Gas Distribution Equipment in Hydrogen Service—Phase II

The hydrogen permeability of three different types of commercially available natural gas polyethylene pipe was determined. The data collected show that the polyethylene pipe is 4 to 6 times more permeable to hydrogen than to methane. Ring tensile tests were conducted on permeability- exposed and as-received samples. Comparisons of apparent strength between exposed and as-received specimens show little difference of any statistical significance. Ring tensile tests of exposed pipe samples, after testing to failure, show more elongation than asreceived samples. Hydrogen-methane gas leakage experiments were conducted. The results show no selective leakage of hydrogen via Poiseuille, turbulent, or orifice flow (through leaks) on the distribution of blends of hydrogen and methane. The leak rates increase with increasing pressure and decreasing specific gravity. Analytical studies show that approximately 13.7 x 106 SCF of natural gas could be lost annually by permeation in the 154,000 miles of installed plastic pipe in the United States; if hydrogen were distributed, 67.4 x 106 SCF could be lost. The unaccounted-for gas loss in the United States is about 236 x 10 9 SCF annually; if hydrogen were distributed, the unaccounted-for gas loss could be 693 x 109 SCF. Permeation losses are less than 0.01% of the unaccounted-for gas losses. A D p' p " q S t Nomenclature = area, cm2 = diffusion constant, cm 2 /s = thickness, cm3 =DS = permeability constant (standard cm3 cm/s cm2 atm) = high pressure, atm = low pressure, atm = volume of gas diffusing, cm3 = solubility constant (standard cm 3 /cm 3 material atm) = time, s