What constrains an ammonia synthesis loop when the hydrogen varies

A green ammonia plant is an electrolyser, an air separation unit, compression, and a synthesis loop. The first three tolerate variation reasonably well. The loop does not, and it therefore sets the flexibility of the whole plant.

The reason is that ammonia synthesis is a catalytic reaction running at high pressure and temperature, with a recycle, and the catalyst has an operating window it has to stay inside. The loop is a system with substantial thermal and pressure inertia whose behaviour at low throughput is qualitatively different from its behaviour at design flow.

Almost everything published on this is simulation, frequently of plants that have not been built, and the models are more optimistic about turndown than operating experience with conventional loops suggests. Treat a published turndown figure as an upper bound until it has been demonstrated on the loop in front of you.

The synthesis reaction is exothermic, and in normal operation the heat it releases is what keeps the bed at temperature. At reduced throughput less heat is released, while heat losses do not fall proportionally, so the bed tends to cool.

Below a certain point the reaction rate falls enough that the bed cannot sustain itself, and recovering from that is not a matter of turning the flow back up. It requires reheating, which takes time and external energy, and this is the practical floor on how far the loop can be turned down.

The window has an upper bound too. Running hot to compensate accelerates catalyst ageing, so the operating strategy that maximises today's production can shorten the interval to the next catalyst change, which is a large planned outage. That trade is worth quantifying rather than resolving by instinct.

Several constraints bind at low flow and the binding one is plant-specific, which is why a generic turndown figure is not useful.

Thermal self-sustainability, as above. Compressor limits, since the synthesis gas compressor and the recycle compressor both have minimum flow requirements and anti-surge protection that will act before anything else does. Heat exchanger behaviour, where the loop's internal heat integration was designed around a flow that is no longer present. And separation, since ammonia is condensed out of the recycle and the condensation duty changes with flow and composition.

The useful question to a licensor or an operating team is not what the turndown ratio is but which constraint binds first at this plant, because that determines whether the limit can be moved and at what cost.

Inert gases, principally argon and methane depending on the feed route, enter with the make-up gas and do not react. They accumulate in the recycle and are removed by a purge stream, which necessarily removes some hydrogen and ammonia with them.

The purge rate is tuned to hold inert concentration at a level that balances conversion against loss. That tuning is done at design flow, and at part load the balance moves: make-up rate falls, recycle behaviour changes, and a purge rate that was correct becomes either wasteful or insufficient.

On a plant whose hydrogen supply varies daily this is not a one-off setting. It is a control problem, and it is one of the more tractable optimisations available because the measurement is straightforward and the economic consequence is direct.

Feed purity from the electrolyser side interacts here. Oxygen carryover and moisture are catalyst poisons, so purification duty matters more when the upstream is varying, and a purity excursion has consequences that outlast the excursion.

Hydrogen storage between the electrolyser and the loop is the standard answer and the main capital decision in a green ammonia plant. It converts a variability problem into a sizing problem.

The sizing question is what fraction of the resource variability the buffer has to absorb to keep the loop above its minimum stable throughput for a target proportion of the year. That depends on the resource profile, the electrolyser turndown, the loop minimum and how much curtailment is acceptable.

It is a coupled optimisation rather than a series of separate ones, which is the part most often got wrong. A larger buffer allows a stiffer loop; a wider electrolyser turndown allows a smaller buffer; accepting more starts on the loop allows both to be smaller and costs catalyst life. Choosing each independently produces a plant where one element is oversized and another is binding.

Ammonia storage downstream is a different and much cheaper buffer, but it decouples production from despatch rather than decoupling the loop from the electrolyser, so it does not help with this problem.

A full loop shutdown and restart is expensive in time before it is expensive in anything else. Bringing the catalyst bed back to temperature, re-establishing pressure and composition, and reaching on-specification product takes hours, and those are hours of hydrogen production that has to go somewhere or be curtailed.

There is a catalyst cost as well, since thermal cycling contributes to ageing, and there is a consumables and utilities cost in the restart itself.

Quantifying the total before choosing an operating strategy is the point. A plant that knows what a stop costs can decide rationally between holding the loop at minimum on stored hydrogen, stopping, or curtailing upstream. A plant that does not know will default to whatever the control room finds least stressful, which is usually not the cheapest option.

The same arithmetic problem appears upstream on the stack, where the cost of a cycle also has to be known before idling and stopping can be compared. Both numbers come from measurement rather than from a datasheet.