Many studies of fires in large compartments reveal that the fire load does not burn uniformly throughout the enclosure. The fire tends to travel and create highly non-uniform temperature distributions, which implies a transient heating of the structure. Travelling fires are not considered in the Eurocodes: the main limit in developing models is the lack of large scale, realistic test results. This project aims to perform such tests and numerical simulations to define the conditions in which travelling fires develop, to build an analytical model which evaluates the thermal effect and to create design guidance which improve structural safety.

3 Trafir 1000 

 

This project aims to realize such tests and to perform numerical simulations in order to develop an analytical model. This model would characterize travelling fire, and design guidance would be provided to improve structural safety.

 

1) First of all (WP 1): knowledge is missing concerning the conditions in which a travelling fire develops. Preliminary studies and numerical simulations will be done in order to define the circumstances leading to such fire scenarios (compartment size and geometry, ventilation conditions, volumes, fire load density, rate of heat released…).

2) Small scale and large scale tests will be performed (WP 2 & 3).

Two series of small scale tests will be performed and will consist of simple tests where the fire dynamics are well controlled. A first series of small scale tests will be performed to determine the fire load arrangement that comes as close as possible to values representative of an office building with 250 m² floor area: fire load density of 511 MJ/m², rate of heat release rate density of 250 kW/m² and medium fire growth rate (tα = 300 s). A second series of small scale tests will be performed to determine if there is a shadow factor related to multiple fires travelling along a certain axis, i.e. if the depth of the burning area has an impact on the heat fluxes received by a structural element aligned to the propagation axis.

The large scale test data represent valuable information since very few large scale tests were realized up to now. Large scale tests will be performed in real building dimensions using a uniformly distributed wooden fire load. The fire path and fire geometry will be observed, and the temperatures, heat fluxes and spread rates will be measured. These tests will also be performed in a “travelling fire environment” and established to be as realistic as possible. Indeed, they will be done with less control over the dynamics, which will allow the characterization of the fire source and, later, the calibration of numerical and analytical models.

3) The fire tests, due to their limited number, are not enough to cover all possible configurations encountered in practice. However, this weakness can be overcome with help of advanced numerical tools. In consequence, CFD numerical models, specifically the FDS (Fire Dynamics Simulator) software, will be calibrated to reproduce the experimental configurations of the fire tests (WP 4).

4) After having calibrated the numerical advanced model, it will be possible to develop an analytical procedure (WP 5) based on the different parameters of the phenomenon (geometry of the fire, fire path, Rate of Heat Release, spread rates, temperatures, etc.). The resulting analytical model will be based on the existing models presented in the literature

The basic parameters of the model will be implemented in a simple calculation tool (Excel Spreadsheet or MATLAB routine) for which free access will be given).

5) Then, the methodology will be introduced in the advanced calculation software SAFIR and OpenSees in order to have a large utilisation of the model by the construction market (WP 6).

6) Finally, design guidance will be provided, including design examples (WP 7). The design guidance will be presented in a document which will provide:

a simplified version of the project scientific content (basic equations)

an accurate and easy to use method to characterise the thermal attack caused by travelling fire

realistic case studies (how to apply the method through worked examples).

Supervisor : Prof. Jean-Marc Franssen

Researchers : Antonio  Gamba, Eric Wellens, Fabien Dumont.

Institutional Partners:  Arcelor Belval & Differdange Sa, SP Sveriges Tekniska Forskningsinstitut AB, University of Edinburgh, University of Ulster.

Funding : RFCS European project ( The Research Fund for Coal and Steel)

Budget : 258.116€

Total project duration : 30 months

Start of project : july 2017

Partagez cette page