Industrial equipment rarely fails without warning.
Before a bearing collapses, a motor overheats, a pump loses hydraulic performance, a gearbox develops serious damage, or a transformer experiences a major fault, the equipment will often produce measurable changes in its physical or operational behavior.
Vibration may begin to increase.
Temperature may gradually rise.
Electrical current may become abnormal.
Lubricant contamination may increase.
Pressure or flow may deviate from normal operating patterns.
Noise characteristics may change.
Efficiency may deteriorate.
These changes are the foundation of condition monitoring.
Condition monitoring allows industrial organizations to observe the health and performance of equipment while it is operating and to identify developing problems before they result in unexpected failure.
Instead of relying only on scheduled maintenance or waiting for equipment to fail, engineers can monitor measurable indicators of equipment health and use those indicators to make better maintenance decisions.
This changes the fundamental maintenance question from:
“When is this equipment scheduled for maintenance?”
to:
“What is the actual condition of this equipment, and when does it need intervention?”
For asset-intensive industries, this difference can have significant consequences.
A critical pump failure can interrupt production.
A transformer failure can affect an entire facility.
A bearing failure can damage a motor shaft.
A cooling fan failure can cause secondary equipment problems.
A compressor malfunction can shut down a production unit.
Condition monitoring helps organizations detect deterioration earlier, understand failure mechanisms, prioritize maintenance resources, and reduce the risk of unexpected downtime.
It is therefore not simply a sensor technology.
Condition monitoring is a structured engineering process connecting:
Assets → Failure Modes → Measurable Parameters → Sensors → Data → Analysis → Diagnosis → Maintenance Decision
Understanding this chain is essential for building an effective condition monitoring program.
What Is Condition Monitoring?
Condition monitoring is the process of observing measurable parameters that indicate the operating condition or health of machinery and equipment.
The objective is to detect changes that may indicate deterioration, abnormal operation, or developing faults.
ISO 17359:2018 provides general guidelines for establishing machine condition monitoring programs and applies broadly to machines. Its framework recognizes parameters including vibration, temperature, tribology, flow, contamination, power, and speed as possible indicators of machine performance and condition.
Examples of condition indicators include:
- Vibration
- Bearing temperature
- Motor winding temperature
- Lubricant condition
- Electrical current
- Voltage
- Power
- Speed
- Pressure
- Flow
- Acoustic signals
- Ultrasonic emissions
- Partial discharge
- Insulation parameters
- Process performance
- Contamination levels
The most appropriate parameter depends on the equipment and the failure mode being monitored.
For example:
A developing bearing defect may initially appear in vibration frequency characteristics.
A lubrication problem may produce temperature changes, vibration changes, and oil contamination.
A pump cavitation problem may appear through vibration, pressure fluctuations, noise, and reduced hydraulic efficiency.
Motor overloading may be visible through electrical current and temperature.
Transformer deterioration may require oil analysis, temperature monitoring, dissolved gas analysis, partial discharge monitoring, and electrical testing.
Therefore, condition monitoring is not simply about collecting one measurement.
It is about selecting the right measurement for the failure mechanism that matters.
Why Does Condition Monitoring Matter?
Modern industrial plants depend heavily on equipment availability.
A manufacturing line cannot produce when a critical motor fails.
A chemical process may need to shut down when a process pump becomes unavailable.
A power plant may lose generating capacity because of rotating equipment problems.
A water utility may lose pumping capability.
A mining operation may suffer production interruptions.
In these environments, equipment reliability directly affects operational and financial performance.
Condition monitoring matters because it provides information about developing equipment problems before failure occurs.
This gives organizations something extremely valuable:
time to respond.
Consider two scenarios.
Scenario 1: Failure Without Condition Monitoring
A bearing begins degrading.
No one detects it.
The bearing continues operating.
Vibration increases.
Temperature increases.
The bearing eventually fails.
The shaft becomes damaged.
The motor trips.
Production stops.
Maintenance now faces an urgent repair.
Spare parts may not be immediately available.
The result can include:
- Emergency maintenance
- Overtime
- Secondary equipment damage
- Production loss
- Safety exposure
- Expedited spare-part costs
Scenario 2: Failure With Effective Condition Monitoring
The same bearing begins degrading.
Vibration monitoring detects a developing abnormal pattern.
Maintenance receives an alert.
Engineers review the trend and confirm deterioration.
The bearing replacement is scheduled for the next planned production shutdown.
Spare parts are prepared.
Labor is planned.
The bearing is replaced before catastrophic failure.
The physical failure mechanism may be identical in both scenarios.
The difference is when the organization becomes aware of it.
That is the central value of condition monitoring.
Condition Monitoring Is Part of Asset Management
Condition monitoring should not operate as an isolated technical activity.
It should support the broader management of physical assets.
ISO 17359 explicitly positions condition monitoring as an important component within asset management and references the ISO 55000 family of standards.
This relationship is important.
Organizations do not monitor equipment simply because monitoring technology is available.
They monitor equipment because they are trying to manage risks such as:
- Production loss
- Safety consequences
- Environmental consequences
- Equipment damage
- Maintenance cost
- Quality problems
- Energy inefficiency
- Regulatory exposure
Therefore, monitoring decisions should be linked to asset criticality and business consequences.
A small non-critical exhaust fan may not justify sophisticated online vibration monitoring.
A critical process compressor shutting down an entire production unit may justify multiple redundant monitoring technologies.
The level of monitoring should reflect the importance of the asset and the consequences of failure.
How Does Condition Monitoring Work?
A complete condition monitoring process typically follows several stages:
- Identify the asset.
- Understand its functions.
- Identify important failure modes.
- Identify measurable indicators of those failures.
- Select appropriate monitoring technologies.
- Establish normal operating conditions.
- Collect data.
- Detect deviations.
- Analyze the cause.
- Decide what maintenance action is required.
A simplified workflow looks like this:
Equipment
↓
Physical Condition
↓
Sensor / Measurement
↓
Signal Processing
↓
Trend / Diagnostic Analysis
↓
Alarm or Condition Assessment
↓
Maintenance Decision
The technology may be sophisticated, but the logic is fundamentally engineering-based.
Start with Failure Modes, Not Sensors
One of the most important principles in condition monitoring is to avoid starting with sensor selection.
A common mistake is:
“We have vibration sensors, so which machines should we install them on?”
The better question is:
“Which failures matter, and what physical indicators would reveal those failures?”
For example, consider a centrifugal pump.
Important failure modes may include:
- Bearing deterioration
- Misalignment
- Mechanical looseness
- Cavitation
- Impeller damage
- Seal failure
- Coupling problems
- Motor electrical problems
- Blocked suction
- Operation away from best efficiency point
Now consider the measurable indicators.
| Failure Mode | Possible Condition Indicators |
|---|---|
| Bearing damage | Vibration, temperature, ultrasound |
| Misalignment | Vibration, axial vibration, temperature |
| Cavitation | Vibration, pressure, acoustic signal |
| Motor overload | Current, temperature, power |
| Hydraulic degradation | Flow, pressure, power |
| Seal degradation | Leakage, pressure, temperature |
| Lubrication issue | Oil condition, temperature, vibration |
This approach connects monitoring directly with reliability engineering.
Types of Condition Monitoring
Condition monitoring includes multiple technologies.
No single technology can identify every failure mode.
The best programs frequently combine several techniques.
1. Vibration Monitoring
Vibration analysis is one of the most widely used technologies for rotating equipment.
Machines naturally generate vibration during operation.
Changes in amplitude or frequency content may indicate changes in machine condition.
Applications include:
- Electric motors
- Pumps
- Fans
- Blowers
- Gearboxes
- Compressors
- Turbines
- Generators
Vibration can reveal problems including:
- Imbalance
- Misalignment
- Bearing defects
- Looseness
- Gear defects
- Resonance
- Structural problems
- Cavitation
ISO 20816-1:2016 establishes general procedures for measuring and evaluating machine vibration and includes criteria related both to vibration magnitude and changes in vibration. The standard specifically links these criteria to reliable, safe, long-term machine operation and operational monitoring.
However, an important engineering principle should be understood:
A vibration alarm value is not automatically a diagnosis.
An increased overall vibration level indicates that something may have changed.
Determining the cause may require:
- Frequency spectrum analysis
- Time waveform analysis
- Envelope analysis
- Phase measurement
- Operating condition information
- Historical comparison
Condition detection and fault diagnosis are related but different activities.
2. Temperature Monitoring
Temperature is one of the simplest and most valuable condition indicators.
Unexpected temperature increases may indicate:
- Excessive friction
- Poor lubrication
- Electrical overload
- Cooling failure
- Mechanical binding
- Process problems
- Bearing deterioration
Temperature monitoring can be performed using:
- RTDs
- Thermocouples
- Thermistors
- Infrared sensors
- Thermal cameras
Typical applications include:
- Bearings
- Motor windings
- Transformer windings
- Gearboxes
- Electrical panels
- Busbars
- Process equipment
Temperature data becomes much more useful when combined with operating conditions.
For example, a motor bearing temperature of 70°C may or may not be abnormal depending on:
- Ambient temperature
- Motor load
- Historical baseline
- Bearing type
- Lubrication
- Cooling condition
Therefore, trends are often more valuable than isolated measurements.
3. Oil and Lubricant Analysis
Lubricating oil contains valuable information about equipment condition.
Oil analysis can examine:
- Viscosity
- Water contamination
- Particle contamination
- Wear metals
- Oxidation
- Additive depletion
Wear debris can provide clues about internal mechanical deterioration.
Applications include:
- Gearboxes
- Turbines
- Compressors
- Large engines
- Hydraulic systems
- Transformers
Oil analysis illustrates an important characteristic of condition monitoring:
Some monitoring is continuous online, while other monitoring is periodic.
A monthly oil sample can still be part of a condition-based maintenance program.
Condition monitoring does not necessarily mean every parameter must be measured every second.
4. Ultrasound and Acoustic Monitoring
Equipment faults can generate ultrasonic or acoustic signals.
Applications include:
- Bearing defects
- Compressed-air leaks
- Steam-trap inspection
- Electrical discharge
- Valve leakage
Ultrasound can be particularly useful for detecting some early-stage faults before they become obvious through other indicators.
5. Electrical Condition Monitoring
Electrical equipment provides a wide range of measurable condition indicators.
For motors these may include:
- Current
- Voltage
- Power
- Power factor
- Current imbalance
- Harmonic content
- Motor current signature
Electrical measurements may help detect:
- Overload
- Voltage imbalance
- Abnormal load
- Rotor problems
- Process changes
- Electrical supply problems
Combining electrical measurements with mechanical condition data can provide a more complete picture.
For example:
Increasing vibration + increasing current + reduced process flow
may provide more diagnostic information than vibration alone.
6. Process Performance Monitoring
Process parameters can also provide evidence of machine health.
For a pump:
- Suction pressure
- Discharge pressure
- Flow
- Power consumption
can be combined to evaluate performance.
For a heat exchanger:
- Inlet temperature
- Outlet temperature
- Pressure drop
can indicate fouling or degradation.
For a compressor:
- Pressure
- Temperature
- Flow
- Power
can indicate changes in efficiency.
This is an important concept:
Condition monitoring does not have to be limited to mechanical sensors.
Existing process data can often be used as condition information.
7. Thermography
Infrared thermography allows equipment surface temperatures to be examined without physical contact.
Common applications include:
- Electrical connections
- Switchgear
- Busbars
- Bearings
- Motors
- Furnaces
- Steam systems
- Building systems
Thermography is particularly effective for detecting abnormal heating patterns.
However, interpretation must consider:
- Emissivity
- Reflections
- Load
- Distance
- Environmental conditions
As with all condition monitoring techniques, measurement quality matters.
Offline, Periodic, and Online Condition Monitoring
Condition monitoring can be implemented at different levels of automation.
Periodic Manual Monitoring
A technician visits equipment and collects data periodically.
Examples:
- Monthly vibration route
- Quarterly thermography
- Monthly oil sampling
Advantages include:
- Lower initial investment
- Flexible inspection
- Expert can observe equipment directly
Limitations include:
- Faults may develop between inspection intervals.
- Large sites require significant labor.
- Remote assets may be difficult to inspect.
Online Condition Monitoring
Sensors remain installed continuously.
Data is collected automatically.
Advantages include:
- Continuous visibility
- Faster detection
- Automatic trends
- Remote monitoring
- Automated alarms
This is particularly useful for:
- Critical equipment
- Remote equipment
- Rapidly developing failures
- Difficult-to-access assets
- 24/7 production
Hybrid Monitoring
Many organizations benefit from combining online and periodic techniques.
For example:
Online
- Vibration RMS
- Temperature
- Motor current
Periodic specialist analysis
- Detailed vibration spectrum
- Oil analysis
- Thermography
This can provide a strong balance between cost and diagnostic capability.
Condition Monitoring vs Preventive Maintenance
Condition monitoring is frequently confused with preventive maintenance.
They are related, but they are not the same.
Preventive Maintenance
Maintenance is generally performed according to:
- Calendar time
- Operating hours
- Number of cycles
For example:
“Replace this bearing every 12 months.”
The maintenance occurs whether the bearing is degraded or not.
Condition-Based Maintenance
Maintenance is triggered by evidence of deterioration.
For example:
“Replace this bearing because vibration trend and spectral analysis indicate progressive bearing damage.”
The difference is important.
Preventive maintenance asks:
How long has the equipment operated?
Condition-based maintenance asks:
What is the actual condition of the equipment?
Neither strategy is universally better.
Certain tasks should remain time-based.
Others can be condition-based.
The appropriate strategy depends on failure behavior and consequences.
Condition Monitoring vs Predictive Maintenance
These terms are also frequently used interchangeably.
A useful distinction is:
Condition monitoring determines or tracks current asset health.
Predictive maintenance attempts to use current and historical condition information to predict future deterioration or estimate when intervention will be required.
For example:
Condition monitoring:
“Vibration is increasing and has entered an abnormal range.”
Predictive analysis:
“Based on the current degradation trajectory, the bearing may reach the intervention threshold within approximately three weeks.”
Prediction requires more information and generally more mature analytics.
Condition monitoring is therefore a foundation for predictive maintenance.
Condition Monitoring vs Predictive Analytics
There can be several levels of maturity.
Level 1 — Data Collection
Equipment measurements are recorded.
Level 2 — Visualization
Measurements can be viewed in dashboards.
Level 3 — Threshold Monitoring
The system detects values outside specified limits.
Level 4 — Condition Assessment
Multiple indicators are interpreted to determine asset health.
Level 5 — Diagnosis
The probable failure mechanism is identified.
Level 6 — Prognosis
The future progression of the failure is estimated.
Level 7 — Prescriptive Maintenance
The system recommends a specific response or optimization.
Many organizations should not immediately jump to machine learning.
Reliable sensing and good data engineering usually need to come first.
The P-F Curve and Condition Monitoring
One of the most useful concepts for understanding condition monitoring is the P-F interval.
Imagine equipment condition deteriorating over time.
At point P, a potential failure becomes detectable.
At point F, functional failure occurs.
The interval between these points provides an opportunity for maintenance intervention.
Different monitoring technologies may detect deterioration at different stages.
For example, a bearing problem might progress approximately like:
Microscopic defect
↓
Ultrasonic signature
↓
High-frequency vibration
↓
Overall vibration increase
↓
Temperature increase
↓
Noise
↓
Functional failure
The exact sequence varies by failure mode, but the general concept is important.
The earlier a meaningful defect can be detected, the more time maintenance has to respond.
However, detecting earlier is only useful if the signal is reliable enough to support decisions.
False alarms can be extremely damaging to user confidence.
Why Trending Is Often More Valuable Than a Single Alarm Limit
Suppose a motor normally operates at:
1.6 mm/s RMS
Measurements over several months show:
1.6
1.7
1.8
2.0
2.3
2.7
3.2
The absolute vibration value may still be below a generic alarm threshold.
But the trend clearly indicates deterioration.
This illustrates why condition monitoring should examine:
- Absolute values
- Rate of change
- Baseline deviation
- Operating conditions
- Historical trends
ISO 20816’s approach also considers both vibration magnitude and changes in vibration when evaluating machine condition.
This is why dashboards should not only show a large current-value gauge.
Historical trend information is essential.
Establishing a Baseline
Good condition monitoring requires an understanding of normal behavior.
Equipment may naturally operate differently under:
- Low load
- Full load
- Different speeds
- Different process conditions
- Different temperatures
- Different seasons
Therefore, the baseline should ideally account for operating conditions.
For example:
A pump vibration reading of 4 mm/s at full load and 4 mm/s at very low flow may represent completely different mechanical conditions.
Context matters.
Important contextual parameters may include:
- Equipment state
- Load
- Speed
- Flow
- Pressure
- Ambient temperature
- Production rate
This is where integration between process monitoring and condition monitoring becomes powerful.
Data Architecture for Modern Condition Monitoring
Traditional condition monitoring often involved handheld instruments.
Modern IIoT systems allow condition data to flow continuously through a digital architecture.
A typical system may look like:
Machine
↓
Vibration / Temperature / Electrical Sensor
↓
Industrial Edge Gateway
↓
Ethernet / RS485 / Wireless / 4G
↓
Database
↓
Condition Monitoring Platform
↓
Dashboard / Alarm / Analytics
↓
Maintenance Team
The edge gateway may:
- Collect sensor data
- Convert industrial protocols
- Perform local calculations
- Buffer data
- Communicate with the server
The monitoring platform may then:
- Store historical measurements
- Display trends
- Calculate health indicators
- Generate alerts
- Compare assets
- Support advanced analytics
ISO 13374 provides a standards framework relating to condition-monitoring data processing, communication, and presentation. ISO 13374-1 remains current following confirmation in 2025, while ISO 13374-3 addresses communication and interoperability within condition-monitoring and diagnostic architectures.
The Importance of Sampling Rate
Not all condition variables require the same sampling rate.
Temperature might only need to be collected every:
- 10 seconds
- 30 seconds
- One minute
High-resolution vibration analysis may require measurements at thousands of samples per second.
This creates a major architectural difference.
Sending every high-frequency vibration sample continuously to the cloud can generate enormous amounts of data.
A better design may perform signal processing at the edge.
For example:
Raw vibration waveform
↓
Edge processing
↓
Calculate:
- RMS
- Peak
- Crest factor
- Frequency features
↓
Send selected condition indicators upstream
Detailed waveform data can be stored or transmitted when deeper analysis is required.
This edge/cloud separation improves scalability.
Alarm Strategy in Condition Monitoring
A condition monitoring system should never be designed as an alarm generator alone.
Poor alarm design creates alarm fatigue.
If maintenance teams repeatedly receive alerts that require no action, they eventually stop trusting the system.
Good monitoring may include several alarm levels.
For example:
Normal
Asset behavior remains within expected conditions.
Advisory
A meaningful deviation is detected.
Continue monitoring.
Warning
Deterioration requires engineering review.
Alarm
Maintenance intervention should be planned.
Critical
Immediate action may be required.
Alarm decisions may consider:
- Absolute threshold
- Relative change
- Rate of change
- Multiple parameters
- Operating mode
- Asset criticality
The correct alarm philosophy should balance:
early detection
against:
false alarm avoidance.
Diagnostics: From “Something Is Wrong” to “What Is Wrong?”
Monitoring indicates whether asset behavior has changed.
Diagnosis attempts to identify why.
For example:
Vibration is high.
Possible reasons include:
- Imbalance
- Misalignment
- Looseness
- Bearing failure
- Resonance
- Cavitation
Overall vibration alone may not differentiate them.
Additional analysis may include:
- Spectrum
- Phase
- Time waveform
- Envelope
- Speed
- Process conditions
The diagnostic process can be represented as:
Detection
Something is abnormal.
↓
Isolation
Which component is involved?
↓
Diagnosis
What failure mode is occurring?
↓
Severity Assessment
How serious is it?
↓
Maintenance Decision
What action should be taken?
This is where technical expertise remains critical.
Role of AI and Machine Learning
Artificial intelligence is increasingly being applied to condition-monitoring data.
NIST’s 2025 systematic review notes the growing potential of condition-monitoring technologies supported by AI and IoT, while also emphasizing the need to evaluate engineering and financial benefits when determining whether investment is justified.
Potential applications include:
- Anomaly detection
- Pattern recognition
- Asset comparison
- Failure classification
- Remaining useful life estimation
- Trend forecasting
One particularly useful approach is anomaly detection.
Instead of defining one static temperature threshold, a model may learn that:
Motor temperature normally depends on:
- Load
- Ambient temperature
- Speed
The model can then identify conditions where measured temperature differs significantly from what is expected.
However, AI does not eliminate engineering fundamentals.
Bad sensors create bad data.
Poor asset context produces misleading analytics.
Missing operating-state information produces false anomalies.
Therefore:
Sensor quality → Data quality → Analytics quality
remains a fundamental relationship.
Which Assets Should Be Monitored?
Not every asset needs continuous monitoring.
Selection should begin with criticality.
Consider:
Failure Consequence
Would failure:
- Stop production?
- Create safety risk?
- Create environmental risk?
- Damage other equipment?
Failure Frequency
Does the equipment fail frequently?
Detectability
Is there a measurable indicator before failure?
Maintenance Cost
Is failure expensive to repair?
Accessibility
Is the equipment difficult to inspect manually?
Redundancy
Is a backup unit available?
These considerations help prioritize investment.
A useful starting point is often the equipment responsible for the largest combination of:
Likelihood of Failure × Consequence of Failure
Condition Monitoring for Electric Motors
Electric motors are excellent candidates because they combine mechanical and electrical failure modes.
Potential measurements include:
- Drive-end vibration
- Non-drive-end vibration
- Bearing temperature
- Winding temperature
- Current
- Voltage
- Power
- Speed
Potential problems include:
- Bearing failure
- Misalignment
- Imbalance
- Overload
- Cooling problems
- Electrical supply imbalance
- Rotor problems
Combining electrical and mechanical condition information produces stronger diagnostics than using either alone.
Condition Monitoring for Pumps
Pump monitoring can combine:
- Vibration
- Bearing temperature
- Suction pressure
- Discharge pressure
- Flow
- Motor power
- Speed
This can help detect both mechanical and hydraulic problems.
Consider the following data:
Vibration increases.
Flow decreases.
Motor power changes.
Discharge pressure fluctuates.
This combination may suggest a process or hydraulic problem rather than simply a bearing fault.
Condition monitoring should therefore consider the machine as a system.
Condition Monitoring for Gearboxes
Gearboxes frequently require:
- Vibration
- Oil analysis
- Temperature
- Speed
- Load
Potential issues include:
- Gear tooth damage
- Bearing damage
- Lubrication problems
- Misalignment
- Excessive wear
Frequency-domain vibration analysis is particularly useful because gear mesh frequencies and bearing frequencies can provide diagnostic information.
Condition Monitoring for Transformers and Electrical Assets
Condition monitoring is not limited to rotating equipment.
Transformers can be monitored using:
- Oil temperature
- Winding temperature
- Load
- Moisture
- Dissolved gases
- Partial discharge
- Insulation indicators
Electrical panels may be monitored through:
- Thermal measurements
- Current
- Voltage
- Harmonics
- Partial discharge
ISO 17359’s current framework also includes power transformers among the equipment considered in its supporting annexes.
Business Benefits of Condition Monitoring
The technical objective is to understand equipment health.
The business objective is to make better decisions.
Potential benefits include:
1. Reduced Unplanned Downtime
Earlier warning allows repairs to be scheduled before functional failure.
2. Reduced Secondary Damage
A bearing can sometimes be replaced before failure damages the shaft, housing, rotor, or coupling.
3. Better Maintenance Planning
Maintenance can prepare:
- Spare parts
- Labor
- Tools
- Contractor support
- Shutdown schedule
4. Lower Inspection Costs
Remote monitoring can reduce the need for routine field inspections, particularly for distributed assets.
5. Improved Spare-Part Planning
Condition information provides better indication of when parts may actually be needed.
6. Increased Asset Availability
Reduced emergency failure can improve production availability.
7. Better Root Cause Analysis
Historical data provides evidence of what happened before an event.
Condition Monitoring Does Not Eliminate Maintenance
Condition monitoring does not mean maintenance disappears.
Its purpose is to improve maintenance decisions.
Equipment will still require:
- Lubrication
- Replacement
- Alignment
- Cleaning
- Inspection
- Overhaul
The difference is that maintenance can increasingly be performed because evidence indicates it is needed.
That can improve resource allocation.
Common Condition Monitoring Mistakes
Mistake 1 — Installing Sensors Without Failure-Mode Analysis
Sensor selection should follow an understanding of how equipment fails.
Mistake 2 — Monitoring Non-Critical Assets First
Choose assets where the monitoring information can create meaningful value.
Mistake 3 — Using Generic Alarm Limits Only
Generic limits can be useful, but asset-specific trends and baselines are also important.
Mistake 4 — Ignoring Operating Conditions
Changes in speed, load, temperature, pressure, or production rate may affect measurements.
Mistake 5 — Collecting Data Without a Response Process
If an alarm occurs, someone must know:
- What it means
- Who reviews it
- What action should follow
Mistake 6 — Expecting AI to Solve Poor Measurement
Artificial intelligence cannot compensate for unreliable sensors and missing context.
Mistake 7 — Monitoring Everything at Maximum Resolution
Data collection should match the physics of the failure mode.
How to Implement a Condition Monitoring Program
A practical implementation can follow the general principles reflected in ISO 17359.
Step 1 — Define Objectives
Examples:
- Reduce motor failures
- Improve pump reliability
- Reduce inspection labor
- Prevent gearbox damage
Step 2 — Perform Asset Criticality Assessment
Prioritize equipment based on consequences of failure.
Step 3 — Identify Failure Modes
Use:
- FMEA
- RCM
- Historical maintenance records
- OEM knowledge
- Reliability data
Step 4 — Map Failure Modes to Condition Indicators
Determine which physical variables change when degradation occurs.
Step 5 — Select Monitoring Techniques
Choose:
- Vibration
- Temperature
- Current
- Oil analysis
- Ultrasound
- Process data
- Other techniques
Step 6 — Establish Baselines
Capture healthy operating conditions.
Step 7 — Define Alarm Logic
Use absolute and relative limits where appropriate.
Step 8 — Build Data Infrastructure
Define:
- Sensors
- Gateways
- Networks
- Database
- Dashboard
- Notifications
Step 9 — Define Response Procedures
Every meaningful alarm should have an owner and expected response.
Step 10 — Measure Results
Track:
- Failure reduction
- Avoided downtime
- Maintenance cost
- Detection accuracy
- False alarms
- User response
- ROI
How IIoT Is Changing Condition Monitoring
Industrial IoT technology is reducing many traditional barriers to online monitoring.
Historically, permanently installed monitoring systems could be justified mainly for very critical or expensive equipment.
Lower-cost sensors, industrial gateways, wireless communication, cellular connectivity, cloud platforms, and time-series databases now allow monitoring to be extended to a larger population of assets.
This creates opportunities for what could be called:
scalable condition monitoring.
For example:
One gateway may collect information from several sensors.
Multiple gateways may communicate with one central platform.
One dashboard may monitor several sites.
The architecture can become:
Assets
↓
Sensors
↓
Edge Gateways
↓
Central Data Platform
↓
Condition Analytics
↓
Maintenance Team
This is one of the areas where Industrial IoT and traditional reliability engineering increasingly converge.
Condition Monitoring Maturity Model
Organizations can implement condition monitoring progressively.
Stage 1 — Manual Inspection
Operators inspect equipment.
Stage 2 — Periodic Measurement
Specialists collect vibration, thermal, or oil data periodically.
Stage 3 — Online Monitoring
Selected measurements are continuously transmitted.
Stage 4 — Centralized Monitoring
Multiple assets are visible from one platform.
Stage 5 — Automated Anomaly Detection
Analytics identifies abnormal behavior.
Stage 6 — Predictive Analytics
The platform estimates deterioration.
Stage 7 — Integrated Maintenance Decisions
Condition information links with maintenance planning and work management.
The correct target depends on the organization.
Not every facility requires Stage 7 immediately.
Standards Relevant to Condition Monitoring
Several international standards provide useful frameworks.
ISO 17359:2018
Condition monitoring and diagnostics of machines — General guidelines
This is the principal general guideline for establishing condition-monitoring programs across machines. ISO confirmed the current edition during its 2023 review.
ISO 20816 Series
Mechanical vibration — Measurement and evaluation of machine vibration
ISO 20816-1 establishes general requirements for measuring and evaluating machine vibration, including operational monitoring and guidance for operating limits. As of September 2026, the 2016 edition remains published, while its replacement is at Final Draft International Standard stage.
ISO 13374 Series
Condition monitoring and diagnostics of machines — Data processing, communication and presentation
This series provides a useful framework for the digital processing and communication architecture behind modern condition-monitoring systems. Part 1 remains current following its 2025 confirmation.
These standards illustrate that modern condition monitoring involves more than the measurement itself.
It includes:
measurement → processing → communication → evaluation → information presentation
From Condition Monitoring to Reliability Intelligence
The real value of condition monitoring develops progressively.
At first, an organization may only know:
Motor vibration = 2.5 mm/s
Later it learns:
Motor vibration has increased 40% in the last three months.
Then:
The increase is concentrated in a bearing-related frequency range.
Then:
The deterioration is accelerating.
Finally:
Maintenance should replace the bearing during the planned shutdown next week.
The information becomes progressively more actionable.
This progression can be represented as:
Data
↓
Information
↓
Condition
↓
Diagnosis
↓
Prediction
↓
Maintenance Decision
That final step is what creates business value.
Conclusion
Condition monitoring is the systematic observation of equipment health using measurable physical, electrical, process, or performance parameters.
Its purpose is not simply to generate more data.
Its purpose is to provide earlier and better information about developing equipment problems.
A successful condition monitoring system connects:
Failure Mode
↓
Condition Indicator
↓
Sensor
↓
Data Acquisition
↓
Trend and Analysis
↓
Diagnosis
↓
Maintenance Action
This helps industrial organizations move from reactive maintenance toward increasingly condition-based and predictive strategies.
The benefits can include:
- Earlier detection of equipment deterioration
- Reduced unplanned downtime
- Better maintenance planning
- Reduced secondary damage
- Improved asset availability
- Better root cause analysis
- More efficient use of maintenance resources
However, technology alone does not create these benefits.
Effective programs require:
- Asset criticality
- Failure-mode analysis
- Correct sensor selection
- Reliable data acquisition
- Appropriate sampling
- Good baseline data
- Well-engineered alarms
- Diagnostic expertise
- Clear maintenance response procedures
The most important principle is simple:
Do not monitor equipment merely because it can be monitored. Monitor the conditions that influence decisions.
For some assets, periodic inspection may remain sufficient.
For others, continuous online monitoring can provide substantial value.
The correct strategy depends on equipment criticality, failure behavior, detection opportunities, and the consequences of failure.
As Industrial IoT technologies continue to reduce the cost and complexity of connecting machines, condition monitoring can increasingly be applied not only to the most critical equipment but also to broader asset populations.
This creates an opportunity for industrial organizations to transform maintenance from a schedule-driven activity into a more data-driven process.
The ultimate objective is not simply to know that equipment is vibrating, heating, consuming power, or producing data.
It is to answer a much more valuable question:
Is this asset healthy, is its condition changing, and what should we do about it?
Condition Monitoring with Siteplore MachineGuard
Siteplore MachineGuard is designed around this principle: transforming equipment measurements into useful information about machine condition.
A typical MachineGuard architecture can connect:
Motors / Pumps / Fans / Gearboxes / Other Assets
↓
Vibration, Temperature, Electrical and Process Measurements
↓
Edge Gateway
↓
Secure Data Infrastructure
↓
MachineGuard Monitoring Dashboard
↓
Trend Analysis + Alerts + Analytics
↓
Maintenance Decision
Potential applications include:
- Motor condition monitoring
- Pump monitoring
- Fan and blower monitoring
- Gearbox monitoring
- Bearing condition monitoring
- Vibration trending
- Temperature monitoring
- Electrical condition monitoring
- Remote rotating-equipment monitoring
- Multi-site asset monitoring
The appropriate monitoring architecture should be determined from the equipment’s failure modes and operating criticality rather than from a predetermined sensor package.
For example, a critical pump may require vibration, temperature, motor electrical measurements, and process performance monitoring.
A smaller motor may only require vibration and temperature.
A remote asset may require an industrial edge gateway with cellular connectivity and local data buffering.
The goal is therefore not to maximize the number of sensors.
The goal is to create sufficient visibility to support better maintenance decisions.
Start with the Assets That Matter Most
An effective condition monitoring program does not need to begin with hundreds of machines.
A practical starting point is to identify:
- Critical rotating equipment
- Repeat-failure assets
- Assets responsible for significant downtime
- Equipment that is difficult to inspect
- Machines with detectable degradation mechanisms
Then determine:
What can fail?
How can we detect it?
How early can we detect it?
What action would we take if deterioration is identified?
Those questions provide a much stronger foundation for condition monitoring than simply asking which sensor to purchase.
Siteplore can then connect the appropriate measurement technologies, edge infrastructure, dashboards, alerts, and analytics into a scalable condition-monitoring architecture.
Measure the condition. Understand the change. Act before failure.
