Hydrogen Fire Safety
Use of hydrogen as an industrial chemical in the manufacture of transportation fuels for upgrading of petroleum and non-petroleum feed stocks is expected to increase as we search for energy solutions in a carbon constrained world. Hydrogen as an energy carrier in stationary and transportation applications requires careful consideration of safety related matters and continuing evolution of design standards. Understanding the sources of and conditions that can lead to ignition is important to advancing safe operating criteria and application of design standards. One such ignition mechanism that is nearly unique to hydrogen is the release of compressed hydrogen into air, as gas release forms a shock wave that heats air in contact with expanding H2 leading to ignition.
The spontaneous ignition caused by the interactions between transient shocks produced by pressure boundary failure and bodies of different shape and location in the emanating transient flow is under study at The Fuels and Combustion Research Laboratory via detailed numerical simulation and compressed hydrogen burst disk failure experiments.
For the numerical studies, an efficient adaptive and multiscale algorithm has been developed and extensively tested to carry out high fidelity numerical simulations using detailed chemical kinetics and diffusive transport. Two-dimensional simulations have been conducted that reveal transient interaction of planar normal shock (simulating that generated by a compressed hydrogen release into air) with wedge-shaped blunt bodies. The computational results demonstrate very rich geometry and pressure dependencies in terms of spontaneous ignition behavior and its transition to turbulent sustained burning. For a blunt body at fixed position and shape, there is a critical driver/driven gas pressure ratio above which spontaneous ignition occurs. The critical pressure ratio was found to strongly depend on the thickness and position of the blunt body. During the spontaneous ignition, a triple flame structure is first initiated close to the blunt body surface, and it propagates through a large volume of stratified mixture. The computational results show promise for investigating conditions and geometries difficult to study experimentally.
Experimental studies have progressed more slowly due to a need to conduct them in open environments. The failure of a burst disk at hydrogen pressures from 800 to 1700 psig has been utilized to generate a transient shocks in air, which are then impacted on two dimensional wedge structures downstream of the point burst disk location. Schlieren imaging has been utilized to observe the initial shock properties, which are found to be far from those typically assumed in all computational modeling work, including the present studies. The shock front is thick and multi-dimensional in character, due to the statistical fracture of the disk, and as a result likely decays more rapidly than is ideally suggested. Small lengths of constant diameter flow downstream of the burst disk location result in producing more planar, ideal character, but do not remove the multi-dimensional effects entirely. Initial experiments, conducted in November 2008 confirm that spontaneous ignition can be generated by blunt body interactions on two dimensional wedges at relatively low pressures <1700 psig bodies placed in close proximity to the burst disk location. A concise summary of this work is available here.