Robustness assessment of an ancient timber roof structure

Pre-printDuring the last decades, it has become a usual practice to replace ancient timber roof structures by concrete or steel roof structures. Despite the durability of ancient timber, the lack of knowledge about how to reinforce or retrofit timber structures may have resulted in this. However, the main reason may be found to be the difficulty in assessing the real condition of timber frames, attending to its level of conservation. Within that scope, two timber collar trusses dating back to more than a hundred years, removed from the Laboratório Chimico in Coimbra (Portugal), were tested up to failure in the structural laboratory of the University of Minho. According to the test results using non-destructive techniques, it was possible to develop different probabilistic analysis considering both exterior and residual cross section dimensions. In this work, a robustness assessment is performed considering the spatial distribution of the strength of timber elements and also the variability of the stiffness of the connections. Permanent loads, snow load and wind load variability were considered, for the structural safety analysis, according to the provisions of the Joint Committee on Structural Safety. The robustness index was obtained based on the reliability indices of intact and damaged roof structure, considering the geometric variation obtained for the exterior and residual cross sections respectively. Crude Monte-Carlo simulation method was used here to estimate the probabilities of failure of the structural system. The comparison between different imposed damage scenarios covering loss of bearing capacity of connections and timber frames is finally presented.The information obtained through training school: Assessment and reinforcement of timber elements and structures (Guimaraes, Portugal) within scope of FP1101 and RILEM TC 245 RTE is acknowledged. The support of the Structural Lab from University of Minho is acknowledged. Financial support of the Portuguese Science Foundation (Fundacao de Ciencia e Tecnologia), through PhD grant PD/BI/52583/2014 included in InfraRisk-PhD program, is gratefully acknowledged