Historical carpentry corner log joints—Numerical analysis within stochastic framework

Abstract

The paper presents the results of numerical analysis performed on historical, traditional carpentry corner log joints of two basic topologies: the short-corner dovetail connection and the saddle notch connection. These types of carpentry joints are commonly used in currently preserved objects of wooden architecture. All connections have been modelled in pinewood, which has been defined in the Finite Element software MSC.Marc/Mentat as an orthotropic material. The numerical calculations have been carried out for two types of connections with two different boundary conditions and load types. The contact phenomenon between the individual elements of the connections has been taken into account. The main purpose of the research is to select the most damage-resistant type of connection and to determine the stress distributions on the contact surfaces, which demines the damage areas. However, a lot of uncertainties appear in the studied models, e.g. due to the natural variability of the material properties of wood and the uncertainty of friction coefficient. Therefore the uncertainty quantification and global sensitivity analysis has been performed in order to include these uncertainties and study their effect on variation of the mechanical response of the connections. A regression-based non-intrusive polynomial chaos expansion method has been employed to complete the task. The state-of-the-art knowledge about the damage-prone zones in the considered connections is immensely important since many wooden buildings, mostly historical, require maintenance, renovation and the reinforcement of existing, especially historical elements. On the contrary, there are not many results of related research published yet.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 18 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.