Event Title

Computational Modeling of Hurricane Wave Forcing on Bridge Decks

Start Date

8-21-2014 3:20 PM

End Date

8-21-2014 3:45 PM

Abstract

Wave forcing experiments previously conducted in the University of Florida wave tank were computationally modeling using CD-adapco's Star-CCM+. Data from the experiments were used to calibrate the computational model. First, the wave signal itself was extensively studied using four wave generation methods - by oscillating fluid motion relative to a stationary wall boundary; by using mesh morphing to model a moving wall boundary similar to a typical piston-style wavemaker; implementation of linear wave theory; and implementation of fifth-order wave theory. Ultimately, the oscillating fluid-motion method was discarded because the latter three methods appeared to reproduce the wave signal more accurately. Associated regression coefficients between modeled and physical wave signal data were 0.81, 0.92, and 0.88 for the piston method, the linear method and the fifth-order method respectively. Once an adequate wave signal had been generated, a comparison was made between experimentally-obtained force data and data from the computational analysis. While the forcing pattern was reproduced with high-levels of accuracy both in terms of amplitude and period, the physical model appeared to behave more non-linearly than the modeled results. Additionally, downward force was not properly simulated because the computationally-modeled bridge deck did not “drain” as it did during the experiments. The computer model was modified to include a simulated “drain,” which significantly improved the model’s accuracy. Finally, the bridge was rotated to study the effects of attack angle for bridges subjected to wave attack.

Contact Information

Mr. Raphael W. Crowley

Assistant Professor

University of North Florida

Department of Building Construction Management

Building 50, Room 2400

Jacksonville, FL 32224

Phone: 904-620-1847

email: r.crowley@unf.edu

Speaker's Biography

Dr. Crowley has been an assistant professor at the University of North Florida since August 2013. Previously, he was a postdoctoral researcher at the University of Florida from January 2011 through August 2013. Dr. Crowley’s research focus is fluid-structure interaction – particularly waves striking bridge decks and bridge scour. As a postdoc/graduate student, he developed a state-of-the-art instrument for measuring cohesive sediments’ rate of erosion – the Sediment Erosion Rate Flume (SERF). This device has been used for scour testing of sediment related to the Tappan Zee Bridge’s replacement in upstate New York and the McMicken Dam in Arizona. Dr. Crowley coded a computational model of the device using CD-adapco’s Star-CCM+ to better-estimate the shear stresses in the instrument. More recently, Dr. Crowley and his students have applied the CFD skills they gained modeling the SERF to model hurricane waves striking bridge decks. Dr. Crowley has taught several courses related to fluid dynamics including Computer and Numerical Methods for Civil Engineers, Maritime Construction, and Coastal Structures. He is a registered professional engineer in Florida with his BS from Bucknell University in Civil and Environmental Engineering, his MS from the University of Florida in Coastal and Oceanographic Engineering, and his Ph.D. from the University of Florida in Civil Engineering.

This document is currently not available here.

Share

COinS
 
Aug 21st, 3:20 PM Aug 21st, 3:45 PM

Computational Modeling of Hurricane Wave Forcing on Bridge Decks

Wave forcing experiments previously conducted in the University of Florida wave tank were computationally modeling using CD-adapco's Star-CCM+. Data from the experiments were used to calibrate the computational model. First, the wave signal itself was extensively studied using four wave generation methods - by oscillating fluid motion relative to a stationary wall boundary; by using mesh morphing to model a moving wall boundary similar to a typical piston-style wavemaker; implementation of linear wave theory; and implementation of fifth-order wave theory. Ultimately, the oscillating fluid-motion method was discarded because the latter three methods appeared to reproduce the wave signal more accurately. Associated regression coefficients between modeled and physical wave signal data were 0.81, 0.92, and 0.88 for the piston method, the linear method and the fifth-order method respectively. Once an adequate wave signal had been generated, a comparison was made between experimentally-obtained force data and data from the computational analysis. While the forcing pattern was reproduced with high-levels of accuracy both in terms of amplitude and period, the physical model appeared to behave more non-linearly than the modeled results. Additionally, downward force was not properly simulated because the computationally-modeled bridge deck did not “drain” as it did during the experiments. The computer model was modified to include a simulated “drain,” which significantly improved the model’s accuracy. Finally, the bridge was rotated to study the effects of attack angle for bridges subjected to wave attack.