Design life is not remaining life

A nameplate says 25 years. The plant is 31 and still running. Somewhere in the organisation, someone decided that was acceptable. The question worth asking is whether anyone wrote down why.

Design life is not a prediction. It is a stack of assumptions made before the plant ever saw a hydrocarbon: a corrosion allowance selected, a corrosion rate assumed, a fatigue spectrum estimated from an operating profile that existed only on paper. Divide the allowance by the assumed rate and you have a design life. It was a planning figure, useful for procurement and for setting an initial inspection philosophy. It was never a measurement.

Remaining life is a measurement. It comes from wall thickness data, from identified damage mechanisms, from what the plant has actually processed rather than what it was designed to process.

But there is a second half to that sentence, and it is the half that gets lost. Remaining life is not a property of the equipment alone. It is a property of the equipment and the duty you intend to give it. Change the duty and the number changes, even though nothing about the steel has changed at all.

Marsh's analysis of the hundred largest hydrocarbon losses puts the average age of the global refining and petrochemical fleet at around 45 years, and observes that failures in ageing plant increasingly stem not from maintenance shortfalls but from inspection coverage, inspection intervals, data interpretation and fitness-for-service decisions. Every item on that list is an analytical failure rather than a physical one: the steel did what steel does, and the assessment of the steel was wrong.

What that looks like in practice

MES recently completed an asset replacement plan for a mature offshore field, ahead of a redevelopment that would roughly double throughput. Thirteen pressure vessels and eighteen pipelines were assessed against two horizons: continued service to the redevelopment start-up, and continued service through the operating period that follows it.

Against today's duty, the fleet looked healthy. Seventeen of the eighteen pipelines had ample margin to reach start-up. One six-inch oil line carried more than eighty years of calculated remaining life on its measured corrosion rate.

Against the planned future duty, the answer is two years

Nothing had been discovered about the pipe. Same wall thickness, same defect population, same inspection data. What changed was the modelled corrosion rate for the new duty, which the bounding production case put an order of magnitude above the present one. Across the fleet the pattern repeated: a set of assets that comfortably passes today's test, and a materially smaller set that passes the test that matters, which is the duty the business has already decided to buy. The static equipment told the same story. Eleven of thirteen vessels could reach start-up; only six could demonstrate the minimum operating period beyond it.

Two further findings generalise well beyond this asset.

The answers were conditional, not absolute. The water injection lines showed no corrosion-driven life limit at all, provided dissolved oxygen control, deaeration, scavenger and biocide performance hold. The oil line results assumed corrosion inhibitor availability and efficiency at ninety percent each; without inhibition, the modelled rate rises roughly fivefold. That is not a caveat at the back of a report. It is the result. Remaining life is a statement about metal and about operating discipline, and the second half is the easier one to lose.

And ten units could not be assessed at all, because the inspection records did not exist. They were reported as unassessed rather than assumed sound. An asset you cannot assess is a finding in its own right, not a gap in the report.

Four terms, in the order they are used

• Design life: The original assumption set, fixed before commissioning.
• Remaining life: What current condition data supports, given identified damage mechanisms, measured rates and a stated future duty.
• Fitness-for-service: A formal assessment of whether equipment containing a known flaw is suitable for continued operation, and under what conditions.
• Integrity operating windows: The process limits within which the assumed degradation rates hold. Operate outside them and the remaining life calculation is void.

A defensible answer rests on four things: damage mechanism identification specific to the actual service, inspection coverage that reflects risk, interpretation by people who understand what generated the data, and integrity operating windows that are current. Coverage is where most programmes are quietly weakest, because it is expensive and its absence is invisible until it is not; windows set at commissioning and never revisited describe a plant that no longer exists.

Life extension is the conclusion those four support, not a starting point. An organisation that begins with "we need another ten years" and works backwards will produce a document. It will not produce a case.

This is a capital question, not only a technical one

Life extension decisions are capital decisions: continue, invest, or exit. Lenders, insurers and acquirers all ask the same question in different vocabulary: what does the condition data support, and who assessed it? An asset with a well-evidenced remaining-life case is cheaper to insure, easier to finance and simpler to sell. One running on a nameplate figure and institutional confidence is none of those things.

Design life told you what the designers assumed. Remaining life tells you what you have, for the duty you are actually planning. If your organisation is using the first number to make decisions that need the second, that is worth correcting before the next turnaround rather than after it.