Elsevier

Fire Safety Journal

Volume 43, Issue 7, October 2008, Pages 495-511
Fire Safety Journal

Numerical simulations of full-scale enclosure fires in a small compartment with natural roof ventilation

https://doi.org/10.1016/j.firesaf.2007.12.003Get rights and content

Abstract

Numerical simulation results are presented of full-scale fire tests in a small compartment (3×3.6×2.3 m). A range of total fire heat release rates (330, 440 and 550 kW), fire source areas (0.3×0.3 m and 0.6×0.6 m) and ventilation roof opening areas (1.45×1, 0.75×1 and 0.5×1 m), is covered. Both two-zone model calculations and field model simulations are considered. For the different configurations, profiles of mean temperatures and temperature fluctuations are reported. Furthermore, the mean flow field and temperature field in the compartment are extensively discussed, providing insight in the entrainment and mixing phenomena in the plume in the compartment. The smoke layer depth is determined for all configurations. The computational fluid dynamics (CFD) simulations agree well with experimental observations. The total fire heat release rate value has the strongest influence on the hot smoke layer average temperature rise, while the influence of the fire source area and the roof opening is smaller. The hot smoke layer depth, determined from the mean temperatures, is hardly influenced by the total fire heat release rate. The roof opening also only has a moderate influence in the range considered. The largest impact on the layer depth is due to the fire source area with an increase of the depth as the fire source area increases. Correlations are given for the average hot smoke layer temperature rise and the total smoke mass flow rate out of the compartment, as a function of the different parameters mentioned. A study of the buoyancy reference velocity leads to the verification of a formula to estimate the total smoke mass flow rate out of the compartment. The performance of different entrainment models in zone model calculations is discussed in relation to the obtained CFD results.

Introduction

In [1], an analysis of an experimental study of full-scale fire tests in a small compartment with a roof opening for natural ventilation is presented. Correlations are presented in [1] of the hot smoke layer thickness and average temperature rise as a function of the total fire heat release rate, roof opening area and fire source area. Three manual calculation methods [2], [3], [4], widely used in the design of smoke and heat exhaust ventilation systems, were evaluated. To that purpose, an estimate had to be made of the total smoke mass flow rate out of the compartment.

The major aim of the present study is to provide insight in the flow phenomena inside the compartment, by means of computational fluid dynamics (CFD) simulations. Therefore, numerical simulation results are presented for all the configurations experimentally studied in [1]. These simulations contain by far more detailed information than the experimental temperature measurements. They provide insight into the entrainment mechanism in the plume, so that the entrainment models of the manual calculation methods [2], [3], [4] can be evaluated.

The NIST code fire dynamics simulator (FDS) [5] is applied. First, it is verified that the mean temperatures and hot smoke layer thickness values, obtained from the CFD simulations, agree well with the experimental data, indicating that the simulation results are reliable. Next, the influence of the thermal wall boundary conditions and the grid sensitivity of the results are investigated.

Finally, two-zone model calculation results are also discussed, by means of the program OZONE [6].

Section snippets

Description of the configurations

Fig. 1 shows the compartment geometry as a ‘Smokeview’ picture. A complete description is provided in [1]. The dimensions are 3.6×3.0×2.3 m with a door opening (0.9×2.0 m) in the middle of the front wall. The two openings in the roof are of size 0.75×1 m each (one opening is covered in Fig. 1). They are centrally positioned around x=1.8 m. The distance between the roof opening centers and the front wall is 1 m. The fire source is positioned in the center of the compartment at 0.3 m height. Two fire

General observations

Fig. 2 shows mean temperature profiles reported as the hot smoke temperature rise with respect to ambient conditions (Ts−Tamb). As in [1], averages are taken over the three thermocouple trees behind the burner. All profile shapes are similar, with a plateau of relatively low temperature below the height z=1.2 m (cold lower layer).

Clearly, the same trends are observed in the simulation results (lines) as in the experiments (symbols). In each figure, the temperatures increase as the fire heat

Conclusions

Numerical simulations have been discussed for large-scale fire tests in a small compartment with a door opening and a ventilation opening in the roof.

CFD simulation results, obtained with FDS, have been presented. The following experimental observations of [1] have been confirmed:

  • The hot layer average temperature rise (Ts,av−Tamb) increases almost linearly with the total fire heat release rate Q˙f, because the total smoke mass flow rate out of the compartment m˙s is hardly affected.

  • (Ts,av−Tamb)

Acknowledgments

This research was partly financed by Ghent University—UGent, through BOF Project 011/013/04.

References (12)

  • B. Merci et al.

    Fire Saf. J.

    (2007)
  • B.J. McCaffrey

    Comb. Flame

    (1983)
  • K.B. McGrattan et al.

    Fire Saf. J.

    (1998)
  • NBN S21-208-1, BIN-IBN, Brussels, Belgium,...
  • CR12101-5, CEN Report, Brussels, Belgium,...
  • J.H. Klote et al.

    Principles of Smoke Management

    (2002)
There are more references available in the full text version of this article.

Cited by (43)

  • Temperature evolution and external flame height through the opening of fire compartment: Scale effect on heat/mass transfer and revisited models

    2021, International Journal of Thermal Sciences
    Citation Excerpt :

    Quintiere and Hu et al. [40] characterized the dynamics of compartment fires under poorly ventilated conditions. Merci et al. [41] investigated the temperature profile, hot layer thickness in the full-scale fire compartment with various fire heat release rates, fire source areas and the ventilation roof opening area by the CFD simulation, where both two-zone model calculations and field model simulations are considered. Merci et al. [42] also investigated the accuracy of the FDS for under-ventilated compartment fire with external flaming according for Lee's experimental work [28–30].

  • Influence of heat tightness of an enclosure fire on ignition risk of unburnt gases in a connected exhaust system – An experimental study

    2019, Fire Safety Journal
    Citation Excerpt :

    The quantitative agreement between experimental data and fully predictive simulation is again fair concerning the burning rate of liquid fuel in an enclosure [16]. Only using the mass loss rate collected experimentally, numerical simulation manages to correctly predict the thermal behaviour [17], visibility [18] and pressure [19] for the mechanically ventilated compartments. In order to contain the potential release of radioactive material and avoid dispersion to the outside, nuclear installations are built with increasingly air/heat-tight envelopes [4].

View all citing articles on Scopus
View full text