Edinburgh Research Explorer

The challenge of scaling-up wood crib fire experiments to travelling fires in large compartments

Research output: Other contribution

Related Edinburgh Organisations

Original languageEnglish
TypeUKCTRF Annual Meeting
Media of outputElectronic
Publication statusUnpublished - 6 Aug 2019

Abstract

“Travelling fires” are important for structural fire engineering design, as they capture elements of the spatial and temporal variation in the thermal boundary condition, albeit it in a highly simplified manner. Characterisation of fire conditions is being studied by means of experiments at a range of scales, spanning single cribs to full-scale compartments [1]. Parallel computational studies, with NIST’s Fire Dynamics Simulator (FDS) [2], are seeking to evaluate the effectiveness of different representations of the fire, from stick-by-stick which exactly match the physical dimensions, to much cruder models which are more practical for large spaces.

The single crib tests were carried out by University of Liège in 2018, using 9 layers of wood sticks (Norway spruce) with cross-sectional dimension 30mm x 35mm, and 120mm stick-to-stick spacing, below a ceiling at height 2.8m (Figure 1). With availability of detailed experimental data, a posteriori simulations were carried out at two levels of model complexity: individual wood sticks and via a simplified ‘chessboard’ representation, having object sizes about 10 times larger (Figure 2). With central ignition, a roughly linear spread converts to the well known “t-squared” fire growth rate, an effect reproduced spatially by both models. However, by enforcing a common heat release profile on each element post-ignition it is found that the evolution of the total heat release rate is a poor match to the experiment, and a progressive reduction of heat release rate per unit area is required to replicate the test (a change which is counter-intuitive but may reflect the conditions in the experiment).

The calibrated ‘chessboard’ model is then applied to a priori modelling of fire spread in a series of three tests in a large compartment of 15 x 9 x 2.8m (Figure 3), with some success. This exercise will ultimately reveal the potential for prediction of travelling fire behaviours in such large-scale experiments, and assist in establishing the links between various design parameters (e.g. ventilation, fire load density, ceiling height, etc.) and their impact on structural response.

[1] X. Dai, S. Welch, and A. Usmani, “A critical review of ‘travelling fire’ scenarios for performance-based structural engineering,” Fire Safety Journal, vol. 91, pp. 568–578, 2017.
[2] K. McGrattan, S. Hostikka, R. McDermott, J. Floyd, C. Weinschenk, and K. Overholt, Fire Dynamics Simulator User’s Guide, Sixth Edit. National Institute of Standards and Technology (NIST), 2017.

ID: 106184143