Operator Rounds: A Practical Guide to What to Check and How Often

Most plants already own their best early-warning system. It walks the site every shift, it can hear a bearing that has started to growl, smell insulation that is running hot and spot a weep at a flange long before any instrument does. It is the operator on the round.

The trouble is that a round only works if it is designed to. Plenty of rounds have drifted into a ritual: the same sheet, the same ticks, the same numbers copied down from yesterday, filed in a binder nobody opens. This guide covers what operator rounds are actually for, what belongs on a route, how often to walk it, and how to make sure the readings turn into action.

What operator rounds are, and what they are not

Operator rounds are regular, structured walks through an area in which operators check the condition of equipment, record a set of readings and report anything abnormal. You will also hear them called operator care, operator-driven reliability or, in plants that follow Total Productive Maintenance, part of autonomous maintenance.

They are easy to confuse with two neighboring activities, and the difference matters when you design them:

• Preventive maintenance is work a technician does on a schedule: change the filter, grease the bearing, replace the belt. It changes the equipment.
• Compliance inspections are checks a regulation or standard requires: fire extinguishers, eyewash stations, forklifts before use. They prove a condition.
• Operator rounds are about condition awareness. They change nothing on their own. Their value is entirely in noticing a developing problem early enough to plan the fix instead of reacting to the failure.

Keeping that purpose in view answers most design questions. A check belongs on a round if it can reveal a developing problem that someone would act on. If it cannot, it belongs somewhere else, or nowhere.

Why rounds catch what instruments miss

Most equipment failures announce themselves before they happen. Reliability engineers describe this with the P-F curve: there is a point (P) at which a developing failure first becomes detectable, and a later point (F) at which the equipment actually fails. The time between them is your window to act.

Different signals appear at different points on that curve. Vibration analysis and oil analysis can pick up some problems early. Heat, noise, leaks and smell usually show up later, but well before the failure, and they are exactly what a person on a round notices. Rounds also cover the large share of equipment in any plant that has no online instrumentation at all, and they sanity-check the instruments that do exist: a local gauge that disagrees with the control room reading is itself a finding.

What to put on a round

Start from the equipment and the ways it fails, not from a blank form. For each asset on the route, ask three questions: what typically goes wrong with it, how does that show itself, and what can an operator observe or measure from where they stand?

That produces a mix of three kinds of check:

• Condition checks. Look, listen, feel, smell. Recorded as acceptable or not, with a note and ideally a photo when not.
• Readings. A number with an expected range: a pressure, a temperature, a level, a flow, a motor current from a local meter.
• Housekeeping and safety items. Guards in place, drains closed, labels legible, no spills, lighting working.

As a starting point, these are the checks that appear on most rounds for common equipment:

• Centrifugal pumps: seal leakage, unusual noise, bearing housing temperature by touch or infrared thermometer, suction and discharge pressure, coupling guard in place. See the centrifugal pump inspection checklist.
• Electric motors: noise, the smell of hot insulation, cooling fins and fan cover clear, current where a local meter is fitted.
• Air compressors: discharge pressure and temperature, oil level, condensate drains working, unusual loading and unloading. See the air compressor daily checklist.
• Boilers: water level in the gauge glass, steam pressure, burner and alarm status, blowdown done, water treatment readings. See the boiler daily inspection checklist.
• Cooling towers: basin level, fan noise and vibration, drift eliminators and fill condition, chemical feed running. See the cooling tower inspection checklist.
• Tanks and secondary containment: level, leaks, bund drain valve closed, corrosion, rainwater in the bund. See the storage tank and containment checklist.
• Heat exchangers: inlet and outlet temperatures, pressure drop where gauges allow, leaks at heads and flanges.

Then cut. A round that can be done properly in the time available beats one that lists everything. Two signs your round is too long: it is routinely finished late, and the readings come back identical shift after shift. Identical readings rarely mean a perfectly stable plant. Usually they mean the numbers were copied.

Readings need limits, or they are just numbers

A reading written on a sheet without an expected range is a number nobody will look at again. Every reading on a round should carry a normal operating range, taken from the manufacturer's documentation, the process design or, where neither exists, the equipment's own history. Next to it, decide what happens when the reading falls outside: re-check it, tell the control room, raise a work order, or stop the equipment.

The best moment to check a reading against its limit is the moment it is taken. The operator is standing next to the equipment. They can look again, take a photo, check a second gauge or get a colleague. A paper sheet cannot tell them the number they just wrote is wrong; a digital form can, and it can ask for a note or a photo before the round moves on.

Limits catch the obvious excursions. Trends catch the subtle ones. A bearing temperature that is still within its range but has climbed a little on every round for two weeks is one of the most valuable early warnings a round produces. It is only visible if readings are stored against the asset over time rather than on separate sheets that are filed by date.

How often should rounds be walked?

There is no universal answer, and anyone who gives you one without asking about your equipment is guessing. Frequency follows two things: how bad the consequence of a failure is, and how quickly a problem develops once it is detectable.

• Equipment where conditions change quickly, or where a failure would stop production or create a hazard, is usually checked every shift.
• Stable utilities and standby equipment may only need a daily or weekly walk.
• Some equipment has frequencies set by local codes, insurers or the manufacturer. Those come first, and the round should respect them.

The P-F curve gives a useful rule: the interval between checks has to be comfortably shorter than the warning window, or the problem can start and finish between two rounds. Then let the results tune the schedule. A round that finds nothing for months is either checking the wrong things or could be walked less often. A failure that showed visible warning signs between two rounds says the round should be walked more often, or should include a different check.

Design the route people will actually walk

Order the checks the way the operator moves through the area, not the way the equipment appears on a drawing. Group them by location so nobody walks back and forth across the unit, and plan safe access: a round should never require someone to enter a guarded area or climb without the right protection just to read a gauge.

Two practical decisions make or break a route:

• Proof of presence. An NFC tag at key checkpoints, tapped from the operator's phone, shows the checkpoint was visited. The point is not surveillance. It makes missed checkpoints visible, so a supervisor can fix the real cause: a route that is too long, a shift that is short-staffed or a checkpoint that is hard to reach.
• Connectivity. Many plant areas have poor or no signal: basements, pipe racks, steel buildings, remote sites. A rounds tool that needs a live connection will be abandoned. It has to work offline and sync when the device is back in coverage.

Make every finding leave the round with an owner

A round that spots a seal leak has done its job. If that finding then sits on a sheet in a binder, the system around the round has failed. Every abnormal finding should leave the round as something with an owner: most often a work order against the asset, carrying the reading, the photo and the operator's note, so the technician knows what was seen and where.

Close the loop in the other direction too. Operators keep reporting what they notice when they can see something happen as a result. When the repair is signed off, it should be visible on the asset's history, where the next operator on the round can find it. We look at this handoff in more detail in why inspection findings stall before they become work orders.

What supervisors should look at

A round produces a lot of data. A few views make it useful:

• Completion by round and shift, including rounds finished late.
• Checkpoints that are repeatedly skipped or marked not accessible.
• Readings outside their limits this week, and readings drifting toward them.
• Open abnormal findings by age.
• Repeat findings on the same asset.

The last two are where the reliability conversation starts. An abnormality that has been open for three weeks, or a pump that has had a seal finding on four of the last ten rounds, is a better agenda item for the morning meeting than a completion percentage.

Common ways rounds go wrong

• Pencil-whipping. Readings copied from the last round or filled in from memory. The signs are identical values and rounds completed faster than anyone could walk them. Limits at entry, checkpoint verification and a shorter, more meaningful list all help. So does acting visibly on what rounds find.
• Too many checks. Every incident adds an item and nothing is ever removed. Review the content of each round at least once a year and delete checks that have never produced a finding or a decision.
• No owner. Equipment is replaced, a line is reconfigured and the round still lists the old pump. Give each round an owner who updates it when the area changes.
• No feedback. Findings disappear into the maintenance backlog and operators never hear back. Reporting quietly stops.

Where to start

Pick one area, ideally one with a recent failure that showed warning signs. List its critical equipment and the ways each item fails. Draft the round from those failure modes, set a normal range for every reading, and walk it with the operators who will do it before anything is finalized. Run it for a month, then review what it found and what nobody looked at.

If you want a head start, our free checklist templates include ready-made rounds for boilers, air compressors, cooling towers and pumps, each downloadable as PDF, Excel or Word. When you are ready to run rounds on mobile, with readings checked against limits as they are entered, checkpoint verification, offline capture and findings that become work orders, see how connected frontline operations work in Vardian.