Understanding Condition Monitoring
Condition monitoring (CM) is the practice of periodically or continuously measuring and trending equipment operating parameters — primarily vibration, temperature, and performance metrics — to assess machine health, detect developing faults early, and schedule maintenance based on actual condition rather than on a fixed calendar. It is the technical engine behind predictive maintenance and condition-based maintenance (CBM): instead of fixing a machine after it breaks (reactive) or overhauling it on a schedule whether it needs work or not (time-based), interventions are timed precisely to the equipment’s measured state.
1. Definition: What is Condition Monitoring?
At its core, condition monitoring turns raw sensor data into a continuously updated picture of machine health. By capturing how a machine behaves when it is healthy, and watching for deviations from that reference over time, an analyst can spot the earliest signature of a fault — often months before failure — and plan a repair around production rather than around a breakdown.
Condition monitoring is fundamental to modern reliability-centred maintenance programs. It provides the data foundation for condition-based decisions that maximise equipment uptime, reduce maintenance costs, prevent catastrophic failures, and optimise spare-parts inventory. The overarching framework for setting up such a program is described in ISO 17359, which lays out the general guidelines for selecting parameters, setting limits, and acting on results.
2. Condition Monitoring vs. Predictive, Preventive and Reactive Maintenance
The terms condition monitoring, condition-based monitoring, condition-based maintenance and predictive maintenance are used loosely and often interchangeably, but they describe different things. Distinguishing them clears up most of the confusion around the subject.
- Condition monitoring (CM) is the measurement activity — collecting and trending parameters such as vibration and temperature to judge a machine’s health. “Condition-based monitoring” and “machine condition monitoring” refer to the same activity.
- Condition-based maintenance (CBM) is the maintenance strategy that acts on those measurements: work is triggered by the machine’s measured condition rather than by the calendar. CM supplies the evidence; CBM is the decision to repair.
- Predictive maintenance (PdM) goes one step further: it extrapolates the condition trend to forecast the remaining useful life, so the repair can be scheduled for the last responsible moment. Predictive maintenance is CBM with a time-to-failure estimate attached.
- Preventive (time-based) maintenance services equipment on a fixed schedule regardless of condition, while reactive (run-to-failure) maintenance waits for the breakdown. Both ignore the machine’s actual state, which is exactly what condition monitoring measures.
In short: condition monitoring is the data, condition-based maintenance is the action, and predictive maintenance is the forecast. All three rest on the same monitoring measurements described below.
3. Core Monitoring Technologies
No single technique sees everything. A mature program layers several complementary measurements so that each confirms and refines the others.
- Vibration analysis (primary): The most comprehensive machinery condition indicator. It detects mechanical defects such as unbalance, misalignment, looseness, and bearing defects, and provides early warning months before failure. Standard techniques include the FFT spectrum, envelope analysis for incipient bearing faults, and long-term trending of overall levels.
- Temperature monitoring: Tracks bearing and winding temperatures and indicates lubrication problems, overload, or cooling issues. It is simple, cost-effective, and a useful way to confirm the severity of a fault already flagged by vibration.
- Oil analysis: Examines wear particles, contamination, and lubricant degradation. Because it samples the actual debris circulating in the oil, it gives early warning of internal wear that surface measurements may miss.
- Thermography: Infrared imaging that reveals hot spots in electrical and mechanical components from a safe, non-contact distance — ideal for surveying switchgear, connections, and bearings.
- Acoustic emission: Listens for the high-frequency stress waves released by crack growth, friction, and the very earliest stages of bearing damage, often detecting a defect before it shows up in the conventional vibration spectrum.
- Motor current signature analysis (MCSA): Electrical signature analysis that detects rotor-bar defects and stator problems without invasive sensors, complementing vibration on electric motors.
The right mix depends on the machine: vibration is the backbone of rotating-machinery monitoring, while oil analysis, thermography and acoustic emission add coverage of failure modes that vibration alone may miss.
4. Condition Monitoring Sensors and Equipment
Every condition-monitoring program is built on the hardware that turns physical change into a usable signal. The choice of sensor follows directly from the parameter being measured and the frequency range of the expected fault.
- Accelerometers are the default vibration sensor — rugged, wide-band, and ideal for the high-frequency signatures of rolling-element bearing and gear faults.
- Velocity sensors (a velometer) are self-generating and well matched to the mid-band where most rotating-machine faults appear.
- Proximity probes are non-contact sensors that measure shaft displacement directly inside fluid-film (sleeve) bearings on large turbomachinery.
- Temperature sensors (RTDs, thermocouples) and infrared cameras support the thermal techniques, while oil-quality and particle sensors support lubricant monitoring.
On the collection side, the equipment falls into two families. Portable data collectors and analysers are hand-carried instruments used to walk a measurement route; a two-channel field unit such as the Balanset-1A both records the data and doubles as a portable analyser and field balancer. Online monitoring hardware consists of permanently wired sensors feeding a rack or edge device that samples continuously and compares every reading against its alarm rules. Selecting equipment is largely a question of criticality, covered in the implementation sections below.
5. Anatomy of a Condition Monitoring System
A condition-monitoring system is more than a sensor on a bearing. Whether portable or permanently installed, every complete system is built from the same logical chain, and it is the later links — not the sensor — that turn isolated readings into actionable intelligence.
- Sensors mounted at consistent, repeatable measurement points.
- Data acquisition — the data collector or DAQ that digitises the signal and computes overall level, spectrum and time waveform.
- A database that stores every reading against the machine and point so a history can accumulate.
- Alarm and analysis logic that compares each new reading to absolute limits and to the machine’s own baseline.
- Reporting and trending dashboards that turn raw numbers into the rising trend lines maintenance teams act on, feeding the work-order system.
The database and trending layers are what separate a true monitoring system from a one-off measurement, and they are the reason consistency of point, unit and procedure matters so much.
6. Implementation Approaches
How the data is collected depends on how critical the machine is and how fast a fault can develop.
Route-based monitoring
A technician walks a fixed route, collecting data from each machine with a handheld data collector or portable analyser on a weekly, monthly, or quarterly cycle. This is cost-effective and scales well across large facilities with many non-critical machines.
Online continuous monitoring
Permanently installed sensors feed an online system that measures continuously or at frequent automatic intervals, with real-time alarming. The cost per machine is higher, so this approach is reserved for critical machinery where unexpected failure is unacceptable.
Hybrid approach
Most real programs combine the two: online monitoring on the few critical assets, route-based collection on the general population. This optimises cost against coverage and is by far the most common arrangement in practice.
7. The Role of a Portable Analyser in the Field
Route-based monitoring lives or dies on the quality of the field instrument. A portable, two-channel analyser such as the Balanset-1A lets a reliability technician capture vibration spectra and overall levels at each measurement point, compare them against the machine’s stored signature, and decide on the spot whether a deviation warrants action. Because the same instrument also measures 1× amplitude and phase, a fault that condition monitoring detects — say, rising 1× vibration from unbalance — can often be corrected immediately by field balancing in the machine’s own bearings, closing the loop from detection to repair without a separate trip or a trip to the balancing shop.
8. Program Implementation and Building a Baseline
A condition-monitoring program is only as good as its setup. Three building blocks matter most.
Equipment criticality analysis
Rank every machine by its impact on production, safety, and cost, then assign a monitoring level accordingly. Critical equipment gets online monitoring; important equipment gets monthly routes; general equipment gets quarterly routes or none.
Baseline establishment
Measure every machine while it is known to be healthy to capture its baseline signature and define its normal operating parameters. This reference is the foundation for all trending — without it, a rising trend has nothing to be measured against.
Alarm limits
Set alert, alarm, and trip levels from the baselines and from recognised severity standards such as ISO 20816 (the modern successor to ISO 10816). Equipment-specific limits beat generic ones, and they should be refined as operating experience accumulates.
9. The ISO 17359 Framework
Setting up a program is not guesswork: the international standard ISO 17359, “Condition monitoring and diagnostics of machines — General guidelines”, defines the procedure that ties every element above together. Its core loop runs from an equipment audit and cost-benefit / criticality review, through selecting the measurement parameters and techniques, establishing the baseline and setting alert and alarm criteria, to data acquisition, diagnosis, and a final feedback step that confirms the maintenance action was effective.
The standard is deliberately technique-agnostic — it governs vibration, thermal, oil and other measurements alike — and it sits within a wider family: ISO 13379 covers data interpretation and diagnostics, ISO 13381 covers prognostics (remaining-useful-life estimation), and ISO 18436-2 defines the training and certification of the people who do the work. Following ISO 17359 is what turns a collection of sensors into a defensible, auditable condition-monitoring program.
10. Benefits and Success Factors
Done well, condition monitoring transforms maintenance from reactive or scheduled to predictive and optimised. The payoffs fall into three groups:
- Operational: increased uptime by preventing unplanned failures, extended equipment life through timely intervention, production continuity by scheduling work during planned outages, and improved safety by heading off catastrophic failures.
- Economic: reduced maintenance cost by eliminating unnecessary preventive work, lower spare-parts inventory by ordering when needed rather than “just in case”, prevention of secondary (collateral) damage through early intervention, and better-targeted labour.
- Knowledge: a deeper understanding of failure modes, feedback into better designs and specifications, and a growing historical database that supports data-driven decisions.
None of this is automatic. Four factors decide whether a program succeeds: sustained management support (resources and a long-term view, since the return on investment takes time); skilled personnel trained in vibration analysis and machinery behaviour — a competence formalised in ISO 18436-2; quality data from consistent procedures and calibrated instruments; and, above all, action on results. A finding that is never acted upon has no value, so condition monitoring must feed the work-order system and include a feedback loop to verify that repairs were effective.
11. Frequently Asked Questions
What is condition monitoring?
Condition monitoring is the practice of measuring and trending equipment parameters — chiefly vibration, temperature and lubricant condition — to judge machine health and detect developing faults early, so maintenance can be timed to the machine’s actual state rather than to a fixed calendar.
What is the difference between condition monitoring and condition-based maintenance?
Condition monitoring is the measurement activity that gathers and trends the data; condition-based maintenance (CBM) is the strategy that acts on it, triggering repairs from the measured condition. Predictive maintenance extends CBM by forecasting how long the machine has before failure.
What techniques are used in condition monitoring?
The main techniques are vibration analysis (the primary indicator for rotating machinery), temperature monitoring, oil and wear-particle analysis, infrared thermography, acoustic emission, and motor current signature analysis. Most programs layer several so each confirms the others.
What sensors and equipment does condition monitoring use?
Accelerometers cover most rolling-element machinery, velocity sensors suit general mid-band readings, and proximity probes measure shaft displacement on fluid-film bearings. Data is gathered either with portable analysers and data collectors on a walking route, or with permanently installed online monitoring hardware on critical assets.
What does a condition monitoring system consist of?
A complete system chains sensors, data acquisition, a historical database, alarm and analysis logic, and trending/reporting dashboards. It is the database and trending layers — not the sensor — that turn isolated readings into the trends a maintenance team can act on.
Which standard governs condition monitoring?
ISO 17359 sets out the general guidelines for a condition-monitoring program — from criticality review and parameter selection through baselines, alarm limits, diagnosis and feedback — supported by ISO 13379 (diagnostics), ISO 13381 (prognostics) and ISO 18436-2 (personnel certification).