Detect machine degradation before it stops production – how a condition monitoring system works

A condition monitoring system is a hardware and software solution that monitors the condition of machines and equipment using parameters such as vibration, temperature, pressure, oil condition, current, voltage and acoustic emission. The goal is to detect technical degradation before it turns into a functional failure, downtime or secondary damage to the asset.

In practice, condition monitoring is not a single technology. It is a set of methods, sensors, measurement rules, analytics and response procedures that help maintenance teams make better decisions: when to keep watching a machine, when to schedule an intervention, when to raise a work order, and when to simply leave the equipment running.

What is a condition monitoring system?

A condition monitoring system is a system for monitoring the technical condition of assets. It collects data from sensors and measuring devices, processes it in an acquisition or gateway layer, analyses it in a software platform, and generates alerts, reports or maintenance actions.

The simplest definition is as follows:

Condition monitoring system = sensors + data acquisition + software platform + analytics + maintenance response.

The system monitors parameters that describe machine health, such as:

• vibration;
• temperature;
• pressure;
• oil condition;
• current and voltage;
• acoustic emission;
• ultrasonic signals;
• alarm statuses;
• process trends;
• deviations from normal operating conditions.

How does a condition monitoring system work?

A condition monitoring system works as a data chain: from the physical measurement, through acquisition and transmission, to analysis, alert and maintenance decision.

A typical flow looks as follows:

Sensor → data acquisition / gateway → communication → software platform → analytics → alert / work order / report

Layer	Components	Function
Sensing	Accelerometers, thermocouples, oil sensors, proximity probes, current transformers, ultrasonic detectors	Measurement of physical parameters
Data Acquisition	DAQ modules, signal conditioners, IoT gateways, edge devices	Digitisation, filtering and data preparation
Communication	4–20 mA, Modbus, Ethernet, Wi-Fi, Bluetooth, LoRa, 5G	Data transmission to the platform
Software Platform	Condition monitoring software, CMMS integration, local or cloud analytics	Data storage, visualisation and processing
Analytics	Trends, alarm thresholds, FFT, envelope analysis, anomaly detection, AI/ML	Detection of degradation patterns
Action	Alerts, work orders, scheduling, reports	Triggering the maintenance response

In an industrial environment the system can run locally, in the cloud or in a hybrid model. For critical assets, online monitoring – continuous or near-continuous measurement – is common. For less critical equipment, a portable approach may be enough: periodic measurements taken by a technician during a route.

Condition monitoring can be integrated with a CMMS, that is a Computerized Maintenance Management System. In that model an alert is not just information on a dashboard: it can trigger a work order, an inspection, an escalation or a decision to schedule a shutdown.

The five elements of condition monitoring

The five basic elements of condition monitoring are parameter selection, sensor selection, data collection, analysis and the maintenance decision.

Element	Meaning	Example
1. Parameter selection	Defining what best describes the condition of a given asset	Vibration for a bearing, temperature for a motor, moisture for oil
2. Sensor or instrument selection	Matching the measuring device to the parameter and operating conditions	Accelerometer, thermocouple, IR camera, ultrasonic sensor
3. Data collection	Deciding how and how often the measurement is taken	Continuous online measurement or periodic route-based readings
4. Data analysis	Interpretation of trends, spectra, thresholds and deviations	FFT, envelope analysis, temperature trend, threshold alarm
5. Decision	Translating the analysis result into action	Alert, inspection, work order, planned shutdown

This model reflects the logic of a condition monitoring rollout: it does not start with buying sensors, but with the question of which failure mode we want to detect and which parameter describes it best.

Types of condition monitoring techniques

The core condition monitoring techniques are vibration analysis, thermography, oil analysis, ultrasonic testing, motor current analysis and acoustic emission monitoring.

Each method measures a different type of signal and detects different types of degradation. That is why an effective condition monitoring system often combines several techniques rather than relying on one.

Vibration analysis

Vibration analysis is one of the most widely used condition monitoring techniques for rotating machinery.

It consists of measuring and analysing vibration patterns in rotating or reciprocating machines. It makes it possible to detect unbalance, misalignment, looseness, bearing damage, gearbox defects, resonance and gear meshing problems.

Typical parameters:

• vibration velocity, usually in mm/s;
• acceleration, usually in g;
• displacement, usually in µm;
• time-domain analysis;
• frequency-domain analysis, including FFT;
• envelope analysis.

Typical sensors:

• piezoelectric accelerometers;
• velocity sensors;
• proximity probes, often eddy-current sensors.

Typical applications:

• motors;
• pumps;
• fans;
• compressors;
• turbines;
• gearboxes;
• highly critical rotating machinery.

Vibration analysis delivers early failure indicators for predictive maintenance. That does not mean, however, that every vibration monitoring system is automatically a predictive system. Prediction requires an additional layer of historical data, models, risk forecasting and a decision-making process.

Thermography and thermal imaging

Thermography, also called thermal imaging, is the non-contact measurement of surface temperature distribution using infrared cameras.

The technique makes it possible to detect unusual thermal patterns that may indicate an electrical, mechanical or process problem.

Typical parameters:

• surface temperature;
• temperature gradient;
• material emissivity;
• comparison with a reference area.

Typical equipment:

• handheld thermal imaging cameras;
• fixed IR cameras;
• thermal sensors.

Typical applications:

• switchgear and electrical panels;
• electrical connections;
• motor windings;
• bearings;
• steam traps;
• heat exchangers;
• insulation;
• HVAC installations;
• refractory linings.

Thermography is particularly useful for detecting hot spots. It can reveal loose electrical connections, overloads, insulation degradation, bearing overheating, steam leaks or restricted heat transfer.

Its major advantage is the non-contact nature of the measurement and the ability to cover a large area quickly.

Oil analysis

Oil analysis consists of laboratory or inline testing of lubricant samples in order to assess oil condition, contamination and wear particles.

It is a particularly important technique for oil-lubricated systems, where the condition of the lubricant can reveal component degradation before clear changes appear in vibration or temperature.

Typical parameters:

• viscosity;
• particle count;
• water content;
• TAN and TBN;
• wear metals, for example Fe, Cu, Al;
• oxidation products;
• contaminant identification;
• spectrographic or ferrographic results.

Data collection methods:

• periodic samples taken manually and sent to a laboratory;
• inline sensors working directly in the system.

Typical applications:

• hydraulic systems;
• gearboxes;
• motors;
• turbines;
• compressors;
• other oil-lubricated systems.

Oil analysis makes it possible to detect bearing and gear wear, the presence of water, dust or other contaminants, oil degradation, oxidation and additive depletion.

Ultrasonic testing

Ultrasonic testing in condition monitoring consists of detecting high-frequency sound, typically in the range of about 20–100 kHz, generated by friction, impacts, turbulence, leaks or electrical discharge.

There are two basic approaches:

• airborne ultrasound – for detecting leaks and electrical discharge;
• structure-borne ultrasound – for assessing bearings and contact signal sources.

Typical equipment:

• handheld ultrasonic detectors;
• contact probes;
• airborne probes;
• headphones and signal recorders.

Typical applications:

• steam traps;
• compressed air leaks;
• vacuum leaks;
• bearing monitoring;
• electrical inspections;
• detection of corona, arcing and tracking in switchgear.

Ultrasonic testing is a highly practical technique for detecting energy losses related to compressed air. It can also reveal early bearing defects before they become visible in temperature or classic vibration analysis.

Motor current analysis

Motor current analysis, also referred to as MCA or MCSA, consists of analysing the signatures of the current drawn by a motor in order to detect electrical and mechanical faults.

The technique can use existing current transformers, so it often does not require additional sensors to be installed directly on the motor.

Typical parameters:

• current spectrum;
• FFT of motor current;
• voltage;
• power factor;
• load deviations.

Typical applications:

• electric motors;
• drives;
• motor-driven equipment;
• electrically driven pumps, fans and compressors.

Motor current analysis can detect broken rotor bars, stator winding faults, eccentricity, bearing problems and load anomalies, among others.

Its advantage is its non-invasive nature. The measurement is based on an electrical signal that is often already available in the plant infrastructure.

Acoustic emission monitoring

Acoustic emission monitoring consists of detecting short-lived elastic waves generated by crack propagation, material deformation, leaks or very early component damage.

The typical frequency range is higher than in classic ultrasonic testing and can be around 100 kHz–1 MHz.

Typical applications:

• pressure vessels;
• pipelines;
• storage tanks;
• structures;
• composites;
• high-speed bearings;
• components exposed to cracking or corrosion.

Acoustic emission monitoring is useful wherever structural integrity matters, not only the condition of a rotating machine. It makes it possible to detect crack initiation and growth, corrosion, leaks and very early symptoms of damage.

Comparison of condition monitoring techniques

Condition monitoring techniques differ in the parameter measured, the type of defects detected, cost, application and the way the measurement is taken.

Technique	Parameter measured	Problems detected	Equipment type	Online / portable	Relative cost
Vibration analysis	Velocity, acceleration, displacement	Unbalance, misalignment, bearing defects, gearbox faults	Rotating machinery	Both models	Medium–high
Thermography	Surface temperature	Hot spots, insulation failures, electrical problems, overheating	Electrical and mechanical equipment	Mainly portable	Medium
Oil analysis	Viscosity, particles, moisture, metals	Wear, contamination, oil degradation	Oil-lubricated systems	Lab or inline	Low–medium
Ultrasonic testing	High-frequency sound	Leaks, bearing defects, electrical discharge	Wide range of assets	Mainly portable	Low–medium
Motor current analysis	Current and voltage spectrum	Rotor and stator faults, eccentricity, bearing defects	Electric motors	Online	Low
Acoustic emission	Elastic waves	Cracks, corrosion, leaks	Structures, vessels, pipelines	Online	Medium–high

This table should not be read as a ranking. Vibration analysis may be the primary method for pumps and turbines, but it will not replace oil analysis in assessing lubricant condition, nor thermography when inspecting electrical panels. The choice of technique should follow from the failure mode, asset criticality and data availability.

Online or portable condition monitoring?

Online condition monitoring means continuous or near-continuous measurement through permanently installed sensors. Portable condition monitoring means periodic measurements taken with a handheld instrument during a planned route.

Dimension	Online condition monitoring	Portable condition monitoring
Installation	Permanently mounted sensors	Handheld instruments used by a technician
Frequency	Continuous or near-continuous	Periodic, for example monthly or quarterly
Data granularity	High, covers transient events	A snapshot at the moment of measurement
Cost per point	Higher: sensor, cabling, DAQ, integration	Lower: one device for many assets
Best use	Critical, expensive, hard-to-access assets	Non-critical, numerous and accessible assets
Detection speed	Immediate, with a real-time alert	Delayed until the next measurement route
Scalability	Cost grows with the number of points	One instrument serves many assets
Typical industries	Power generation, oil & gas, critical manufacturing	General manufacturing, facility management

Online is not automatically “better” than portable. For a critical turbine, compressor or strategic pump, continuous monitoring may well be justified. For hundreds of less critical motors, periodic route-based monitoring may be the more rational choice.

The best programmes often combine both models: online for critical assets and portable for a large population of lower-criticality equipment.

Condition monitoring sensors and equipment

Condition monitoring sensors are selected according to the parameter, the failure mode, the operating conditions and the expected response time.

Sensor or device type	Signal measured	Typical application
Accelerometer	Vibration acceleration	Bearings, gearboxes, motors, pumps
Velocity sensor	Vibration velocity	General condition assessment of rotating machines
Proximity probe	Shaft displacement	Turbines, compressors, large rotating machines
Thermocouple	Contact temperature	Bearings, windings, housings, process
IR camera	Surface temperature	Electrical panels, insulation, hot spots
Oil particle counter	Particle count and size	Hydraulic systems, gearboxes, turbines
Oil moisture sensor	Water content	Lubrication and hydraulic systems
Ultrasonic detector	Sound at 20–100 kHz	Leaks, bearings, electrical discharge
Current transformer	Motor current	Motor current analysis
Acoustic emission sensor	Elastic waves	Vessels, pipelines, structures

Sensor selection should take into account the measuring range, environmental robustness, mounting method, sampling rate, power supply, communication, calibration and the ability to integrate with a data platform.

Benefits of condition monitoring

The main benefits of condition monitoring are earlier detection of degradation, fewer unplanned downtime events, better work planning, longer asset life, improved safety and data-driven maintenance decisions.

Benefit	Mechanism
Less unplanned downtime	An early degradation signal makes it possible to plan the response
Longer asset life	Detecting abnormal operating conditions limits further wear
Lower maintenance costs	Fewer emergency repairs and fewer replacements carried out unnecessarily
Improved safety	Earlier detection of overheating, leaks or structural degradation
Better maintenance decisions	Data supports work prioritisation and shutdown planning
Fewer secondary failures	Intervening before a functional failure limits the domino effect
Better use of teams	Technicians work according to risk and asset condition, not the calendar alone
Support for Availability in OEE	Fewer unplanned stops can improve equipment availability

Condition monitoring can shorten MTTD, that is mean time to detect, and by identifying the problem sooner it supports a reduction of MTTR, that is mean time to repair. Over a longer horizon it can also improve MTBF, that is mean time between failures, because degradation is caught earlier and does not lead to repeated functional failures.

In manufacturing, condition monitoring mainly affects equipment availability, that is the Availability component of OEE. It does not, however, replace a full OEE analysis or broader production KPIs.

Limitations of condition monitoring

Condition monitoring does not detect every failure mode and should not be deployed in the same way across all assets.

The main limitations:

• the initial cost of sensors, cabling, gateways and software;
• the complexity of installation on existing machines;
• the need for technically competent people to interpret the data;
• the risk of false alarms when thresholds are set incorrectly;
• limited effectiveness for very rapidly developing failures;
• limited value for cheap, redundant or non-critical assets;
• the need to maintain sensors, calibration and data quality;
• difficulties with data context if signals are not linked to the asset, operating mode and event history.

Condition monitoring delivers the greatest value when it is connected to a decision-making process. An alert on its own does not improve machine availability if the organisation does not know who should respond, how to classify the risk and when to schedule the action.

Integrating condition monitoring with a data platform, analytics and CMMS

Condition monitoring generates technical data that should be connected with asset context, failure history, work orders, the production plan and maintenance decisions.

In practice this means integration with several layers:

• an industrial data platform;
• a CMMS;
• a historian;
• a SCADA system;
• an MES;
• dashboards;
• analytical tools;
• the maintenance workflow.

A data platform helps to collect signals from different sources and give them a shared context. Production and maintenance analytics use condition monitoring data for trending, comparisons, anomaly detection and reporting. A CMMS can receive alerts or recommendations and turn them into work orders.

Condition monitoring also supports decision intelligence, because asset condition data can be one of the inputs to a decision: repair now, keep watching, schedule a shutdown, order a part or change the maintenance strategy.

In the context of Smart RDM, what matters is the ability to combine condition monitoring data with other industrial data: production, quality, energy, events, the work plan and the history of actions. In this way condition monitoring does not remain a separate technical screen, but becomes part of a broader operational process.

Standards related to condition monitoring

Condition monitoring should rely on consistent terminology, measurement principles and diagnostic procedures. The ISO standards for machine condition monitoring and diagnostics are particularly relevant here.

Standard	Scope	Relevance for condition monitoring
ISO 13373	Condition monitoring and diagnostics of machines using vibration	Applies to vibration condition monitoring
ISO 13379	Data interpretation and diagnostic and prognostic techniques	Supports the move from monitoring to diagnostics and prognostics
ISO 17359	General guidelines for condition monitoring and diagnostics of machines	Helps structure the rollout of a CM programme
ISO 10816 / ISO 20816	Evaluation of machine vibration by measurements on non-rotating parts	Supports severity assessment for vibration monitoring

Standards do not replace engineering judgement, but they help to harmonise the approach to measurement, signal interpretation and the organisation of a condition monitoring programme.

Condition monitoring examples

Condition monitoring examples show that each technique addresses a different type of problem and a different degradation mechanism.

Pump bearing monitored with vibration analysis

The pump runs continuously and its failure would stop a critical process step. An accelerometer measures vibration, and spectrum analysis together with envelope analysis reveals a developing bearing defect.

The team can schedule the replacement during a service window instead of waiting for a failure and unplanned downtime.

Electrical panel monitored with thermography

An IR camera inspection detects a hot spot on an electrical connection. The temperature at that point is higher than on comparable connections under a similar load.

The action is to schedule an inspection and correct the connection before overheating, insulation damage or a power failure occurs.

Gearbox monitored with oil analysis

Periodic oil analysis shows an increase in the number of metal particles and a change in viscosity. The result points to possible wear of internal components or lubricant contamination.

Maintenance can combine the result with the gearbox operating history and decide whether an oil change, filtration, borescope inspection or a planned overhaul is required.

Compressed air monitored with ultrasonic testing

A technician walks a route with an ultrasonic detector and identifies a leak in the compressed air system. The leak is neither visible nor audible to the operator under normal production conditions.

Fixing the leak reduces energy losses and improves the stability of the installation.

Motor monitored with motor current analysis

Analysis of the motor current spectrum indicates symptoms of rotor bar damage or an abnormal load. No sensor needs to be installed on the motor housing, because the analysis uses the electrical signal.

The technique makes it possible to detect drive and load problems without direct intervention in the machine.

Pipeline monitored with acoustic emission

Acoustic emission sensors detect signals indicating crack growth, corrosion or a leak in the structure. The technique is especially relevant where structural integrity is critical and the asset is not a classic rotating machine.

Cost of a condition monitoring system

The cost of a condition monitoring system depends on the number of assets, the techniques chosen, the online or portable model, the type of sensors, the scope of installation, software, integration and the analytics required.

The main cost drivers:

• the number of measurement points;
• the type of sensors;
• the need for cabling;
• DAQ modules and gateways;
• mechanical mounting;
• environmental requirements;
• software licences;
• integration with a CMMS, SCADA, MES or data platform;
• configuration of alarms and dashboards;
• user training;
• system maintenance and calibration;
• data analysis by specialists.

For critical assets, the higher cost of online monitoring can be justified by the risk of downtime. For a large population of less critical equipment, a portable or route-based model may deliver a lower cost per asset.

The cost of the system should be assessed against the cost of failure, the safety impact, spare part availability, repair time and production losses.

Vendor landscape 2026

The market for condition monitoring software and equipment includes sensor manufacturers, vibration monitoring systems, portable tools, asset management platforms, IIoT solutions and industrial data platforms.

Vendor	Area	Deployment model / context	Target segment
ifm	Industrial sensors, IO-Link, vibration and temperature monitoring	Sensors, gateways, IIoT solutions	Manufacturing and industrial automation
Baker Hughes / Bently Nevada	Machinery protection and vibration monitoring	Monitoring systems for rotating machinery	Power generation, oil & gas, heavy industry
IBM	Enterprise asset management and analytics	Maximo, CM analytics, AI-driven insights	Large organisations and enterprise environments
Tractian	Wireless vibration and temperature sensors	AI-driven platform for maintenance teams	Production plants and maintenance teams
Fluke	Portable test and measurement instruments	Handheld vibration testers, thermal cameras, ultrasonic detectors	Technicians, maintenance teams, route-based inspections
Siemens	Monitoring integrated with an automation ecosystem	SIMATIC, MindSphere and industrial solutions	Plants using Siemens automation
SKF	Bearing monitoring and rotating equipment	CM systems for bearings, vibration, temperature and oil	Rotating machinery and maintenance teams
Emerson	Asset management and condition monitoring for processes	AMS, CSI and solutions for process plants	Process industry and continuous plants
Smart RDM	CM data integration, dashboards, workflow and operational context	Industrial data platform and operational monitoring	Industrial organisations combining data from many sources

The table is not a ranking. The choice of solution depends on the asset types, criticality, available data, the online or portable model, cybersecurity requirements, team competence and the need to integrate with existing systems.

How to start a condition monitoring rollout?

A condition monitoring rollout is best started with critical assets and known failure modes, not with a mass installation of sensors.

A practical sequence:

1. Identify the critical assets.
2. Define the consequences of failure: safety, production, quality, environment, cost.
3. Identify the most important failure modes.
4. Match the measurement parameter to the failure mode.
5. Choose the technique: vibration, thermography, oil, ultrasonic, MCA or acoustic emission.
6. Decide whether monitoring should be online or portable.
7. Select the sensors and the data acquisition method.
8. Set thresholds, trends and alarm rules.
9. Integrate alerts with the CMMS or the maintenance process.
10. Measure the effectiveness of the programme and update the rules.

The most important implementation question is: which decision is this measurement meant to support? If the organisation does not know what it will do with an alert, the sensor alone will not solve the problem.

FAQ

What is a condition monitoring system?

A condition monitoring system is a hardware and software solution that monitors the technical condition of machines and equipment using sensor data such as vibration, temperature, pressure, oil condition, current, voltage or acoustic emission.

The system collects data, analyses trends and deviations, and generates alerts or information for the maintenance team in order to detect degradation before a functional failure occurs.

What is condition monitoring?

Condition monitoring is the process of monitoring asset condition on the basis of technical parameters that may indicate wear, degradation or abnormal operating conditions.

A condition monitoring system is the tool or architecture that makes this process possible, using sensors, data acquisition, software and analytics.

What are the four main types of monitoring?

The four main types of condition monitoring are:

1. Vibration analysis.
2. Thermography.
3. Oil analysis.
4. Ultrasonic testing.

More advanced programmes also use motor current analysis and acoustic emission monitoring.

What are the different condition monitoring techniques?

The most important condition monitoring techniques are vibration analysis, thermography, oil analysis, ultrasonic testing, motor current analysis and acoustic emission monitoring.

Vibration analysis is used mainly for rotating machinery. Thermography works well for thermal and electrical problems. Oil analysis assesses lubricant condition and internal wear. Ultrasonic testing detects leaks, electrical discharge and early bearing defects. Motor current analysis examines the electrical signals of motors. Acoustic emission monitoring supports the assessment of structural integrity.

What are the 5 elements of condition monitoring?

The five elements of condition monitoring are:

1. Parameter selection.
2. Sensor or instrument selection.
3. Data collection.
4. Data analysis.
5. The maintenance decision.

These elements form a practical implementation framework: from defining what to measure, through selecting the technology, to the maintenance response.

How does condition monitoring differ from predictive maintenance?

Condition monitoring collects and analyses data about the current condition of equipment. Predictive maintenance uses that data, operating history and analytical models to predict future failure risk.

Put simply: condition monitoring provides the data, and predictive maintenance uses it to forecast and plan interventions.

How does condition monitoring differ from condition-based maintenance?

Condition monitoring is the layer that collects and analyses asset condition data.

Condition-based maintenance is a maintenance strategy in which actions are carried out when condition data indicates the need to intervene. CM provides the data; CBM uses it to make decisions.

How does condition monitoring reduce downtime?

Condition monitoring reduces downtime through earlier detection of degradation. Instead of reacting only to a functional failure, the maintenance team can schedule an inspection, repair or replacement in a convenient production window.

The mechanism is simple: an earlier signal → a better diagnosis → a planned intervention → lower risk of unplanned downtime.

Which sensors are used in condition monitoring?

Typical condition monitoring sensors are accelerometers, velocity sensors, proximity probes, thermocouples, IR cameras, oil sensors, particle counters, ultrasonic detectors, current transformers and acoustic emission sensors.

Sensor selection depends on the asset, the failure mode, the operating environment and the required measurement frequency.

What does online condition monitoring mean?

Online condition monitoring means continuous or near-continuous monitoring of an asset through permanently installed sensors.

It is used above all for assets that are critical, expensive, hard to access, or whose failure causes major production, safety or environmental losses.

What does portable condition monitoring mean?

Portable condition monitoring means periodic measurements taken with a handheld instrument during a planned measurement route.

It is useful for a large population of equipment that is accessible to a technician and does not require permanent online monitoring.

How much does a condition monitoring system cost?

The cost of a condition monitoring system depends on the number of measurement points, the type of sensors, installation, the online or portable model, software licences, integration, analytics, training and system maintenance.

Cost should not be judged by the sensor price alone. What matters is the total cost of implementation compared with the cost of failure, downtime, lost production or safety risk.

Which industries use condition monitoring?

Condition monitoring is used in manufacturing, power generation, oil & gas, mining, process industries, pulp & paper, facility management, transport and infrastructure.

It is most often deployed where asset failure causes high costs, safety risk, loss of production or a difficult process restart.

Sources

1. ISO 13373 – vibration condition monitoring.
2. ISO 13379 – diagnostics and prognostics in condition monitoring.
3. ISO 17359 – general guidelines for condition monitoring and diagnostics of machines.
4. ISO 10816 / ISO 20816 – evaluation of machine vibration.
5. Technical materials from manufacturers of sensors, vibration monitoring systems, portable tools and asset management platforms.
6. Internal Smart RDM materials on industrial data, asset monitoring and maintenance.