The cost of an unplanned stoppage in a crushing and screening plant is not just the price of the failed part: a stopped primary crusher stops the whole plant, shipments and often the customer relationship too. Predictive maintenance aims to detect failures before they occur by measuring the condition of machines, and to carry out maintenance “according to need” rather than “according to the calendar”.
Thanks to falling sensor prices, wireless communication and cloud-based monitoring platforms, predictive maintenance is now within reach not only of large mining companies but also of medium-sized aggregate plants. In this article we explain which methods are used for predictive maintenance of crushing and screening equipment, what to measure and how to get started step by step.
The Difference Between Reactive, Preventive and Predictive Maintenance
| Approach | Logic | Weakness |
|---|---|---|
| Reactive (run to failure) | Repair when it breaks | Unplanned stoppage, secondary damage, high cost |
| Preventive (periodic) maintenance | Replace or inspect at set hours/intervals | Sound parts are replaced early; failures developing between intervals can be missed |
| Predictive (condition-based) maintenance | Measure the condition, follow the deterioration trend, intervene in time | Requires a measurement infrastructure and data interpretation skills |
In practice, the best results come from a smart mix of all three: predictive monitoring on critical equipment, a preventive maintenance schedule for routine tasks (greasing, bolt checks), and run-to-failure for non-critical, cheap parts.
Which Failures Can Be Seen in Advance in Crushing and Screening Plants?
- Bearing failures: Crushers, screen vibrators, conveyor pulleys; can be detected weeks in advance from vibration and temperature.
- Bushing and gear wear: Cone crusher eccentric bushing, countershaft, pinion and gear; seen as metal particles in oil analysis.
- Oil contamination: Dust seal damage, water ingress; identified by oil analysis and particle counting.
- Imbalance and looseness: VSI rotor, impact crusher rotor, screen body; clearly visible in the vibration spectrum.
- Wear parts running out: Can be predicted from liner thickness, CSS drift and the power-capacity relationship.
- Electrical faults: Motor windings and panel connections; by thermography and current analysis.
- Conveyor idler failures: Can be found before they seize with acoustic and thermal scanning.
The Main Methods of Predictive Maintenance
1. Vibration monitoring
Vibration analysis is the most powerful early-warning method for mechanical problems in rotating equipment. Bearing damage, imbalance, misalignment and looseness produce different frequency signatures in the vibration spectrum.
- Periodic measurement: Weekly or monthly route measurements with a portable analyser.
- Continuous monitoring: Fixed wireless sensors on critical machines (primary crusher, VSI, screen vibrators); an alarm when a threshold is exceeded.
- Screen “health” measurement: Measuring stroke, acceleration and motion angle at the four corners of the screen body to detect a broken spring, loose weights or a cracked body early.
2. Oil analysis
Oil is the machine’s “blood test”. For cone crushers, large gearboxes and hydraulic systems, samples taken at intervals of 250–500 hours are monitored for the following:
| Parameter | What it shows |
|---|---|
| Silicon (Si) | Dust ingress; seal or breather filter problem |
| Copper, tin, lead | Bronze bushing wear (e.g. eccentric bushing) |
| Iron, chromium | Gear, shaft and bearing wear |
| Water (ppm) | Water ingress, condensation; the oil film breaks down |
| Viscosity | Oil degradation, the wrong oil being added |
| Particle count (ISO code) | General cleanliness level, filter performance |
The trend of results matters more than a single analysis: copper rising continuously over successive samples is a strong signal to plan a bushing change.
3. Temperature and thermography
- Continuous monitoring of bearing temperatures (PT100 sensors) and tracking of oil return temperature.
- Monthly rounds with a thermal camera: motor housings, panel connections, conveyor idlers, couplings.
4. Process data: power, pressure, CSS
The crusher itself is also a sensor. When power (kW), hydraulic pressure, CSS and capacity data are monitored together:
- The wear rate of liners and the time to replace them can be predicted,
- Rising power at the same capacity signals internal friction or an incorrect setting,
- Frequent overload events show feed or metal separation problems.
5. Ultrasonic and acoustic monitoring
Bearings that need lubrication and conveyor idlers about to seize can be detected from high-frequency acoustic signals. Ultrasound-assisted greasing prevents both under- and over-greasing.
6. Wear measurement
Measuring liner profiles with a laser scanner or templates shows crushing chamber performance and remaining life numerically. Liner changes can then be made at a planned shutdown rather than after capacity has fallen. See mantle and concave replacement.
A Step-by-Step Predictive Maintenance Programme
- Criticality analysis: Which machine stops the whole plant when it stops? The primary crusher, main conveyor and secondary crusher are usually the most critical group.
- Failure modes: For each critical machine, identify the possible failures and the measurement that will reveal them early (e.g. cone crusher → oil analysis + temperature + power).
- Baseline measurements: Record vibration spectrum, temperature and oil analysis values while the machine is healthy.
- Threshold values: Define warning and alarm thresholds based on manufacturer values and your own baselines.
- Data collection: Start with portable instruments and regular routes; add fixed sensors at the points that produce results.
- Decision and action: Define who will do what, and within what time, when an alarm comes in; otherwise the data just remains a record.
- Feedback: After every failure, ask “could we have seen this coming?” and update the programme.
How Is the Return on Investment Calculated?
The return on predictive maintenance is usually covered by a single major failure prevented. The calculation should take into account:
- The cost of an hour of unplanned downtime (lost production × contribution margin),
- Avoided secondary damage (e.g. a bearing failure destroying the shaft and housing),
- Fewer unnecessary early replacements,
- Better preparation of labour and spare parts for planned shutdowns.
Common Mistakes
- Installing sensors without defining alarm management,
- Trying to monitor too many points at once and following none of them properly,
- Taking an oil sample from the wrong point (tank bottom, stopped machine),
- Deciding on a single measured value rather than following the trend,
- Leaving the maintenance team out of the process.
Frequently Asked Questions
Where should a small aggregate plant start with predictive maintenance?
The lowest-cost start is regular oil analysis for cone crushers and gearboxes, a monthly thermal camera round and portable vibration measurement on critical bearings.
Are wireless vibration sensors reliable on crushers?
They are reliable when mounted correctly (flat, clean surface; suitable attachment) and when the right measuring range is chosen. The sensor must have a protection rating suited to dust and impact conditions.
Does predictive maintenance replace preventive maintenance?
Not completely. Routine tasks such as greasing, cleaning and bolt checks continue; predictive maintenance adds information to these routines about “when and on what to intervene”.
Monitoring Matrix by Equipment
| Equipment | Priority measurement | Recommended frequency |
|---|---|---|
| Jaw crusher | Bearing temperature, vibration, run-down time, current | Continuous temperature; monthly vibration route |
| Cone crusher | Oil analysis, oil return temperature and pressure, power, CSS | Continuous process data; oil analysis every 250–500 hours |
| Impact crusher | Rotor bearing vibration and temperature | Continuous or weekly |
| VSI | Vibration, oil temperature/level | Continuous |
| Vibrating screen | Stroke/motion analysis, bearing temperature | Monthly motion analysis; continuous temperature (on critical screens) |
| Conveyors | Slip, misalignment, idler temperature (thermal/acoustic) | Continuous switches; monthly round |
| Gearboxes | Oil analysis, vibration, temperature | Oil analysis every 3–6 months |
| Electrical panels and motors | Thermography, current imbalance | Thermal round every three months |
How to Take an Oil Sample Correctly
- The sample should be taken while the machine is running or immediately after stopping, while the oil is still mixed and warm.
- It should always be taken from the same point (preferably from the sampling valve on the return line, before the filter).
- The sampling valve and hose should first be flushed by letting some oil run through.
- Use clean, labelled sample bottles; note the machine, operating hours, how many hours the oil has been in use and any recent top-ups.
- Samples should be sent to the laboratory without delay.
A wrongly taken sample leads to a wrong decision. For example, a sample taken from the bottom of the tank can look much dirtier than the oil actually circulating.
How Are Threshold Values Set?
- Manufacturer values: For parameters such as bearing temperature and oil pressure, the limits in the operating manual are the first reference.
- Standards: The classifications in the ISO 10816/20816 series provide a general framework for vibration severity; however, for machines that work under shock loads, such as crushers, machine-specific references are more meaningful.
- Your own baseline: Values measured while the machine is healthy and their statistical spread. A marked deviation from the baseline (e.g. doubling) should be a reason for a warning.
- Trend: Regardless of absolute value, a parameter showing a continuous increase should be investigated before it reaches the alarm threshold.
An Example from the Field: A Silent Rise in Copper
In regular oil analyses of a cone crusher, the copper value showed a slow but continuous rise over several samples, while the oil temperature was still below the alarm limit. The maintenance team checked the eccentric bushing at the next planned liner change shutdown and found early-stage damage on the bushing surface. The bushing was replaced in the same shutdown. It was judged that if the damage had progressed, the eccentric and main shaft could also have been damaged, meaning an unplanned stoppage lasting days. This example shows the value of following the trend in oil analysis rather than a single value.
Digital Tools and Data Platforms
Predictive maintenance data comes from sensors, laboratory reports, operator checklists and maintenance work orders. Collecting this data on a single platform (maintenance management software or a cloud-based monitoring system) allows:
- Equipment history to be seen in one place,
- Alarms and work orders to be linked automatically,
- Failure root cause analyses to be recorded,
- Spare parts consumption and costs to be monitored.
For details on remote monitoring and digital twins, see our article on automation and digital twins.
Team and Competence
The success of predictive maintenance depends more on people than on technology. Operators need to notice and report unusual sounds, smells and vibration; maintenance technicians need to use measuring instruments correctly and interpret the data; and managers need to allocate resources to act when an alarm comes in. A simple training programme and clearly defined responsibilities can bring more benefit than an expensive sensor investment.
More Questions
Why is vibration measurement difficult on crushers?
Crushers by nature work under shock loads; their vibration levels and spectra are therefore “noisier” than those of rotating machines. This is why comparison with the same machine’s own history and trend tracking matter more than absolute values.
What problems does a thermal camera show?
It can show overheating bearings, loose or oxidised electrical connections, unevenly loaded motor phases and conveyor idlers about to seize.
Does predictive maintenance reduce spare parts stock?
Seeing failures in advance allows parts to be ordered close to when they are needed. However, safety stock is still required for critical parts with long lead times.
Maintenance Support from CSP Mühendislik
CSP Mühendislik provides service, maintenance and repair for crushing and screening plants and supports failure analysis, overhauls and spare parts supply. Contact us to create a maintenance programme for your plant.
