What Is Centerlining in Manufacturing? Ending Setting Drift Between Shifts
Two operators run the same filler on the same product on different shifts, and the machine behaves like two different machines. One runs the diaphragm pump at 38 psi, the other at 46 because it 'fills better.' The pump seals can last a different amount of time depending on who is on shift. That is the problem centerlining solves, and the answer to what is centerlining in manufacturing starts with one question: for each setting on this machine, what is the approved target or range for this product and operating condition?
A centerline is the documented best-known setting for each key controllable variable for a given product and operating condition: speeds, pressures, temperatures, tensions, gaps, timer values, recipe parameters. Centerlining is the discipline of finding those values, marking them where the operator can see them, and running to them until a controlled change proves a better one.
The practice complements total productive maintenance by giving operators a standard to check and maintain. It is one of the few reliability practices a plant can adopt without buying anything.
What Is Centerlining in Manufacturing?
Put a plant manager on the spot with the question What is centerlining in manufacturing and the useful answer is this: running each key process variable at its documented target or within its approved range for the product being made, with a visible mark, so that any drift away from it is obvious at a glance.
Three things make a centerline: a value, with upper and lower limits where the process allows a range. A reason, written down, so the next person knows why 42 psi and why the limit is 45. And a visual mark, a painted band, a scribed line, a tag on the gauge, so the setting can be checked without opening a binder.
A centerline is a value, a written reason, and a mark the operator can see from where they stand.
The scope reaches past machine knobs. Centerlines apply to raw material specifications, to utility conditions where they matter (chiller water temperature, compressed air dew point), to tooling dimensions, and to the order in which a changeover is performed. Anything that can vary and that affects output or equipment health is a candidate.
What separates centerlining from a settings sheet in a drawer is the audit. Centerlines get checked at startup, after changeovers, and at set intervals during the run, and a deviation triggers a question before it triggers an adjustment. That order, question first and adjustment second, is the whole point.
The reliability payoff follows from there. Consistent settings under comparable product and operating conditions reduce one source of load variation. That makes wear trends and condition monitoring baselines easier to interpret, though a behavior change can still come from the material or the process.
Why Settings Drift in the First Place
Settings move for reasons that make sense to the person moving them. Understanding those reasons is how a plant stops the drift without turning the floor into a rules dispute. The usual drivers:
A worn component: the operator turns up the pressure to compensate for a leaking seal or a slipping belt, and the setting now masks a developing failure
Product variation: a stiffer batch of film or a colder batch of resin gets tuned around instead of flagged to quality
Personal preference: 'I always run it a little hot' survives for years because nobody can prove it wrong
Undocumented fixes: a technician changed a timer during a 3 a.m. breakdown and the new value became the setting by default
No reference point: with nothing marked on the gauge, there is no such thing as drift, only the current setting
The first driver is the one reliability engineers should care about most. A setting that has to move to keep product in spec is a symptom, and it is often the earliest symptom available. A regulator cranked from 40 to 52 psi over six months can signal leakage, restrictions, or changed process conditions, and centerlining is what makes that announcement audible.
The other drivers are habits, and the fix is the same for each: a documented value, a visible mark, and an agreement that nobody moves it without saying so.
Finding the Centerline
The settings a plant runs today are usually the settings that survived, which is a different thing from the settings that work best. Finding the real centerline takes evidence, and there are three ways to get it, in rising order of effort.
The golden run. Pull the process data from the best run of the last year, the one with the fewest stops, the lowest scrap, and the steadiest rate, and record every setting from that run. That becomes the first draft of the centerline sheet. It is quick, and it already beats three operators' opinions.
The best run of the past year gives you a starting point; the job is to write it down and confirm it can be repeated.
The OEM baseline. Equipment manuals and commissioning records hold the design settings, including the ones that protect the machine: maximum speeds, pressure limits, torque values on tensioners. Where the golden run and the manual disagree, the manual wins on anything that affects equipment life, and the disagreement goes on a list for engineering to resolve.
Structured trials. For the handful of variables that drive the most loss, a short designed experiment, three or four settings each, run for a defined period with output and quality measured, settles the argument with data. This is where centerlines earn their reputation, because the proven value often sits nowhere near where the folklore put it.
Whichever route produced the number, write down how. A centerline with a source can be defended on the floor at 2 a.m.; a centerline that came from 'engineering said so' lasts until the first bad batch.
Documenting and Marking the Settings
A centerline sheet lives at the machine. One page per machine or process step, identifying the applicable product or recipe and listing each variable, its centerline value, its limits, the unit, where the setting is read, and the reason. Laminate it, post it at the operator station, and version-control it with a date and an owner.
Then mark the settings on the equipment itself, where visual management best practices do their best work. A green band on the pressure gauge. A scribed line on a valve handle with a matching line on the body. A label on the temperature controller showing the setpoint and the limits. The goal is a settings check that takes one walk around the machine and no binder.
Digital parameters need the same treatment. Recipe files on a PLC or HMI get a locked master, a visible version number, and a comparison at startup, since a changed recipe parameter has no gauge to paint. A usable centerline sheet contains:
Variable name and where it is read (gauge tag, HMI screen, physical location)
Centerline value, upper and lower limits, and units
The reason the value was chosen and its source (golden run, OEM, trial)
Check frequency: startup, post-changeover, hourly, or per batch
Owner, revision date, and the change request that last modified it
Keep it to the variables that matter. A sheet with 60 lines gets skimmed. A sheet with the 12 variables that drive output and equipment stress gets checked.
Operators Own the Centerline
A centerline that engineering owns and operators tolerate drifts within a quarter. The practice holds when the people who run the machine check the settings, record the deviations, and have a say in the trials that change the values.
That fits naturally inside autonomous maintenance, where operators already carry the clean, inspect, and tighten routines. Adding a settings check to the startup inspection costs a few minutes and gives the operator a concrete reason for each setting, which is the difference between following a sheet and defending one.
A settings check at startup costs a few minutes and turns following a sheet into defending one.
Training is the lever. An operator who understands that unnecessary pressure can shorten seal life, and why, stops turning the regulator up on their own. Ten minutes on the failure mechanism does more than any amount of policy.
Deviations get recorded, including the ones the operator corrected. A log that says 'pressure found at 48, reset to 42, third time this week' is a maintenance work request in disguise, and the plants that get the most from centerlining route those logs to the reliability engineer weekly.
Give operators a path to challenge a centerline. If someone believes 44 psi runs better than 42, the answer is a trial with data, with the operator involved, and a changed sheet if they're right. Centerlines that can be improved get respected. Centerlines that are frozen get ignored.
Change Control Without Bureaucracy
Centerlining fails in two directions. Too loose, and settings drift the way they always did. Too tight, and the floor stops running trials because the paperwork takes longer than the experiment.
A one-page change request handles it: which variable, current centerline, proposed value, reason, trial period, and who signs off. Temporary changes get a time limit and a machine tag. Permanent changes update the sheet, the markings, and the PLC master the same day they are approved.
The tag matters more than it looks. A temporary change with no tag becomes a permanent change the moment the person who made it goes on vacation.
Maintenance belongs in the loop. A technician who replaces a worn pump or retensions a belt should check whether the centerline had been compensating for that wear, and reset it. Otherwise, the new component runs at the setting the old one needed, and the first failure of the new part is already being set up.
Measuring What Centerlining Returns
Ask what is centerlining in manufacturing worth, and the gains show up in three places. Setup and adjustment time, which sits among the TPM six big losses for good reason, drops when changeovers end at a known setting instead of a tuning session. Quality losses fall as the process stops wandering. Equipment life can also improve when approved settings remove chronic overload.
Track it simply. Centerline conformance (the percentage of checks where every variable sat inside its limits) and the count of deviations found per week are enough to show whether the practice is alive. If conformance climbs from 60 percent to 95 percent over a quarter, check whether minor stops and quality losses fall with it.
Higher conformance matters when it reduces the losses the centerlines were chosen to control.
Reliability sees it in the condition monitoring data. Vibration trends become easier to compare when speed, load, and measurement conditions are comparable. Bearing and seal life can improve where centerlining removes chronic overload.
Initial costs can be modest: engineering and trial time per machine, paint and labels, and the training sessions. Against one avoided seal failure on a critical pump, the arithmetic is short.