Benchmarking and visualization of building portfolios by applying text analytics to maintenance work order logs

Nojedehi, Pedram; O’Brien, William; Gunay, H. Burak

doi:10.1080/23744731.2021.1913957

Cited by 12 publications

(10 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…First, it is proposed that the CMMS databases are taken as input and processed by the methodology previously developed. This way, the textual and other types of data are preprocessed and normalized, and the spatial properties of each WO are detected and extracted from the WO descriptions [103], [106] via a programming language like Python. Then, the processed dataset is used for further text analysis explained in section 3.…”

Section: Methodsmentioning

confidence: 99%

“…Although they tend to be standalone applications and cannot represent buildings' components and relationships in a structured manner, their functionalities can expedite the digital twin development process. CMMS is the conventional method to record and manage maintenance work orders as well as occupants' complaint management [6], [103]. At the same time, EMA and micro-EMA are more recent methods to capture real-time occupants' data in buildings.…”

Section: Methodology For Digital Twin Developmentmentioning

confidence: 99%

“…Although these approaches have their advantages and disadvantages, the CMMS data are recorded in textual form, hence the text analytics, as well as more intuitive and less complex processes/tools are still essential when the researchers intend to develop a new tool for the FM tasks. Apart from our previous paper regarding this [103], McArthur et al…”

mentioning

confidence: 85%

“…For instance, the main classifications are mechanical, electrical, construction, and safety in this study. This methodology was described in detail in our previous papers defining general dictionaries to classify the WOs more broadly with the help of rule-based methods and an algorithm counting and comparing the words from each dictionary [103], [106].…”

Section: Text Analyticsmentioning

confidence: 99%

“…(key performance indicators) are defined to map and visualize these non-geometric attributes if existed [103]. As shown in Figure 3.6, a set of simple nodes, Tool 1, that exchanges the data is built on Dynamo.…”

Section: Data Paths Toward Bimmentioning

confidence: 99%

See 4 more Smart Citations

Towards Smart Building Maintenance: Leveraging Multi-Source Data and Digital Twins with Ecological Momentary Assessment

Nojedehi

View full text Add to dashboard Cite

Complaint management, fault detection and diagnosis (FDD) are critical for building performance and occupants' well-being. Facility management addresses these services with work order logs containing occupant complaints and FDD identifying system faults. This thesis enhances these aspects using existing and new data sources.The first section maximizes existing data (Chapters 2 and 3). Text analytics (Chapter 2) is used to create benchmarks for maintenance performance evaluation and transforms data into visual insights. Additionally, facility management-enabled building information models (FM-BIMs) (Chapter 3) integrate maintenance logs with BIM, enhancing decisionmaking through contextual data exchange. A case study is presented to demonstrate the effectiveness of this methodology.The second section employs novel data collection methods and their application in digital twins (Chapters 4 and 5). Cozie smartwatch app representing micro-ecological assessment gathers occupants' thermal comfort and air quality feedback. This data source is integrated with building system sensor data. FDD rulesets utilize this data for fault detection, validated by occupants' accurate feedback. Surveys revealed that occupants seldom report discomfort unless the temperature deviates by 3°C. This approach reduces reporting thresholds, enhancing issue reporting by 29%.Chapter 5 introduces a digital twin framework for maintenance, integrating diverse building system data and real-time insights. Occupant feedback collection methods are diversified through online surveys and a smartwatch app, enhancing indoor quality assessment. The framework supports two tiers of analysis and user interfaces for facility management, demonstrated via a prototype that leverages existing infrastructure. This iii methodology allows facility management agents to develop digital-twin-enabled workflows bringing multiple real-time data streams together to make more informed decisions.In conclusion, this thesis advances complaint management, FDD, and other maintenance practices by innovating existing and new data sources. The proposed methodologies, from text analytics to digital twin frameworks, empower facility management to enhance building performance and occupant satisfaction proactively. iv PrefaceThis integrated thesis consists of five journal papers, four published and one under review.The articles included in this thesis are as follow:

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodology For Digital Twin Developmentmentioning

confidence: 99%

mentioning

confidence: 85%

Section: Text Analyticsmentioning

confidence: 99%

Section: Data Paths Toward Bimmentioning

confidence: 99%

See 3 more Smart Citations