VFD Monitoring: How to Monitor Variable Frequency Drives
Key Takeaways
- VFDs complicate condition monitoring because they produce non-sinusoidal current waveforms and variable operating speeds, requiring adapted analysis techniques such as order tracking and VFD-aware motor current signature analysis.
- Bearing current damage is a VFD-specific failure mode that standard vibration or current monitoring alone cannot detect; it requires dedicated high-frequency current clamps or shaft voltage probes.
- A complete VFD monitoring program covers four domains: electrical parameters (output voltage, current, DC bus, THD), thermal conditions (drive panel and motor), mechanical behavior (vibration via order tracking), and bearing discharge activity.
What Is VFD Monitoring?
A variable frequency drive converts fixed-frequency AC supply power into variable-frequency, variable-voltage output to control motor speed. This power conversion introduces failure modes that do not exist in direct-on-line motor installations: capacitor degradation on the DC bus, IGBT switching failures, harmonic distortion injected back into the power system, and high-frequency bearing currents that pit bearing surfaces.
VFD monitoring addresses each of these by tracking the right parameters at the right points in the drive-motor circuit. It combines electrical measurements at the drive input and output, thermal measurements across the drive cabinet and motor frame, vibration measurements on the motor and driven load, and specialized bearing current measurements at the motor shaft. Together, these signals give reliability teams early warning across the full drive-motor system rather than reacting to failures after the fact.
Why VFDs Complicate Condition Monitoring
Standard condition monitoring techniques assume sinusoidal supply voltage and fixed motor speed. VFDs violate both assumptions simultaneously, which breaks several conventional measurement approaches.
Non-Sinusoidal Current Waveforms
VFDs synthesize output voltage using pulse-width modulation (PWM). The resulting current waveform contains the fundamental frequency the motor needs plus a wide range of harmonic frequencies at multiples of the PWM carrier frequency, which typically runs between 2 kHz and 16 kHz. These harmonic currents add to winding and iron losses, raise operating temperature above nameplate ratings, and accelerate insulation aging.
For vibration analysis and motor current signature analysis, harmonic content from the drive masks the lower-amplitude fault sidebands that analysts look for. Without filtering or VFD-aware signal processing, analysts may miss rotor bar defects, eccentricity, or bearing faults that would be clearly visible on a direct-on-line motor.
Variable Speed Operation
When a motor runs at a fixed speed, its fault frequencies are fixed. Bearing defect frequencies, gear mesh frequencies, and unbalance appear at predictable locations in the vibration spectrum. When a VFD changes motor speed in response to process demand, all these frequencies shift proportionally. An alarm set at a fixed frequency will trigger false positives at some speeds and miss real faults at others.
This variable-speed challenge is the primary reason that conventional route-based vibration monitoring struggles with VFD-driven assets. Measurements taken at different times may capture the motor at different speeds, making trend comparisons meaningless unless speed is recorded alongside every data point.
Key Parameters to Monitor on a VFD
Effective VFD monitoring covers parameters at three points: the drive input from the power supply, the drive output to the motor, and the motor itself.
Output Voltage and Current
Output current is the most direct indicator of motor load. Trending RMS output current against motor nameplate full-load amperage reveals overloading before thermal damage occurs. Phase imbalance between the three output phases points to winding degradation or connection problems. Sudden current spikes indicate mechanical overload events, locked rotor conditions, or short circuits.
Output voltage waveform quality, particularly peak voltage and rise time, determines the stress placed on motor winding insulation. PWM drives with fast switching IGBTs generate high peak voltages at the motor terminals, especially when cable lengths between drive and motor are long. Exceeding the insulation's voltage rating accelerates partial discharge and shortens winding life.
DC Bus Voltage
The DC bus sits between the drive's input rectifier and its output inverter stage. DC bus voltage directly reflects the health of the rectifier diodes and the electrolytic capacitors that smooth the rectified waveform. Capacitors are the most failure-prone component in a VFD; they degrade with operating temperature, ripple current, and time. As capacitance falls, DC bus voltage ripple increases, which introduces low-frequency harmonics into the output and stresses IGBTs. Monitoring DC bus ripple voltage, not just average DC voltage, provides the earliest indication of capacitor aging before the drive trips on overvoltage or undervoltage faults.
IGBT Temperature
Insulated gate bipolar transistors (IGBTs) are the switching elements that generate the PWM output waveform. IGBT junction temperature is the primary determinant of switching device lifetime. Manufacturers specify maximum junction temperature (typically 125 to 175 degrees Celsius) and thermal resistance from junction to heatsink. Monitoring heatsink temperature via built-in thermistors or the drive's own temperature register, combined with knowledge of ambient temperature and load current, allows calculation of estimated junction temperature. Trends showing heatsink temperature rising at a given load indicate cooling system degradation: blocked filters, failing cooling fans, or fouled heat exchangers.
Total Harmonic Distortion
VFDs draw non-sinusoidal current from the supply network, injecting harmonic currents at characteristic frequencies (5th, 7th, 11th, 13th harmonics for a six-pulse rectifier). Total harmonic distortion (THD) quantifies the ratio of harmonic content to the fundamental. High THD at the drive input heats transformers and cables, creates voltage distortion that affects other equipment on the same bus, and may exceed the limits set by IEEE 519 at the point of common coupling. Monitoring input THD over time identifies whether the aggregate harmonic load from multiple drives is approaching the threshold where active harmonic filters or drive upgrades become necessary.
Motor Bearing Temperature
Motor bearing temperature, measured with resistance temperature detectors (RTDs) or thermocouples embedded near the bearing housings, provides a direct indicator of lubrication degradation, overloading, misalignment, and bearing current damage. Temperature-based thermal monitoring gives advance warning before elevated vibration from a damaged bearing becomes detectable. For VFD-driven motors, trending bearing temperature against speed is essential: bearings on motors running at reduced speed receive less lubrication splash from the oil ring and may run hotter than expected at low load.
Vibration Analysis on VFD-Driven Equipment
Conventional vibration analysis uses Fast Fourier Transform (FFT) to convert a time-domain signal into a frequency-domain spectrum. For a fixed-speed motor, bearing defect frequencies, unbalance (1x), and misalignment (2x) appear at fixed, predictable Hz values that analysts can trend over time.
The Variable Speed Problem
On a VFD-driven motor, every speed change shifts every spectral peak proportionally. If bearing outer race defect frequency is 83 Hz at 1,480 RPM, it will be at 56 Hz at 1,000 RPM. Automated alarms keyed to fixed Hz bands will generate false alarms or miss real faults depending on where the motor happens to be running when the measurement is taken.
Continuous monitoring systems that capture data without recording simultaneous speed face the additional problem that spectra from different acquisitions cannot be meaningfully compared unless operating conditions were identical.
Order Tracking
Order tracking solves the variable-speed problem by expressing the vibration spectrum in multiples of running speed (called orders) rather than absolute Hz. A tachometer or encoder provides a once-per-revolution pulse that the analyzer uses to resample the vibration time waveform at constant angular increments rather than constant time increments. The resulting order spectrum places unbalance at exactly 1x, bearing outer race at a fixed order position calculated from bearing geometry, and all other fault signatures at consistent locations regardless of operating speed.
With order tracking, alarm bands can be set at fixed order positions and trend reliably across all operating speeds. This makes automated, continuous monitoring of VFD-driven equipment practical in a way that fixed-frequency FFT analysis cannot achieve.
Bearing Current Monitoring
Bearing current damage is specific to VFD-driven motors and is one of the most underappreciated causes of premature bearing failure. The PWM switching action of the drive's output stage induces common-mode voltage on the motor shaft relative to the motor frame. When shaft voltage builds to the dielectric breakdown threshold of the bearing lubricant film (typically 5 to 30 volts), current discharges through the rolling elements and races.
Repeated discharge events cause pitting on bearing surfaces that develops into a characteristic washboard pattern called fluting. Fluting produces elevated vibration at bearing defect frequencies, but by the time vibration rises measurably, the bearing surface is already significantly damaged. A fluted bearing may fail within weeks.
How to Monitor for Bearing Currents
Shaft voltage measurement uses a probe riding on the motor shaft or slip ring to measure the common-mode voltage waveform directly. Peaks above 5 to 10 volts indicate high discharge risk. High-frequency current clamps around the motor shaft ground strap or bearing housing detect the discharge current pulses directly.
Mitigation options include shaft grounding rings (carbon fiber brushes that provide a low-impedance path for discharge currents), insulated non-drive-end bearings (which block the discharge current path through the bearing), and reduced PWM carrier frequency (which reduces common-mode voltage generation).
Thermal Monitoring of VFD Panels
VFD enclosures generate significant heat from switching losses and I-squared-R losses in input and output components. Drive manufacturers specify maximum ambient temperature (typically 40 degrees Celsius for most industrial drives) and require derating above that temperature. Enclosures that exceed rated ambient shorten capacitor and IGBT life, increase failure probability, and may trigger drive protection shutdowns that interrupt production.
Infrared thermography of open VFD panels during scheduled maintenance identifies hotspots on bus bars, fuse connections, terminals, and heat exchangers. Connections that run hotter than their neighbors indicate high-resistance joints from loose fasteners or corrosion. IGBT modules running above their expected heatsink temperature indicate partial failure, driver circuit problems, or cooling system degradation.
Continuous thermal monitoring via drives' internal temperature sensors, with data logged to a monitoring platform, enables trending and allows overtemperature alarms to trigger before the drive shuts itself down on a thermal fault.
VFD Monitoring Parameters and Methods
| Parameter | What It Detects | Measurement Method | Failure Warning Lead Time |
|---|---|---|---|
| Output current (RMS + phase balance) | Overload, winding faults, phase loss | CT clamps or drive internal meter | Days to weeks |
| DC bus voltage ripple | Capacitor aging, rectifier degradation | Oscilloscope or drive diagnostics | Months |
| IGBT heatsink temperature | Cooling failure, IGBT degradation | Drive internal thermistor, IR camera | Hours to days |
| Input THD | Harmonic distortion to supply, compliance risk | Power quality analyzer | Weeks to months |
| Motor bearing temperature | Lubrication loss, bearing current damage, overload | RTD or thermocouple at bearing housing | Hours to weeks |
| Vibration (order-tracked) | Bearing defects, unbalance, misalignment | Accelerometer + tachometer, order tracking analyzer | Weeks to months |
| Shaft voltage / bearing current | Bearing current discharge risk | Shaft voltage probe or HF current clamp | Months (before fluting visible) |
Integration with Predictive Maintenance
VFD monitoring data has limited value when treated as isolated readings. The real diagnostic value emerges when electrical, thermal, vibration, and bearing current data are combined and trended over time in a predictive maintenance platform.
Correlating rising motor bearing temperature with elevated vibration at bearing defect frequencies, for example, confirms an active bearing fault far earlier and more reliably than either signal alone. Correlating DC bus ripple growth with IGBT temperature trends identifies a drive nearing end-of-life capacitance, allowing a planned replacement during a scheduled outage instead of an emergency repair during production.
Data Integration Requirements
To support predictive maintenance decisions, VFD monitoring systems need to capture operating speed alongside every electrical and vibration measurement. Without speed data, vibration spectra from different times cannot be compared, and thermal data cannot be normalized to load. Modern monitoring platforms pull speed directly from the drive via fieldbus protocols (Modbus, EtherNet/IP, PROFINET) or from a dedicated tachometer channel on the sensor, eliminating the need for manual speed recording.
Real-time monitoring with continuous data streaming enables event-triggered analysis. When an overtemperature alarm fires, the monitoring system can automatically pull the vibration record, current waveform, and speed data from the preceding 60 seconds to give the analyst full context without requiring a separate manual measurement. This capability is especially important for intermittent or transient faults that do not persist long enough for a technician to reach the equipment with a handheld instrument.
Alarm Strategy for Variable-Speed Assets
Effective alarm configuration on VFD-driven assets uses speed-dependent alarm bands rather than fixed thresholds. Vibration alarms in order-tracked spectra use order-relative limits. Thermal alarms account for load current and ambient temperature so that a motor running at full load in a hot environment is compared to appropriate baseline, not to a cold-running light-load baseline. Current alarms normalize to rated full-load amperage at the actual operating frequency rather than applying fixed amperage thresholds that would generate false overload alarms at reduced speed.
VFD Monitoring and Motor Current Signature Analysis
Motor current signature analysis (MCSA) analyzes the frequency content of motor supply current to detect mechanical and electrical faults. For induction motors running direct-on-line, MCSA is a mature technique. Fault sidebands appear at precise frequencies relative to the supply fundamental (50 or 60 Hz) and can be automated with fixed alarm bands.
On VFD-driven motors, the fundamental output frequency is itself variable and the current waveform carries the drive's carrier harmonics. MCSA applied to VFD motors requires pre-filtering to isolate the fundamental from the carrier harmonics, then analysis relative to the actual output frequency rather than the supply frequency. Systems designed specifically for VFD-aware MCSA can detect rotor bar faults, eccentricity, and bearing defects in variable-speed operation. Standard MCSA instruments that assume a fixed 50 or 60 Hz fundamental are not suitable for VFD-driven motors without this adaptation.
Frequently Asked Questions
What parameters should you monitor on a VFD?
The most important parameters include output voltage and current (to detect load imbalance and overload), DC bus voltage (to catch rectifier degradation and capacitor aging), IGBT junction temperature (to prevent thermal runaway), total harmonic distortion at the drive output, and motor bearing temperature. Each parameter targets a distinct failure mode and together they give a complete picture of drive and motor health.
Why is vibration analysis harder on VFD-driven motors?
VFD-driven motors operate at variable speeds, so the fault frequencies in the vibration spectrum shift with every speed change. Traditional FFT analysis assumes stationary conditions and fixed fault frequencies, so peaks move around and are easy to miss. Order tracking solves this by normalizing the spectrum to shaft rotational orders rather than absolute frequencies, making fault signatures appear at consistent locations regardless of speed.
What is bearing current damage in VFD-driven motors?
High-frequency switching in VFD power stages induces common-mode voltages on the motor shaft. When this voltage exceeds the dielectric strength of the bearing lubricant film, current discharges through the bearing rolling elements and races. The resulting pitting and fluting on bearing surfaces causes premature failure, often within months on motors without shaft grounding rings or insulated bearings. Bearing current monitoring uses high-frequency current clamps on the motor shaft to detect discharge pulses before visible damage occurs.
How does order tracking differ from standard FFT analysis for VFD applications?
Standard FFT analysis produces a spectrum in fixed frequency units (Hz). When a motor speed changes, all fault frequencies shift proportionally, making automated alarm detection unreliable. Order tracking uses a tachometer or encoder signal to resample the vibration data relative to shaft rotation, expressing the spectrum in multiples of running speed (orders) rather than absolute frequency. Fault signatures then appear at consistent order positions regardless of the speed at which the measurement was taken.
Can you use motor current signature analysis on VFD-driven motors?
MCSA works on VFD-driven motors but requires extra signal processing. VFDs generate non-sinusoidal current waveforms with multiple carrier frequency harmonics that mask the fault sidebands MCSA relies on. Advanced MCSA systems designed for VFD applications apply filtering to isolate the fundamental output frequency and analyze sidebands around it. Direct online (DOL) MCSA tools generally cannot be used without this VFD-aware processing.
How often should VFD thermal monitoring be performed?
Continuous thermal monitoring is preferred for critical VFD installations. Infrared thermography can be used as a periodic check during scheduled downtime or when cabinet doors are safely accessible, typically every three to six months. Continuous monitoring with built-in thermistors or external thermocouple inputs on the drive provides real-time overtemperature alarms and load-based trending. Thermal limits set by the drive manufacturer specify maximum ambient temperatures and derating curves that inform alarm thresholds.
What is total harmonic distortion (THD) and why does it matter for VFDs?
Total harmonic distortion (THD) is the ratio of harmonic content to the fundamental frequency in a voltage or current waveform, expressed as a percentage. VFDs draw non-sinusoidal current from the supply, injecting harmonics back into the power distribution system. High THD heats transformers, cables, and capacitors beyond design ratings, shortens motor insulation life, and can cause nuisance tripping of protective relays. IEEE 519 sets THD limits at the point of common coupling. Monitoring THD at the drive input helps identify when harmonic mitigation, such as line reactors or active front-end drives, is needed.
The Bottom Line
VFD monitoring is not optional for facilities running critical variable-speed assets. The same drive technology that delivers energy savings and process control creates failure modes that standard condition monitoring techniques are not equipped to detect: bearing currents that destroy bearings before vibration rises, capacitor aging that escalates to catastrophic drive failure, and harmonic distortion that shortens the life of every motor and transformer on the same supply bus.
A complete monitoring program addresses all four domains: electrical parameters at the drive input and output, thermal conditions across the drive panel and motor, order-tracked vibration on the motor and driven machine, and bearing current activity on the motor shaft. Each domain targets distinct failure mechanisms. Used together and integrated into a predictive maintenance platform, they convert what would otherwise be reactive emergency repairs into planned, scheduled interventions at a fraction of the cost and production impact.
Monitor VFD-Driven Assets Before They Fail
Tractian's condition monitoring platform combines continuous vibration, current, and temperature data from VFD-driven motors with built-in speed-normalized alarming. See faults early, plan repairs, and protect every critical drive in your facility.
See How Tractian Monitors VFD-Driven AssetsRelated terms
Industrial Maintenance: Types
Industrial maintenance is the set of activities that keep manufacturing and industrial equipment operating reliably. Learn the types, key strategies, performance metrics, and how to build an effective program.
Industrial Automation: Types
Industrial automation uses control systems, robotics, and software to perform manufacturing tasks with minimal human input. Learn the types, key technologies, benefits, and how automation affects maintenance.
In-Process Control: Definition
In-process control (IPC) is the real-time monitoring and testing of production processes to catch defects before they compound. Learn how it works, key methods, and its role in regulated industries.
In-House Maintenance: Definition
In-house maintenance is when a company uses its own employees to handle maintenance tasks. Learn how it compares to outsourcing, its advantages, challenges, and how a CMMS supports internal teams.
Hold for Disposition Stock: Definition
Hold for disposition stock is quarantined inventory awaiting an authorized decision before use, return, rework, or scrap. Learn the process, outcomes, compliance requirements, and how a CMMS helps.