Key Points
- Most failures on a plastics line build up over weeks, and they give off measurable signals the whole time. Vibration usually changes first, well before heat, noise or a visible problem.
- Each critical asset has its own signature: gear mesh changes on extruder gearboxes, 1x and 2x running-speed peaks for misalignment, broadband noise for pump cavitation and bearing defect frequencies across almost everything.
- No single signal tells the whole story. Vibration leads on mechanical wear, but motor current, temperature, speed and process load catch what vibration misses and confirm when a change is real.
- A signal only matters if it becomes a planned repair, and that takes a clear diagnosis and a team ready to act on it.
Failures Don't Come Out of Nowhere
When a plastics line goes down unexpectedly, it can feel like the failure came out of nowhere. The gearbox was running fine on Tuesday. On Wednesday, it seized.
But in my experience evaluating equipment in the field strictly a route-based procedure only provides a small data point to the complex world we deal with every day as maintenance and reliability professionals. Bearings, gears, pumps and motors wear in predictable ways, and that wear leaves a trail in the data long before anyone hears it on the floor.
That trail matters because of what a surprise failure costs a plastics plant. The line stops, material in the barrel gets scrapped and orders slip. Industrial plants report a median of $125,000 per hour of unplanned downtime (ABB, 2023). Reading the signs early is how a plant turns that from an unscheduled repair to routine maintenance.
Here's what those signs look like, where they show up on a plastics line and how to tell a real problem from noise.
How a Failure Develops
Reliability engineers use a simple idea called the P-F interval. "P" is the point where a developing problem first becomes detectable. "F" is functional failure, when the machine stops doing its job. The time between the two is your window to plan a repair.
A typical bearing failure moves through stages like these:
- Failure begins. A small defect forms inside the bearing, often from a lubrication problem, contamination or a microscopic crack under the surface. The machine runs normally, and nothing shows up yet, even in the data.
- Subtle vibration changes. High-frequency energy starts to appear at the bearing's defect frequencies. Nothing is visible or audible yet. This is where continuous vibration monitoring catches it.
- Clearer vibration patterns. The defect grows, and the signal becomes stronger and easier to identify.
- Audible noise. A technician walking by may hear it now.
- Heat. Friction raises the temperature. A handheld thermometer or the operator's hand can pick it up.
- Failure. The bearing seizes, and often damages the shaft, gears or housing around it.
The earlier you detect a problem, the longer your window, the cheaper the cost to intervene. That's why temperature alone isn't enough. By the time a bearing is running hot, you may have days left instead of weeks. Vibration buys you the time to order the part, schedule the work and plan around production.
That timeline describes a mechanical fault. Not every failure on a plastics line starts there. A motor winding breaking down, a heater band drawing uneven current or a screen pack clogging upstream of the die may barely move the vibration needle at first. Those problems show up first in electrical current, energy draw or process load.
So the real goal is to watch several signals that each sit at a different point on that curve, not only the one that moves first. When vibration, temperature, current and speed are read side by side, each one fills in what the others can’t see.
The Signals, Asset by Asset
Every machine on a plastics line has its own way of telling you something is wrong. Here's a quick reference for the assets that matter most.
| Asset | Common failure | Early warning sign | What confirms it | Why it matters |
|---|---|---|---|---|
| Extruder gearbox | Gear tooth wear, thrust bearing wear | Rising gear mesh frequency with growing sidebands; bearing defect frequencies | Oil temperature trending above baseline; motor current rising at the same screw speed | Stops the whole line; parts often have long lead time |
| Extruder drive motor | Misalignment, imbalance, bearing wear, winding issues | Elevated 1x and 2x running speed (misalignment), 1x radial (imbalance) | Current imbalance across phases; winding temperature rise | Same impact as the gearbox it drives |
| Molding hydraulic pump | Internal wear, cavitation | Broadband high-frequency noise (cavitation) | Higher current draw per cycle; oil temperature climbing | Slower cycles, inconsistent parts, more scrap |
| Chiller compressor | Bearing wear, mechanical looseness | Bearing defect frequencies; multiple running-speed harmonics (looseness) | Longer run times and higher energy use for the same cooling load | Can stop every press it cools |
| Air compressor | Bearing wear, overheating | Bearing defect frequencies; temperature trending up | Rising energy per unit of air delivered; short cycling | Stops controls, handling and blow molding |
| Packaging conveyors and palletizers | Gearbox and motor wear, misalignment | Gear mesh changes; 1x and 2x running speed | Current spikes at startup or under load | Finished product can’t ship |
A Closer Look at the Extruder Gearbox
The extruder gearbox deserves extra attention because it carries the heavy axial load from the screw pushing melt through the die. That load lands on the thrust bearing, which makes it one of the most common and costly failure points on the line.
Two patterns are worth watching. As gear teeth wear, the gear mesh frequency (the number of teeth multiplied by shaft speed) changes, and sidebands appear around it at the shaft's running speed. As the thrust bearing wears, energy shows up at its defect frequencies, often first in the axial direction. Either pattern trending upward over a few weeks is a strong sign to start planning.
Why One Signal Just Isn’t Enough
Vibration is the best early warning for mechanical wear, and it’s the backbone of everything above. But a plastics line fails in more ways than bearings and gears, and some of the costliest problems hide in signals vibration doesn’t carry.
- Electrical faults. Insulation breakdown, phase imbalance and drive problems show up in current and voltage before they show up as vibration or heat.
- Process-driven load. A clogging screen pack, a material change or a cold barrel zone raises the load on the extruder motor. That shows up in current and torque, not as a mechanical fault, but it can still trip the line or wear the drivetrain faster.
- Energy drift. A chiller or air compressor working harder to deliver the same output is often the first sign of a leak, fouling or internal wear. It also costs money every hour it runs that way.
Signals become most useful when they’re read together. A vibration rise with climbing oil temperature and higher current at the same speed is a developing fault worth planning for. A vibration rise that lines up exactly with a speed change or a new product run is probably the process, not the machine. Each signal is a check on the others.
That’s the idea behind unified asset health: one view of each machine that combines vibration, temperature, speed, electrical data and maintenance history. You see the whole machine instead of one symptom at a time.
Why Plastics Lines Are Tricky to Monitor
The signals above are well understood. Reading them correctly on a plastics line is where it gets harder. There are four challenges I see most often.
Variable speeds. Extruders often run on variable frequency drives, and line speed changes with product and gauge. Because fault frequencies move with shaft speed, an analysis that assumes a fixed speed can miss real problems or flag false ones. Good monitoring tracks speed and adjusts for it.
Cyclic loads. An injection molding machine doesn't run at a steady state. Injection, hold, cooling and ejection each put a different load on the hydraulics. A reading taken at the wrong moment in the cycle can look alarming when nothing is wrong, or normal when something is. Trending over many cycles separates real change from normal variation.
Heat. Barrels, dies and molds run hot by design, so a high temperature on its own tells you very little. What matters is change against each machine's own baseline. That's another reason vibration is the more reliable early signal.
Neighboring equipment. Plastics plants pack a lot of machinery into a small space. Vibration from a nearby grinder or compressor can travel through the floor and show up on the wrong asset. Sensor placement and frequency analysis help tell the difference.
None of this means plastics equipment can’t be monitored well. It means the analysis has to account for how these machines actually run. That’s where combining signals pays off. Speed data separates a VFD change from a real fault. Current and cycle timing show where a molding machine is in its cycle. A machine’s own temperature and energy baselines show whether “hot” means “normal for this die” or “something changed.”
Turning a Signal Into a Planned Repair
Detecting a problem is only half the job. The value comes from what happens next. A useful alert answers four questions:
- Which asset, and which component? "Gearbox, output shaft bearing" is actionable. "Abnormal vibration" is not.
- What's the likely fault? Bearing wear, misalignment, looseness and cavitation each call for different parts and different fixes.
- How urgent is it? The team needs to know whether this is a next-changeover job or a this-week job.
- What should we do? A clear recommendation, like "inspect lubrication" or "plan bearing replacement," saves hours of guesswork.
The more signals behind an alert, the better it answers those questions. “Bearing defect frequencies rising, oil temperature up, current steady” points to a different fault and a different urgency than vibration alone. Tying that view to the asset’s work order history (when it was last greased, which part was swapped last quarter) is often what takes a diagnosis from likely to certain.
Where Tractian Fits
This is the work my team does every day at Tractian. Our condition monitoring sensors capture vibration, temperature and running speed on critical equipment around the clock. Our energy monitoring adds the electrical side: current, voltage and power draw. That data lives in the same platform as our CMMS, so the diagnosis, the work order and the asset’s repair history sit in one place. Our AI benchmarks each machine against millions of historical machine-hours to deliver diagnoses that name the component, the likely fault and what to do next.
The sensors run on 4G/LTE straight to dedicated receivers, with no plant Wi-Fi or IT project required. That makes them practical on long extrusion lines and in hot, noisy areas where network coverage is hard.
We've seen this work on equipment much like a plastics line. At tire maker Pirelli, a gearbox flagged a possible gear-wear signal. The team investigated, found an undetected oil leak and corrected it before it could cause a failure.
In an independent study, Verdantix modeled Tractian’s financial impact on a mid-sized manufacturer. The result: a 401% ROI over three years, an 8-month break-even, and about $1.4 million saved per year, with savings starting in year one (Verdantix Independent Study).
If you're deciding where to begin, read Predictive Maintenance for Plastics Manufacturing: Where to Start. And if you'd like to know what your own equipment is saying, my team would be glad to walk your line with you.
FAQ
What are the early warning signs of equipment failure in plastics manufacturing?
Changes in vibration are usually the earliest sign, often weeks before a failure. Common patterns include bearing defect frequencies, gear mesh changes on gearboxes and elevated running-speed peaks from misalignment or imbalance. Rising temperature, audible noise and visible damage tend to appear later.
Why is vibration a better early indicator than temperature?
Most mechanical faults change a machine's vibration before they generate enough friction to raise its temperature. By the time a component is running hot, the failure is often close. Vibration gives you a longer window to plan the repair.
What is unified asset health monitoring?
It means combining several data sources into one view of each machine: vibration, temperature, speed, electrical current, energy use and maintenance history. Each signal catches different failure modes, and reading them together confirms real problems and filters out false alarms.
Can vibration monitoring catch electrical problems?
Sometimes, but usually late. Motor winding degradation, phase imbalance and drive issues tend to show up first in current and voltage data. That’s why pairing vibration with electrical monitoring gives better coverage on motors and drives.
How do you monitor an extruder that runs at variable speeds?
Fault frequencies shift with shaft speed, so the analysis has to track speed and adjust for it. Without that, real faults can be missed and normal changes can trigger false alarms.
What is the P-F interval?
It's the time between the point when a developing failure first becomes detectable (P) and the point when the equipment fails (F). The earlier you detect a problem, the longer that window, and the easier it is to plan the repair.
Can injection molding machines be monitored effectively?
Yes. Because molding machines run in cycles, the key is trending data across many cycles and comparing each machine to its own baseline. That separates real wear in pumps and motors from normal changes in load.