Text and Information Analytics for Fully Automated Energy Code Checking

Energy compliance checking aims to check the compliance of design information embedded in building information models (BIMs) with applicable building energy codes. To fully automate the compliance checking process – without the need to manually code the rules in a compliance checking software, there is a need to automatically process and understand the text in the codes, extract the relevant requirements from the text, and match the extracted requirements to the information in the BIMs. This is challenging because of the text complexities of the codes, including longer provisions, requirement exceptions, hierarchically-complex sentence structures, and different terminologies and levels of detail compared to BIMs. To address these challenges, a set of text and information analytics methods were proposed and implemented in an energy compliance checking prototype. The prototype includes four main modules: text classification, information extraction, semantic information alignment, and compliance checking and reporting. For testing the prototype, a BIM of an educational building was checked for compliance with three energy codes – the 2012 International Energy Conservation Code, the 2013 Building Energy Efficiency Standards (known as the California Energy Code), and the Ontario Building Code Supplementary Standard SB-10. A promising performance of 91.7% recall and 84.6% precision in noncompliance detection was achieved. This study could benefit different participants in the domain of architecture, engineering and construction, by streamlining the energy code compliance process in a fully-automated manner. Future research may test the proposed methods and prototype on other energy topics (e.g., fenestration) and different types of documents (e.g., contract specifications). The challenges for developing a compliance checking method that is fully-automated and generalized across different types of documents could also be studied.