Performance trends and control strategies for the Schatz Solar Hydrogen Project

This thesis reports on the long-term performance and compares two system arrangements of the Schatz Solar Hydrogen Project (SSHP), an off-grid solar hydrogen research project with twenty years of history. The SSHP is designed to continuously power a 500 W load and includes a 7 kW mono-Si photovoltaic (PV) array, a 6 kW alkaline electrolyzer, a 1 kW PEM fuel cell, and 3.8 kg of H2 storage in tanks with a volume of 5.7 cubic meters. For the PV array, I observed array-level performance degradation of 1.3%/year. For the electrolyzer, I observed an unexpected improvement in the voltage efficiency over time (0.8%/year) and identified a source of measurement error that may be endemic to alkaline electrolyzers in research applications. In 2006 my colleagues and I redesigned the power-handling and control system, shifting from a "direct connect'' arrangement (utilizing relays for equipment connections) to a "decoupled'' arrangement (utilizing DC/DC converters for connections). After correcting for long-term performance trends, I made comparisons between the "control efficiency" of the arrangements, a measure of system control effectiveness compared to the ideal case. The decoupled system is more effective than the original SSHP direct-connect system based on control efficiency (78% vs. 74%). The counter-intuitive trend I found in electrolyzer performance decreased the effectiveness of the direct connect system, which would have performed better with a correct prediction of the trend in the design phase. For decoupled systems, the choice of buffer batteries and control software for controlling allocation of power on a second-to-second timescale are critical to the performance.