Digital library of construction informatics
and information technology in civil engineering and construction


Paper lc3-2017-227:
Relevant Properties and Relationships for Construction Defect Data Mining on Building Information Models

Facilitated by the SciX project

Joyce Delatorre and Eduardo T. Santos

Relevant Properties and Relationships for Construction Defect Data Mining on Building Information Models

Abstract: A Building Information Model can be considered a construction data source which contains, among others, information about geometry and topology of construction elements. These information-rich structures can be a particularly good source for data mining systems on their goal of discovering hidden patterns if these are related to the size, shape and/or position of construction elements. That kind of pattern may be useful in execution quality control and productivity, maintenance, and Post Occupancy Evaluation, among other analyses. However, most of this information is in implicit form (e.g., which elements are in front of a given window) and need to be extracted from the model to be used in data mining systems.Data mining systems need certain properties from construction components as well as to identify some relationships between objects to be able to identify execution problem patterns related to the geometry and topology of these components. This work aims to determine which are the main properties and relationships needed for performing data mining from BIM models augmented with construction defect data. Execution quality data and after-sales maintenance records from two companies, comprising 26 projects, were collected and analysed. The identified problems were grouped in five classes. From this analysis, the relevant properties and relationships were proposed. This information can be used as basis for the development of a BIM data mining platform able to identify patterns in construction defects.

Keywords: BIM, Building Information Model, Data Mining, Quality Control, Construction Defect

DOI: https://doi.org/10.24928/JC3-2017/0227

Full text: content.pdf (562,267 bytes) (available to registered users only)

Series: jc3:2017 (browse)
Similar papers:
Sound: N/A.


hosted by University of Ljubljana



© itc.scix.net 2003
Home page of this database login Powered by SciX Open Publishing Services 1.002 February 16, 2003