Träbärverk i äldre svenska kyrkor: Verkningssätt hos takstolar utan bindbjälke

Doctoral thesis, 2005

The study concerns the structural behaviour of old wooden roof trusses in Swedish churches. FE-modelling of their properties as based on seven different trusses was used in evaluations aimed at the creation of a tool for gaining a better understand¬ing of such trusses. The trusses investigated all lacked tie beams and were of wide span. A great number of roof structures from different periods are preserved in Swedish churches. These represent a still largely unexplored cultural heritage. No clear picture is available of the construction principles employed or of how their manner of carrying the load follows from the design. Due to this lack of knowledge, anti¬quarian authorities lack means at present of judging how well motivated proposed measures for the restoration of load-carrying roof structures are. Alterations can readily be thought to be too extensive and to damage the historic constructions involved by altering their structural behaviour. The study aims at increasing interest in these old trusses and insight into the principles behind them, as well as at facilitating their preservation. Answers to the following questions are sought: - How do old roof trusses in Swedish churches carry their load? - How is this load bearing influenced by the design? - Which designs are most rational? - How can comparative qualitative results in analyzing these trusses best be achieved? An approach was selected in which three basic "tasks" for roof trusses were defined. For each of these, the structural behaviour of different truss designs was explored using methods based on structural mechanics. Truss designs reflecting the variations in form observed in seven different church trusses studied in an earlier field survey were selected for this purpose. Results of two different types were obtained. The first and most important result was the creation of a tool for interpreting the structural behaviour of roof trusses. Interpreting such behavior took account of the possibility that the design in ques¬tion provided desirable load paths that could be identified. Results of a second type involved obtaining both general and specific answers to the questions posed.