Relay feedback excitation for identification of Fuel Cell performance parameters

Fuel cell technology is emerging as a zero-emission power generation alternative. Durability is a critical factor and it is therefore important to monitor degradation parameters that can be used for prognostics of remaining useful life and to take precautions to optimize the lifecycle. The Giantleap project is founded by EU through the Horizon 2020 programme and focuses on these issues for automotive applications in Fuel-Cell Electric buses. One interesting parameter is the so-called low frequency intercept resistance (LFR). From basic studies of performance degradation, it has been observed that changes in this parameter correlate well with ageing, making it an interesting indicator to measure (Pivac et.al., 2018). A key issue is that dynamic excitation is needed to identify this parameter, since it is not observable in steady state. It is simply the low frequency point of the Nyquist diagram of the Fuel Cell impedance where the phase crosses through zero while passing from positive to negative phase angle. In a laboratory, this can be measured by performing Electrochemical Impedance Spectroscopy (EIS). However, this requires special equipment and is not practical (or possible) to do for a fuel cell stack that is an integrated component in an application.
In the control community, the celebrated relay feedback excitation method (Åström and Hägglund, 1995) is used to directly identify the ultimate- gain and frequency of an unknown process: the result can e.g. be used for tuning of PID controllers. Here, this excitation technique is adapted to obtain direct identification of the interesting low frequency intercept point (Halvorsen et. al., 2019). Results from simulations and from testing on fuel cell stacks will be presented. The simple relay excitation method can be implemented as a software component in typical fuel cell control systems without any extra hardware equipment.