摘要： The National Renewable Energy Laboratory and Acciona Energy North America are developing and implementing a process that addresses the issue of hydrogen buildup in parabolic trough power plants. We are pursuing a method that selectively removes hydrogen from the expansion tanks of the power plant to control hydrogen levels in the circulating heat-transfer fluid. As part of this effort, we are developing a sensor that measures hydrogen partial pressure in the expansion-tank headspace gas. During previous work, we demonstrated that our sensor design measures hydrogen levels over a wide partial pressure range from 1.33 mbar down to 0.003 mbar with an accuracy of ±20%. We acquired these results under limited conditions because we measured hydrogen partial pressure in gas mixtures of hydrogen and nitrogen only. In this paper, we report our most recent results in which we measured hydrogen partial pressures in gas mixtures that contain hydrogen, nitrogen, biphenyl, and diphenylether. Our results for this gas mixture were essentially the same as those obtained for the hydrogen/nitrogen mixture only, and they also had the same accuracy of ±20%. This level of accuracy is sufficient to evaluate the performance of a hydrogen mitigation process because hydrogen partial pressure in the expansion tank decreases more than an order of magnitude when the mitigation process is operating. We examined the palladium/silver membrane after completing these measurements and did not observe any signs of membrane deterioration or contamination due to exposure to biphenyl and diphenylether at 405oC. The membrane for these tests was 25 μm thick. For the next sensor design that we are developing in 2018, we are using a membrane that is 127 μm thick. This thickness is required for the membrane to withstand the 150 psi pressure drop across the membrane when exposed to actual headspace gas from the expansion tank. We performed initial tests in which we exposed a 127-μm-thick palladium alloy membrane to 400oC and 150 psi pressure drop and observed no signs of failure or gas leakage. Finally, we developed a transient model that accounts for hydrogen transport mechanisms within the sensor when measurements are made. Our model accounted for hydrogen diffusion through the boundary layer that is adjacent to the membrane and for hydrogen permeation across the membrane and into the permeate volume. Our modeling results agreed very closely with our experimental data when we assumed a boundary-layer thickness of 0.2 cm. This thickness is within the range of expected thicknesses for this gas composition, temperature, and pressure at minimum flowrate. Our model results showed that hydrogen diffusion through the boundary layer adjacent to the membrane is the rate-determining step for the transient pressure response when measuring hydrogen partial pressure.