On this page7 sections
Start with the asset register
Scope begins with a list, not a sentence. Every asset under the contract should be identified by tag, with enough information to plan work and spares:
- Tag number, service description and location.
- Asset type, make and model, and firmware or software version where relevant.
- Calibrated range, required tolerance and the reference method used to check it.
- Criticality — the safety, environmental, production or quality impact if the asset fails or reads wrongly.
- Maintenance history, known issues and spare parts held.
Group the register by asset class — field instruments, analyzers, control valves, controllers and I/O, operator stations and servers, networks, power supplies and UPS, panels — because preventive tasks, skills and test equipment all follow asset class.
Preventive maintenance by asset class
Preventive tasks should be specific enough to be verified. Frequencies are set per site from manufacturer recommendations, regulatory or quality requirements, criticality and experience; the table below sets out what the tasks and records typically include.
| Asset class | Typical preventive tasks | What the record should show |
|---|---|---|
| Pressure, DP and level transmitters | Visual inspection, impulse line draining or blowdown, zero check, calibration check | As-found and as-left readings; condition of impulse lines and seals |
| Temperature instruments | Sensor and thermowell inspection, loop check, comparison against a reference | Deviation found; whether the sensor was replaced |
| Control valves | Stroke test, positioner diagnostics, packing and actuator inspection | Stroke time, deviation, diagnostic alerts |
| Analyzers and sample systems | Validation, consumables, filter and pump service, flow and temperature checks | Validation results, parts replaced, sample flows |
| PLC, DCS and I/O | Diagnostics review, redundancy switchover test where permitted, battery replacement, program backup | Backup version and location; diagnostic events |
| Operator stations, servers and networks | Health checks, disk and log review, backup and restore test, switch diagnostics | Backup verified; faults found |
| Panels and power | Fans and filters, terminations, earthing, UPS battery test | UPS test result; thermal or visual findings |
Calibration intervals from criticality and drift
There is no universal calibration interval. Start from the manufacturer's recommendation, any regulatory or quality requirement and the asset's criticality, then let recorded drift decide. The key is recording the as-found reading before any adjustment and the as-left reading after it. Without as-found data, drift cannot be measured and intervals cannot be justified.
Calibrate against reference equipment traceable to national standards, with an uncertainty well below the tolerance being checked; a test uncertainty ratio of about 4:1 is a common rule of thumb. A worked example of an interval review:
Step 1: Set the tolerance
Two DP transmitters have a required tolerance of ±0.5 % of span and a current calibration interval of 12 months.
Step 2: Review transmitter A
As-found errors over four calibrations: 0.10, 0.12, 0.15 and 0.14 %. All sit well within tolerance — at no more than a third of it — and the drift is small. A site rule might allow the interval to be extended — for example to 18 months — with the next as-found result reviewed.
Step 3: Review transmitter B
As-found errors: 0.20, 0.35 and 0.62 %. The error is growing, and the last result is out of tolerance. Shorten the interval, assess the effect of the out-of-tolerance period on the process, and look for a cause — impulse lines, overpressure, ambient temperature or a failing sensor.
Step 4: Record the decision
Document the rule applied, the data and the new interval, so that the next review starts from evidence rather than memory.
Instruments that form part of safety instrumented functions follow the proof-test intervals set by their safety requirements. The AMC must respect those intervals rather than optimize them.
Response tiers and breakdown support
A response commitment means something only when the contract defines each tier. Define tiers by consequence, then agree the response for each explicitly: time to acknowledge, time to attend on site or connect remotely, and the hours covered.
| Tier | Definition | Typical expectation |
|---|---|---|
| 1 — Critical | Safety, environmental or production-stopping failure; loss of control or loss of view of the process | Shortest agreed acknowledgement and attendance, including outside normal hours if contracted |
| 2 — Degraded | Redundancy lost, a critical measurement unavailable, or operation continuing on workarounds | Attendance within the agreed period, with interim measures agreed |
| 3 — Routine | A fault with no immediate production impact | Scheduled attendance, often combined with preventive visits |
| 4 — Planned | Improvements, modifications and minor changes | Quoted and scheduled separately unless included in the scope |
State how remote support works — who authorizes a connection, through which secure route, and how sessions are logged — and what happens when a fault needs a spare that is not held.
Spares and obsolescence
- Critical spares — items whose failure stops production or disables a safety or control function, and whose lead time is longer than the plant can tolerate. Hold these on site.
- Ownership — state whether spares are held by the site or the contractor, who pays for consumption, and how stock is replenished.
- Shelf life — batteries, electrochemical sensors, electrodes, reagents and some calibration gases have limited shelf lives; rotate stock and record expiry dates.
- Obsolescence — review the lifecycle status of controllers, I/O, operator stations and analyzers at least annually, and plan replacement before spares run out.
- Software — keep licences, installation media and program backups under version control, with a copy stored separately from the control system.
Records and KPIs
Records turn maintenance into evidence. Each visit and job should leave a record tied to the asset tag: what was found, what was done, parts used and time taken. From those records, a small set of indicators shows whether the contract is working:
- MTBF
- Mean time between failures: total operating time of a population of assets divided by the number of failures. Useful for spotting asset types or services that fail more often than their peers.
- MTTR
- Mean time to repair: total repair time divided by the number of repairs. Define when the clock starts and stops — at report, at arrival or at restoration — so the figure means the same thing every period.
- PM compliance
- Scheduled preventive tasks completed within their planned window, as a share of those due.
- As-found out-of-tolerance rate
- The share of calibrations found outside tolerance before adjustment. A rising rate suggests intervals are too long or a condition is degrading.
- Repeat failures
- Assets with more than one failure in a period. A short list of these ‘bad actors’ focuses root-cause work where it matters most.
Exclusions and responsibilities
Most AMC disputes are about what was never written down. For each of the following, state who is responsible and who bears the cost:
- Permits, isolations and access, including shutdown windows for work that needs them.
- Spares, consumables, calibration gases and reagents.
- Third-party equipment and packaged units that have their own suppliers.
- Software licences, operating-system updates and security patching.
- Modifications and additions, and how they enter the register and the scope.
- Damage from process upsets, misuse or unauthorized changes.
- Replacement of obsolete equipment, as distinct from repair.
Engineering note
Review the contract with data
Schedule a periodic review of the register, calibration intervals, response records and KPIs. An AMC renewed without review keeps every gap it started with.
Engineering questions
- What is included in an instrumentation AMC?
- Typically preventive maintenance, calibration or calibration support, breakdown response and troubleshooting for the assets listed in the contract, with a record of each. What is included in practice depends on how precisely the asset register, tasks, response tiers, spares and exclusions are defined.
- How often should process instruments be calibrated?
- There is no single interval. Start from the manufacturer's recommendation, any regulatory or quality requirement and the instrument's criticality, then adjust based on as-found results. Instruments that consistently stay well within tolerance may justify longer intervals; those that drift need shorter intervals and investigation.
- Should PLC and SCADA systems be in the same AMC as field instruments?
- They can be, and one contract covering the full signal path from field device to operator screen has advantages. The tasks and skills differ, so the scope should list control-system assets, software versions, backup duties and patching responsibilities explicitly.
- What is the difference between preventive and breakdown maintenance in an AMC?
- Preventive maintenance is scheduled work that keeps assets in condition and finds problems early. Breakdown maintenance is the response to failures. A useful AMC covers both, and uses breakdown records to improve the preventive schedule.