Natural Fire Safety Concept

The Development and Validation of a CFD-based Engineering Methodology for Evaluating Thermal Action on Steel and Composite Structures



EC Proposal N : P4169/ Contract N : 7210-PR-184

Project Co-ordinator :

Dr Suresh Kumar, BRE - Watford (GB)

Project Partners :

Building Research Establishment Ltd (GB)

ProfilARBED S.A. (LU)

VTT Building Technology (FIN)


Arbeitsgemeinschaft Brandsicherheit AGB (DE)



Design methods for steel/composite structures are described in the structural Eurocodes, which are in the process of being converted to full EN standards. The methods now include elements relating to the 'natural fire safety concept' (developed under previous ECSC research programmes - NFSC1 and NFSC2) and provision is also made for use of "advanced calculation models" such as "Computational fluid dynamics" (CFD).

The natural fire safety concept provides a more realistic description of the thermal exposure of structures under "natural fire conditions". The current project builds on this approach by exploiting the techniques of computational fluid dynamics (CFD) to develop, verify and apply a comprehensive engineering methodology for predicting the thermal response of steel and composite components.

The main challenge in this application lies in that fact that the component of interest is typically very small relative to that of the whole computational domain. When structured computational meshes are used, as is commonly the case, this means that unacceptably large numbers of cells are required for accurate modelling of the heat transfer to and within structural members.

In order to overcome this problem, a "multiblock" approach can be adopted in which an "embedded" mesh, with a sufficiently fine grid resolution, is used in regions of interest. This model has been implemented within the SOFIE CFD code enabling the solution for the solid-phase heat transfer to be performed simultaneously with the full gas-phase solution of the fire flowfields.

The new solver is accessed by means of new pre-processing windows in the fire engineering interface. Provision is made for the selection and placement of different beam and column sections, together with a variety of methods of application of protection material. Temperature-dependent material properties are used for the materials, so that heat transfer inhibition due to moisture effects can be accommodated.

A progressive model validation exercise is nearing completion. This involves simulation of the thermal response of steel/composite members in fire tests for which experimental data is available - a localised beam fire test, standard fire-resistance furnace tests and full-scale tests involving natural fires.

The effect of a range of relevant numerical and physical models has been examined in some detail. These include the radiation model options (discretization scheme, number of rays and method of incorporating the absorption/emission characteristics of the participating gas-phase media), the effects of soot loading, the simulation of thermocouple temperatures and the influence of the numerical grid. No significant weaknesses were found in the model representations and overall, the predictions are found to be fair and physically plausible.

Ultimately, model predictions will be used to identify the critical design parameters affecting the thermal action on the steel/composite structures and where possible, to determine equivalent values for the key parameters used in the simple methods. Recommendations will be made on improvements to the standard procedures and/or extension of the methodology as necessary.



2.1. Introduction

Various approaches for prediction of thermal actions on steel/composite structures are provided for in the structural Eurocodes, now being converted to final EN standards. The methodologies include provision for the treatment of structures exposed to "natural fires" derived from recently completed ECSC research programmes (NFSC1, NFSC2). The current project is seeking to build on this approach by exploiting the techniques of computational fluid dynamics (CFD) to model the thermal response of structures under natural fire conditions, with a view to the development and extension of the Eurocodes methodologies.

The modelling technique of computational fluid dynamics (CFD) is being extensively used in a range of engineering disciplines for simulation of fluid flow and heat transfer processes. In the past twenty years, CFD has found increasing application in fire modelling, and this has contributed to the establishment of the discipline of fire safety engineering. However, there has so far been no coordinated attempt to exploit the potential of the technique in prediction of thermal behaviour of boundary materials due to the effects of fire in buildings. This project is seeking to apply CFD techniques to develop, validate and apply a comprehensive engineering methodology for prediction of thermal actions on steel and composite structures, and to contribute to the development of the fire-related Eurocodes.

2.2. Objectives

The objective of the project is to develop an engineering methodology, exploiting the advanced capabilities of computational fluid dynamics (CFD), for determining the thermal behaviour of structural elements in steel/composite-framed buildings. Specific objectives of the project are as follows:


3.1. Model development

SOFIE is an ordinary single-block structured code, which means that the entire modelled geometry is represented inside a single three-dimensional curvilinear mesh. This means that the same numerical grid is used for the whole domain and the grid resolution in areas of interest is restricted by the need to maintain reasonable cell aspect ratios (the ratio between the sizes of the largest and smallest cells) and within this constraint, the limit on the overall number of cells arising from computation time considerations. Typically, the structural component of interest has flange and web size of only a few millimeters, and when protection material is applied, or the member is a composite component, an exact model may require the representation of additional detailed geometrical features.

A method of bypassing these problems is to use a multi-block structured mesh, so that each sub-block is discretised using a separate mesh of its own. The overall CFD solution is obtained by the appropriate coupling between meshes.

The coding for the new model is now complete and the fire engineering interface (GUI) for pre- and post-processing of the CFD analysis has been extended to allow selection and orientation of the structural components, including the specification of applied protection materials. Figure 1 illustrates some of the interface windows associated with the section geometry specification and figure 2 shows the model predictions for an idealised test case used for internal verification of the solid-phase model - exposure to a constant temperature environment at 500 oC.

3.2. Model validation

During this period, the emphasis of model verification work (WP2) has been to improve gas-phase predictions of fire characteristics such as gas temperatures, heat fluxes, smoke layer depth and flow velocities in the door opening, and to explain any discrepancies between predictions and measurements. The validation test cases studied are the BRI localised beam fire (WP2.1), VTTs furnace tests (WP2.2), and the VTT large room and BRE large compartment tests (WP2.3). All project partners have attempted to model one or more of the above test cases by the CFD model SOFIE and most have now reported modelling results for all of them.

A fairly detailed analysis of the localised beam fire test case (WP2.1), performed by BRE, was reported previously (c.f. last TRA-EFCT report). The original findings have been reconfirmed by the results reported by other partners. In summary, despite some discrepancies between prediction and experiment, no major weaknesses were found in the model representations and a thorough understanding of the key physical models, in particular the radiation heat transfer, was achieved. By this means, the requirements for radiation model representation of such a localised fire were clarified and insight was gained into the relevant physical phenomena.

Building on this achievement, the main focus of the model validation study moved on to the full-scale compartment test cases, the VTT room and the BRE large compartment (see Figure 3). In these cases, it has been shown that it is important to consider thermocouple "errors" due to radiative transfer. A number of approaches have been used to study this phenomenon, including a simplified model reported in the literature (e.g. Blevins & Pitts, 1999), a simple one-dimensional analysis implemented in a spreadsheet and the existing full "thermocouple simulator" in the SOFIE CFD code, the implementation of which was improved.

The findings show clearly that the thermocouple errors are normally relatively small in the upper hot layer, but that they can be very large in the lower colder layer and also significant in the hot layer in the doorway opening. Figure 4 shows a prediction of the thermocouple temperature distribution in the plane of the doorway centreline in the BRE large compartment.

The difference between the results from within the compartment and the doorway is due to the fact that in the upper part of the compartment the hot layer is relatively deep and also optically thick so that a thermocouple bead is not influenced greatly by temperature gradients or remote radiation. The reported values are therefore representative of the local gas temperatures. The good correlation found between the heat flux measurements and the thermocouple results in the region of the meter supports this finding.

Another aspect of these simulations studied is the effect of the numerical grid. ProfilARBED examined results obtained using 19 different grids, clearly demonstrating that improved grid design is crucial for faster convergence of the CFD solution and improved predictions.

BRE and LABEIN looked at the influence of the soot loading on the temperature within the fire flowfield. Some of the results obtained are shown in figures 5 and 6.

These results show that the soot tends to smooth out the temperature profile within the compartment, whilst yielding slightly higher temperatures in the hot layer. Overall, quite a reasonable agreement is achieved with the experimental thermocouple data.

For the VTT large room, AGB have demonstrated the sensitivity of the predictions to choice of plume entrainment expression.

Finally, for the BRE large compartment, initial simulations were performed of the protected indicative used in the test. A 2D finite-element model was used, with the thermal exposure defined by test data. Very good agreement with experimental values was obtained when the default (constant) thermal properties were adopted for the sprayed fibre protection material. The effect of allowing for the influence of moisture in the protection material was shown to be insignificant in this case where the thermal exposure is very severe.



The new solid-phase solver is now complete and functionality has now been added to the user interface to allow definition of structural members, with and without protection. The model has been internally verified. For the range of test cases chosen, the project partners have now managed to reproduce the qualitative features of the experimental tests. With improved skills in the use of the models, continued improvement in convergence of the model predictions by different partners has been achieved and consequently, better agreement with experimental data generally achieved.



Blevins, L.G. & Pitts, W.M. (1999) "Modelling of bare and aspirated thermocouples in compartment fires", Fire Safety Journal, vol. 33, pp. 239-259