Loop checking that is actually a check

Loop checking verifies that each measurement and each output is connected to what the drawings say it is connected to, reads correctly across its range, and behaves as intended. It is the last opportunity to find wiring and configuration errors before the plant runs.

It is also performed under maximum schedule pressure, at the end of construction, by people who are being asked when they will be finished. The result is that on many projects it degenerates into a signature exercise.

What a complete check covers

• Continuity: the field device is wired to the terminal, the card and the channel the drawings specify.
• Range and scaling: a known input at several points produces the correct engineering value in the control system, at zero, mid-scale and full scale.
• Direction: increasing the process variable increases the reading, and an increasing output opens the valve if that is the intent.
• Tag agreement: the tag on the device, the drawing, the control system and the display all match.
• Alarm settings: thresholds are configured at the values the rationalisation specifies, and annunciate.
• Failure behaviour: on loss of signal, the value goes where it should and quality is reported bad.
• For outputs, valve travel: the element actually moves, through full stroke, in the correct direction, with the correct fail position on loss of air or signal.

The last two are the ones most often skipped, and they are the ones that matter during an abnormal event.

Support teams serving several sites face coordination problems similar to consultants managing multiple clients. Further details are available via this resource.

For process facilities, the surrounding management framework is covered in OSHA process safety guidance.

Two people, two ends

A loop check requires someone at the field device and someone at the control system, communicating. Checks performed by one person simulating at the marshalling cabinet verify the system side and not the field side, which is where a substantial proportion of errors live.

The classic error a field-end check catches is the transposed pair: two similar transmitters wired to each other's channels. Everything reads plausibly and the plant runs with two measurements swapped, sometimes for years.

Three points, not one

Checking a single value verifies that something is connected. Checking zero, mid-scale and full scale verifies the scaling, and it catches the common errors — a range configured as zero to a hundred against a device ranged zero to sixty, a square root extraction applied twice or not at all.

The record is the deliverable

The loop check sheet should record the values observed rather than a tick. A sheet showing an applied input of 12 mA and an observed value of 50.2 percent is evidence. A sheet with a tick in a box is an assertion.

Recorded values also make the sheets useful later: when a measurement is suspected of drift years afterwards, the commissioning record establishes what it read when it was known good.

Handling the ones that fail

Loop checks find faults. The value of the exercise depends on what happens to them.

A punch list entry, an owner, a retest after correction, and a record of the retest. The failure mode is a fault noted, corrected informally, and never retested — which frequently means corrected incorrectly.

Scheduling it realistically

Loop checking is estimated by dividing the loop count by an optimistic rate. The realistic rate depends on access, on how far apart the two ends are, and on the fault rate, which is unknown until the work starts.

The practical approach is to check a representative sample first, establish the actual rate and fault rate, and re-plan. A project that discovers in the first fifty loops that one in six has a fault is in a different situation from one finding one in fifty, and the schedule should reflect that rather than the original assumption.

Partial and phased checking

Where a plant is commissioned in sections, loops are checked as sections become available, which spreads the work sensibly.

The risk is loops that cross section boundaries and loops modified after their check. Any loop touched after checking needs rechecking, and tracking that requires the punch list and the change process to be connected.

The checks that get deferred

Under schedule pressure, certain checks are routinely postponed to be completed during commissioning proper: valve stroking, failure mode verification, alarm testing.

Deferral is sometimes unavoidable. What matters is that it is recorded as an open item with an owner rather than quietly dropped, because the items deferred are consistently the ones that verify abnormal behaviour.

Sequencing against construction

Loop checking cannot start until the field device is installed, the cable is terminated at both ends, the control system is powered and the configuration is loaded. On a real project those four milestones arrive at different times for different loops.

Planning by loop rather than by system reflects that. A schedule organised around which loops are actually ready avoids the pattern where a team arrives to check a system and finds a third of it incomplete.

The instruments that cannot be simulated easily

Simple analogue loops are straightforward to check with a signal source. Others are not: analysers requiring a calibration gas, radar level instruments needing a target, flow meters where no flow exists, weighing systems requiring test weights.

These need planning and frequently need vendor attendance. Identifying them early, rather than discovering them when the checking team reaches them, is what keeps the schedule intact.

Recording what could not be completed

Some checks cannot be finished before startup: a valve that cannot be stroked because the line is not ready, a measurement that cannot be verified without process fluid.

Each becomes an open item with a defined completion point during commissioning, and the register of them needs to be visible to whoever is authorising startup. A loop that has passed a partial check should not be recorded as checked.

Who signs, and what the signature means

Loop check sheets carry signatures from the parties present, and the meaning of those signatures is frequently undefined.

A contractor's signature normally attests that the work was performed. A client witness signature attests that the client observed it. Those are different statements and conflating them causes difficulty later, when a loop found to be wrong has a client signature on its check sheet.

Stating on the sheet what each signature means resolves it, and it also tends to improve the rigour of witnessing, because a witness who understands what they are attesting to watches more carefully.

Retesting after late changes

Loops modified after their check — a range change, a rewire, a card replacement — are no longer checked, and tracking that requires the change process and the check register to be connected.

A simple rule works: any change to a loop invalidates its check, and the check status is visible on the punch list until it is repeated.