Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Uncertainty quantification of aeroacoustic power sources in corrugated pipes

Publication files not online:

Author: Swamy, M. · Shoeibi Omrani, P. · González Díez, N.
Type:article
Date:2015
Publisher: American Society of Mechanical Engineers ASME
Source:ASME 2015 Pressure Vessels and Piping Conference, PVP 2015, 19-23 July 2015, Boston, MA, USA, 4
Identifier: 533051
ISBN: 9780791856970
Keywords: Mechanics · Acoustic noise measurement · Aeroacoustics · Computational fluid dynamics · Distribution functions · Electric power transmission networks · Fluid structure interaction · Geometry · Orifices · Periodic structures · Pressure vessels · Risk assessment · Risk management · Sensitivity analysis · Tensors · Confidence levels · Corrugated pipes · Geometrical dimensions · Predictive modelling · Probabilistic models · Stochastic collocation method · Uncertainty quantifications · Industrial Innovation · Fluid & Solid Mechanics · HTFD - Heat Transfer & Fluid Dynamics · TS - Technical Sciences

Abstract

Gas transport in corrugated pipes often exhibit whistling behavior, due to periodic flow-induced pulsations generated in the pipe cavities. These aero-acoustic sources are strongly dependent on the geometrical dimensions and features of the cavities. As a result, uncertainties in the exact shape and geometry play a significant role in determining the singing behavior of corrugated pipes. While predictive modelling for idealized periodic structures is well established, this paper focusses on the sensitivity analysis and uncertainty quantification (UQ) of uncertain geometrical parameters using probabilistic models. The two most influential geometrical parameters varied within this study are the cavity width and downstream edge radius. Computational Fluid Dynamics (CFD) analysis was used to characterize the acoustic source. Stochastic collocation method was used for propagation of input parameter uncertainties. The analysis was performed with both full tensor product grid and sparse grid based on level-2 Clenshaw-Curtis points. The results show that uncertainties in the width and downstream edge radius of the cavity have an effect on the acoustic source power, peak Strouhal number and consequently the whistling onset velocity. Based on the assumed input parameters distribution functions, the confidence levels for the prediction of onset velocity were calculated. Finally, the results show the importance of performing uncertainty analysis to get more insights in the source of errors and consequently leading to a more robust design or risk-management oriented decision.