Forecasting smoke transport during compartment fires using a data assimilation model. (January 2015)

• Record Type: Journal Article
• Authors: Lin, Cheng-Chun; Wang, Liangzhu (Leon)
• Is Part Of: Journal of fire sciences. Volume 33:Number 1(2015:Jan.)
• Page Start: 3
• Page End: 21
• Publication Date: 2015-01
• Subjects: Forecast -- smoke -- compartment fire -- Ensemble Kalman Filter -- CFAST -- OpenDA; Flammable materials -- Periodicals; Combustion gases -- Toxicology -- Periodicals; Fireproofing -- Periodicals; 628.922
• DOI: 10.1177/0734904114548837
• Languages: English
• ISSNs: 0734-9041

Abstract

Forecasting simulation of an unknown compartment fire is challenging and usually accompanied with a large number of uncertainties. As the simulation progresses in time, the forecasted physical conditions such as fire heat release rate, room temperature, and vent airflow rate may sway from reality in a highly dynamic environment. Conventional deterministic fire simulation tools using one set of initial inputs to predict fire smoke transport may not easily generate satisfactory results. In this article, a new application of Ensemble Kalman Filter to forecast smoke dispersion during compartment fires is presented. The model utilizes measurement data from multiple sensors in multi-room compartments and is able to predict the fire heat release rate and smoke dispersions within several minutes. In addition, detailed formulation of the Ensemble Kalman Filter model and three case studies are also discussed in this article. The resulting model can be considered as a prototype forecast simulation system to assist occupant evacuation, firefighting, and smoke extraction in a building fire accident.