Part load operation of alkaline electrolysers

Alkaline water electrolysis is a well-established and relatively low-cost technology to convert electric power into hydrogen. Coupled with a drying and oxygen purification step, it easily produces hydrogen with a purity above 99.9%. However, this technology has been developed for nominal (near full-load), steady-state operation and does therefore struggle with part-load and intermittent operation. The increasing share of renewable electricity into the grid and the intention to store renewable electricity through hydrogen storage are consequently challenging for alkaline water electrolysis.

Challenges linked to part load operation

During electrolyser operation, oxygen is produced on the anode side and hydrogen on the cathode side, with the separator preventing most transfer of oxygen towards the cathode side and hydrogen towards the anode side. The produced gases are therefore not pure, but the contamination level is low. Problems arise when operating at lower loads: the leakage through the separator roughly stays the same, while the production rate decreases, so that the relative influence of leakage increases, thereby increasing the contamination level.

Under nominal operation, the low level of contamination is dealt with by adding a purification step, taking oxygen out in the hydrogen stream and taking hydrogen out in the oxygen stream. So, why does it become a problem when contamination increases? Because more contamination means mixtures on both sides coming closer to the explosivity range leading to safety concerns. Oxygen and hydrogen are mutually highly reactive with combustion/explosion possible as soon as a ratio of 4 %vol O2 + 96 %vol H2 or 96 %vol O2 + 4 %vol H2 is reached. For safety reason, an electrolyser is therefore shut down as soon as any of the minority species reach the 2 %vol mark in order to allow for an orderly purging at a low risk level.

When operating at part load, the contamination level is increasing with lowering load and the minimal allowed load corresponds to the 2 %vol limit in contamination in any stream. Depending on the electrolyser design, such a limit is reached between 10% and 25% of nominal operation. This means that any fluctuation of power supply below 25% of nominal is a potential shutdown of the electrolyser. Beyond the obvious cut in hydrogen production, such shutdowns are unwanted for 2 main reasons:

  1. Restarting the electrolyser takes time, so that the shutdown will be longer than the time required to get the power source back to 25% of nominal
  2. Each shut-down/start-up ages the electrodes, so that frequent on/off events reduce the expected lifetime of the equipment, thereby increasing maintenance and capital costs. Nickel electrodes, for example, are known to degrade significantly after 5000 to 10,000 start/stop cycles

Why is the part load limit range of alkaline water electrolysers spanning 10% to 25%?

This is related to the management of the alkaline electrolyte: either through separate circuits for the oxygen and hydrogen streams or through a common path.

The mixture of gas and electrolyte for each electrode is always routed separately to separators to extract most gas but some residual gas (intended gas and impurities) remains embedded in the liquid. If the electrolyte streams leaving the separators are mixed before being fed back to each electrode, there is a mixing of residual gases and an increase in residual hydrogen on the oxygen-producing side as well as an increase in residual oxygen on the hydrogen-producing side, thereby increasing contamination level and imposing a shutdown at a higher load. This setup would typically call for a shutdown near the 25% limit.

Keeping the electrolyte streams independent after the separators prevents such added contamination and pushes the safety limit closer to the 10% load limit. However, this solution comes with its own drawbacks: water consumption on the hydrogen-producing side (cathode) increases electrolyte concentration, while water production on the oxygen-producing side (anode) decreases electrolyte concentration. Some liquid level balancing and electrolyte concentration balancing is still required to maintain pumping capability (liquid level) and optimal electrolyte concentration (electrolyser efficiency depends on electrolyte conductivity). For example, using NaOH as electrolyte, peak conductivity is ~65 S/m at 50°C and reached for a concentration just below 20%mas, while using KOH, peak conductivity is ~95 S/m at 50°C and reached for a concentration just above 30%mas.

Effect of pressure

Operating the electrolyser under higher pressure is a common strategy to save some compression work down the line. However, a higher pressure within the electrolyser also increases the level of contamination as:

  1. Diffusion through the separator increases
  2. More gas can be dissolved in the liquid

Part load operation under pressure is therefore the most difficult type of conditions. Literature review shows that the mixed electrolyte cycle is highly sensitive to the combination of pressure and current density:

  • At atmospheric pressure, the current density can be reduced down to 0.05 A/cm2 before the safety limit of 2 %vol H2 in O2 stream is reached
  • At 10 bar operating pressure, the same limit is reached for a current density of 0.5 A/cm2 and any lower load operation becomes very dangerous
  • At 20 bar operating pressure, with a maximum current density of 0.7 A/cm2, operation below the 2 %vol safety limit is impossible with gas impurity already at 2.5 %vol

Part-load, pressurised operation therefore calls for separated electrolyte circuits with mandated additional strategies to balance fluid level and electrolyte concentration.

Voltage management at power cuts

If gas impurity can be maintained below the safety limit, short-duration excursions below the low limit (generally defined as the minimum acceptable load for continuous operation) can be tolerated but only with a very careful management of electrode voltage. Cathode degradation is significantly accelerated below a voltage of 0.25 V, so that a shutdown can be ordered when closing on this value even if gas impurity is still low. The complex multi-layer electrodes can temporarily act as capacitors therefore delaying the voltage drop after a complete power loss and allowing for maintained operation if power is quickly recovered. Experiments have shown possible transients as long as 10 min.

Temperature management

Part-load operation also brings new difficulties with temperature management. Best efficiency is reached when the electrolyte temperature is between 50 and 80°C (higher electrolyte conductivity, while not inducing component degradation). To avoid wasting electricity in a separate heating unit, heating exclusively relies on capturing wasted heat from electrolysis reactions and a cooling system avoids exceeding 80°C during full load operation in a hot ambient. Yet, when reducing load, heat of reaction is low and reaching 50°C can become challenging.

Dealing with part load through operational strategy

To deal with power fluctuations and excursions below the 10% to 25% limit, external strategies are required:

  • Either absorbing power fluctuations through energy storage devices
  • Or implementing an electrolyser stack sub-division

The latter consists in operating a sub-set of the total electrolyser so that the load remains high on that part of the electrolyser. The management of load split and thermal state as well as the recording of individual usage history for each stack sub-set when dealing with electrolyser ageing and maintenance become the difficulties.