Tailored bench tests are vibration experiments specifically developed for a product to reproduce the severity of a customized application, generally within an accelerated timeframe.When structures are dynamically excited, the response is extremely sensitive to input(s). The tailored test parameters, (a) input degree of freedom, (b) magnitude, (c) frequency content and (d) exposure time, must be determined not only based on mission characteristics but also considering the component’s modal properties. Since inevitable deviations occur between the operational and test environments, a consistent tailoring method is essential to avoid uncorrelated failure modes and worthless tests. Herein, computational tools (FEA) were combined with traditional signal processing to enhance durability assessment and test specification. The studied case is a durability failure of a pneumatic brake chamber support, a typical unsprung structure attached to a vehicle’s axle. The crack occurred during an off-road application, a worldwide ordinary chaotic mission.The tailored input is given by PSD(s) for each DoF. They are inherently ergodic and stationary signals. Therefore, reproducing the off-road wide variable multiaxial events, whether deterministic or random, is challenging to start with. There are, however, other questionable assumptions during PSD computation according to the mission’s fatigue damage spectrum (FDS). The first is the election among SDoF responses, displacement or velocity, as the stress-proportional parameter. The second is the flat frequency, a white noise base-driven acceleration. Such a condition is only rarely empirically satisfied. Last, multiaxial loaded components are not properly addressed, since FDSs are calculated individually for each input DoF despite their simultaneous imposition on operational application.Two tailoring test approaches were compared in terms of life contour correlation, durability prediction, failure mode reproduction and acceleration factor. While the first assumes simultaneous application of each input DoF (x, y and z) and equalizes the overall damage level, the second adjusts the exposure time (and damage level) specifically for each input DoF, composing a successive test to match the mission total damage. Counterintuitively, the latter resulted in a superior correlated result.
Read full abstract