Every running machine is constantly broadcasting information about its internal health — through vibration, heat, sound, and fluid chemistry. A technician who can read those signals can intercept a failure days or weeks before it destroys equipment, halts production, or creates a safety hazard. This lesson introduces the four primary Predictive Maintenance (PdM) technology families and teaches you how to match the right instrument to the right failure mode.
By the end of this lesson, you will be able to:
- Objective 1 (Cognitive): Differentiate between Preventive Maintenance (calendar/run-time based) and Predictive Maintenance (condition-based).
- Objective 2 (Diagnostic): Match specific machinery failure modes to the correct predictive technology — Vibration, Thermography, Ultrasound, or Fluid Analysis.
- Objective 3 (Practical): Complete a basic field inspection using a handheld thermal imaging camera to detect high-resistance electrical or mechanical friction faults.
Prerequisites: TECH-2.1 (Machinery Assets Overview), TECH-2.5 (PM Routing Execution). Related: TECH-2.4 (Lubrication Chemistry), TECH-2.10, TECH-11.2 (Root Cause Analysis).
A tech running a standard visual PM route would check "OK" and move on.
However, you pull out an infrared thermal camera and look at the inboard motor bearing housing. The screen displays a bright, glowing white-hot node reading 185°F (85°C), while the identical outboard bearing reads a cool 115°F (46°C). You caught a major failure hours before the metal expanded enough to cause a physical lockup. You diagnosed an invisible failure mode — without touching the machine, without stopping production, without any physical evidence visible to the naked eye.
While Preventive Maintenance (PM) relies on arbitrary time intervals or machine cycle counts to guess when an asset needs servicing, Predictive Maintenance (PdM) — also known as Condition-Based Maintenance (CBM) — uses non-destructive testing (NDT) instruments to measure the actual physical condition of the machine while it is running under normal production loads.
Instead of replacing a perfectly healthy bearing every 12 months just because "it's on the schedule," a PdM model tracks the asset's active health signatures, triggering a repair order only when internal variables cross a defined structural warning boundary.
PdM asks: "What is the machine telling us right now?" — and acts on the data.
This shift is significant. A calendar-based approach will service a healthy bearing unnecessarily (wasting time and introducing re-assembly risk) while sometimes missing a bearing that fails faster than expected. A condition-based approach only acts when the data demands it — and it catches failures that no calendar schedule can predict.
Technicians must master the application areas of the four primary predictive technology families. Each intercepts failure signals in a different part of the physics spectrum:
① Vibration Analysis
Measures rhythmic oscillating movement of running machinery components. Excellent for detecting mechanical unbalance, shaft misalignment, loose mounting bolts, and internal ball-bearing raceway defects. Data is collected via accelerometers or handheld vibration pens mounted to the machine casing.
② Infrared Thermography
Captures electromagnetic radiation emissions to map surface temperature distributions. Excellent for locating high-resistance electrical connections, friction in dry bearings, and insulation degradation inside ovens or steam pipes. Uses a thermal imaging camera as a completely non-contact measurement.
③ Airborne & Structure-Borne Ultrasound
Listens to high-frequency acoustic waves (above 20 kHz) outside the range of human hearing. Excellent for isolating microscopic compressed air or vacuum line leaks, early-stage bearing friction events, and electrical arcing inside high-voltage panels. A heterodyne meter translates the signal into audible tones.
④ Oil Analysis / Tribology
Fluid samples are sent to a specialized laboratory to track chemical degradation, additive depletion, and precise concentrations of microscopic metallic wear particles floating in the fluid. Provides long-horizon trending data on internal wear rates that no surface sensor can capture.
Machines always emit physical energy signatures as they operate. As a component degrades, its emission patterns change predictably long before a functional breakdown occurs. PdM technologies intercept these changing signals across different physics spectrums:
The degradation cascade follows a predictable four-stage sequence — and each stage is detectable by a different technology:
The Chemical Trace (Oil Analysis)
Microscopic metal particles flake off a bearing race and enter the lubricant. A fluid sample sent to a lab reveals abnormal particle counts and composition — often weeks before any other technology can detect the fault.
The Acoustic Ping (Ultrasound)
As the metal surface worsens, rough raceway contact generates high-frequency micro-friction anomalies — captured as elevated dBμV readings by an ultrasound detector. Still invisible to the human ear.
The Wave Spike (Vibration Analysis)
The structural damage expands, causing the running shaft to physically displace on its axis — registering as measurable velocity or acceleration spikes on a vibration sensor. The bearing fault frequency appears in the spectrum.
The Thermal Blast (Thermography)
In the final stages before complete destruction, intense raceway friction drives localized component temperatures up rapidly — visible on a thermal camera as a bright, white-hot node. At this stage, failure is imminent.
Matching the right instrument to the right failure mode is the core skill of a PdM technician. Applying the wrong tool wastes time and gives a false "no fault found" result:
| Machine Fault Condition | Primary PdM Technology | Secondary Validation Tool |
|---|---|---|
| Loose electrical screw terminal | Infrared Thermography (thermal camera) | Ultrasound (checks for micro-arcing noise) |
| Misaligned coupling shafts | Vibration Analysis (phase / spectrum) | Infrared Thermography (hot coupling joint) |
| Pinhole pneumatic line leak | Airborne Ultrasound Meter | Traditional soap-bubble spray test |
| Gear tooth profile degradation | Lubricant Analytical Ferrography | Vibration Analysis (gearmesh frequency) |
| Early-stage rolling element bearing fault | Ultrasound (earliest detection) | Vibration Analysis (BPFI/BPFO frequencies) |
| Motor winding insulation degradation | Infrared Thermography | Motor Circuit Analysis (MCA) |
An industrial machine displaying healthy, normal predictive parameter baselines registers the following general benchmarks (verify asset-specific limits against OEM documentation and site standards):
- Vibration Boundary: Overall velocity measurements settle safely below 0.15 in/s RMS on general utility motor assemblies operating under normal loads.
- Thermal Boundary: Electrical terminal connections under load display zero distinct "hot-spot" thermal gradients relative to adjacent matching phases. Bearing housings run within 20°F of ambient air temperature on properly lubricated assemblies.
- Acoustic Boundary: Ultrasound dBμV readouts track consistently within a flat-line historical average for that specific asset and measurement point — no trending upward over consecutive readings.
- The Emissivity Error (Thermography): Attempting to read the temperature of a bright, mirror-shiny stainless steel or copper busbar without adjusting the camera's emissivity setting. Polished metals reflect background heat like a mirror, giving a falsely low reading on screen. Always set emissivity to the correct value for the target material before logging any temperature data.
- Sensor Mounting Laxity (Vibration): Holding a vibration sensor loosely against a motor casing by hand instead of mounting it via a rigid magnetic base or screw-in stud. A loose contact dampens high-frequency waves, completely erasing vital bearing defect indicators from the spectrum.
- Missing the Baseline: Collecting data once and assuming you understand the asset's health profile. A bearing reading 140°F might be normal if it has run at 140°F for two years — and alarming if it read 90°F last week. PdM relies entirely on tracking trends over time. A single data point is not a trend.
The Reality: A thermal camera or vibration pen is a tool that extends your senses — it cannot fix a machine. If a technician captures an advanced bearing vibration anomaly but does not possess the structural alignment skills or torque verification practices to install the replacement component correctly, the new asset will fail just as quickly as the old one.
PdM instruments are diagnostic tools. The repair still requires hands, tools, mechanical understanding, and precision. The camera tells you what is wrong. You still have to fix it right.
🔒 Checkpoint — Confirm Before Continuing
Check all boxes to unlock the remaining sections.
When assigned an infrared thermography survey route across active machine control panels:
- Don your appropriate baseline NFPA 70E Arc Flash face shield and voltage-rated safety apparel before approaching any energized enclosure.
- Power on your infrared thermal imaging camera and verify the battery cell charge registers above 30%.
- Open the outer door latching mechanisms of the target electrical cabinet carefully to expose internal components under active running factory load.
- Hold the camera steady and scan the panel layout from top to bottom — inspecting main line fuses, circuit breakers, terminal bus block strips, and motor starter contactor pads.
- Identify any wiring or connections displaying a bright, high-contrast white or yellow thermal signature compared to surrounding conductors.
- Use the camera's crosshair cursor tool to capture the exact maximum temperature value of any thermal anomaly.
- If an abnormality is located, take a matching digital visual reference photograph using the camera's standard lens mode for the service record.
- Secure the enclosure doors tightly, return to your log interface, record the thermal delta (ΔT) value relative to adjacent phases, and flag the component for torque verification or replacement.
- Never open electrical enclosures without proper arc flash PPE for the panel's calculated incident energy level.
- Keep all instrument cables and straps away from rotating equipment shaft extensions and cooling fan guards.
- Follow the camera manufacturer's minimum safe working distance from energized components — this varies by lens type and panel voltage class.
- Confirm that all access doors you open during an infrared survey are fully secured upon departure.
Immediately stop routing procedures and log an emergency critical escalation if:
- An infrared scan of an active main breaker reveals an electrical terminal connection temperature exceeding 250°F (121°C) — indicating an imminent catastrophic electrical fire or arc fault event threat.
- A machinery vibration signature jumps suddenly during your live route check, accompanied by heavy structural groaning or clicking noises from internal gearbox components.
- An ultrasound meter reading jumps more than 12 dBμV above its established baseline on a bearing measurement point in a single collection cycle — indicating rapid, accelerating internal damage.
- Log the exact numerical high-temperature readings and the associated baseline ambient room value inside the service log.
- Document the specific instrument model number and serial number used to verify calibration compliance tracking parameters.
- Record the asset tag, measurement point ID, date, time, technician name, and shift number for every data collection event.
- Attach both the thermal image file and the visible-light reference photograph to the associated work order in the CMMS.
Infrared Thermography Cameras (e.g., FLIR E-series): Optical matrix sensors that convert invisible infrared light energy into readable color-coded thermal maps. Resolution, sensitivity, and lens FOV determine the minimum detectable temperature difference and maximum safe inspection distance.
Acoustic Ultrasound Heterodyne Meters: Instruments that intercept high-frequency ultrasound waves (20–100 kHz) and translate them down into lower-frequency tones audible through a technician's headphones. Used for leak detection, bearing monitoring, and electrical inspection.
Handheld Vibration Analyzers: Instruments ranging from simple single-value overall meters to full-spectrum FFT analyzers that display bearing defect frequencies, harmonics, and phase relationships. Paired with calibrated accelerometers mounted to defined measurement points.
Online Continuous Monitoring Senders: Permanently mounted wireless sensors that stream vibration and temperature metrics to a central data dashboard 24 hours per day — eliminating the data collection gap between manual route visits.
- TECH-2.4: Basic Lubrication Chemistry & Fluid Dynamics — understanding the lubricant that oil analysis samples
- TECH-2.5: PM Routing Execution & Inspection Techniques — the manual inspection layer that PdM extends
- TECH-11.2: Root Cause Analysis Techniques — what to do with the failure mode once PdM identifies it
Standards Referenced: ISO 17359 — Condition monitoring and diagnostics of machines, general guidelines. ASNT Recommended Practice for Personnel Qualification in Predictive Maintenance Disciplines. NFPA 70B — Recommended Practice for Electrical Equipment Maintenance.
An industrial fan skid has a hidden failure somewhere in the system. The machine is running under load. Deploy the correct instrument from your virtual toolbox to isolate the root failure class before the asset health drops to zero. You have three attempts.
Select your instrument:
§17 · Knowledge Check
You are checking an electric motor using a handheld thermal camera. The main terminal block wiring connections appear cool and uniform. However, you notice that the small flexible coupling connecting the motor shaft to a centrifugal pump is glowing bright yellow, reading 165°F (74°C). The nearby shafts read 110°F (43°C).
What specific failure mode does this thermal profile indicate?
✅ Lesson 2.6 Complete
You can now identify the four pillars of industrial PdM, match failure modes to the correct detection technology, understand the failure cascade from chemical trace to thermal event, and conduct a basic infrared enclosure audit.