IsoBIM project
2022/02/02 - 2025/01/02
the ANR ISOBIM project aims to propose a collaborative process for renovation by external ISOlation based on
Lean and BIM paradigms. In this context, one of the objectives is to be able to ensure the management of
project data in a formal repository capable of covering the design, scheduling, planning and implementation
monitoring phases. In addition, this repository will have to support the digital building twin allowing, in
particular, 4D simulation in the execution phase of a building site.
Building Information Modelling (BIM) can be defined as "a modeling technology and associated set of
processes to produce, communicate and analyse building models" (Sacks, Eastman, Lee, & Teicholz, 2018).
Because of its multidisciplinary nature, the construction field brings together different trades and therefore
different types of data that need to be grouped together within the same modelling (classic data models,
formal knowledge models, such as ontologies, taxonomies, concept hierarchies).
Objectives and issues:
The modelling of information related to the design phase is largely dealt in the BIM. Indeed, beyond the
standards commonly used in mechanical design (STEP (Pratt, 2005), IGES (Association & others, 1996), DXF
(Autodesk, 2010), ...), standards have been defined for the modelling of construction data such as IFC (ISO,
2013) or CIS/2 (Reed, 2003). Within the framework of the IsoBIM project, it is envisaged to integrate 3 business
points of view (design, planning and construction monitoring) and their associated constraints to support the
renovation process. All of IsoBIM's activities require an adequate information system implementing various
functionalities, including efficient clustering algorithms, a data model capable of storing multimodal data:
geometric (panels from the layout possibilities), planning (resources, stocks, ranges, etc.) and site monitoring
(task status, estimated time, person in charge of the task, etc.). This information system could be used to
support the digital twin of the renovation process.
We are therefore interested in the following research questions:
1) Which data structure would be the most suitable for our digital twin?
2) How can this structure be obtained automatically to reduce the modelling time/difficulty?
To answer these two questions, this work will seek to explore the use of data mining tools to propose the
following contributions:
1) A formalization of knowledge in ontological form to structure the concepts of the digital twin of the
renovation process.
2) A semi-automatic methodology to obtain this formalisation, based on clustering algorithms.
This last point will be based on data mining, with two methods already envisaged. Firstly, the Formal Concept
Analysis (FCA) method (Ganter, Stumme, & Wille, 2004) would be used for knowledge modelling based on a
set of examples or structured data. In a second step, its extension Relational Concept Analysis (RCA) (Rouane-
Hacene, Huchard, Napoli, & Valtchev, 2013) (Wajnberg, Valtchev, Lezoche, Panetto, Blondin Massé, 2019)
would be used for the extraction of relationships between each knowledge domain.

Bibliography :
Association, U. S. P. D., & others. (1996). Initial Graphics Exchange Specification IGES 5.3. ANSI. Retrieved July,
12, 2008.
Autodesk. (2010). DXF Reference. Consulté à l'adresse
Ganter, B., Stumme, G., & Wille, R. (2004). Formal Concept Analysis: Theory and applications-J. UCS Special
Issue. Journal of Universal Computer Science, 10(8), 926.
ISO. (2013). Industry Foundation Classes (IFC) for data sharing in the construction and facility management
industries. Consulté à l'adresse
Pratt, M. J. (2005). ISO 10303, the STEP standard for product data exchange, and its PLM capabilities.
International Journal of Product Lifecycle Management, 1(1), 86.
Reed, K. (2003). The role of the CIMSteel integration standards in automating the erection and surveying of
structural steelwork. NIST SPECIAL PUBLICATION SP, 15‑20.
Rouane-Hacene, M., Huchard, M., Napoli, A., & Valtchev, P. (2013). Relational concept analysis: Mining concept
lattices from multi-relational data. Annals of Mathematics and Artificial Intelligence, 67(1), 81‑108.
Sacks, R., Eastman, C., Lee, G., & Teicholz, P. (2018). BIM handbook: a guide to building information modeling
for owners, designers, engineers, contractors, and facility managers. John Wiley & Sons.
Wajnberg, M., Valtchev, P., Lezoche, M., Panetto, H., Blondin Masse, A. (2019). Mining process factor causality
links with multi-relational associations. 10th International Conference on Knowledge Capture, K-CAP'19, Nov
2019, Marina Del Rey, CA, United States. pp.263-266, ⟨10.1145/3360901.3364446⟩. ⟨hal-02377662v2⟩
Data mining, formalization of knowledge, ontology, BIM, DIGITAL TWIN
Modeling and Control of Industrial Systems