Online monitoring detects what is developing. Offline testing confirms what online monitoring has flagged and detects conditions that cannot be measured while the machine runs. Neither replaces the other; together they cover the full range of failure modes.
Online Monitoring Methods
Online monitoring measures the machine while it is running under normal operating conditions. This is not just a scheduling convenience: many failure modes only manifest under load, at temperature, and at the operating speed. A bearing defect that produces a detectable vibration signature at 1,800 RPM under full load may be invisible when the machine runs slowly during a startup test.
Vibration Analysis
Vibration measurement at bearing housings while the machine runs under load is the most widely applied online monitoring technique. Spectrum analysis identifies bearing fault frequencies, imbalance (1x), misalignment (1x and 2x), looseness, and gear mesh frequencies. The key advantage of online measurement is that it captures the fault under the exact load and speed conditions where it develops. A bearing defect that is load-sensitive will only be apparent in online data.
Motor Current Signature Analysis (MCSA)
MCSA measures the current draw of an induction motor using a clamp-on current transducer while the motor runs under load. Electrical and mechanical faults modulate the current waveform in characteristic ways: broken rotor bars produce sidebands around line frequency at slip frequency multiples; eccentricity produces sidebands at twice line frequency; load-side mechanical faults modulate the current at the fault frequency. MCSA is non-contact and requires no sensor mounting on the machine itself, which makes it useful for motors that are difficult to access for vibration measurement.
Oil Analysis on Running Equipment
Oil samples drawn from the lubrication system while the equipment runs capture particles and contaminants generated under operating conditions. Wear particle count and morphology indicate bearing and gear wear rate and mode. Contamination (water, process fluid) is detected by physical and chemical analysis. Viscosity and additive depletion indicate lubricant degradation. Running samples are more representative of actual wear conditions than oil drained during an outage.
Thermal Imaging
Infrared thermography performed while equipment runs detects heat generated by friction (failing bearings, misaligned couplings), electrical resistance (loose connections, failing components), and fluid flow irregularities. Thermal imaging of motor control panels and switch gear is necessarily online because the heat is generated only under load. Bearing temperature by IR requires line-of-sight access to the bearing housing surface.
Acoustic Ultrasound
Ultrasonic measurement in the 20 kHz to 100 kHz range detects high-frequency energy from bearing defects at earlier stages than vibration measurement in the standard audio frequency range. Ultrasound is also used to detect compressed air or steam leaks in running systems. The technique requires the machine to be running and load-carrying to generate the ultrasonic emission from a defective bearing contact zone.
What Online Methods Detect That Offline Cannot
- Load-dependent faults: Some bearing defects, misalignment conditions, and looseness only appear under operating load. An alignment check performed on a cold, static machine does not reflect the thermal growth and load-induced deflection present during operation.
- Intermittent conditions: Process upsets, varying load, and temperature transients can trigger faults intermittently. Continuous online monitoring captures these events. A monthly route-based measurement is statistically unlikely to coincide with an intermittent fault condition.
- Speed-specific resonances: Structural resonances excited at specific operating speeds are only detectable during operation at those speeds. Coast-down vibration analysis captures resonances encountered during speed change, which is an online technique.
- Electrical fault signatures: MCSA and motor insulation monitoring under load detect conditions that do not appear in static electrical testing.
Offline Testing Methods
Offline testing is performed during planned maintenance windows when the machine is stopped, isolated, and accessible. These methods detect conditions that are either impossible or unsafe to measure while the machine runs, or that require physical access to internal components.
Insulation Resistance Testing
A megohmmeter applies a DC test voltage (typically 500V to 5,000V depending on equipment voltage class) to the motor winding insulation while the motor is de-energized and isolated. The ratio of applied voltage to leakage current gives insulation resistance in megohms. Low insulation resistance indicates moisture absorption, contamination, or physical insulation degradation. Polarization Index (PI) testing, which compares the 10-minute to 1-minute resistance reading, indicates insulation quality more reliably than a single resistance measurement.
Surge Testing (High-Potential Testing)
Surge testing applies a high-voltage impulse to motor windings and measures the waveform reflected back. Turn-to-turn shorts within a winding coil cause waveform asymmetry that identifies which coil and which phase contain the defect. Surge testing detects insulation degradation between adjacent winding turns that insulation resistance testing does not detect.
Gearbox Oil Analysis (Drained Sample)
Oil drained from a gearbox sump at an outage captures accumulated wear particles from the gear tooth and bearing surfaces. Drained oil analysis is complementary to running samples: the concentration of large wear particles in drained oil, settled at the sump bottom, provides different information than the small particles suspended in circulating oil drawn from a running sample port.
Precision Shaft Alignment
Laser shaft alignment is performed with the machine stopped and cooled. It measures the positional relationship between coupled shaft centerlines and identifies angular and parallel offset misalignment in two planes. The measurement is necessarily offline because alignment correction requires physical adjustment of the machine position. Online vibration at 1x and 2x running speed indicates misalignment is present; precision alignment measurement quantifies it and guides correction.
Visual Inspection and Borescope
Physical access to internal machine components during an outage allows inspection of bearing condition, gear tooth surface, coupling element wear, and seal condition. Borescope inspection of compressor internals (impeller, diffuser, labyrinth seal) detects damage not visible from outside the machine. Visual findings confirm what online monitoring has indicated or identify damage that was not detected by external sensors.
Bearing Installation Quality Checks
During bearing replacement at a planned outage, bearing installation quality directly affects subsequent service life. Interference fit verification, preload measurement, and lubrication quantity at installation are offline activities that affect what online monitoring will detect going forward. A bearing installed with incorrect preload or an interference fit outside tolerance will fail earlier and differently than one installed correctly.
The Handoff from Online to Offline
Online monitoring identifies a suspect asset. Offline testing at the next planned outage confirms or rules out the fault before repair is committed.
This handoff matters because repair decisions based on online monitoring alone are probabilistic, not certain. A rising BPFO trend on a bearing indicates developing outer race damage with high probability. But the same vibration pattern can result from contamination, overloading, or a measurement artifact. The cost of an unnecessary bearing replacement is significant. The cost of a planned outage to inspect and confirm is lower.
Effective programs use online monitoring to prioritize which assets receive attention at the next planned outage, not to trigger emergency repairs. Exceptions are assets where the failure consequence justifies immediate action or where online data indicates the fault has advanced beyond a point where a planned outage is realistic.
Scheduling Offline Testing
Offline testing intervals should be driven by asset criticality, service severity, and the feedback from online monitoring, not by calendar schedule alone.
NFPA 70B recommends intervals for motor insulation testing based on environment severity: 6 months for severe environments (high humidity, corrosive atmosphere), 12 months for normal industrial environments. API 670 specifies calibration and testing intervals for protection system components on large rotating machinery. NFPA 70E and NFPA 79 define testing requirements for electrical switchgear during planned maintenance.
The most effective approach is to use online monitoring data to adjust offline testing intervals. A machine showing stable online monitoring results over multiple periods can extend its offline testing interval. A machine showing a deteriorating trend warrants an earlier inspection outage.
Closing the Loop After Repair
After a bearing replacement, seal replacement, or alignment correction, establishing a new baseline measurement before the machine returns to service is essential. The new baseline captures the machine condition immediately after repair, which allows subsequent online monitoring to detect any deviation from the post-repair state. Without this step, the monitoring system is comparing against a baseline that reflects the pre-repair condition rather than the repaired machine.
Post-repair baseline measurements should include: vibration spectrum at rated speed and load, bearing temperature at stabilization, and motor current at rated load. These measurements confirm the repair was successful and provide the starting point for ongoing monitoring.
Online vs Offline Comparison
| Method | Online / Offline | What Is Detected | Production Impact | Equipment Required | Best Application |
|---|---|---|---|---|---|
| Vibration analysis | Online | Bearing defects, imbalance, misalignment, looseness, gear faults | None | Accelerometer, data collector or permanent sensor | Rotating machinery, all criticality levels |
| Motor current signature analysis | Online | Broken rotor bars, eccentricity, load-side mechanical faults | None | Current clamp transducer, spectrum analyzer | Induction motors where vibration access is limited |
| Oil analysis (running sample) | Online | Wear particles in suspension, contamination, viscosity change | None (sample drawn from port) | Sample bottles, laboratory analysis | Gearboxes, large bearings, compressors with oil systems |
| Thermal imaging | Online | Friction heat at bearings and couplings; electrical resistance heat | None (line of sight required) | Infrared camera | Electrical panels, motor surfaces, coupling guards (where accessible) |
| Acoustic ultrasound | Online | Early bearing defects, compressed gas/steam leaks | None | Ultrasonic detector | Early-stage bearing detection; leak detection in pressurized systems |
| Insulation resistance testing | Offline | Moisture, contamination, bulk insulation degradation | Machine must be de-energized and isolated | Megohmmeter | Motors, transformers, switchgear at planned intervals |
| Surge testing | Offline | Turn-to-turn insulation shorts within winding coils | Machine must be de-energized and isolated | Surge tester | Motors with suspected winding defects; pre-rewind confirmation |
| Shaft alignment (laser) | Offline | Angular and parallel shaft misalignment; quantifies what vibration indicates | Machine must be stopped; coupling disconnected | Laser alignment system | After bearing replacement, coupling replacement, or alignment-indicated vibration |
| Visual and borescope inspection | Offline | Gear tooth surface condition, internal wear, seal condition | Full outage required for access | Borescope, basic inspection tools | Confirming online fault indications; planned major maintenance intervals |
| Oil analysis (drained sample) | Offline | Accumulated large wear particles; settled contamination | Oil drain required | Sample containers, laboratory analysis | Gearboxes at planned oil change; confirming wear severity |
Online Condition Monitoring for Rotating Equipment
Wireless vibration and temperature sensors that run continuously, detect bearing faults automatically, and connect to your maintenance workflow. Covers the online half of a complete monitoring program.
See Mechanical Monitoring