For aesthetic reasons, architects or owners often want to use exposed wood products in large, tall open structures and in atria where non-combustible materials are generally specified by building codes. In order to demonstrate that such applications can be made safe provided extra fire safety measures are taken, fire protection engineers often must employ computer modelling or full-scale fire-testing. Unfortunately though, the use of wood products in many of these projects is not pursued because there is not sufficient time, commitment or resources available to undertake the lengthy and expensive analyses required to get approval for the non-compliant designs.
This study was initiated to examine the impact on fire safety presented by combustible ceiling and/or wall coverings as the floor area and ceiling height of both sprinklered and unsprinklered compartments increase. It was proposed that existing or newly developed fire models could be employed to predict under what circumstances wood linings can be used safely in tall and/or large-area compartments. It was further suggested that simple design criteria could be developed for the use of the design community or, possibly, for inclusion in building codes.
A literature review identified two computer fire models that were potential candidates for use in this project: BRANZFIRE and Fire Dynamics Simulator (FDS). BRANZFIRE is a zone model; that is, in the event of fire, it divides a room (atrium) into two zones: a hot upper layer and a cooler lower layer. While is was found that BRANZFIRE is ideally suited for predicting fire development involving combustible wall and ceiling linings in small rooms, its predictions in large rooms, such as atria, are poor. The two zone approximation is simply too coarse.
FDS is a computational fluid dynamics model in which the room (or atrium) of fire origin can be broken up into a large number of cells (potentially thousands), rather than just two zones. Although this was promising, it was learned that recent attempts by others to use FDS to model upward flame spread on combustible walls had not been successful. The pyrolysis and combustion models in FDS are not up to the challenge.
Attempts made to attract a student at Carleton University, where considerable computing resources are available, to develop improved pyrolysis and combustion models for use in FDS have not been successful. Coupled with the reduced funding available for General Revenue projects, a decision was taken to terminate work in this project.