Towards data-driven sustainable design: decision support based on knowledge discovery in disparate building data

Abstract

Sustainable building design requires an interplay between multidisciplinary input and fulfilment of diverse criteria to align into one high-performing whole. BIM has already brought a profound change in that direction, by allowing execution of efficient collaborative workflows. However, design decision-making still relies heavily on rules of thumb and previous experiences, and not on sound evidence. To improve the design process and effectively build towards a sustainable future, we need to rely on the multiplicity of data available from our existing building stock. The objective of this research is, therefore, to transform existing data, discover new knowledge and inform future design decision-making in an evidence-based manner. This article looks specifically into this task by (1) outlining and distinguishing between the diverse building data sources and types, (2) indicating how the data can be analysed, (3) demonstrating how the discovered knowledge can be implemented in a semantic integration layer and (4) how it can be brought back to design professionals through the design aids they use. We, therefore, propose a performance-oriented design decision support system, relying on BIM, data mining and semantic data modelling, thereby allowing customised information retrieval according to a defined goal.

Semantic Enrichment of Association Rules Discovered in Operational Building Data for Reuse of Building Performance Patterns

“…Included in this context is also the direct mining of formalized knowledge through the use of novel semantic data mining methods. In this article, we rely on results from a previously proposed method for combination of knowledge discovery (motif discovery and Association Rule Mining (ARM)) (Agrawal et al 1993) in operational building data and semantic data modelling for knowledge representation of a performance enriched semantic building graph (Petrova et al 2018;Petrova et al 2019). Association rules indicate to what extent certain events (patterns) are related to, or are potentially caused by other events (patterns).…”

“…Advanced knowledge discovery methods aid the extraction of high-level knowledge from low-level data (Fayyad et al 1996). Such knowledge allows higher level analyses and has the potential to redefine the way buildings are designed by serving as an evidence base in performance-oriented design decision-making (Petrova et al 2019).…”

“…In depth research has been performed to identify how to transform raw data into building performance insights and make use of the multiplicity of collected, but usually rarely reused data. Included here are efforts aiming to define the various building data types (Petrova et al 2019), as well as corresponding machine learning methods for data pre-processing, mining, visualization and use of discovered knowledge (Fan et al 2015;Fan et al 2018a;Fan et al 2018b;Miller et al 2018;D'Oca et al 2018;Fan et al 2019). Research has also shown that publishing data effectively, breaking up information silos, integration of data across domains and making data readable and understandable by both machines and humans is equally important.…”

BIM for sustainable project delivery: review paper and future development areas

“…This raised due to the fact that the design of projects in practice is a complex process and includes multiinterdepend decisions that can't be simplifies into linear one as the tools deal with a certain problem. It can be described as a series of cycles where diagnosing, planning, taking action and evaluating the action take place by iteratively analysing a problem to reach a suitable solution (Petrova et al, 2018) (Bueno, Pereira, & Fabricio, 2018). Designer and engineer go through iterative, nonlinear, non-single model and multiple source of information and analysis to optimize decision making.…”

Clustering of designers based on building information modeling event logs

“…According to a survey on a website (Lowe, 2019), BIM modeling software is found to approximately decrease 61% errors and 55% times, and enhance 52% quality of a design project. Due to the abovementioned merits, BIM‐based collaborative modeling has increasingly attracted attention in the recent 5 years for more sustainable design (Petrova, Pauwels, Svidt, & Jensen, 2019). Furthermore, the advanced modeling software will automatically save details of the whole design process as logs in journal files.…”