Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Chemical Engineering


Tom Spicer

Committee Member

Heather L Walker

Second Committee Member

Karthik Nayani

Third Committee Member

Keith Walters

Fourth Committee Member

Jim Leylek


Gas Release;Jack Rabbit II;Laser Doppler Velocimetry;Particle Image Velocimetry;Wind tunnel modeling


Accidental release of toxic chemicals can put workers and nearby populations at significant risk. Consequence assessment of accidents in urban environments is of particular interest. Urban geometries create wind channels between buildings, along with particularly dangerous areas downwind of buildings, where near-stagnant flow is present. Furthermore, recirculation zones can be formed between buildings, trapping high concentrations of toxic gases at ground level, particularly for denser-than-air gases. In 2015, the Department of Homeland Security conducted the Jack Rabbit II Field Test (JR-II) at Dugway Proving Ground, UT. During JR-II, chlorine was released at the center of a Mock Urban Environment (MUE) and three types of data were collected: visual recordings of the releases, concentration measurements at specific locations, and wind velocity and direction. Due to the toxicity and corrosive nature of chlorine, some measurements could not be made during the JR-II trials. In a subsequent study called Jack Rabbit II – Special Sonic Anemometer Study (JRII-S), wind velocity and turbulence were measured at points of interest within the MUE. Simulations of hazardous gas releases can be made in wind tunnel models, which provide for repeated tests with greater statistical analysis. Moreover, changes to test conditions can be made with relative ease and at lower costs. A 1:50 scale model of the JR-II MUE was constructed in the wind tunnel of the University of Arkansas. Laser Doppler Velocimetry (LDV) measurements show agreement between the approach wind characteristics in the tunnel model and field test. LDV velocity measurements within the MUE wind tunnel model also agree with the anemometry results obtained in JRII-S. Visualization of simulated chlorine releases in the wind tunnel shows good reproduction of release characteristics observed in the JR-II field tests. Mathematical analysis investigated if it was possible to obtain accurate 3D velocity measurements from a simplified 2D LDV setup. Error quantification showed that the simplified setup yields valid approximations for the vertical velocity and turbulence directions depending on the flow field characteristics, but high error is observed in the velocity and turbulence in the cross-wind component. After demonstrating that the wind tunnel models the JR-II and JRII-S field tests, Particle Image Velocimetry (PIV) experiments were conducted to provide velocity measurements that could not be obtained during JR-II. The wind tunnel model yields more comprehensive results than the pointwise measurements taken during JRII-S using sonic anemometers. Velocity measurements in a horizontal plane parallel to the ground were used to compare to computational fluid dynamics (CFD) simulations of the JRII-S field tests found in the literature. PIV measurements taken in the wind tunnel model pointed out concerns with those simulations. Reproduction of the JR-II field tests showed the velocity field in a section of the MUE before and during the chlorine release. The effect of the added momentum of the release in the flow field and in the height of the dividing streamline at CONEX 11.4 was investigated, with bigger impact observed at lower atmospheric momentum. Key Words: Wind tunnel modeling, gas release, Jack Rabbit II, Laser Doppler Velocimetry, Particle Image Velocimetry