Stochastic process generation from relaxed power spectra utilising stochastic harmonic functions

verfasst von

Marco Behrendt, Marius Bittner, Michael Beer

Abstract

In order to design and safely construct buildings and structures that are exposed to environmental processes such as earthquakes and wind loads, simulations are essential in advance. Although simulations are an approximation of reality, they are still dominated by uncertainties, that must be taken into account. These uncertainties can arise for various reasons, such as incorrectly recorded data or inaccurate simulation models. One widely used approach for generating and simulating environmental processes is the power spectral density (PSD) function. It establishes a relationship between the time and frequency domains and determines the relevant frequencies and their magnitude of the transformed signals. Since the model of the PSD function provides discrete values of the density for each frequency, the idea arises to model this density uncertain. For this purpose, statistical values are extracted from an ensemble of similar PSD functions that differ slightly in shape and peak frequency, for example. These values are used to derive a relaxed model of a PSD function. By using such a model, the response processes of buildings and structures are also uncertain, resulting in a range of possible values instead of discrete values. However, a disadvantage of this model is a higher number of random variables that can affect the simulation results. In this work, the stochastic harmonic functions (SHF) are used instead of the spectral representation method (SRM). Stochastic processes generated with both SRM and SHF are compared in a benchmark simulation to assess the advantages.

Organisationseinheit(en)

Institut für Risiko und Zuverlässigkeit

Typ

Aufsatz in Konferenzband

Seiten

52-58

Anzahl der Seiten

7

Publikationsdatum

09.2022

Publikationsstatus

Veröffentlicht

Peer-reviewed

Ja

ASJC Scopus Sachgebiete

Sicherheit, Risiko, Zuverlässigkeit und Qualität, Managementlehre und Operations Resarch

Elektronische Version(en)

https://doi.org/10.3850/978-981-18-5184-1_MS-01-220-cd (Zugang: Geschlossen)