No abstract

No abstract

Frequency Response Functions (FRF) haves been used to aid noise and vibration designs in various industries. Those design considerations include resonance avoidance, vibration reduction, etc. Numerical methods have been widely applied to predict Frequency Response Functions (FRF) of structures. However, the computational resources (i.e., CPU time, memory, disk space) needed to solve large and detailed numerical models are getting large. Furthermore, the need to resolve resonant response peaks can drive up the number of FRF calculations required. Lately, advanced numerical techniques based on a Krylov subspace and Galerkin Projection (KGP) and Pade Approximation have been demonstrated that they can significantly accelerate the overall process by approximating the frequency dependent response (calculating the forced response at only a few frequencies, then using KGP or Pade approximates to reconstruct the FRF for the rest of desired frequency points.) This paper will present the latest enhancements to the KGP: modeling of viscoelastic material via its complex modulus representation and adaptive capability (AKGP) in automating the frequency sweep process via calculated tolerance error. To illustrate the accuracy and efficiency of the new enhancements, a numerical example will be exercised and used as a benchmark to compare different numerical tools with and without the AKGP.

No abstract