Title of Dissertation: DEVELOPMENT OF A PHYSICS OF FAILURE MODEL AND QUANTITATIVE ASSESSMENT OF THE FIRE FATALITY RISKS OF COMPRESSED NATURAL GAS BUS CYLINDERS

The research presented in this dissertation details the development of a new probabilistic fracture mechanics model of corrosion fatigue failure of steel CNG bus cylinders. This model was used to estimate the frequency of leakage or catastrophic rupture, due to the propagation of a micro-crack on the inside, outside or transition surface, of hemispherical and flat-bottom cylinder designs, in assessing the fire and explosion fatality risks associated with a typical CNG bus. Quantitative assessment of the fire and explosion fatality risk was completed by analytically modeling the postulated fire scenarios from initial release of natural gas from a failed cylinder. The frequency of the initiating events, likelihood of subsequent events leading to a fire or explosive event was combined with the consequence of each event in a Probabilistic Risk Assessment (PRA) model to estimate the overall risk. Epistemic and aleatory uncertainties in the approach was evaluated using a combination of parametric modeling, conservative estimation and engineering judgment.

Direct computation of the fire fatality risk associated with diesel powered buses is possible because these are mature technologies for which historical performance data are available.  Due to the limited experience, fatal incident data for CNG buses fleets are minimal.  This study therefore had to rely on analytical modeling of failures, dynamics of fire initiation and propagation along with the subsequent events in this PRA approach.  The new methodology provides guidance on performing risk assessment of other novel technologies presently being developed or for which actuarial performance data is not available. 

This study predicts that the mean fire fatality risk for a typical CNG bus is approximately 23 fatalities per 100-million miles for all persons involved, including bus passengers.  Estimated CNG bus passengers mean risk is 14.4 fatalities per 100-million miles or 63% of fire fatalities.  Based on historical data, diesel school bus mean fire fatality risk is 0.091 and 0.0007 per 100-million miles for all people and bus-passengers respectively.  One can therefore reasonably conclude that CNG school buses are expected to be more prone to fire fatality by 250 times that of diesel buses, with the bus passengers being more at risk by over four orders of magnitude.  Explosion due to detonation and deflagration of a flammable vapor cloud within a bus or building, for which there is some historical events, is a major contributor, to this increased risk, a phenomenon not normally associated with diesel fuel.

The overall mean fire risk frequency has also been estimated at 2.23 x 10-3 fatalities/bus/year.  The 5% and 95% uncertainty bounds are 1.18 x 10-4 and 8.83 x 10-3 respectively.  These results provide the foundation for doing comparative analysis of CNG with other technologies by combining the estimated mean fire fatality risk, with the expected health and environmental benefits of using CNG powered buses.