Hydrogen Leak Detection Method Derived using DCOV Methodology

Hydrogen is the most abundant element in the universe, accounting for more than 90% of the molecules and more than 75% of the mass [1]. However, due to the small molecule size and high buoyancy, it is not available in it’s free form on Earth. In recent years, hydrogen has gained the attention of the automotive industry [2–12] as an environmentally friendly alternative fuel. As a fuel, hydrogen is unique - it is odorless, colorless, tasteless, and burns invisibly in sunlight. Detection solutions such as the odorants used in natural gas are not yet feasible for automotive hydrogen because the available additives can poison the fuel cell catalyst. Additionally, the lower flammability limit of hydrogen is lower, and the flammability range wider, than fuels such as gasoline. For these reasons, hydrogen leaks must be detected reliably so that appropriate action can be taken [13]. Placing hydrogen sensors in areas of potential accumulation is presently the most widely used solution for detecting hydrogen leaks in vehicle applications. However, cost, sensitivity to environmental factors, and reliability concerns are driving leak-detection alternatives that do not require chemical sensors. Ford Motor Company, in cooperation with Daimler and NuCellSys, has implemented a control strategy utilizing pressure decay measurements inside the hydrogen system to identify leaks. Although this strategy has shown to be very reliable and procedurally very simple, it temporarily overrides normal control of the fuel cell system and is time intensive in order to achieve the desired sensitivity and accuracy. In this paper, a leak detection method based on monitoring control signals in the hydrogen subsystem is proposed. Hydrogen system pressure is controlled based on system conditions and load. The setpoint is maintained utilizing closed-loop control. At idle, for example, the control effort required under defined environmental conditions can be predicted and shown to be repeatable. Deviation from the predicted range can be an effective indicator of leakage from the hydrogen subsystem. Detection can be fast and transparent to fuel cell system operation. Six-sigma methodologies were applied to develop the necessary transfer functions, define the thresholds at which different failure modes can be assumed, and determine the necessary mitigation actions.