Polyurethane Conveyor Blade: Setting a Replacement Interval

Maintenance Planning · Blade Life

Most sites change a polyurethane conveyor blade on a calendar interval that somebody set once and nobody has revisited. That interval is wrong in two directions at the same time across a year, and both errors are invisible, because a changeout that happened on schedule never gets audited. A belt cleaner blade replacement interval built from your own wear rate gives you a date you can plan a shutdown around and a reorder date your store can work to. This post sets out how to measure the wear rate, how to choose a threshold, how to turn the two into a date, and what breaks the calculation.

01Why does a calendar interval fail?

A calendar interval assumes the blade wears at one rate for its whole life and that the rate holds across the year. Neither assumption survives contact with a real conveyor.

Your product changes with the season. A Queensland coal operation running wet product through the summer and dry product through the winter is presenting two different duties to the same blade on the same conveyor. Your production rate changes. Your belt gets replaced and the new cover behaves differently. None of that reaches the interval, because the interval lives in a maintenance schedule rather than in anything that measures the blade.

The interval then goes wrong in two directions. Set it conservatively and you pull blades with a third of their life left, which spends money on consumables and puts a person at a conveyor more often than the work requires. Set it optimistically and the blade runs past the point where it was protecting the belt, so carryback returns in the back half of every interval and your crew winds on tension to compensate.

Sites usually find out about the second error and never find out about the first. Carryback on the floor generates a complaint. A blade thrown away at 40% wear generates nothing.

Answer A calendar interval assumes a constant wear rate across a year that has seasons in it. It discards good blade in the easy season and runs spent blade in the hard one, and only the second failure is visible.

02What actually sets your wear rate?

Blade wear responds to work done, and calendar time is a poor proxy for work done. Ranked by how closely they track the thing that is actually consuming your blade:

  • Tonnes through the position. The closest available measure of the work the blade has performed. A blade at eight weeks on a conveyor that ran four days a week has done roughly half the work of the same blade at eight weeks on a conveyor that ran continuously.
  • Running hours. A workable substitute where tonnage at the position is not readily available, which on most sites it is not. Running hours capture the idle periods that calendar days hide.
  • Calendar days. The weakest of the three, and the one nearly every site uses, because it is the one that needs no data at all.

Underneath the exposure measure sit the conditions that set wear per tonne. The abrasive character of the product governs most of it, which is why a dry fines duty and a wet sticky duty produce different rates on identical blades. Belt speed multiplies sliding distance. Contact pressure, set by tensioner load and by how wide the tip contact band has become, sets the load term. Splices and mechanical fasteners deliver impact events that take material off the edge in a way steady abrasion does not.

None of that needs to be modelled. It is the reason your wear rate is specific to one blade position on one conveyor in one season, and the reason a rate borrowed from another position will mislead you.

Answer Wear tracks work done, so tonnes through the position beats running hours, and running hours beat calendar days. Conditions such as product abrasiveness and belt speed set the rate per tonne.

03How do you measure wear without a measuring tool?

Your service tech assesses blade condition by eye at the conveyor. That is how the work is done on every site FM8 visits, it is done by people who have looked at thousands of blades, and it is a reasonable method. The obstacle to building an interval from it is consistency rather than competence.

Two techs looking at the same blade on the same shift return two different numbers. That spread is often larger than the difference you are trying to detect between one inspection and the next, so a history assembled from eyeball calls has more noise in it than signal, and the slope you would draw through it means very little.

The FM8 Conveyor Cleaner Health app removes the variation without adding an instrument to the tech's kit. They photograph the blade in place on their phone, the app reads FM8 blade geometry off the image and returns an estimated percent wear, and the tech confirms or adjusts it. The figure that lands in the history is the figure the next tech would have produced, which is the property a wear rate calculation needs.

What a usable history looks like

Three inspections spaced across a blade's life is the practical minimum before a slope means anything. Two points draw a line through whatever noise is in them. Capture the surrounding facts at the same visit, because the covariates are what let you explain a rate that changes later: date, inspector, work order reference, measured tensions, blade model and tensioner type. The app captures those on the same form at the conveyor, so there is no clipboard and nothing to re-enter in the office.

Answer Wear is a visual assessment captured as a phone photo, with the app estimating percent wear for the tech to confirm. Three or more inspections across a blade's life give you a slope worth trusting.

04What wear threshold should you change at?

The threshold is a business decision rather than a property of the blade, and it is the input sites think least about.

Running a blade to exhaustion means running it well past the point where it stopped protecting your belt. As the tip wears back into thicker section, the contact band widens and cleaning pressure falls, so the last portion of a blade's physical life is a period where carryback is climbing and your crew is compensating with tension. The threshold worth setting is the point where cleaning performance falls off, which arrives earlier than the point where the blade runs out of material.

Two site-specific factors move it. Access is the first. A blade on a shiploader boom or in an underground roadway justifies a more conservative threshold than one on a walkway-accessible transfer, because the cost and the risk of an unplanned changeout are higher and the window to do it is narrower. Consequence is the second. Where the return run crosses a road, a water course or occupied ground, the cost of a late changeout is not confined to the conveyor.

FM8 does not publish a threshold number, because a value correct for a Pilbara boom conveyor is wrong for an underground trunk and prescriptive figures get applied where they do not belong. Set it per position, review it against what your own history shows about when cleaning fell away, and keep it adjustable.

Answer Set the threshold at the point where cleaning performance falls off rather than at blade exhaustion, and set it per position, since access and consequence differ across a site.

05How do you turn a wear rate into a date?

The arithmetic is deliberately simple, and simple is what makes it survive contact with a maintenance planner's week.

Take the slope through your inspection history as percent wear per week. Subtract the current wear from your threshold and divide by the rate. A blade sitting at 62% wear, wearing at 11% per week, against an 85% threshold, has roughly two weeks left. That is the whole calculation.

Two adjustments turn the answer into something a planner can use. Subtract your procurement lead time from the date to get a reorder-by date, which for a remote site is the number that actually prevents a stockout. Then move the changeout to the nearest planned shutdown or access window ahead of the date, because a scheduled changeout three days early is worth considerably more than an unplanned one on the exact day.

Treat the result as a planning window. Blade wear is not perfectly linear across a life, since bed-in behaves differently to steady state and a widened tip wears differently again, so the extrapolation gets more reliable as you approach the threshold. Re-inspect as the date comes up and let the estimate tighten.

https://cchapp.internal/conveyor/CV-118/primary/history
Ridgeback Coal Operations (demonstration data) CV-118 primary History
Wear rate
11.4%
per week, last 5 inspections
Predicted changeout
2 Oct
at 85% threshold
Reorder by
8 Sep
24 day lead time applied
Inspection history
DateInspectorEst. wearEvidenceNote
21 Jul 2026J. Okonkwo4%2 photosFitted, WO 117201
04 Aug 2026D. Reihana19%3 photosBed-in complete
18 Aug 2026J. Okonkwo42%2 photosTension unchanged
01 Sep 2026M. Whitlam51%3 photosWet product, slower week
15 Sep 2026D. Reihana62%3 photosEdge square, no chipping
Changeout threshold85% wear
Five inspections, each a phone photo confirmed at the conveyor, become a wear rate and a predicted changeout date. The reorder date carries the site's own lead time. Demonstration data. Note the 01 September reading sitting below the slope, which is the wet week the tech recorded in the note.
Answer Subtract current wear from your threshold and divide by the weekly wear rate. Take procurement lead time off that date for a reorder date, then move the changeout to the access window ahead of it.

06What breaks the calculation?

A wear rate belongs to one blade position, one blade model and one set of conditions. Several ordinary events invalidate it, and a rate that has quietly stopped applying is worse than no rate at all, because it carries the authority of a number.

  • The season turns. The largest single disruptor on Australian coal operations. Wet and dry product wear at different rates, so a rate carried across the change will mislead you for the first weeks of the new season.
  • Production changes. A campaign, a shutdown, a rate increase. Where the rate is expressed per week rather than per tonne, any change in running time moves it.
  • A tensioner develops a fault. A seized or over-wound tensioner changes contact pressure, which changes the rate directly. This one is common at coastal terminals where salt attacks the mechanism.
  • The belt is replaced. A new cover at full hardness presents a different surface to the blade than the worn one it replaced.
  • A different blade model is fitted. Rates are not transferable across models or materials, and a FRAS blade will not follow the rate its standard equivalent produced.
  • Mistracking develops or a splice is repaired. Both change what the tip meets on each pass.

The practical control is to keep a separate wear curve per position and to record the covariates at each inspection, so a rate that shifts can be explained rather than argued about. Where a site keeps one blended figure across a fleet of conveyors, every one of the events above is averaged into invisibility.

Answer Seasonal product change, production changes, tensioner faults, a new belt and a different blade model each invalidate a rate. Keep one curve per position and record the conditions so a shift can be explained.

Field example: a Moranbah area coal operation

A Bowen Basin operation near Moranbah ran a six week changeout on the primary cleaners across its plant conveyors. The interval had been set years earlier, it was in the maintenance system, and it was met reliably. By the usual measures the site was doing this well.

Building a wear history per position over two seasons produced two rates rather than one. Through the dry months the primary on the main plant feed conveyor wore at roughly 7% per week. Through the wet it ran closer to 16%.

Against a six week interval, that meant the dry season blades were coming out at around 40% wear, with well over half their life discarded. In the wet the same blade passed the point where cleaning fell away at about week five, so the last week of every interval ran a spent blade, which is where the wet season carryback the crew had always attributed to the weather was coming from.

Splitting the interval by season, roughly eleven weeks through the dry and five through the wet, moved the annual changeout count by about one. The changeout count was never the gain. What changed was that the site stopped throwing away half-worn blades for six months of the year and stopped running spent blades through the back of every wet season interval, and the wet season carryback stopped being treated as a fact of life.

The planner's comment was the useful part. The value was not the calculation, which is arithmetic anybody can do. It was having a wear history consistent enough that the two rates could be told apart at all.

Conveyor blade replacement interval practice in international bulk handling

Conveyor maintenance mining Australia practice on changeout intervals is not unusual by international standards, and neither is the calendar habit. Operations across South African coal, Indonesian thermal export and North American aggregate run the same schedules for the same reason: a calendar interval requires no measurement, and measurement of blade wear has historically been inconsistent enough that nobody trusted the trend it produced.

Where the exposure measure is available, tonnes beats time in every jurisdiction. Sites running a conveyor scraper blade longwall mining duty or a high tonnage export circuit already have tonnage at the position in their control system, and expressing wear per hundred thousand tonnes rather than per week removes the production variability from the rate in one step. Where that data is not accessible at the position, running hours will carry most of the benefit.

Polyurethane conveyor blade life: what this is worth

A conveyor blade replacement cost calculator built on unit price answers the smallest question in the problem. The interval reaches four other numbers. Blade consumption falls when you stop discarding half-worn blades. Belt cover life extends when you stop running spent blades that have lost contact pressure. Carryback clean up and the idler damage behind it fall with the same change. Each changeout you avoid is one fewer occasion where a person is working at a conveyor.

Against that, the input the calculation needs is one photo per blade per inspection visit, taken by a tech who was already standing at the conveyor looking at it. That is the least expensive data collection FM8 can point a site at, and it is the reason the app exists in the form it does.

Contact FM8 for a demonstration

FM8 will show you the Conveyor Cleaner Health app building a wear curve on a site like yours, and will work through the interval calculation against your own positions and lead times.

Australian operations are supported through your FM8 Distributor. International enquiries are handled by FM8 Engineering directly.

Contact FM8
FM8 · Engineered to Exceed · Proprietary Engineering Content
This article is published by FM8 for the technical information of maintenance, reliability, engineering and procurement personnel. It describes engineering principles and FM8 design intent, and does not constitute a design specification or a maintenance standard for any particular conveyor installation. Wear thresholds, replacement intervals, wear rates and field scenarios described here are indicative of typical operating ranges and are not guarantees of performance. Changeout intervals should be set against your own site conditions, inspection history, risk assessment and applicable standards. App screens shown are demonstration data and do not represent any FM8 client site. All FM8 product designs, blade geometries, material formulations and the FM8 Conveyor Cleaner Health application remain the intellectual property of FM8. Reproduction, adaptation or reverse engineering of FM8 designs or content is not permitted without written consent.
Next
Next

FRAS Conveyor Scraper Blade: What Actually Makes It FRAS