Designed for Steady State, Damaged In Between

Every plant has a design case — the temperatures, flows and compositions under which the equipment is sized and the performance figures are guaranteed. That case is almost always steady state.

Steady state is the easy case. The damage happens in the transitions: starting up from cold, shutting down, changing load, recovering from a trip. These are the periods when the plant passes through conditions it was never specified for, and in most designs they get a fraction of the attention given to the hours in between.

WHAT A START-UP ACTUALLY DOES

The reactor begins cold and is brought up to temperature over hours. During that time conversion is incomplete, and tar forms in quantities the downstream equipment never sees in normal operation.

That tar moves into equipment that is also cold and condenses on surfaces that are not yet hot enough to keep it in the gas phase. A film forms. Some of it bakes on later and stays.

Sorbent beds see gas with the wrong composition at the wrong temperature, and saturate early. Catalysts pass through the temperature windows where they degrade fastest. Refractory expands and contracts, and every cycle costs it something that is not visible until the lining fails.

Then the plant reaches steady state, and everything that happened on the way there is embedded in the equipment.

SHUTDOWN IS NOT THE REVERSE

A planned shutdown can be controlled: load reduced gradually, temperatures brought down at a rate the refractory tolerates, the gas path purged before it cools.

A trip does none of this. Feed and heat stop, the plant cools at whatever rate it cools, and whatever gas was in the system condenses wherever it happens to be. Trips are where the sharpest damage occurs, and they are the transitions least under the operator's control. A design that assumes shutdowns will be planned has not planned for the ones that matter.

WHY IT IS AN EMISSIONS QUESTION TOO

During transitions the conditions that suppress unwanted compound formation — sufficient temperature, residence time, full oxidation — are precisely the ones not yet achieved, and the cleaning stages are outside their design window. Permits typically include specific start-up and shutdown provisions for exactly this reason. A plant that cycles often spends more of its life in that window.

THE CYCLING PENALTY

Two plants with the same total operating hours can age at very different rates depending on how many times they stopped and started. Long continuous campaigns accumulate wear slowly. Frequent cycling accumulates it in concentrated bursts, and the concentration does the damage.

Cycling frequency is therefore a design input, not an operating statistic. A plant designed for continuous operation and then run in cycles is outside its design basis every time it starts.

THE HONEST PART

Designing for transients costs money: auxiliary heating, purge systems, bypass routes so cleaning stages can be isolated until the gas is fit for them, refractory chosen for cycling rather than peak temperature, and control logic that sequences all of it automatically.

That is capital on a plant whose budget is being scrutinised at exactly the point when these systems are specified — and because transient wear shows up years later, the cost of leaving them out is invisible at that moment. Some transients will also always be uncontrolled. The honest design goal is not to eliminate their cost but to bound it.

WHAT TO ASK OF A DESIGN

How many start-stop cycles per year is the plant designed for, and what is that based on?

What happens to the gas path on an unplanned trip, and is the sequence automatic?

Which components are most exposed during warm-up, and how are they protected?

Are cycles counted and inspected against, or only hours?

Steady-state figures describe the condition the plant is in for most of its hours. The transitions decide how many of those hours it will have.

If the start-up and shutdown philosophy is a page at the back of the design document, that is worth a conversation.

Talk to us at zero-x.co