Experimental testing of door panel boundary conditions to determine NVH variability

The variability of measured frequency response functions (FRFs) from nominally identical structures is a well-known phenomenon and trying to eradicate it increases the design challenge for automotive manufacturers. In this paper a vehicle door is experimentally tested in order to assess the effect of variability in the attachment boundary conditions between the door structure and trim components upon measured FRF data. Plastic clips can be used to hold the trim to the door panel, so individual clips were systematically removed and then replaced in order to generate a set of measured FRFs that demonstrate how individual property changes can influence the global structure. Point and transfer FRFs, with corresponding normalised standard deviations, were measured by exciting the door panel and measuring the response both on the door panel and on the attached trim. The door response was found to vary by up to 10% over all clip combinations, and this is compared to the test variability. A newly developed function that predicts the FRF variability due to measurement test error was also applied. The results of this prediction match closely with the normalised standard deviation calculated from repeat FRF measurements taken on the structure. This will therefore enable test-to-test variability to be separated from structure-to-structure variability.