New possibilities in structural design and testing of wooden structures?

Until recently there was a standard way of designing structures. The accepted way involved a standard calculation and static analysis, seldom a dynamic one, the calculation of stability and the application of accepted standards: the proofs of safety and serviceability of the whole and all parts and details, according to materials applied in the structure. The calculations and analysis were carried out in different ways, frequently manually and lately using FE modelling and analysis. The calculations of safety, stability and serviceability have lately been mostly performed by use of various closed programs and mighty FE based and material based computer codes. In this paper new possibilities are shown in designing GLULAM timber roof structures based on the use of artificial neural networks trained on samples. The networks are trained on samples generated by rule based expert systems and/or historical design samples (from engineering practice).