HyStack

Challenge

Today, challenges arise in hydrogen production using alkaline electrolysis systems when the process is scaled up. Dissolved hydrogen can diffuse through the membrane that separates the oxygen and hydrogen sides, potentially creating an explosive mixture of oxygen and hydrogen. Without early detection, local H₂-in-O₂ levels can exceed safety limits and trigger minor explosions, damaging the cell stack and causing costly downtime.

Currently, monitoring typically occurs in a pressure-reduced side stream after the separator, which mixes gas flows from multiple cell stacks and introduces delays of several minutes. This solution is far too slow and makes it difficult to pinpoint a single faulty stack or take action before explosive mixtures build up.

Solution

The HYSTACK project is developing a robust, in-line hydrogen sensor that measures H₂ concentration in the O₂ solution directly in the circulating alkaline electrolyte for each cell stack. The sensor concept combines SulfiLogger’s microsensor platform, which is being developed to withstand 40 wt% KOH, temperatures up to 90 °C, and pressures up to 55 bar, with a measurement range of 0–5% H₂ and a response time of less than 30 seconds for fast, stack-specific diagnostics.

HYSTACK is a direct continuation of the EUDP project SafePtX: SafePtX demonstrated SulfiLogger’s pressure-tolerant H₂ sensor in gas lines and paved the way for SIL certification; HYSTACK extends this to in-line liquid-phase measurement inside alkaline stacks. The same safety mission, but with a different location and operating conditions. Stiesdal Hydrogen supports the development by, among other things, testing the sensors at their facilities.

Impact

Increased safety and uptime. Immediate, stack-specific detection of damage reduces explosion risk and enables isolation of a single faulty cell stack instead of shutting down and disassembling all stacks, significantly reducing maintenance time (demonstrated potential up to 90%) and avoiding replacements that can cost up to ~1 million DKK per incident. In practice, downtime of 1–2 weeks can be reduced to about one day.
As the sensor is designed for modular MW-scale systems, it supports safe scaling of industrial PtX.