What Does Shaft Misalignment Look Like in Machine Data?

Shaft misalignment commonly appears as changes around 1X and 2X running speed, often combined with elevated axial or radial vibration. However, a single frequency peak cannot confirm misalignment. Reliable diagnosis requires evaluating frequency, direction, location, operating state and how the signal changes relative to the machine’s historical baseline.

A maintenance team may therefore see:

2X vibration increasing

without immediately knowing whether the cause is:

• Shaft misalignment
• Mechanical looseness
• Structural behaviour
• Coupling-related forces
• Another developing mechanical condition

The objective is not simply to detect a peak.

It is to understand the pattern around that peak.

What Is Shaft Misalignment?

Shaft misalignment occurs when the rotational centre lines of two coupled shafts do not correctly align.

There are three common configurations.

Angular Misalignment

The rotational centre lines of the two shafts intersect at an angle.

Angular misalignment can create increased vibration in the axial direction, particularly around running-speed frequencies.

Parallel or Offset Misalignment

The shafts remain parallel, but their centre lines are displaced.

Parallel misalignment can produce stronger vibration in the radial or tangential direction, often with activity around 2X running speed.

Combined Misalignment

Real equipment may experience both angular and offset errors simultaneously.

The resulting vibration spectrum can therefore contain:

1X + 2X + additional harmonics

across several measurement directions.

What Does 1X Mean in Vibration Analysis?

1X represents the rotational frequency of a shaft.

For example, consider a motor running at:

1,500 RPM

The shaft rotates:

1,500 ÷ 60 = 25 times per second

Therefore:

1X = 25 Hz

2X = 50 Hz

3X = 75 Hz

A 1X component is normal in rotating machinery.

Its presence alone does not indicate a fault.

The diagnostic value comes from evaluating:

• Amplitude
• Direction
• Rate of change
• Relationship with other harmonics
• Measurement location
• Historical machine behaviour

Does 2X Vibration Mean Shaft Misalignment?

Not necessarily.

An elevated 2X running-speed component is commonly associated with shaft misalignment, especially offset misalignment.

But 2X is a diagnostic clue, not proof.

Maintenance teams should also ask:

• Is 1X increasing?
• Is the vibration axial or radial?
• Are harmonics developing?
• Is the same behaviour visible on both sides of the coupling?
• Has machine speed changed?
• Has machine load changed?
• Is this pattern new?
• How does it compare with the historical baseline?

This broader context makes fault identification more reliable.

What Does Angular Misalignment Look Like in Vibration Data?

Angular misalignment can create elevated vibration in the axial direction.

A simplified signature could appear as:

1X ↑

2X ↑

Axial vibration ↑

Additional harmonics may also be visible depending on severity and machine configuration.

However, there is no single universal signature that applies to every machine.

What Does Parallel Misalignment Look Like?

Parallel or offset misalignment can create stronger vibration in the radial or tangential direction.

A simplified pattern might be:

1X present

2X ↑

Radial vibration ↑

Additional harmonics possible

Again, this pattern should be considered together with machine history and operating conditions.

Angular vs Parallel Misalignment: Quick Comparison

Condition	Common Signal Clues	Direction Often Examined
Angular misalignment	1X and 2X activity	Axial
Parallel misalignment	Elevated 2X	Radial / tangential
Combined misalignment	1X + 2X + harmonics	Axial + radial
Imbalance	Dominant 1X	Radial
Mechanical looseness	Multiple harmonics	Depends on location

These are diagnostic patterns rather than absolute fault rules.

Why Can Overall Vibration Miss Developing Misalignment?

Overall vibration compresses a complex machine signal into one value.

That is useful for general monitoring.

But it can hide frequency-specific changes.

Consider the following illustrative example.

Monitoring Stage	1X	2X	Axial Vibration	Overall Vibration
Baseline	1.00	1.00	1.00	1.00
Stage 1	1.02	1.18	1.08	1.02
Stage 2	1.07	1.48	1.31	1.06
Stage 3	1.16	1.92	1.63	1.13
Stage 4	1.29	2.48	2.02	1.28

Illustrative normalized data only. These are not customer measurements.

At Stage 2:

Overall vibration = 1.06× baseline

while:

2X = 1.48× baseline

A traditional overall alarm might still consider the machine normal.

But frequency analysis tells a different story.

A specific component of the vibration signal is changing much faster than the overall value.

Suggested Figure 1

Title: Illustrative Development of a Misalignment-Related Vibration Pattern

Plot:

• 1X vibration
• 2X vibration
• Axial vibration
• Overall vibration

Caption:
Illustrative normalized example showing how 2X and axial vibration may begin increasing faster than overall vibration as a potential misalignment-related pattern develops.

Why Should Measurements Be Compared Across the Coupling?

Misalignment exists between two connected shafts.

For example:

Motor

↓

Coupling

↓

Pump

If the alignment condition changes, vibration behaviour may appear at measurement locations on both sides of the coupling.

Instead of asking:

“Is Motor Sensor A high?”

maintenance teams can examine:

This pattern provides significantly more diagnostic context.

Why Does the Machine’s Baseline Matter?

Two nominally identical machines may not produce identical vibration signatures.

Their behaviour can differ because of:

• Foundation stiffness
• Installation conditions
• Coupling condition
• Structural resonance
• Machine loading
• Process conditions
• Temperature
• Historical wear

That means a single universal vibration threshold cannot always tell the full story.

A machine-specific baseline changes the question from:

“Is this vibration value generally high?”

to:

“Is this vibration abnormal for this particular machine?”

That becomes especially important when faults develop gradually.

Operating State Matters Too

Machine vibration should also be interpreted according to whether the equipment is:

• Off
• Starting
• Idling
• Operating normally
• Operating under heavy load
• Changing speed

A vibration level that is normal during active operation could appear highly abnormal compared with an off-state measurement.

This is why operating-state separation is an important part of machine-specific condition monitoring.

A Groundup.ai deployment with Etihad Rail provides a practical example.

Real-World Example: Etihad Rail Sandfighter

Groundup.ai’s Cognitive Maintenance solution monitored critical drivetrain components on Etihad Rail’s Kershaw Sand Remover / Sandfighter.

Monitoring covered components including:

• Differential gearboxes
• Transmission
• Other drivetrain measurement locations

Importantly, this was not a shaft misalignment case.

It demonstrates a broader principle:

How machine-specific vibration behaviour can be used to distinguish normal operation from developing abnormal behaviour.

Understanding Normal Operation First

Before attempting to identify developing faults, Groundup.ai first established how the monitored asset behaved under different operating states.

The actual vibration data below shows a clear difference between when the machine is operating and when it is switched off.

Actual Etihad Rail Differential Gearbox Data

Etihad Rail Sandfighter — differential gearbox vibration signature during machine operation versus off-state.

The purple-highlighted section shows increased fundamental-frequency activity while the Sand Remover is operating.

The red-highlighted section shows the vibration signature when the Sand Remover is switched off.

This distinction provides essential context.

Without operating-state information, the system could compare:

machine operating

against:

machine idle

and incorrectly interpret normal operational vibration as an anomaly.

Instead, the monitoring process establishes:

The question becomes:

“Is the vibration high?”

versus the more useful:

“Is this vibration abnormal for this machine while it is operating?”

Why Is Operating-State Classification Important in Predictive Maintenance?

Operating-state classification helps prevent normal changes in machine behaviour from being mistaken for mechanical faults.

For example, vibration may naturally increase when:

• Production load increases
• A gearbox becomes engaged
• A pump begins moving fluid
• Machine speed changes
• The drivetrain starts transmitting torque

Without context, these changes can look anomalous.

With an established operating baseline, the analysis can distinguish:

Expected operational variation

from:

Unexpected mechanical deviation

This reduces noise in condition-monitoring data and gives anomaly detection a more meaningful reference.

When the Etihad Rail Pattern Started Changing

Once the Sandfighter’s normal behaviour had been established, abnormal deviations became easier to identify.

Groundup.ai observed an aggressive upward trend in vibration behaviour across monitored differential gearbox data.

The analysis then moved deeper into the signal.

Rather than stopping at:

“Vibration increased.”

the system analysed the underlying frequency behaviour and identified a repetitive impact pattern.

This provided additional information about what was changing inside the drivetrain.

The workflow became:

The pattern was assessed as being consistent with a lubrication-related mechanical condition.

Why Might the First Manual Inspection Find Nothing?

Early-stage machine degradation may change vibration behaviour before producing obvious physical symptoms.

A machine may still:

• Continue operating
• Produce no unusual audible noise
• Show no visible mechanical damage
• Remain below conventional vibration alarms

while the internal mechanical condition has already begun changing.

This is one of the central advantages of continuous condition monitoring.

Machine behaviour can provide an earlier indication than visual inspection alone.

Further investigation of the Etihad Rail asset later identified supporting evidence within the lubrication condition, including contamination and wear-related indicators.

From Signal to Maintenance Evidence

The Etihad Rail example can be simplified into six stages.

1. Establish the Baseline

Understand normal machine operation.

↓

2. Detect the Change

Identify vibration behaviour that deviates from the baseline.

↓

3. Analyse the Signal

Determine which frequency components and measurement directions changed.

↓

4. Interpret the Pattern

Compare the behaviour against possible mechanical conditions.

↓

5. Guide the Inspection

Direct maintenance teams toward the relevant components or systems.

↓

6. Validate

Use physical inspection, lubrication analysis or other maintenance evidence to confirm the condition.

This represents a much broader process than simply generating an alarm.

What Does the Etihad Rail Case Have to Do With Misalignment?

The specific fault mechanism was different.

The diagnostic principle is the same.

For misalignment:

2X alone is not enough.

For the Etihad Rail drivetrain:

one frequency component alone is not enough.

The useful information comes from combining:

This is the difference between detecting an anomaly and providing meaningful diagnostic context.

Anomaly Detection vs Fault Diagnosis

Anomaly detection answers:

“Is something different?”

Fault diagnosis tries to answer:

“What could be causing the difference?”

Maintenance decision support goes one step further:

“What should be inspected next?”

A useful condition-monitoring workflow therefore looks like:

Can AI Detect Shaft Misalignment?

AI-based condition monitoring can help identify behaviour consistent with shaft misalignment by analysing multiple vibration features and comparing them with machine-specific historical behaviour.

Rather than relying only on:

2X > threshold

the system can evaluate:

• 1X amplitude
• 2X amplitude
• Running-speed harmonics
• Axial vibration
• Radial vibration
• Rate of change
• Sensor location
• Machine operating state
• Historical baseline
• Related machine signals

For example:

2X ↑

alone provides limited context.

But:

• 2X ↑

• Axial vibration ↑

• Related behaviour on both sides of coupling

• Deviation from historical baseline

provides a much stronger pattern for maintenance investigation.

What Should Maintenance Teams Check When Misalignment Is Suspected?

If machine data indicates a potential misalignment pattern, maintenance teams can work through the following checks.

1. Compare Both Sides of the Coupling

Review measurements from the drive end of both connected machines.

2. Separate Axial and Radial Vibration

Directional behaviour provides diagnostic context.

3. Examine 1X and 2X

Review current amplitude and historical trends.

4. Check Other Harmonics

Additional running-speed harmonics may help differentiate between fault conditions.

5. Review Operating Conditions

Determine whether speed, load or process state changed.

6. Compare Against the Machine Baseline

Identify whether the pattern is truly unusual for that asset.

7. Inspect the Coupling

Check for wear, damage or abnormal mechanical behaviour.

8. Check Soft Foot

Machine-foot conditions can influence alignment.

9. Consider Thermal Growth and Pipe Strain

Alignment may change as the machine reaches operating conditions.

10. Verify Shaft Alignment Directly

Use appropriate alignment measurement techniques to confirm the mechanical condition.

Vibration monitoring identifies where to investigate.

Physical alignment measurement confirms whether the shafts are actually misaligned.

Frequently Asked Questions

What frequency indicates shaft misalignment?

Misalignment is commonly associated with vibration around 1X and 2X running speed, sometimes with additional harmonics.

The exact signature varies according to machine configuration, fault type and severity.

Does 2X vibration always mean misalignment?

No.

A strong 2X component can be associated with misalignment, but other mechanical conditions can also create running-speed harmonics.

Direction, historical trend, machine location and operating conditions should also be evaluated.

Does angular misalignment cause axial vibration?

Angular misalignment commonly produces elevated vibration in the axial direction, particularly around running-speed frequencies.

However, actual machine signatures may be more complex.

What is parallel misalignment?

Parallel misalignment occurs when two shafts remain parallel but their rotational centre lines are offset from one another.

It can be associated with increased radial vibration and elevated 2X running-speed behaviour.

Why is a vibration baseline important?

A baseline provides a reference for how a specific machine normally behaves.

This allows condition monitoring to identify changes relative to the machine itself rather than relying only on generic vibration limits.

Why does machine operating state matter?

Machine vibration changes naturally when equipment transitions between off, idle, startup and operating conditions.

Separating these states helps prevent expected operational vibration from being mistaken for a mechanical anomaly.

Can vibration monitoring detect a fault before maintenance sees physical damage?

Potentially, yes.

Mechanical changes can alter vibration patterns before they create obvious visible or audible symptoms.

This is why continuous condition monitoring can provide an earlier indication that a machine requires investigation.

Can AI confirm shaft misalignment?

AI can identify vibration patterns that are consistent with misalignment and help direct maintenance teams toward further investigation.

Physical shaft-alignment measurements should still be used to confirm the actual alignment condition.

The Takeaway

Misalignment does not have one universal vibration signature.

Sometimes the strongest clue is:

1X

Sometimes:

2X

Sometimes:

Axial vibration

Sometimes several harmonics begin changing together.

And sometimes the first meaningful indication is simply:

This machine has started behaving differently from its own normal operating pattern.

The Etihad Rail example demonstrates why that distinction matters.

The chart does not simply show:

high vibration

It shows the importance of understanding:

when the machine is operating

versus:

when the machine is off

That operating context creates the baseline needed to identify meaningful deviations later.

Effective condition monitoring therefore needs to move beyond:

“Is the vibration high?”

and toward:

• What changed?

• Where did it change?

• Which part of the signal changed?

• Under what operating condition?

• How does it compare with the machine’s baseline?

• What mechanical condition could explain the pattern?

• What should maintenance inspect next?

Because predictive maintenance is not simply about collecting more machine data.

It is about turning that data into maintenance decisions.