Experimental On-Line Damage Detection For Industrial Structural Components

Different numerical-experimental methods have recently been introduced in literature for the diagnosis of structural damages by using location-dependent changes in the modal parameters (natural frequencies, damping factors, mode shapes, etc.). The main task is the determination of presence, location and extent of the damage. This work is concerned with the possibility of introducing alongside a manufacturing process an on-line quality control for structural components prior to assembly using modal parameters. Basing on the natural differences among different components, the problem of error measurements, connected with available modal data, could become more noticeable than the health monitoring of an only particular system. Indeed, the modal data that should be used in diagnosing routines, belong to different systems which are obtained by different molding cycles carried out over different metal sheets. For this reason a deviation of the modal data, with respect to a mean value, should be assessed to appreciate the possibility to introduce an online quality control. Moreover, an experimental work benchmarking the deviation of such a modal data in the case we are dealing with is not available. In this work a certain number of structural components is experimentally tested in order to appreciate the deviation of the modal data. To this end, comparisons among mode shapes, natural frequencies and damping factors have been carefully examined and their relevant usability are discussed.