Abstract

Very often analysts are faced with making design choices related to durability before adequate stresses can be obtained from the prototype. Sometimes, revisions to subsystems are needed when only Adams type MBDA models are available which do not, or cannot, include stress information from the flexible bodies within the system. In these circumstances, if the analyst wants to assess durability improvements, he or she, is forced to use tools like the Fatigue Damage Spectra (FDS) approach. The FDS approach is a highly simplified assessment of damage potential for a hypothetical system. The biggest limitation of the approach is that a “pretend” one degree of system is used to transform the input loading (e.g. acceleration) into stresses which can be used in a fatigue calculation. The resonant frequency of the 1 DOF system is swept through the required frequency range and the damage plotted as a function of frequency. In effect, the approach assumes zero knowledge about the real system. By far the best option in determining damage potential is to pass the loading through a real system, or a simplified version of it, in which case we have almost 100% knowledge about the system. The frequency domain approach offers the possibility to do this because of the separation of loads, system properties (transfer functions) and response (damage). In this paper, something in-between is proposed (which we call Pseudo Damage) where partial knowledge about the system is available via an Adams MBDA model. The MBDA model is used to generate a set of (acceleration) responses on the model. These acceleration responses are then transformed by switching the units into stress and scaling the magnitudes until the damage caused (for a specified material) is equal to 1.0. The scaling factors (different for each channel) are then stored as the “reference case for the model. Any proposed design changes can then be evaluated by seeing if the scale factors needed on the modified design go up, or down when compared to the reference case. This is the Pseudo Damage calculation. This approach is currently implemented in the time domain and results from a typical analysis will be shown. Further work to extend it into the frequency domain is ongoing.