PSI - Issue 58

Kay Büttner et al. / Procedia Structural Integrity 58 (2024) 95–101 Kay Büttner et al. / Structural Integrity Procedia 00 (2019) 000–000

97 3

Fig. 1. Process graphic V-model with research areas for elastomeric bushing development.

2. Target property computing The method of target property computing will be presented using the example of vehicle drive train mounts. During real driving operation, mounts are spatially displaced, i.e. multi-axial load conditions may exist. The loads, which are required for a durability component design, also have high amplitudes in the range of higher frequencies. Due to the complex component design, strong couplings exist between the individual directions. Nonlinear springs, several stops and hydraulic damping elements are used to resolve the previously described target conflicts between individual vehicle properties. When looking at the standard modelling of aggregate mounts, for example an adapted model of Dronka and Rauh (2007) (see Fig. 2 (a)), the following is apparent:  Regarding to Ernst et al. (2019) the standard experimental characterization of aggregate mounts for model parameterization is usually performed uniaxial. This means that the coupling effects between the axes are not taken into account. Additionally, the quasi-static stiffness measurement detects excitations with high amplitude and low excitation velocity.  The dynamic properties of the mount are determined by loss angle and dynamic stiffness, especially at low force amplitudes in the linear stiffness range. Thus, the operating range relevant for fatigue strength or durability design is not considered.  For virtual load data acquisition using current models like Rizov (2022), peak loads are not represented well. As a result, the signal damage of the simulated forces is calculated to be significantly lower than in real driving operation. In the example shown here (see Fig. 3), a signal damage of approx. 23% is obtained for the z-direction of a gearbox mount. Consequently, with these virtually determined loads for a component validation, for example in an operating load test, the durability is predicted to be significantly too low. A revised empirical rheological model with a parameterization process was developed by Ernst et al. (2021) to improve the accuracy of virtual load data acquisition (see Fig. 2 (b)). This complex model represents the transfer behaviour of the mount for one spatial direction. It consists of a uniaxial part for effects the main direction and two parts for the influences of the secondary directions. With this approach, mount- displacement-dependent-relaxation-properties can be modelled. The change of uniaxial static and dynamic mount properties due to multiaxial spatial excitations can be displayed as well. For the parameterization, a Particle Swarm Optimization was used with stochastic multi-axial spatial excitation signals. The signal damage and the mean simulation error are used as optimization criteria. By this method, the predicted signal quality can be significantly