Fire Resistance of Thermally Insulated Log-House Timber Walls

This paper presents the results of standard test results and Finite Element numerical modelling for a thermally insulated log-house timber wall loaded in compression and exposed to fire. A key aspect in the design of log-house systems, being susceptible to possible local failure mechanisms and buckling phenomena, is represented by geometrical details like the cross-section of logs (typically characterised by depth-to-width ratios in the order of two or more) and the corner joints features. This is also the case of log-house structures exposed to fire, were the premature collapse due to combined thermo-mechanical loads should be properly prevented. In this context, insulation claddings can markedly delay the fire propagation and temperature increase in the timber components, as to increase their overall fire resistance. Such an aspect is important especially for commercial buildings or accommodation facilities, according to international design requirements. To this aim, the paper reports on a full-scale, ≈ 3 m × 3 m and 90 mm thick, thermally insulated log-house wall specimen (‘W90-insulated’) loaded in compression (R N = 0.13 the ratio, as in a 2-storey building) and exposed to fire, in accordance to the EN/ISO temperature–time standard curve. The most important test results are commented, including comparisons towards an unprotected wall with identical nominal geometry (‘W90′), which was previously tested in the same facility. It is shown, in particular, that the insulation package can extend the fire resistance of the wall up to more than 150 min (with limited charring and deformations of logs), that is three times the unprotected sample. The furnace experiment herein presented, even in presence of intrinsic limitations for the testing method, emphasises the need of a wide set of instruments to properly capture the key temperature and deformation results of log-house assemblies, as well as the lack in design standards of specific performance parameters. The W90-insulated test predictions are hence explored in the paper via FE numerical models, where major advantages for the timber components are derived from past literature efforts and special care is spent for the thermal insulation package. Even in presence of well known simplifications (i.e., thermo-mechanical boundaries/loads and thermo-physical properties of materials), as shown, the FE method can offer interesting correlation with the full-scale predictions, for more than 100 min of exposure. In addition, the crucial role of cladding layers on the fire resistance and failure time/mechanism of log-house walls is further assessed, based on FE parametric studies of technical interest, inclusive of different configurations for the thermal insulation layers. © 2018, Springer Science+Business Media, LLC, part of Springer Nature.