Predicting the thermal response of timber structures in natural fires using computational ‘heat of hydration’ principles

Abstract

The thermo‐physical response of timber structures in fire is complex. For this reason, debate still exists today as to the best approaches for simulating thermal response in fire using tools such as finite element analysis (FEA) modelling. Much of the debate is concerned with the thermal properties of timber, for example, conductivity, specific heat and density, at elevated temperature and how such properties should be implemented or interpreted in numerical calculations. For practitioners intending to use modelling as a fire design tool for timber buildings, guidance exists on the thermal properties of softwood in Annex B of EN 1995‐1‐2. These properties are limited for use under standard fire exposure conditions because of the way in which they were derived from calibration against focussed test data. As a result, they cannot be applied to non‐standard fires, which are more representative of real fires due to a combination of varying heating rates and the decay phase of fire development. The limitations of the standard fire test (and associated curve) are widely understood. As a result, much recent structures in fire research has focussed on the ‘performance based design’ of buildings subject to increasingly realistic fire conditions. Such an approach allows engineers to quantify the level of safety that can be achieved in a building should a fire occur. In addition, the design of buildings to withstand fires proportionate to the risks foreseen and also the geometry present results in better value buildings that are inherently more robust. For the same approaches and associated benefits to be realised for timber buildings, then a number of barriers must be overcome. The most obvious of these is engineers' ability to determine timber structure temperatures as a result of fires other than the standard fire curve. This however presents a number of challenges. Upon heating, the moisture bound within begins to evaporate, volatiles begin to flow from the heated surface and char forms. The rate of which these behaviours occur and the nature of the char that forms depends on a number of factors, but most notably the rate of heating. Upon cooling, the timber member continues to generate heat energy as the surface oxidises. As a result, any models intended to simulate temperature development must consider the relationship not only between temperature and thermo‐physical characteristics but also between heating rate and the process of heat generation. Many models have been developed for this purpose; however, they are extremely complex and are some way from being ready for implementation as design tools. This paper proposes implementing ‘heat of hydration’ routines, intended for the curing of concrete structures, to simulate the heating and cooling process in timber structures. Such routines are available in many commercial FEA software packages. The adoption of the hydration routines allows the heat generation process, as a result of oxidation, to be considered in parallel with solid phase heat transfer using ap...

Sustainability Improvement in the Design of Lightweight Roofs: A New Prototype of Hybrid Steel and Wood Purlins

“…There are many lines of research on this last point, as it is not easy to predict what will happen to wood in cases of fires at temperatures that differ from standard values [35]. In this regard, Hopkin proposed a very interesting method [36]. However, it is worth noting that fire resistance can be further improved through appropriate flame-retardant treatments with special surface treatments [37].…”

A performance indicator for structures under natural fire

“…This delayed charring process is generated by the energy received from the radiative and convective heat flux still produced by the cooling surrounding fire and from the combustion of the structural component itself [27]. As a consequence of the resulting increase in charring depth, the effective cross-section continues to decrease.…”

OpenSeesdevelopment for modelling timber structural members subjected to realistic fire impact

“…During the charring process, the combustion heat of timber is largely released to the room environment as the burning of pyrolysis gas, whereas a small portion of this heat is absorbed by the timber section and transferred to the deeper layer, and this dominates the charring propagation in low‐intensity fires. Therefore, the combustion heat model for the char layer proposed in the author's other work 23 based on Hopkin's model 25 is activated to consider the heat generation due to combustion.…”

“…Facilitated by the layer state tag, the thermal properties of each layer of timber consisting of density , specific heat and thermal conductivity k (Figure 3B) are defined based on the EC5 thermal property models 1 and data in, 7,26–28 as well as the consideration of heat generation due to timber combustion, 25 which now is dominated by the material temperature and layer states. In the case that the layer transition does not occur, these thermal properties vary with elevated temperature and are linearly interpolated between the key control input parameters at ambient temperature, dry state (100°C), and char state (300°C).…”

“…Therefore, the combustion heat model for the char layer proposed in the author's other work 23 based on Hopkin's model 25 is activated to consider the heat generation due to combustion.…”