On the influence of crack closure on strength estimates of wood

Three well-known duration of load models (Gerhard, Barrett/Foschi, DVM) are considered in this note with respect to their ability to predict lifetime of wood subjected to harmonically varying loads. The result obtained is that they practically predict the same lifetime—which for low frequency loading can be considered approximately true. For higher frequencies, however, this result can be far too overestimated. The reason is that the models considered do not take into account the effect of the crack closure phenomenon (which are the main mechanisms of energy dissipation causing fatigue failure in metals). It is suggested that any of the simple models can be used in practice when low frequency load variations are considered. The DVM model, however, should be preferred because of its ability to predict residual strength, and because of its ‘build in’ flexibility with respect to wood quality and ambient climatic conditions. For high frequency load histories more refined models are required. The extended DVM model, recently developed by the author, is suggested as such a model—especially because it has the potentials of being further developed to consider arbitrary load variations (such as earthquakes). Finally, the widely spread concept of estimating long-term strength by multiplying short time strength with a codified factor (so-called k _MOD factor) is discussed. It is concluded that the k _MOD -method can be justified in practice with low frequency load variations. When high frequency load histories or unexpected peak loads are considered, the k _MOD -method may cause considerably overestimated lifetimes.