Monday  |  Salon 9  |  03:20 PM–03:40 PM

#16923–Unveiling New Insights for Uncertainty Propagation in Frequency-based Substructuring

Frequency-based substructuring (FBS) has been widely used in the industry to (de-)couple experimental models, which are prone to measurement errors. Understanding the propagation of measurement uncertainties within the FBS framework is crucial for assessing sensitivity to errors and identifying critical components in virtual acoustic prototypes (VAP). This work provides new insights into state-of-the-art methods based on the delta method for uncertainty propagation in FBS, especially applying the virtual point transformation (VPT). A novel evaluation of confidence bounds on the absolute value level of the frequency response function output, as well as an uncertainty indicator employing error-covariance ellipses in the complex plane is introduced, simplifying the assessment of uncertainties in FBS. Additionally, a second-order Taylor approximation is derived to estimate bias errors, considering the nonlinear nature of the dual coupling in the frequency domain. The proposed methods are numerically validated through Monte-Carlo simulations, and their effectiveness is demonstrated through an application to an experimental example, considering the influence of the VPT. These insights have significant implications for enhancing the accuracy and reliability of VAP during their development phase.

Javier del Fresno Zarza   KU Leuven
Francesco Trainotti   Tecnical Universität München
Frank Naets   KU Leuven