Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Engineering (PhD)

Degree Level



Chemical Engineering


Jerry Havens

Committee Member

Edgar Clausen

Second Committee Member

Robert Beitle

Third Committee Member

Rick Couvillion

Fourth Committee Member

Tom Spicer


Pure sciences, Applied sciences, Health and environmental sciences, Accidental releases, Dense gas, Explosions, Gravity currents, Molecular diffusion


The accidental release of hazardous, denser-than-air gases during their transport or manufacture is a vital area of study for process safety researchers. This project examines the importance of molecular diffusion on the developing concentration field of a gas gravity current released into a calm environment. Questions which arose from the unexpectedly severe explosion in 2005 at Buncefield, England were of particular interest. The accidental overfilling of a large tank with gasoline on a completely calm morning led to a massive open air explosion. Forensic evidence showed that at the time of ignition, a vapor cloud, most of which now appears to have been within the flammability limits, covered approximately 120,000 m2. Neither the severity of the explosion, nor the size of the vapor cloud would have been anticipated.

Experiments were conducted in which carbon dioxide was released from a sunken source into a one meter wide channel devoid of any wind. These experiments were designed in such a way as to mitigate the formation of a raised head at the front of the gravity current which would have resulted in turbulent entrainment of air. This was done to create a flow in which molecular diffusion was the controlling form of mixing between the carbon dioxide and air. Concentration measurements were taken using flame ionization detection at varying depths and down channel locations.

A model of the experiments was developed using COMSOL Multiphysics. The only form of mixing allowed between carbon dioxide and air in the model was molecular diffusion. In this manner the accuracy of the assertion that molecular diffusion was controlling in our experiments was checked and verified.

Experimental measurements showed a large variation of gas concentration with depth of the gravity current at the very beginning of the channel where the gas emerged up from the sunken source and began flowing down channel. Due to this variation, molecular diffusion caused the vertical concentration profile to get more uniform as the gravity current flowed down the channel. A COMSOL model was developed which showed an overall increase in the depth of the flammable region of a cloud with increasing time, due to this effect.