Picture of a production series rear muffler (top) and CAE model (bottom). The system is depicted upside-down. The number in the lower plot mark accelerometer positions (cf. section II B). The inner structural parts of the rear muffler are removed.
Contour plot of the mode shapes determined by an experimental modal analysis.
Experimental setup. The structure is excited by an acoustic source. Structural vibrations are measured by accelerometers [cf. Figure 1(b) for positioning].
Magnitude of transfer function between the acoustic pressure at the inlet and the normal structural displacement on the lower shell [see Fig. 1(b) for positioning]. A strong excitation via the acoustic path is observed close to the structural eigenfrequencies (dotted vertical lines).
Fluid-structure coupled problem. The interior acoustic domain and the structural domain intersect on . The interior and exterior acoustic domains are coupled on . The fluid normals of the interior and exterior domain, and , point out of the fluid.
Simulation scheme. The interior acoustic problem and the structure are fully coupled. The feedback of the acoustic pressure of the exterior acoustic problem on the structure is neglected. The two acoustic problems are coupled by impedance conditions on the outlets.
Comparison of the reduced-order models to the full-order model.
Pressure-induced structural vibrations. Comparison of simulation and experimental results (solver variant Ru-CB).
Contour plot of the operational deflection shape at obtained by the simulation model CB-Ru. At this frequency the dynamic behavior is dominated by mode 10 [cf. Figure 2(d)].
Noise radiation of the rear muffler as computed by the FMM-BEM. The effect of surface sound radiation is clearly seen on the plane above the main part. Close to the orifice a circular spreading is observed.
Eigenfrequencies and modal damping ratios of the rear muffler obtained by an experimental modal analysis.
Parameters of the rear muffler after model updating.
Computation times (CPU time) of the reduced-order models compared to the full-order model. The reduced-order models outperform the full-order solution.
Article metrics loading...
Full text loading...