Condition Monitoring: Data Signals and Condition Monitoring #
In modern warehouse operations, condition monitoring is the practice of turning machine data into maintenance decisions. Conveyors, sorters, palletizers, stretch wrappers, and automated storage machines all emit measurable signals while they run: motor current, vibration, temperature, position feedback, cycle times, and control-system fault codes. These signals are not just operational noise; they are evidence of component health. When captured, organised, and interpreted correctly, they tell a maintenance team whether a unit is stable, deteriorating, or approaching failure. This article explains how warehouse operators, maintenance engineers, and controls teams can use data signals to design better inspections, code failures accurately, plan spares, and reduce the number of repeat faults.
What Condition Monitoring Means in a Warehouse Environment #
Condition monitoring is often contrasted with time-based preventive maintenance. A time-based plan replaces a bearing every six months because the calendar says so. A condition-monitoring approach replaces the bearing when its vibration, temperature, or running sound indicates degradation. In practice, the two approaches complement each other. Time-based intervals provide a safety net, while condition monitoring provides precision and often extends component life.
Warehouse equipment is particularly suited to condition monitoring because most assets are fixed, operate on predictable cycles, and are already connected to programmable logic controllers (PLCs), drives, and sensors. The data signals that control the machine are the same signals that reveal its condition. A conveyor that begins taking longer to move a carton, a sorter that requires additional air pressure, or a vertical lift that draws more current on every cycle are all communicating through their normal operating signals. The maintenance challenge is not a lack of data; it is a lack of structured interpretation.
Data Signals: Origin, Type, and Meaning #
Data signals relevant to condition monitoring fall into four broad families: electrical, mechanical, thermal, and operational. Each family provides a different view of the same physical asset, and no single family tells the whole story.
Electrical Signals #
Motor current, voltage, power factor, and drive output frequency are common electrical signals. Current draw is one of the most versatile indicators available in a warehouse. A conveyor motor that draws consistently higher current may be handling heavier loads, experiencing mechanical drag, or suffering from a deteriorating bearing. A drive that reports repeated overcurrent events may point to a mechanical jam, a failing brake, or an electrical insulation problem. Electrical signals are best interpreted in the context of load: the same motor may draw 6 amps when carrying a full carton and 4 amps when empty. Understanding the load profile is essential before judging the signal.
Mechanical Signals #
Vibration and acoustic signals are the primary mechanical data streams. Vibration can be measured continuously, periodically, or on demand using portable instruments. In a warehouse environment, continuous vibration monitoring is usually reserved for critical assets such as large fans, high-speed sorters, or refrigeration compressors. Periodic vibration measurement is more economical and is often carried out as part of a scheduled inspection route.
Acoustic signals include the sound emitted by a bearing, gearbox, or coupling. Ultrasound inspection is a useful maintenance technique because high-frequency sound from a failing bearing or an air leak is often audible before low-frequency vibration becomes measurable. Even the human ear, when trained, remains a useful sensor: experienced operators frequently notice a change in the “tone” of a sorter or conveyor before any alarm appears.
Thermal Signals #
Temperature is a lagging indicator. It rarely reveals a problem at the moment the problem begins, but it is a reliable confirmation of an ongoing fault. Gearbox oil temperature, motor winding temperature, and control-panel ambient temperature can all be monitored through existing drive sensors or dedicated thermocouples. Thermal imaging cameras are also valuable during inspection rounds. They can quickly reveal a hot motor terminal, an overloaded electrical contact, or a bearing housing that is running significantly hotter than its neighbour.
Operational and Control Signals #
Operational signals include cycle times, throughput rates, position tolerances, sensor registration, and fault-code histories. These are often the most accessible data because they already exist inside the control system. A palletizer that takes an extra two seconds per cycle, a shuttle that misses its target position more frequently, or a photoeye that registers non-consecutive cartons are all producing data signals that indicate degradation. Control-system trends are especially valuable because they record the exact operating conditions at the moment of a fault.
Raw signals differ from derived indicators. A raw signal is the direct measurement, such as current in amps or temperature in degrees. A derived indicator is a mathematical interpretation, such as the trend slope, the rate of temperature rise, or the harmonic content of a vibration measurement. Derived indicators are usually more stable and informative than isolated raw values. A motor may run at 80 degrees on one day and 81 degrees the next; both measurements are within limits. But if the trend shows a steady rise from 68 to 81 degrees over three weeks, that rate of change is the true condition signal.
Component Interactions: Reading the System, Not Just the Part #
A common mistake in condition monitoring is focusing on a single component without considering its neighbours. Warehouse equipment rarely fails in isolation. A conveyor system is a chain of interacting elements: the drive motor, coupling, gearbox, belt or chain, pulleys, bearings, sensors, and the controller that coordinates them. A fault at one point produces a ripple of abnormal signals across the rest of the system.
Consider a typical belt conveyor with a mechanical misalignment. The misaligned belt rubs against the conveyor frame, increasing friction and creating belt debris. The additional friction increases the torque demand on the gearbox and motor. Motor current rises. The gearbox may begin to run warmer because it is transmitting more torque than normal. The PLC may record intermittent load spikes that trigger an overtorque warning. A maintenance engineer who only looks at motor current might conclude the motor is failing and replace it. The motor, however, is merely the messenger. The root cause is belt alignment. If the evidence trail is followed from electrical signal back to mechanical cause, the correct repair is realignment and not motor replacement.
Component interactions also matter when using multiple data signals to confirm a diagnosis. A motor with a failing bearing will usually show both increased vibration and increased temperature, but the vibration signal will change before the temperature signal. A motor with a failing supply cable may show intermittent faults and voltage imbalance without any vibration signature. The wise approach is to collect evidence from several signal families and look for agreement. When electrical, mechanical, and operational signals all point in the same direction, the diagnosis is much stronger than when only one signal is abnormal.
Observable Symptoms and Their Diagnostic Weight #
Observable symptoms are the human-visible or system-visible signs that accompany a developing fault. They include abnormal noise, excessive heat, drifting position tolerances, increased cycle time, intermittent error codes, and physical contamination such as oil stains or belt debris. Not all symptoms carry the same diagnostic weight.
- Loss of repeatability is a high-weight symptom. A robot or shuttle that used to place a load within 5 mm of a target and now varies by 20 mm is showing a clear degradation trend. Position drift often precedes hard faults.
- Intermittent faults are medium-to-high weight. An intermittently tripping photoeye or photo sensor can indicate a contaminating lens, a failing cable, or a vibration-induced loose connection. Because they are intermittent, they are easily misattributed to software or external interference.
- Abnormal noise is medium weight and needs corroboration. A slight change in gearbox whine might indicate a lubrication issue, or it might simply be a change in ambient temperature. A persistent knocking sound is more significant.
- Low-weight symptoms include isolated fault codes, one-off excursions in current, or a single hot point on a thermal image. These require confirmation through a second measurement.
Diagnostic weight increases when a symptom repeats on a regular cycle. A current spike that occurs every time a carton passes a particular diverter, or a vibration peak that appears only when a sorter runs at high speed, points directly to a load-dependent, position-dependent, or speed-dependent cause. Recording the conditions of each symptom is as important as recording the symptom itself.
A Practical Diagnostic Table for Common Warehouse Signals #
The following table summarises common data signals, their typical meaning in a warehouse context, and common interpretation errors. It is intended as a training aid, not as a substitute for OEM documentation or site-specific engineering analysis.
| Data Signal | Typical Healthy Behaviour | Early Warning | Common Misinterpretation |
|---|---|---|---|
| Motor current (conveyor) | Stable draw for a given load; small cycle-to-cycle variation | Slow upward trend over days; spike at a particular position | Blamed on product weight when it is actually caused by belt drag or bearing wear |
| Drive overcurrent / torque faults | Absent or very rare | Intermittent faults during acceleration or at specific speeds | Treated as a drive problem when the real cause is a mechanical jam or a failing brake |
| Vibration velocity on a motor or gearbox | Low and steady; no new frequency peaks | Rise in broadband vibration; a distinct bearing or gear mesh frequency | Compared against a generic limit from memory instead of the equipment baseline |
| Operating temperature | Stable warm-up then plateau within a normal range | Continued rise during normal running; warmer than parallel identical units | Dismissed as seasonal when the rise is driven by mechanical friction |
| Cycle time of a shuttle or lift | Consistent cycle duration regardless of load position | Gradual lengthening; controller accumulates extra time waiting for sensors | Attributed to operator requests or traffic when the cause is slower mechanical response |
| Position error / registration drift | Small, stable deviation from set point | Increasing variance; occasional re-registration events | Blamed on sensor calibration when the cause is a loose coupling or stretched belt |
Evidence Collection: From Timestamps to Physical Samples #
Condition monitoring is only as good as the evidence behind it. Collecting evidence properly is a discipline that separates a reliable diagnosis from a guess. The following principles apply across all warehouse asset types.
Record context with every reading. A data signal without context is nearly useless. A motor current reading of 7.5 amps means little unless you also know the load, the speed setting, the ambient temperature, and the time of day. For rolling equipment, record the zone of the conveyor, the product type, and the direction of travel. Context allows future readers to compare like with like.
Use timestamps and event sequences. WHEN a fault occurred is as important as the fault itself. The exact order of control-system events, the time between an input command and a response, and the duration of an overcurrent event all become part of the evidence. A fault that occurs only on a cold morning start has a different cause than one that occurs after three hours of continuous running.
Capture physical and visual evidence. A photograph of a belt edge wearing, a small oil sample from a gearbox, a metal debris sample from a magnetic plug, or a thermal image taken from a consistent position can all be retained in the maintenance record. These physical artefacts often resolve disputes between engineering teams about what happened first. Store them methodically with a clear reference to the asset.
Tie evidence to the operating state. For example, when collecting vibration data from a sorter, record whether the measurement was taken during idle, partial speed, or full speed. The same machine will produce very different vibration signatures in different operating states. A vibration spike that only appears at full speed is a clue to a resonance or imbalance issue, not a general deterioration problem.
Common Interpretation Errors #
Whatever the quality of the data, interpretation can still go wrong. The same signal can lead to very different conclusions depending on the interpreter’s assumptions. The following errors are common in warehouse maintenance teams.
- Mistaking correlation for causation. Two signals change at the same time, and the engineer assumes one caused the other. For example, motor temperature rises and throughput also rises; the temperature rise may be caused by increased throughput, not by a mechanical fault. Isolating variables is difficult in a live operation, but it is necessary.
- Over-relying on fixed thresholds. Generic threshold limits from a manual are useful starting points, but they do not account
Related Pearl Gateway Guides #
Site-Specific Review Worksheet #
This educational worksheet supports a structured review of condition monitoring: data signals and condition monitoring. Begin by identifying the equipment boundary, control ownership, operating modes, material characteristics, upstream dependencies and downstream consequences. Record what the system is expected to do, what was actually observed and which evidence is time-aligned. Avoid changing several variables at once, because simultaneous changes make cause and effect difficult to establish.
Evidence to collect #
- Operating mode, active mission or route, and the exact sequence state.
- Alarm history, device state changes and controller timestamps.
- Physical observations such as alignment, contamination, wear, obstruction and load condition.
- Recent maintenance, software changes, parameter changes and recurring work orders.
- Upstream and downstream readiness, including blocked, starved and unavailable conditions.
Decision boundaries #
Use approved site procedures and competent engineering judgment before intervention. General information in the Maintenance & Reliability library cannot determine whether a specific machine is safe to enter, restart or modify. Preserve original settings, document authorized adjustments and establish a rollback point before controlled testing. When evidence conflicts, stop and resolve the timestamp, naming or measurement discrepancy before drawing a conclusion.
Closeout record #
A useful closeout record states the symptom, confirmed cause, evidence, corrective action, validation method, residual risk and follow-up owner. It should also identify whether the event exposed a design weakness, maintenance gap, training issue, spare-parts issue or monitoring blind spot. This turns a single recovery into reusable reliability knowledge without treating one observation as universal.
