Industrial IoT and Business Operations · Study deck

Predictive Maintenance: Signals and Vibration

This first route connects maintenance strategy and business evidence to a signal pipeline, then develops vibration features into reviewable fault evidence.

Ada is your guide for this deck.

iiotpredictivemaintenance

Ada, the module guide, in a scene from this chapter.

After studying this chapter

Learning objectives

You will be able to:

Explain: The gateway may compute FFT features locally, forward time-series data through OPC UA or MQTT, and store results in a historian, AVEVA PI System, InfluxDB, TimescaleDB, or cloud data lake.
Explain: The edge stage turns that raw window into a few named features and records how they were produced, so a later reviewer can tell a real change from a settings change.
Explain: The path to watch in Figure: Predictive maintenance pipeline with example data volumes is the red one along the bottom, which shows what is lost when a stage is skipped.

Infrared thermal-imaging camera with blue housing, mounted vertically on a stand, connected to cables and electronic devices.

Major section

Start With the Story

Heat, sound, current, and vibration may give early clues.

A change during heavy work may be normal; the same change at light work may matter.
The simple story has a limit.
A pattern can show that something changed; it may not prove why.
Safe action still needs physical checks and maintenance judgment.

Major section

PdM Turns Signals into Work

Predictive maintenance is valuable when a signal changes a maintenance action before the asset fails.

The goal is not to collect every possible waveform.
The goal is to detect a developing fault early enough to order parts, schedule labor, protect safety, and avoid unplanned downtime.
A useful deployment starts with one named asset and one named fault.

A predictive-maintenance path preserves the link from vibration sensing through edge analysis and prediction to a maintenance alert.

Major section

PdM Turns Signals into Work (continued)

The pipeline succeeds only when a technician can confirm or reject the alert and close the loop.

The strongest candidates are assets with expensive failure modes and measurable degradation patterns: motors, pumps, compressors, fans, gearboxes, spindles, conveyors, chillers, turbines, and critical bearings.
The signal may come from vibration, temperature, acoustic emission, motor current, pressure, oil debris, lubricant analysis, cycle counts, or control-system operating context.
The business case should also include what will not be predicted.

White dome-shaped thermal camera with a dark lens, attached beneath a metal mounting arm on a helicopter.

Major section

PdM Turns Signals into Work (continued)

The last step is not another chart: a maintenance alert must identify the asset, evidence, urgency, and action owner.

If the signal loses its units, sample conditions, or asset context along the way, the prediction cannot be checked.
For example, a drive-end bearing on a cooling pump might use a triaxial accelerometer, motor current, and discharge pressure.
The first proof is whether the system can distinguish normal load changes from a developing bearing defect, then raise an alert early enough for a planned inspection.

Uniformed personnel examine a handheld yellow thermal camera inside a ship during a demonstration.

Major section

Design the Signal Chain First

Process equipment may use pressure differential, flow, valve position, energy use, startup time, or compressor discharge temperature.

A practical PdM pilot starts with failure modes and effects analysis, asset history, and maintenance records.
A poorly mounted accelerometer can produce clean-looking data that is useless for diagnosis.
Useful features depend on the asset.

Major section

PdM Needs Physics and Workflow

Condition data needs a traceable path from sensor to action.

An IEPE accelerometer, MEMS sensor, current transformer, oil particle counter, or temperature probe may connect to an edge DAQ, PLC, or gateway.
Sampling choices change what faults can be seen.
Model operations are part of the design.

Major section

PdM Needs Physics and Workflow (continued)

The gateway may compute FFT features locally, forward time-series data through OPC UA or MQTT, and store results in a historian, AVEVA PI System, InfluxDB, TimescaleDB, or cloud data lake.

Nyquist limits, anti-alias filtering, window length, spectral resolution, tachometer references, sensor orientation, mounting stiffness, and unit conversion all affect the signal.
Downsampling a 10 kHz vibration waveform to a 1 Hz dashboard trend can erase the bearing information that maintenance needed.
Alerts need severity, confidence, asset id, feature values, baseline comparison, recent maintenance state, and recommended inspection.
A PdM system that cannot learn from technician outcomes becomes an expensive alarm list.

Vibration analysis workflow from sensing to defect detection

Major section

PdM Needs Physics and Workflow (continued)

Teams must handle false positives, false negatives, seasonal operation, new product mixes, replaced parts, firmware changes, sensor drift, and concept drift.

The data model should keep raw windows, derived features, and maintenance events connected.
A feature such as vibration RMS or envelope energy should carry the source sensor, asset id, timestamp, window size, filter settings, firmware version, and operating state.
The work-order outcome should then link back to the same feature window.

Major section

Sammy Listens to Machines

The repair took 3 days because nobody knew it was about to break.

Temperature Terry has a new job at a candy factory!
Sammy's Solution: Be a Machine Doctor!
Monday: Sammy notices the mixer is shaking a tiny bit more than usual.
Monday: The mixer is back to making chocolate perfectly!

Major section

Key Concepts

Asset Criticality: Ranking equipment by production impact, safety consequence, repair cost, spare-part lead time, and whether failure stops a constrained process.

Remaining Useful Life (RUL): An estimate of time or cycles until a failure threshold is likely, valid only for the modeled fault mode and operating context.
The dollar values are illustrative inputs, not universal benchmarks.
CMMS/EAM Integration: Routing alerts into maintenance systems so inspection, parts, labor, and technician disposition are captured.

Illustrative comparison of reactive, preventive, and predictive maintenance strategies

Major section

Key Concepts (continued)

This scenario timeline contrasts how the same equipment can behave under three maintenance regimes.

The reactive example combines a $50,000 repair with $150,000 of lost production, so the unplanned event costs $200,000 before other effects.
Preventive maintenance buys predictable inspections and an eight-hour stop, but may replace healthy parts.
The point is timing, not totals.

Major section

Key Concepts (continued)

Figure: Timeline comparing three maintenance strategies across 12 months puts the same machine on three tracks, so you can see when each strategy spends.

Predictive work adds sensing, analysis, and review so a four-hour repair can be planned from condition evidence.
Its lower downtime is conditional: alerts must be trusted, actionable, and linked to maintenance capacity rather than treated as universal savings.
Each regime moves the decision earlier, and the figures are illustrative inputs.

Timeline comparing three maintenance strategies across 12 months: reactive work carries outage risk, preventive work can replace healthy parts, and predictive work uses condition evidence to schedule inspection and repair.

Major section

Predictive Maintenance Pipeline

The next question is how that evidence travels from a physical asset to a maintenance decision without losing context.

The edge gateway derives RMS, crest factor, and FFT bands while the source conditions are still known.
An anomaly alert alone is not maintenance; value appears when evidence becomes owned work and improves the next assessment.
This diagram uses concrete example values to show where data volume changes.
The path to watch in the Predictive maintenance pipeline with example data volumes: sensing captures raw condition signals, edge processing extracts features, analytics estimate severity or RUL, and maintenance systems create reviewable actions diagram is the red one along the bottom, which shows what is lost when a stage is skipped.

Predictive maintenance data pipeline from condition sensing to maintenance action

Major section

Predictive Maintenance Pipeline (continued)

Carry: Predictive Maintenance Data Pipeline into predictive maintenance pipeline; use 1.

The path to watch in Figure: Predictive maintenance pipeline with example data volumes is the red one along the bottom, which shows what is lost when a stage is skipped.
The edge stage turns that raw window into a few named features and records how they were produced, so a later reviewer can tell a real change from a settings change.
Sampling straight at one reading per second removes the frequencies the diagnosis depends on, and no later stage can bring them back.

Predictive maintenance pipeline with example data volumes: sensing captures raw condition signals, edge processing extracts features, analytics estimate severity or RUL, and maintenance systems create reviewable actions.

Deck summary

Key takeaways

Heat, sound, current, and vibration may give early clues.

Predictive maintenance is valuable when a signal changes a maintenance action before the asset fails.
The pipeline succeeds only when a technician can confirm or reject the alert and close the loop.
The last step is not another chart: a maintenance alert must identify the asset, evidence, urgency, and action owner.
Process equipment may use pressure differential, flow, valve position, energy use, startup time, or compressor discharge temperature.

Retrieval practice

Recall check 1 of 3

Ada says: answer from memory, then check your reasoning.

Q1A motor runs normally but its vibration trend changes. Which maintenance response fits the evidence?

AWait for a breakdown before looking at it

BReplace every motor on the same calendar date

CIgnore the trend because the motor is still running

DThe trend warrants inspection and planned motor work.

Show answer

Answer: D Condition evidence can justify a timely check before failure.

Retrieval practice

Recall check 2 of 3

Ada says: answer from memory, then check your reasoning.

Q2A pump motor has a stable dominant 2x shaft-frequency vibration in axial and radial directions, without prominent bearing-defect frequencies or multiple looseness harmonics. A mechanical check finds mounting bolts tight. Which defect is most consistent with this evidence and should be checked next?

ABearing wear - inner-race fault without bearing frequencies

BRotor imbalance - the rotor has uneven mass distribution

CLooseness - there are loose bolts or mounting issues

DShaft misalignment - motor and pump shafts are not aligned

Show answer

Answer: D A strong, stable 2x component with axial and radial motion is consistent with shaft misalignment.

Retrieval practice

Recall check 3 of 3

Ada says: answer from memory, then check your reasoning.

Q3A team has time-stamped bearing feature windows and wants to train and check a remaining-useful-life model. Which data link is needed?

AUse only the latest unlabelled dashboard score, with no asset or repair date link.

BTreat every anomaly as a confirmed failure date, because the model score establishes the repair outcome.

CDrop asset IDs and operating context, then compare generic failure dates with feature windows.

DLink asset feature windows to fault dates; compare predicted life with observed outcomes.

Show answer

Answer: D Confirmed event dates and asset context provide labels to train and check remaining-useful-life estimates.

Print reference

Answers

Answer key.

D · Condition evidence can justify a timely check before failure.
D · A strong, stable 2x component with axial and radial motion is consistent with shaft misalignment.
D · Confirmed event dates and asset context provide labels to train and check remaining-useful-life estimates.