Correlating a finite element model with tests from a set of sample structures

Whatever the updating procedure employed, a careful testing is mandatory to provide significant experimental data and achieve a reference model. When the updating concerns a set of nominally-identical structures from an industrial line, a critical point is the definition of a sample that should be representative of the whole set. Another troblesome point concerns the chance of correcting the theoretical model by using tests from a particular sample structure. To explore in detail these questions, three nominally identical frame structures were constructed, experimentally tested and updated. Cross-comparisons are performed on the obtained results to draw some preliminary but interesting indications.