A data-driven dynamic fire risk analysis tool based on building information modeling

Date

2024-7-16

Author

Kızılkaya Öksüz, Nilüfer

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

344
views

0
downloads

Cite This

Conventional fire safety inspections and information management practices, which often exclude building information modeling (BIM) integration, limit assessing and documenting dynamic and ever-changing fire risks that inherently occur in the buildings' operation lifecycle. To address this limitation, there is a need for a systematic and ongoing evaluation of fire risks based on changing fire performance metrics. The continuous fire risk assessment necessitates structured spatial and semantic information modeling based on real-time data analytics to manage the changing fire risks. This study develops a data-driven dynamic fire risk assessment tool with a specifically developed five-step methodology to evaluate the fire risk factors during the building operation lifecycle. Initially, a web-based questionnaire targeting fire safety and hotel management professionals was conducted to understand the fire-starting impacts of behavioral and spatial factors in hotel buildings. A data-driven clustering methodology was introduced in the second phase, based on a statistical analysis of questionnaire results and data analytics of the hotel guest dataset. In this phase, a real-world dataset was processed using K-means clustering analysis to structure fire risk level categories of hotel guest rooms. In the third phase, the Industry Foundation Classes (IFC)-based information modeling framework of the dynamic fire risk assessment methodology was constructed based on the hierarchical and related structure of the entities. For fire risk information modeling, the semantic data of a Hotel Building was exported from the BIM tool with the IFC 4.3.2.0 schema. In the next phase, a web application named FireBIMetrics was developed to provide a standard interface for fire safety professionals and building operation managers to upload, process, download, and visualize hotel guest room fire risk assessment information based on IFC files. The web application code development was ensured by the IfcOpenShell Python library, and the Flask was used for the user interface and database integration. The last phase includes testing and validating the methodology using design fire scenarios generated by performance-based fire safety evaluation. During the testing phase, the data-driven dynamic fire risk assessment methodology was verified by evaluating the fire dynamics and evacuation performances of a hotel building using four distinct scenarios, which varied based on fire location and occupant behavior changes. In the validation phase, scenarios involving random and strategically managed guest room assignments were comparatively evaluated for evacuation time and tenability factors. The results revealed a notable reduction in evacuation times and enhanced tenability limits when the room assignments were strategically managed according to the fire risk levels of rooms. This outcome highlights the critical importance of continuously reassessing fire risks in response to changing building conditions and occupant behavior risks. In addition, IFC file processing enables facility managers and fire safety professionals to visualize risks for informed decision-making and to identify potential fire ignition areas in design-fire scenarios. As a result, the proposed fire risk assessment methodology and web application tool significantly enhance the management of dynamic fire risks throughout the operational lifecycle of buildings, contributing to their fire resilience.

Subject Keywords

Building information modeling, Data analytics, Dynamic fire risk assessment, Human behavior in fire, Industry foundation classes

URI

https://hdl.handle.net/11511/110436

Collections

Graduate School of Natural and Applied Sciences, Thesis