| Popis: |
Physically correct and invariant inputs are important in vehicle development to perform virtual optimization at the suspension or body-level. These inputs can be applied to vehicle models in order to predict interior vibration or noise-levels and evaluate NVH performances without having a physical prototype. The applied inputs can be wheel-center loads or, even more invariant, tire-patch displacements. These tire-patch inputs, applied to a tire model, allow accurate predictions up to 300 Hz. A limitation of this approach however, is that no accurate inputs can be obtained at lower frequencies, therefore application for comfort analysis is not possible. In this study it is investigated if the low-frequency quality loss in the input identification is related to the measurement approach, the data post-processing steps, or non-linearities in the tire or suspension. An alternative measurement setup including the use of strain-sensors is evaluated and compared with the original acceleration-based methodology. A comparison of predicted target vibrations at the seat rail as well as obtained inputs is presented for the different methods. The results indicate that the proposed alternative methodology increases the quality of the estimated inputs in the low-frequency region. With these improved low-frequency inputs, the methodology of using tire-patch displacements as inputs to hybrid models can now also be applied for lower frequency (ride, comfort) applications. |
