Pre-Processing IFC Building Models for Code Compliance Checking based on Visual Programming
Abstract
The use of building information models for automatically checking building designs for compliance regarding codes and regulations promises a significant reduction of today’s laborious and error-prone manual processes.... [ view full abstract ]
The use of building information models for automatically checking building designs for compliance regarding codes and regulations promises a significant reduction of today’s laborious and error-prone manual processes. Therefore, the digitalization of regulations forms a key part in many national or international efforts, such as the French Plan de Transition Numérique du Bâtiment. However, the challenges involved are manifold. On the one hand side, the existing rules and regulations have been written for interpretation by humans and do in most cases not allow for a direct translation into machine-readable representation. At the other hand, Building Information Models require a careful preparation to provide information in a manner that is usable for code compliance checking procedures. In this regard, a research project is conducted by the CSTB, which aims at proving the general feasibility of digital rule checking and developing a methodology which works under real-world practical conditions. The methodology allows to capture the domain’s knowledge concerning the interpretation of the regulations and to transfer it into a representation that is processable by computers. It is based on the application of a semi-formal language and the identification of key terminology. The domain expert groups analyze the existing regulation texts and translate it into semi-formal rule definitions consisting of if-then clauses. The project is based on a strong involvement of experts from the different domains and the regulation-setting committees. In many cases key concepts identified in the regulatory texts do not have a direct counterpart in the IFC data schema. Accordingly, a pre-processing step is required to process the original IFC building model and derive or compute the required information and add it to the IFC model. Available tools for perform preprocessing steps as part of the code compliance have major drawbacks: (1) the computing code is hidden and cannot be validated by domain experts, and (2) they are inflexible with respect to adapting the procedures to national requirements or adding new ones. For this reason, the authors propose the application of a visual programming language for defining IFC pre-processing procedures. The language allows domain experts to define the procedures on their own and to check the outcomes of individual processing steps in an interactive way. The paper discusses the features and the syntax of the language in detail and illustrates its application by way of practical examples.
Authors
-
Cornelius Preidel
(Technical University of Munich)
-
Nicolas Bus
(Centre Scientifique et Technique du Bâtiment)
-
André Borrmann
(Technical University of Munich)
-
Bruno Fies
(Centre Scientifique et Technique du Bâtiment)
Topic Area
Building Information Modeling (BIM)
Session
O22 » Intelligent Computing (10:15 - Wednesday, 6th June, Small Auditorium)
Paper
ICCCBE2018_Preidel_final.pdf
Presentation Files
The presenter has not uploaded any presentation files.