Abrasion Resistant Plastic Conveyor Belts: Material and Selection

Ask a plant engineer to describe an abrasion resistant plastic conveyor belt and you tend to get one of two reactions. Either nobody has ever specified one on purpose, or the belt arrived bolted to a machine and has been quietly outlasting everything the maintenance team expected. Both reactions are reasonable, because almost all the wear advice written in English is written about rubber. Cover grades, DIN 22102 abrasion classes, RMA ratings, wear-resistant compounds — genuinely useful material, and none of it transfers cleanly to a belt whose body is a polymer such as POM, PA, PE or UHMWPE.

That gap is what this page closes. We build belting rather than machines, and we have watched enough plastic-bodied belts run to the end of their life — and a few fail early — to say plainly where the material wins and where it does not. What follows stays at the level of the belt body, the polymer it is made from, the abrasion test that is quoted for it, and the way a buyer should turn a duty into a written specification. If you want the general argument about covers and carcasses instead, our abrasion resistant conveyor belt page covers the rubber-cover story in depth, and it is the right place to start if your belt has a rubber top cover.

Send us your product, speed and abrasive — get a plastic belt material recommendation back

One boundary before the engineering, because it matters for how you read every table below. A plastic belt resists wear through the bulk properties of its own body. A rubber belt resists wear through a sacrificial top cover whose single job is to be abraded slowly so the fabric or cord underneath never sees the load. Those are two different design philosophies, and they fail in two different ways. When this page says a polymer "wears well", it means the body itself loses very little material, keeps its surface, and does not shed a cover because there is no cover to shed. Where the abrasive is heavy, hot, or arriving as a fast sharp impact, that same philosophy becomes the weak point, and we say so directly in section 07 rather than pretending the material is universal.

01Two Kinds of Wear: Plastic Body Versus Rubber Cover

A conveyor belt is asked to do two separate jobs at once, and it is worth separating them before any material is chosen. The first job is to carry a load across a span of rollers or a slider bed without stretching beyond recovery. The second job is to let the material slide across its surface, day after day, without the surface being cut away. On a traditional heavy-duty industrial conveyor belt those two jobs are split deliberately. The carcass of EP or NN fabric, or steel cord, carries the tension, and the rubber cover takes the abrasion. The cover is a consumable, and the day it is gone the belt is scrap even though the carcass was never the problem. That consumable logic is built into how a rubber conveyor belt is specified, and it is worth understanding before you set it beside a plastic body.

A plastic belt collapses those two jobs into one component. There is no cover to sacrifice, because the same polymer that carries the load is the surface being abraded. That single change has consequences a buyer feels immediately. The belt cannot hide a worn surface by wearing through a cover slowly; the moment the body itself starts to lose material, the whole belt is losing its section. But the flip side is that many engineering thermoplastics have a genuinely low coefficient of friction and a high resistance to the rubbing that wears rubber down, so on the right duty the same belt runs for years with almost no measurable change in thickness. I have pulled plastic belts off food lines after thirty months of service and found the top surface still smooth, still flat, with edge wear under half a millimetre.

The reason the two materials behave so differently comes down to what abrasion actually is at the surface. Rubber is an elastic material. Under a sharp asperity it deforms, and a great deal of rubber abrasion is a fatigue process that tears small pieces away at the point where the material can no longer recover. That is why rubber cover grades are ranked by volume loss against a rotating drum, and why a soft, grippy rubber can be a poor wear performer. A hard thermoplastic behaves differently. It resists the asperity by not deforming much at all, so the abrasive has less material to pull up and tear. On a sliding duty with fine to medium abrasive, that difference alone can double belt life, which is exactly why plastic bodies show up on sorting decks, parcel chutes and small-parts conveyors where a rubber belt would be replaced far more often.

Where the analogy breaks is impact. Rubber absorbs energy by bending a long way and coming back; a plastic that resists deformation is, by definition, bad at absorbing energy. Drop a 20 kg lump of limestone with a sharp edge onto a rigid POM belt from 600 mm and you do not wear it, you crack it. This is the central trade-off of the whole article and the reason section 07 exists. A belt can excel at sliding wear and be entirely wrong for your line.

Notice what the table is not saying. It is not claiming one material is better. It is saying the enquiry you send to a supplier should name the wear mode, because "abrasion resistant" on its own describes a symptom, not a mechanism. Send the same word to a conveyor belt manufacturer twice, once for a fine powder and once for a 300 mm sharp lump, and you should receive two completely different recommendations. If you get the same answer both times, the quotation was not engineered.

02The Engineering Plastics Behind a Plastic Belt Body

When a buyer asks for a plastic belt, the useful question is not "is it plastic" but "which polymer". The word covers a range of materials whose abrasion behaviour differs by a factor of five or more, and whose prices differ by an order of magnitude. The families that actually appear in conveyor belt bodies are a small group. Acetal, sold as POM, is the workhorse for modular and link belts and for machined sprocket-driven belts. Polyamide, PA6 or PA66, trades some of that rigidity for toughness and is common in thin flat belts that have to survive a fold. Polyethylene, and especially ultra-high-molecular-weight polyethylene, is the surface champion of the group and shows up as wear strips, scraper edges and low-friction belt faces. Polypropylene and polyester appear in light-duty and food-contact belts, and PVC and polyurethane appear as coated-fabric constructions where a polymer face is carried on a woven or corded backing.

The property that governs abrasion resistance in this group is not a single number. It is a combination of surface hardness, the coefficient of friction, the material's ability to recover from a scratch, and how that combination changes as the belt warms up in service. A polymer that is hard and slippery at 20 °C can become soft and grippy at 60 °C, and the wear rate follows the surface, not the datasheet. I have leaned on this repeatedly on packaging lines. A POM belt that gave no trouble on a dry ambient line began to show polishing and thinning when the same duty moved downstream of an oven and the belt face sat at 55 °C for three shifts running. Same material, same load, different temperature, and the wear mechanism changed from abrasion to a slow smear.

UHMWPE deserves its own sentence because it is so often the correct answer and so often overlooked. Its molecular weight is measured in millions rather than thousands, and that long-chain structure gives it outstanding resistance to sliding abrasion plus one of the lowest friction coefficients of any thermoplastic. On a chute liner or a belt that rubs against a stationary edge, it can outlast a mild steel wear plate. What it will not do is carry much tension, because it creeps under sustained load. That is why UHMWPE usually appears in a plastic belt as a wear surface, a guided edge or a low-friction slider rather than as the load-bearing body.

Cost and availability then rearrange the shortlist. POM and PE are commodity grades, quoted from stock in sheets and rods, and priced so that a replacement belt rarely needs a budget conversation. Polyamide sits a little above them. PVDF and the high-temperature grades sit far above, and a single belt in those materials can cost more than every other belt on the line combined, which is fine if the alternative is a quarterly replacement but painful if the duty did not need them. A conveyor belt supplier who quotes you the exotic grade first has either not read the duty or is selling on margin.

Two abbreviations in that table cause more quotation errors than anything else. PVDF and PEEK belong to a higher temperature class than the commodity polymers and are priced accordingly, so they only enter the conversation when a line runs hot or chemically aggressive. And the coated-fabric row is not really a plastic body at all in the sense used by this page. On a coated belt the polymer is a face on a fabric that carries the tension. That construction is covered at length on our PVC conveyor belts and modular pages, and it wears by a slightly different rule because the load path and the wear surface are once again separated.

03Sliding, Impact and Gouging: Three Failure Modes That Need Different Answers

Wear is not one thing, and the single most common mistake in specifying a plastic belt is to buy for the average and lose on the worst case. In the field the three modes are easy to tell apart once you know what you are looking at, and each one points to a different fix. Reading the failure correctly is worth more than any datasheet number, because a belt that failed in mode two cannot be rescued by a material chosen for mode one.

Sliding wear is the friendly one. Material rubs across the belt surface under light to moderate load. The symptoms are a polished or lightly scratched top face, a slow reduction in thickness, and material dust that matches the conveyed product rather than the belt. This is the mode where a plastic body wins comfortably. The low friction that makes a polymer slippery also means the abrasive has less purchase, so the wear rate stays low for a long time and then changes only gradually. On a plastics-recycling line I followed, a POM carrying belt handling shredded regrind was still within a tenth of a millimetre of its original thickness after fourteen months, while the adjacent rubber belt on the same duty had already been rotated once. That is the plastic belt doing exactly what it was bought to do.

Impact wear is where the story turns. A lump with mass and a sharp edge arrives at speed, and the belt has to absorb energy it was not designed to absorb. A rigid polymer cannot bend far enough to soak up that energy, so the failure appears as crazing, a star-shaped crack under the impact point, or a clean split that runs from the impact. The tell is that the damage is localised to the loading zone and looks like a fracture rather than a wear scar. Once impact is in play, the material conversation ends and the geometry conversation begins, because the fix is to reduce drop height, add impact idlers or a cushioning bed, or change the belt family entirely rather than to hunt for a harder polymer.

Gouging sits between the two and is the mode buyers under-specify most often. It happens when a sharp, hard object is dragged along the belt under pressure rather than dropped onto it, cutting a groove instead of bruising the surface. A plastic belt resists gouging better than its impact performance would suggest, because the same rigidity that cracks under a drop also resists being carved, but it is not immune. Where the conveyed material includes broken glass, sharp scrap, or tramp metal, a cut-resistant construction and an inspection routine for the splice become mandatory and are covered on our impact and cut resistant belt page.

04Abrasion Tests: What the Numbers Do and Do Not Prove

Suppliers quote an abrasion figure the way bakers quote a price, which is to say selectively. Two belts can carry the same milligram loss number and perform completely differently in your plant, because the number was produced under a lab condition that may or may not resemble your duty. Knowing which test produced a figure, and what that test was originally written for, is the difference between comparing materials and comparing brochures.

The dominant method in belting is the rotating-drum test, standardised as ISO 4649 and, in its German form, DIN 53516. A cylindrical specimen is pressed against a rotating abrasive drum under a fixed load for a fixed travel, and the result is reported as volume loss in cubic millimetres, or sometimes as a relative abrasion index. It is the test behind the familiar DIN 22102 cover classes, and it is a reasonable proxy for sliding abrasion under moderate pressure. Its weakness, for a plastic buyer, is that it was written to rank rubber covers. A polymer that cracks under a point load or smears when warm can score well on the drum and still fail on your line, because the drum never delivers a sharp impact and never runs the specimen hot.

Two other families of test turn up when the abrasive is more aggressive. The Taber abraser, standardised for plastics in general and often cited as ASTM D4060, runs the specimen under two abrasive wheels and reports mass or volume loss per thousand cycles. It is flexible enough to model different severities by changing the wheel grade, which makes it useful when you want to compare one polymer against another rather than one belt against a standard. Then there is a plain field method that costs nothing and often tells you more than either lab test. Weigh a new belt section and a worn section from the same production batch, and divide the mass lost by the running hours and the tonnage carried. That figure is specific to your plant, your product and your operators, and no supplier datasheet can argue with it.

Here is the discipline that matters when you receive a quotation. Ask which standard produced the number, on which specimen geometry, at what load and what temperature, and whether the abrasive was renewed between runs. A supplier who can answer those four questions has actually tested the belt. A supplier who answers with a single number and a confident adjective has copied it from a resin datasheet that describes the raw polymer, not the finished belt. Those are not the same thing, because extrusion, moulding and any post-finish all change the skin that does the wearing.

If the duty is genuinely abrasive and the volumes are large, ask your wholesale conveyor belts supplier to run the two candidate materials side by side on the same rig and to send the raw readings rather than the summary. A conveyor belt factory that owns its own press can usually do this, and the resulting comparison is far more persuasive than a catalogue table, because both belt bodies were made under the same process and only the polymer changed.

05Turning Duty into a Specification: Load, Speed, Particle, Heat and Chemicals

Selection is a conversation between five variables and one constraint, and the constraint is that all five must be satisfied at once. Buy for the average and the worst variable will find you. What follows is the order in which we run that conversation with a customer, and the reason each step sits where it does.

Load is the first gate because it is the one that most often eliminates plastic outright. A plastic body must carry the working tension across the support span without creeping, and engineering polymers vary enormously in how much sustained load they will hold before they deform permanently. Get the working tension and the support spacing on the table before discussing polymers. If the belt is long, steep, or heavily loaded, the honest answer may be that a plastic body is the wrong family and that the load should go onto a fabric or cord carcass with a plastic or rubber face instead.

Speed then refines the answer rather than replacing it. Higher speed raises the frequency of every contact and pushes belt temperature up, so a polymer that is comfortable at 40 m per minute can be marginal at three times that figure, purely because it never gets to cool. Where a line runs fast and the product is light, a low-friction plastic face is an asset rather than a liability. Where the same line runs fast and carries a heavy or hard product, the same face becomes a heat generator at the pulley and drums.

Particle size and sharpness decide whether the abrasive is polite or vicious, and they matter more than the material's datasheet hardness. A fine, rounded powder polishes a belt slowly over years. A coarse, angular or jagged material cuts it. Two loads described in identical tonnage can sit at opposite ends of belt life on this variable alone, which is why the enquiry should describe the material rather than just naming it. Then temperature and cleaning chemistry arrive together, because in food, pharmaceutical and electronics duty they usually travel as a pair. A belt that must be washed at 60 °C with a caustic detergent is not simply a belt that runs warm; the chemistry can attack the polymer surface, remove any plasticiser, or lift a coating, and none of those effects shows up in an abrasion number.

Once the five variables are written down, the material usually selects itself and the remaining argument is about price and lead time. If the answer lands on a coated-fabric belt rather than a solid polymer body, the transmission side of the plant often needs attention at the same time, and the same engineering logic applies to drives, which is where a transmission belt manufacturer and a belt body supplier are asked the same question about wear and expected life. The same wear logic runs through the drive side of a plant, where a slipping belt abrades its sheave and itself in the same slow, patient way, and a V-belt manufacturer will say the same thing we do about matching belt to duty rather than to price. On a site that buys through one source, it is usually the conveyor belt distributor holding the stock and the joint kits who catches a mis-specification before it ships, which is worth more than any discount.

06Joints, Tensioning and Tracking on a Plastic Belt

A plastic belt asks for more care at three places than a rubber belt of the same width, and all three are installation issues rather than material issues. Get them wrong and an abrasion-resistant belt fails early for reasons that have nothing to do with abrasion, which is exactly the kind of failure that gets a good material blamed unfairly.

The joint comes first, and the good news is that many plastic belts do not have one. A modular link belt is assembled from moulded segments and pinned across the width, so there is no vulcanised splice to fail and no adhesive line to lift. Single links can be replaced in minutes, and the belt length is set by the number of rows rather than by a press operator's skill. Where the belt is a continuous thermoplastic sheet rather than a linked belt, the joint options are a butt weld, a finger or stepped weld, or a mechanical fastener. A properly welded joint in a compatible polymer can approach the parent material's strength and is far more chemical-resistant than a glued joint, which matters on any washdown line. A cheap mechanical fastener is acceptable for light duty and quick repairs, but it concentrates stress and becomes the point where edge cracking begins.

Tensioning is the second place plastic differs, and it is where most preventable failures start. Engineering thermoplastics creep under sustained load, so a belt tensioned hard on day one will relax, and the take-up has to be long enough to absorb that relaxation rather than merely to hold the initial figure. Over-tensioning to stop a tracking problem is a common field error and a costly one, because it raises the load on the polymer, accelerates creep and drives belt temperature up through the pulley. If a belt will not track at a reasonable tension, the answer is almost always geometry rather than more tension, which is why we usually send a customer to look at their pulleys and guides before discussing any change of material. Crowned pulleys, correctly set guides, and square-cut belt ends solve most tracking complaints, and the hardware side of that argument lives on our conveyor pulleys and conveyor components pages.

Tracking itself then behaves differently on a low-friction plastic surface. The same slipperiness that reduces abrasion also means the belt has less grip to correct itself against a misaligned roller, so a plastic belt on a badly aligned frame mistracks earlier and more sharply than a grippy rubber belt would. On wide, fast belts this is serious enough to justify an active guide, which brings a trained or self-aligning roller into the frame and takes the correction away from belt friction entirely. It is a small addition to the build and it removes a whole class of downtime complaint.

Two practical notes before we leave the subject. First, every joint on a plastic belt that does have one should be treated as a named asset with its own inspection interval, because the joint is where water, product and cleaning chemistry concentrate and where failure is most likely to start. Second, when a plant standardises on one belt family across several lines, the joint method is the detail most often left off the purchase order, and it is the detail that most often causes a substitution argument later. Put it in writing. Many of our customers consolidate all of this under one relationship, placing it as a standing programme with their conveyor belt distributor so that the joint, the tension figure and the spare links are ordered together rather than piecemeal, and the same discipline that keeps a belt running keeps the V-belt manufacturer side of a plant's drives predictable too.

07When Plastic Loses: Impact, Heat, Heavy Lumps and When to Return to Rubber

This is the section that keeps the rest of the page honest. A plastic belt body is a superb solution to a narrow problem, and it is the wrong answer to several that are easy to mistake for that problem. Four conditions should send you back to a rubber-covered belt, and recognising them before you order saves a year of avoidable replacements.

Fast or heavy impact is the clearest of them. If material drops more than roughly half a metre onto the belt, or if the lumps are heavy and sharp, a rigid polymer will crack where a rubber cover would simply flex. There is no polymer on the commodity shelf that fixes this, because the requirement is energy absorption and rigidity is the property that makes a plastic good at sliding wear. Adding an impact bed and lowering the chute may rescue the duty, but if neither is practical the belt family is wrong. I have watched a POM belt at a stone yard fail in six weeks under a drop that a standard rubber belt had been absorbing for two years, and no amount of material shopping would have changed that outcome.

Continuous heat above about 80 °C rules out most commodity polymers, and intermittent heat spikes are worse than a steady low temperature because they cycle the material and promote cracking. Hot clinker, warm castings, freshly fired ceramics and hot food straight from an oven all belong to a different material class. If the product is genuinely hot, the discussion belongs on our heat resistant conveyor belt page rather than in a polymer datasheet.

Heavy lumps under load are the third condition, and they are a load problem rather than purely a wear problem. When a belt has to move large, dense pieces across a long span, the working tension climbs and the polymer has to resist creep while also surviving whatever the lumps do at the transfer points. This is the duty where a cord or fabric carcass earns its place, because the carcass carries the tension and lets the wear surface be chosen freely. Aggressive chemicals are the fourth condition, and they are often mixed up with heat, since many hot processes also involve acids, oils or solvents. A polymer that resists abrasion beautifully can be dissolved or swollen by a solvent it was never checked against, and the failure looks like rapid softening rather than wear. Where oils and fats are present as well, the belt choice belongs alongside our oil resistant conveyor belt and chemical resistant conveyor belt guidance rather than being forced onto a plastic body.

None of this contradicts section 03. It completes it. A plastic belt wins on sliding and light rubbing, and hands over the moment the duty demands impact absorption, high temperature or heavy tension. The skill in specifying one is knowing which of those describes your line, and the two words "abrasion resistant" on an enquiry form do not tell you.

08Writing the Specification, Verifying Delivery and the Cost of Wear

A material decision only survives if it reaches the supplier as a specification rather than as a conversation. Most substitution problems on plastic belts are not malicious. They happen because the purchase order described the belt with three words and left the rest to whoever answered the phone. A specification that names the polymer and the grade, states the dimensions and their tolerances, fixes the joint method and carries a measured wear expectation removes almost all of that risk.

Start with the material line, and be specific enough to be checkable. "Plastic" is not a specification. "Acetal homopolymer, natural, 12 mm thick, FDA-compliant grade" is. Name the polymer family and the grade or trade name where a specific grade matters for food contact, chemical resistance or temperature. Then add the dimensions with tolerances, because a belt cut twenty millimetres narrow on a wide line will mistrack no matter how good the polymer is. Width tolerance of around plus or minus two millimetres and a cut that is square to within about one millimetre are ordinary expectations, and they belong in writing. Add the colour, the surface finish and whether the edges are closed or open, since an open edge on a washdown line is a harbour for product and a starting point for cracking.

The joint then needs its own line, and it should say welded or mechanical, and which of the two weld styles if welded. On a modular belt, give the pitch, the pin material and the number of rows, because a replacement link has to match the original moulding exactly. On a coated belt, name the coating grade rather than just the coating polymer, since two PU belts from two suppliers can differ more than a PU and a PVC belt. Once the belt is described, the verification that the right belt arrived is straightforward and should not be skipped, because the whole point of specifying is to be able to prove compliance at goods-in.

At goods-in, measure rather than trust. Check thickness at several points across the width and along the length, weigh a known area to confirm the material thickness and density are consistent with what was ordered, and on any welded belt ask for a tensile test result on a sample of the joint and compare it with the parent material. On food, pharmaceutical or electronics duty, confirm the compliance documentation matches the grade ordered, not merely the polymer family. Our quality assurance process is built around exactly this sequence of measurement and paperwork, and a supplier who is uncomfortable with it is telling you something useful.

None of that matters as much, however, as the one number that turns a materials argument into a financial one, which is cost per tonne carried. A cheap belt that lasts a year and an expensive belt that lasts three years are not competing on price at all once the maintenance hours, the downtime and the scrap at each change are counted. That is the honest way to compare a plastic body against a rubber-covered belt, and it is why we ask customers to track the replacement history rather than the invoice. Industry context matters here too. The operating pattern on a recycling line, where belts run long shifts with mixed sharp scrap, is nothing like the pattern on a logistics and warehousing deck handling clean cartons, and the same cost-per-tonne calculation will point to different materials on the two. Packaging and food handling add their own constraints on hygiene, which is why a coated belt in a food packaging hall is often chosen from the wider coated range, including the tough PU-faced options described on our PVC belt suppliers for food pages.

Heavy industry deserves a separate sentence because the numbers there are unforgiving. Where a plastic body cannot carry the tension or survive the impact, the answer is not a harder polymer but a different construction, and the choice of belt on a mining and quarrying pit or a cement plant almost always sits with a carcass belt wearing a chosen cover. Plastic earns its place at the edges of those plants, on the clean, light, fast jobs that a solid polymer body handles beautifully, and it is a mistake to stretch it into the main haul. If the broader picture of belt families is unfamiliar, the comparison of polymer and rubber options on our conveyor belt materials page and the selection logic behind conveyor belt thickness both make useful reading before you commit to a build. Where the belt is a linked modular type, the construction details and link-level spares are set out on our modular conveyor belt page, and where a PU-faced belt is the answer, the manufacturing questions worth asking a PU conveyor belt manufacturer are collected there.

Finally, plan for the end of the belt's life before you order the first one. Decide who will inspect it, at what interval, and what measurement will trigger replacement, because a plastic belt wears gradually and gives plenty of warning if someone is watching the thickness number. A plant that replaces on a fixed calendar regardless of condition wastes money. A plant that waits for a failure on a load-bearing plastic belt risks far more than a belt, because a break at speed on a handling line drops product across the floor and stops everything downstream. Set the trigger at a measured percentage loss of thickness and let the calendar retire to a backup role. That single habit, more than any material choice, is what separates plants whose belts quietly last for years from plants whose belts are always in the middle of a problem.

Talk to our engineers about a plastic belt specification

09Frequently Asked Questions

What is actually meant by a plastic belt body, and how is it different from a belt with a plastic cover?

A plastic belt body means the load-carrying structure and the wear surface are the same polymer, with no separate rubber cover and no fabric carcass behind a coating. The material does both jobs at once. A belt with a plastic cover is a different animal, because there the polymer is only a face sitting on a woven or corded backing that carries the tension. The distinction matters because the two fail for different reasons, and a specification written for one will not protect you on the other.

Why do these belts often outlast rubber on the same sliding duty?

Because they wear by a different mechanism. Rubber is elastic, so much of its abrasion is a fatigue process that tears small pieces away where the surface can no longer recover. A hard engineering thermoplastic barely deforms under the same rubbing, so the abrasive gets less purchase and loses less material. On a fine to medium abrasive sliding across the belt under light to moderate load, that difference alone can double service life, which is why tray and parcel handling lines lean on it.

Where do plastic belts fail early compared with a rubber alternative?

Three places. Fast or heavy impact, where a rigid polymer cracks instead of flexing. Sustained heat, where most commodity polymers soften above roughly eighty degrees Celsius and lose their surface. And heavy tension over long spans, where the polymer creeps because nothing behind it is carrying the load. Any one of those conditions usually means a rubber-covered, carcass-backed belt is the better answer, even if the abrasion itself looks mild.

Which polymer should I ask for, and does the grade really matter?

The family comes from the duty, and the grade comes from the environment. Acetal is the general-purpose choice for linked and machined parts, nylon is tougher and suits thin belts that fold, and ultra-high-molecular-weight polyethylene is the sliding-wear champion where the polymer is a wear surface rather than a load body. The grade matters because food contact, chemical resistance and temperature are properties of the grade, not just the polymer name, so two suppliers quoting the same family can still deliver two very different belts.

How do I compare two abrasion figures from two different quotations?

Ask which standard produced each figure, on what specimen geometry, at what load and temperature, and whether the abrasive was renewed between runs. A rotating-drum result and a Taber result are not directly comparable, and a number copied from a raw resin datasheet describes the polymer rather than the finished belt. If a supplier cannot answer those four questions, treat the figure as marketing and ask for a side-by-side test on the same rig with the raw readings.

What should go onto the purchase order so the belt is not substituted?

Name the polymer family and grade, the thickness and width with tolerances, the edge and surface finish, the colour if it matters for hygiene, and the exact joint method including the weld style. For a linked belt, add the pitch, pin material and row count. Then agree a wear trigger as a percentage loss of thickness so replacement is planned rather than reactive. Most substitution arguments disappear once those lines are on the order, because the belt can be checked against them at goods-in.