Introduction
Most of my career was spent operating other people's designs. The calculations were rarely the problem — the sizing held, the metallurgy was right, the trip settings did what the datasheet promised. What surfaced instead, usually at three in the morning, were the decisions nobody had recognised as decisions: a drain that almost reaches the header, an isolation that cannot be proven, a sample point positioned for the model rather than the human.
This is the list I wish every design team taped to the wall — not a criticism of engineers, but the operator's half of a conversation that mostly happens too late, at the pre-startup review, when the steel is already welded.
Isolation Is an Operating Task, Not a Drawing Symbol
Every valve on the P&ID will one day be someone's isolation plan. Design for that day:
- First line break needs proven isolation. A single valve between a person and hydrocarbon inventory is a promise, not a proof. Where maintenance will be routine — filters, instruments, pumps, pig traps — the operator needs double block and bleed or a spade position, and the bleed needs somewhere safe to go.
- Spectacle blinds and spades need handling. A 12-inch spade at the top of a rack with no davit, no platform, and no swing clearance is an isolation that will be deferred — and deferred isolations shape how plants eventually get hurt.
- Think in isolatable maintenance envelopes. Can the produced-water pumps be isolated without dropping the separator? Can one filter be changed at full rate? The envelope boundaries decide whether maintenance costs an hour or a shutdown.
The Plant Lives at Turndown
The design case gets the attention; the plant spends its life somewhere else — early wells ramping up, late life declining, one train down, summer versus winter. Facilities are operated across their envelope, not at their nameplate:
- Control valves sized elegantly for design flow can be nearly closed at year-one rates — hunting, eroding, and teaching the operators to distrust the loop. Check the rangeability across the real production profile, not just the design point.
- Margins between operating and trip points get consumed by the drift of real operation. A trip set close enough to normal that a compressor start dips through it converts routine actions into unit trips — and, soon after, into pressure to move the trip.
- Minimum-flow behaviour matters as much as maximum: separators that cannot hold level at low rate, heaters cycling on and off, flow meters below their floor. Ask what the panel looks like at 30% throughput.
Commissioning Provisions Are Cheaper in the Model
Once hydrocarbons arrive, every missing provision is either a hot-work permit or a workaround. Designed-in from the start, commissioning and start-up provisions cost almost nothing:
- High-point vents and low-point drains that actually cover the flushing, drying, and purging sequences — traced against the procedure, not sprinkled by habit.
- Spool pieces and break flanges where temporary strainers, blinds, and test spades will live.
- Utility stations, nitrogen connections, and instrument-air drops where the work will actually happen.
- A start-up recycle or pressurisation route that does not improvise through equipment never meant for it.
Give the Field Its Instruments
The control system sees the plant; the operator walking it should too. Local pressure gauges either side of filters, temperature indication where hands cannot judge it, level gauges readable from the deck, and sample points that a person can reach with both feet on grating, at working height, with a drain for the flush volume — positioned where the sample answers the process question, not where the pipe was convenient. If taking the routine sample requires a ladder and three joints of tubing patience, the data quality tells you so before the lab does.
Documentation the Night Shift Can Use
The operating organisation inherits thousands of documents and uses perhaps twenty of them weekly. Make those twenty excellent:
- A cause-and-effect matrix readable in one page per system — when the unit trips at 3 a.m., the question is "what tripped it and what else just closed", answered in seconds.
- Set-point and alarm registers that match the plant — a register that disagrees with the DCS breeds exactly the informal changes it was meant to prevent.
- As-builts that are actually as built. The red-line discipline of the last month of construction determines the trustworthiness of every drawing for the next thirty years.
Walk It Down Before You Call It Done
Actions closed on paper have a way of reopening in the field. The HAZOP said a restriction orifice; the field has a spool. The drawing shows a drain to closed system; the pipe stops at a plug. The final walkdown — engineering, operations, and the punch list together — is where the paper plant and the welded plant reconcile, and it is the cheapest audit the project will ever run. The PSSR exists as the formal last gate, but it works far better as confirmation than as discovery.
Conclusion
None of this argues for gold-plating. Operators are pragmatists; most of the list above is layout, valve selection, and drawing discipline — near-zero-cost decisions when made in the model, and permanent irritations when discovered after first oil.
The design habit that captures all of it: for each system, imagine operating it — the isolation for the first filter change, the panel at half rate, the sample round in the rain, the trip diagnosis at 3 a.m. Designs produced by people who have run that film are the ones operators quietly thank for decades.
