- Online only
- New

Random vibration loading commonly induces non-proportional multiaxial stress states. Their evaluation requires sophisticated methods, which often entail high computational demands. To address this, statistical characterizations of loading and stresses have emerged as a powerful approach, enabling highly efficient evaluations across entire finite element (FE) models, including parametric analyses of critical planes.
These statistical characterizations combine two key aspects: frequency-domain characterizations, which account for structural dynamics and cycle counting, and correlation statistics, such as the covariance matrix, to describe interactions between stress components. However, traditional covariance matrix analysis fails to capture the full phase information necessary…
Datenschutzbedingungen (bearbeiten im Modul "Kundenvorteile")
Lieferbedingungen (bearbeiten im Modul "Kundenvorteile")
Rücksendebedingungen (bearbeiten im Modul "Kundenvorteile")
Random vibration loading commonly induces non-proportional multiaxial stress states. Their evaluation requires sophisticated methods, which often entail high computational demands. To address this, statistical characterizations of loading and stresses have emerged as a powerful approach, enabling highly efficient evaluations across entire finite element (FE) models, including parametric analyses of critical planes.
These statistical characterizations combine two key aspects: frequency-domain characterizations, which account for structural dynamics and cycle counting, and correlation statistics, such as the covariance matrix, to describe interactions between stress components. However, traditional covariance matrix analysis fails to capture the full phase information necessary to accurately describe these interactions in a multiaxial stress state.
To overcome this limitation, we introduce a complex-valued covariance matrix as an extension of the well-know covariance matrix. We demonstrate its utility in characterizing multiaxial stress states through a comparative study of time- and frequency-domain approaches. Our contribution will focus on accurately considering the interactions between shear and normal stresses on critical planes and proposing an extension of a widely used critical plane criterion for random vibration fatigue.
In summary, this work highlights the potential of statistical-based fatigue evaluations as a fast alternative to time-domain assessments of multiaxial stress states. Such an approach can serve as a pre-processing step to identify critical nodes and plane orientations for subsequent time-domain analyses or as input for fatigue damage estimations using statistical damage models.