A theoretical model of fully developed fire in mass timber enclosures
Thesis DisciplineCivil and Natural Resources Engineering
Degree GrantorUniversity of Canterbury
Degree NameDoctor of Philosophy
Existing practices for the fire design of mass timber buildings based on traditional fire resistance frameworks previously developed for non-combustible enclosures are inadequate. The contribution of mass timber surfaces to a fully developed enclosure fire is coupled to the design fire such that timber charring rates determined from standard fire resistance tests or parametric time temperature relationships may not apply. This is particularly important when considering structural fire performance of tall mass timber buildings.
This thesis describes a theoretical fire model for calculating the thermal environment within enclosures constructed from fully or partially exposed mass timber elements such as cross-laminated timber. The fire model includes two new pyrolysis submodels to enable calculation of the mass loss rate, energy release and char depth within wood surfaces burning in the enclosure. Phenomena such as debonding of lamellae in engineered wood panels is included and discussed. The pyrolysis submodels are coupled to the two-zone fire model B- RISK enabling the fire dynamics in small mass timber enclosures to be predicted. Model predictions for heat release rate, gas temperatures and/or char depths are compared with data from 19 full-scale fire experiments previously published in the literature.
The thesis also describes a submodel for predicting the enclosure thermal enhancement and ventilation effects on the mass loss rate of a burning fuel package. Model predictions have been evaluated with good agreement for an inert reduced-scale enclosure based on a series of heptane pool fires and a series of upholstered chair fires in a full-size enclosure. The application of the submodel to mass timber enclosures in combination with the previously developed pyrolysis submodels is discussed however additional experiments with well characterised fuel sources are required to more thoroughly evaluate this feature of the model. Potential applications of the model include generating thermal boundary conditions for a more advanced thermal/structural finite element code and for deriving modified fire load energy density values applicable to mass timber for use in simplified fire severity formula.
It is concluded that where mass timber structures must be designed to ’not collapse’ in fire then satisfying prescriptive time periods in standard fire resistance tests is not sufficient. The fire performance of these structures need to be specifically engineered considering the expected fire growth, duration and decay periods. This requires a coupled interaction between the moveable fire load and the combustible enclosure surfaces to be considered.