In the context of noise and vibration control, constrained viscoelastic materials can be used to reduce structure-borne noise of vibrating structure. This passive damping treatment finds many applications due to its robustness and low cost. While several approaches are proposed in the literature to optimise the placement of constrained viscoelastic patches, few studies has been dedicated to the optimisation of the viscoelastic material's properties. The difficulty stems from their frequency-dependency. In this work, a fractional derivative model is used to describe the complex frequency-dependent viscoelastic properties. This study illustrates the effectiveness of using surrogate modeling to predict the damping performance of viscoelastic patches. A multi-model reduction method combined with a modal parameter identification method is used to build a surrogate model at a reasonable computational cost. Once trained, this metamodel is explored to optimise the parameters of the fractional derivative model, thus allowing the design optimisation of viscoelastic properties. This approach could be extended to optimise the material properties of viscoelastic foams in the low frequency range.
Read full abstract