ACCORD

Event name and speaker

7 September 2023

Formalisation of building codes and regulations in knowledge graphs

Gonçal Costa

2024.04.19

​

Gonçal Costa, Edlira Vakaj, Thomas Beach, Rita Lavikka, Maxime Lefrançois, Antoine Zimmermann, Thamer Mecharnia,

Vladimir Alexiev, Amna Dridi, Hansi Hettiarachchi, Nataliya Keberle

Goncal Costa - Digital Building Permit conference 2024

9 April 2024

Current situation in the building permit application process

different interpretations,

depending on their context and purpose, and therefore a

difficulty in their application

Building codes and regulations

Complex language and technical vocabulary

The use of BIM and digital checking tools

Have been paving the way for compliance checking to become more automated and less dependent on interpretation

However, none of current methods have proven close to being a definitive “automation” solution

More research is necessary

How can we help facilitate automation?

Event name and speaker

9 April 2024

How can we help facilitate automation?

Most regulations are in PDF

Main ingredients of our approach

NLP

Event name and speaker

9 April 2024

AEC3PO Knowledge

Graphs

Selection of a regulation

Formalization in clause forms

Regulation documents in PDF

NLP process

Pre-Processing

Information Extraction

Rule-Generation

Conversion to data structured Format

bSDD Mapping

Tagging / Annotation

via RASE method

Expression Authoring

Manual process

Rule Formalisation Process

1. Content extraction and conversion to structured data format

4. Mapping words to execution context

2. Identify clauses and formalise logical relationships

3. Rule formalization

Rule formalization process

Event name and speaker

9 April 2024

Extract from the Finnish regulations on the operational safety of buildings

Event name and speaker

9 April 2024

Event name and speaker

9 April 2024

Results

• We have generated machine-readable documents from selected regulations of the ACCORD project pilot cases.
• Non-English regulation has been translated into English to have a common linguistic base to simplify the complexities in a first development version of the process.
• From regulations, a set of formalised regulatory documents has been generated in an open, machine-readable format (YAML-DL/RDF).

Conclusions and next steps

• Semantic Web / NLP 🡪 has proven to be useful for solutions in the field of building compliance checking. However 🡪 how to provide accurate and complete regulation data, the complexity of modelling compliance requirements, and the difficulty of interpreting natural language texts still require more research.
• Developing 🡪 Web tool to graphically carry out the Rule Formalization Process: Rule Formalization Tool.
• Among the issues to be addressed is finding an appropriate method for transforming the content provided in images and tables into structured data.

​

Gonçal Costa

Edlira Vakaj

Thomas Beach

Rita Lavikka

Maxime Lefrançois

Antoine Zimmermann

Thamer Mecharnia

Vladimir Alexiev

Amna Dridi

Hansi Hettiarachchi

Nataliya Keberle

​

Follow us

Access our website

This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement no. 101056973

Thank you!

UK Participants in Horizon Europe Project ACCORD are supported by UKRI grant numbers 10040207 (Cardiff University), 10038999 (Birmingham City University and 10049977 (Building Smart International)

This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement no. 101056973

Partners