Pure oxygen systems pose significant fire and explosion risk to equipment, operations, and personnel if a loss of oxygen containment occurs. In industrial oxygen systems, combustion of metal can be initiated by a variety of ignition sources. The combustion of process vessel walls, components, or piping can result in loss of containment of pressurized oxygen. Automatic shutdown and isolation systems are designed to detect loss of containment, incipient loss of containment due to process parameters (e.g., pressure drop), or internal fire conditions. The effectiveness of these isolation systems is dependent upon the speed with which they activate after the initiating event.
The prediction of failure modes and time scales can be an important tool in the design, operation, and maintenance of industrial high purity oxygen systems. Modeling of metal consumption in oxygen fueled fires requires consideration of multiple ignition mechanisms. Previously published experimentation has been motivated by the need to compare the relative burn resistance of metals and nonmetals under a variety of operating conditions, but has not produced a tool to inform the potential for loss of containment in oxygen systems based on the different ignition mechanism failure modes.
This work aims to develop a series of theoretical models using experimental results and kinetic relationships in order to predict loss of containment in piping for industrial systems for combinations of different failure modes. Using this type of analysis, process designers can make informed decisions on the design of high purity oxygen systems, the timescale required by emergency shutdown systems, and schedules for routine maintenance and inspection. © 2014 American Institute of Chemical Engineers Process Saf Prog 33: 299–304, 2014