Automated energy compliance checking in construction

Zhou, Peng

Permalink

https://hdl.handle.net/2142/102907 Copy

Description

Title

Automated energy compliance checking in construction

Author(s)

Zhou, Peng

Issue Date

2018-12-07

Director of Research (if dissertation) or Advisor (if thesis)

El-Gohary, Nora

Doctoral Committee Chair(s)

El-Gohary, Nora

Committee Member(s)

• El-Rayes, Khaled A.
• Girju, Corina Roxana
• Liu, Liang
• Golparvar-Fard, Mani Y.

Department of Study

Civil & Environmental Eng

Discipline

Civil Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Automated Energy Compliance Checking
• Building Information Modeling
• Semantic Information Alignment
• Natural Language Processing
• Ontology
• Text Classification
• Information Extraction
• Machine Learning
• Energy Codes
• Contract Specifications

Abstract

Automated energy compliance checking aims to automatically check the compliance of a building design – in a building information model (BIM) – with applicable energy requirements. A significant number of efforts in both industry and academia have been undertaken to automate the compliance checking process. Such efforts have achieved various levels of automation, expressivity, representativeness, accuracy, and efficiency. Despite the contributions of these efforts, there are two main gaps in existing automated compliance checking (ACC) efforts. First, existing methods are not fully-automated and/or not generalizable across different types of documents. They require different degrees of manual efforts to extract requirements from text into computer-processable representations, and matching the concept representations of the extracted requirements to those of the BIM. Second, existing methods only focused on code checking. There is still a lack of efforts that address contract specification checking. To address these gaps, this thesis aims to develop a fully-automated ACC method for checking BIM-represented building designs for compliance with energy codes and contract specifications. The research included six primary research tasks: (1) conducting a comprehensive literature review; (2) developing a semantic, domain-specific, machine learning-based text classification method and algorithm for classifying energy regulatory documents (including energy codes) and contract specifications for supporting energy ACC in construction; (3) developing a semantic, natural language processing (NLP)-enabled, rule-based information extraction method and algorithm for automated extraction of energy requirements from energy codes; (4) adapting the information extraction method and algorithm for automated extraction of energy requirements from contract specifications; (5) developing a fully-automated, semantic information alignment method and algorithm for aligning the representations used in the BIMs to the representations used in the energy codes and contract specifications; and (6) implementing the aforementioned methods and algorithms in a fully-automated energy compliance checking prototype, called EnergyACC, and using it in conducting a case study to identify the feasibility and challenges for developing an ACC method that is fully-automated and generalized across different types of regulatory documents. Promising noncompliance detection performance was achieved for both energy code checking (95.7% recall and 85.9% precision) and contract specification checking (100% recall and 86.5% precision).

Graduation Semester

2018-12

Type of Resource

text

Permalink

http://hdl.handle.net/2142/102907

Copyright and License Information

Copyright 2018 Peng Zhou

Owning Collections