Krylov-Based Model Order Reduction in Vibroacoustics
* Presenting author
Vibroacoustic simulations at an early stage of product development are crucial in anticipating a model’s behavior in its real application. Structure-borne vibration problems can be numerically modeled with Finite Element Methods (FEM). However, the complexity of models in practical applications often leads to large dimensions of linear equations that demands huge computational efforts for solving. Therefore, it is advisable to perform a reduction in model order by using model order reduction techniques, thereby yielding faster computations. However, conventional reduction techniques like the CMS method faces challenges in reducing a model in the presence of localized damping. Therefore, the contribution investigates another technique of efficient Krylov-based Model Order Reduction (KMOR) for accurately reducing the models, which are subjected to structure-borne vibration and high-localized damping. In addition, one suitable way to approach coupled system response of complex model assemblies is by evaluating the transfer functions of each FE model representing a single substructure using KMOR techniques and coupling them within a Frequency Based Substructuring (FBS) framework. The procedure thereby yields flexibility and faster computations without compromising on accuracy. Consequently, the vibration energy flowing through the structure is evaluated and compared with other conventional methods.