Service life planning of wooden structures: Mathematical prediction models versus professional experience

IRG/WP 19-20663

C Brischke, J Niklewski, M Humar, G Alfredsen

During the last 15 years, enormous efforts have been made in developing models for predicting the service life of wooden structures and components. Currently, a framework of how exposure, dimension, design details and the material-intrinsic and the ability to take up and release water can be linked to model the moisture risk in wood products is in principle available. The aim of this study was to compare the ‘predictive power’ of such a service life prediction model with that of different groups of wood users. Besides professional wood-related craftsmen, architects, wood scientists and customers of so-called ‘DIY –markets’ were asked to estimate the time span between the beginning of exposure and the occurrence of fungal decay on case examples with known history and service life. A set of models predicted the service lives of the different example cases fairly well with the exception of one playground structure, which was made from preservative treated wood. Material-specific data on the resistance and wetting ability are needed to further improve the model fit. In many cases, the average service life estimates of all survey respondents met the real service life fairly well, but suffered partly from high scattering. Groups of experts such as wood-related craftsmen and wood scientists did not provide significantly better estimates than non-experts. The need for comprehensive and elaborate prediction instruments became evident, since neither ‘laymen’ nor ‘experts’ were able to provide sufficient statistically reliable estimates for service lives of wooden components. The predictive power of any expert is dependent on personal experience, which is often restricted to local climatic conditions and regional construction techniques and traditions.


Keywords: design guideline, expert knowledge, knowledge transfer, performance-based design, service life prediction

Conference: 19-05-12/16 Quebec City, Canada


Download document (1.2 Mb)
free for the members of IRG. Available if purchased.

Purchase this document