Using the hybrid FE-SEA method to predict structure-borne noise transmission in a trimmed automotive vehicle

You are here: Homepage > Search

Using the hybrid FE-SEA method to predict structure-borne noise transmission in a trimmed automotive vehicle

Charpentier, A. / Cotoni, V. / Fukui, K.

A Hybrid model of a BIW with focus on the floor and dash sub-assemblies was created. By combining FE and SEA, this model is capable of predicting structure-borne noise transmission over the range [200-1000] Hz. A process to analyse the dynamics of a complex structure and define the most appropriate model partitioning was established. Based on this partitioning, the regions of the structure that exhibit short wavelength behaviour were described with SEA. Only the stiff regions were left as FE, making the resulting Hybrid model computationally efficient. It was shown how detailed local FE models of the BIW components could be used to efficiently calculate the SEA subsystem properties and SEA couplings loss factors across complex junctions. The SEA subsystems were coupled to the rest of the structure which was modelled as FE through Hybrid point and line junctions. The structure-borne dynamics of the Hybrid model were validated both numerically and experimentally. The Hybrid model predictions compared well with the predictions of the original FE model. The Hybrid model is capable of predicting power inputs due to point force excitation at the engine mounts and shock tower locations within 3dB of tests for most frequency bands. Additionally, the Hybrid model is capable of predicting the velocity distribution on the floor and dash, not only for point force excitation on the front frame (shock towers and engine mounts), but also for spatially delta-correlated excitation (rain on the roof) of any component of the floor. Key components for the propagation of structure-borne vibration in the floor and dash were identified. In particular, the main rails supporting the engine sub-frame carry a large portion of the energy to all floor subsystems. Additionally, predictions were made to identify the contribution of the various subsystems to interior noise based on their respective radiated sound power. These predictions were performed in both bare and trimmed configurations over the mid-frequency range [200-1000] Hz.