PSI - Issue 57

Marco Bonato et al. / Procedia Structural Integrity 57 (2024) 799–809 / Structural Integrity Procedia 00 (2019) 000 – 000

802

4

Q=10 (5% damping). Similarly, for electronic sub components (solder joints of PCB) the typical damping is 2% hence Q = 25. The parameters defining the slope of the relative Wohler curve (the Basquin b exponent) follows similar criteria. In case of fatigue of metals or hard plastic, b = 8. If the vibration failure is expected to occur on electronic parts, the Basquin b considered is b = 4. All these Q values have been experimentally obtained by Valeo from experimental measurements; the Basquin coefficients considered are both the results of experimental analysis (tests to failure and interpolations of the dataset to extrapolate the SN curve) as well as values suggested from the literature. It is noticeable that the zone of failure (mechanical or electrical) highly influences the parameters estimation assumed for the simplification of the validation signal. This is challenging because in recent years the automotive industry has seen the rise in products related to the electrification of the powertrain. More and more components belong to the category of mechatronic systems. These innovative components show an integration of mechanical, electrical and electronic engineering systems. It is very important to apply a physics of failure (PoF) approach and engineering judgment to properly select the parameters for the simplification of the vibration signal previous to FEA simulation, especially if the result is the cumulative damage. In this framework, we have investigated the effect of these two parameters on the predicted simulation fatigue on 2 mechatronic systems: the high voltage coolant heater (HVCH) and the alternator. 1.2. Simulation Fatigue Damage Design validation in the automotive industry was usually carried out by performing FEA simulations on the predicted zone of maximum stress. Conservative criteria related to the UTS (as UTS/2 or UTS/3) of the material are applied to verify that the max stress will not generate any failure. Such criteria are basically targeting an infinite fatigue life endurance. Recently, the product validation based on fatigue design criteria have found appreciation in a framework of automotive component downsizing and design to cost. An ever increasing research has been dedicated to fatigue life simulation tools, that allows the calculation of the cumulative fatigue damage on the component weak parts (called “hot spots”). Finite Element Analysis is performed in order to identify critical locations on the product. Stress and strain tensors are calculated in these hot spots. Dedicated post processing on using these stress/stress tensors and fatigue data (as Wohler curves) are then performed to calculate damage or fatigue life of the product. These estimations can be done on different kinds of materials as metals, polymers, and composite. The design validation is obtained in a more realistic way, because the results provided not only the answer to the question "Will it pass the test?" but also what is the design margin, and this before any prototype is even made and physically tested. 1.3. Influence the of signal cumulation parameters If the simplification of the complex vibration signals into a unique PSD or sine sweep allows to run preliminary FEA simulations, additional investigations are needed to assess the effect of this signal cumulation. As mentions, numerous parameters play a major role in the output: - the nature of simplified signal (random or harmonic) - the fatigue coefficient - the damping ratio Recently, Chabod et al. performed a comparative study on the FEA simulation fatigue of an automotive support bracket (Chabod et al. (2022)). The original sweep sine on random is converted to damage equivalent PSD or sine sweep. While the objective of this investigation was to point out the advantages and disadvantages of frequency -based and time-based methods for simulating swept-sine on random durability tests using finite element models, the starting point of the generation of the equivalent signals is th e calculation of the FDS and its “inversion” to PSD and sine sweep. Their results clearly show that for this particular example, the most suitable approach is to perform the finite element fatigue analysis via the stress time series, because of the transient nature of the starting signal. Both stochastic and deterministic FDS equivalent signals (the PSD and the sine sweep, respectively) fail to converge to the same fatigue life predictions. The paperhowever does not investigate the effect of the dampingra tio on the severity of the equivalent PSD and sine sweep.