P-Flash – A machine learning-based model for flashover prediction using recovered temperature data
Jun Wang,Wai Cheong Tam,Youwei Jia,Richard Peacock,Paul Reneke,Eugene Yujun Fu,Thomas Cleary
DOI: https://doi.org/10.1016/j.firesaf.2021.103341
IF: 3.78
2021-06-01
Fire Safety Journal
Abstract:<p>Research was conducted to examine the use of Support Vector Regression (SVR) to build a model to forecast the potential occurrence of flashover in a single-floor, multi-room compartment fire. Synthetic temperature data for heat detectors in different rooms were generated, 1000 simulation cases are considered, and a total of 8 million data points are utilized for model development. An operating temperature limitation is placed on heat detectors where they fail at a fixed exposure temperature of 150 ̊C and no longer provide data to more closely follow actual performance. The forecast model P-Flash (<u>P</u>rediction model for <u>Flash</u>over occurrence) is developed to use an array of heat detector temperature data, including in adjacent spaces, to recover temperature data from the room of fire origin and predict potential for flashover. Two special treatments, sequence segmentation and learning from fitting, are proposed to overcome the temperature limitation of heat detectors in real-life fire scenarios and to enhance prediction capabilities to determine if the flashover condition is met even with situations where there is no temperature data from all detectors. Experimental evaluation shows that P-Flash offers reliable prediction. The model performance is approximately 83 % and 81 %, respectively, for current and future flashover occurrence, considering heat detector failure at 150 ̊C. Results demonstrate that P-Flash, a new data-driven model, has potential to provide fire fighters real-time, trustworthy, and actionable information to enhance situational awareness, operational effectiveness, and safety for firefighting.</p>
materials science, multidisciplinary,engineering, civil