Predictive Maintenance of Motors and Drives

Short answer

Predictive maintenance of motors and drives measures variables that change before failure: vibration, temperature, current, insulation and the condition of the drive capacitors. Techniques are chosen by equipment criticality, a baseline is set and action is taken on trend, not on a single reading.

From correcting to anticipating

There are three ways to maintain a motor. Corrective waits for the failure and repairs; preventive intervenes at fixed intervals, needed or not; predictive intervenes when a measurement shows the condition is degrading. The last avoids both the unexpected shutdown and the replacement of healthy parts, and rests on a simple principle: almost every failure gives signs before it happens.

In a manufacturing or process plant, where a pump, fan or compressor motor stops an entire line, that anticipation pays for itself. The electrical predictive maintenance guide covers the general idea; this one focuses on motors and drives.

Why motors and drives fail

Knowing the causes guides what to measure:

Equipment	Usual failure	Sign that appears first
Motor	Worn or poorly lubricated bearings	Vibration at bearing frequencies, ultrasound, temperature
Motor	Unbalance, misalignment, looseness	Vibration at one and two times running speed
Motor	Insulation deterioration	Falling insulation resistance, partial discharges
Motor	Broken rotor bars	Sidebands in the current spectrum
Motor	Overheating from ventilation or overload	Temperature, current, thermography
Drive	Aged DC bus capacitors	Internal temperature, voltage ripple, operating hours
Drive	Blocked fans and filters	Heat sink temperature, unit warnings
Drive	Loose connections	Hot spots in thermography
Motor on a drive	Currents through the bearings	Fluting on the races, noise, early failure

With drives, a particular mechanism appears: the common-mode voltage generated by switching can make current flow through the bearings and mark the races. It is mitigated with proper grounding, shielded cabling, filters and, where applicable, insulated bearings. For that reason, on a drive-fed motor the bearing condition deserves closer attention.

The techniques and what each measures

Vibration analysis. Velocity or acceleration is measured at the supports with an accelerometer. The overall value is compared with severity limits such as those of the ISO 20816 series, which classifies by zones A to D according to machine type and size. The spectrum identifies the cause. It is the basis of most programs.

Thermography. An infrared camera shows temperature differences at terminals, junction boxes, frame, bearings and on the drive. A guide on thermography in substations and panels details the capture criteria; they apply equally to motor panels.

Motor current signature analysis (MCSA). Examines the spectrum of the supply current. It detects broken rotor bars, eccentricity and some load problems, with no instruments on the machine. For bearings it usually detects damage at advanced stages, so it does not replace vibration for that purpose.

Ultrasound. Picks up the high-frequency sounds produced by friction and lack of lubrication, and is also used for leaks and discharges. It is useful to decide when to lubricate and how much grease to apply.

Insulation and electrical tests. Insulation resistance, polarization index and winding resistance with the equipment de-energized give the condition of the winding. The guide on electrical tests on low-voltage motors explains what each one detects. On medium voltage, other tests are added.

Power quality and load. Voltage unbalance, harmonics and undervoltage degrade the motor. The guides on drive harmonics and power quality in industry show how they influence.

Data from the drive itself and the motor relay. Current, temperature, hours, number of starts and historical alarms are already available over industrial communication and can be taken to the SCADA or an asset management system. A motor protection relay also provides start counters and thermal state.

Frequency and decision criteria

Frequency depends on how fast the failure develops. A bearing defect advances over weeks or months, so a monthly or quarterly route detects it with margin. A lubrication problem on a high-speed machine can progress in days and justifies continuous monitoring. A loose connection that heats evolves erratically and is checked with thermography on every round and after every intervention in the panel.

To decide what to do with a finding, a simple three-state scale is useful: normal, watch (the interval is shortened and work is planned) and action (the intervention is scheduled in the next available window). The decision weighs the severity of the symptom, the equipment criticality, spare part availability and the next scheduled shutdown. A severe finding on equipment with a spare available and a nearby shutdown is acted on. The same finding on equipment with a six-month spare lead time calls for close watching and ordering the spare now.

Building the program

Rank equipment by criticality: consequence of the shutdown, repair cost, replacement time, safety. That decides the technique and the frequency.
Choose the technique by failure mode, not by habit. A critical compressor gets online vibration and temperature monitoring; an auxiliary fan, a monthly or quarterly route.
Measure the baseline with the equipment in good condition and repeatable operating conditions: same speed, same load. On drive-fed motors, speed changes the spectrum, and comparisons must be made under the same condition.
Set alarms by trend and by level, with an early warning and an action alarm.
Record and close the loop: each finding generates a work order and, after the intervention, a new measurement.
Review the program: if there are failures that were not detected, adjust the technique or the frequency.

Illustrative numerical example

Data invented to show the economic reasoning.

A process pump, with an estimated line stoppage of 6 hours per failure and a loss of USD 4,000 per hour, suffers two bearing failures a year: USD 48,000 in lost production, plus repairs. A vibration and temperature monitoring system costs, say, USD 6,000 installed and USD 1,000 a year in upkeep. If monitoring allows the change to be scheduled in a planned 1-hour shutdown, the loss per event falls to about USD 4,000, and secondary damage to the shaft is avoided. The annual saving would be on the order of USD 40,000 and the investment is recovered in a few weeks.

For a small motor that is cheap to replace and has no production impact, the same analysis would give another result: a periodic route or no measurement. Criticality decides.

Drives: what to watch on the unit

A drive does not wear like a rotating machine, but its components age with temperature. The electrolytic capacitors of the DC bus lose capacitance with heat and hours, and their life depends on operating temperature. A poorly ventilated panel shortens it noticeably. Cooling fans have a limited life and their failure causes overtemperature trips. Dust on filters and heat sinks reduces dissipation. That is why the plan includes cleaning, checking the panel ambient temperature, retightening terminals with the equipment de-energized and reading the internal counters. Many manufacturers recommend replacing fans and capacitors by operating hours or by condition; follow the manual for the model.

Common mistakes

Measuring without a baseline and without repeatable conditions.
Reacting to an isolated reading instead of the trend.
Taking vibration on a drive-fed motor without noting the speed.
Changing bearings without investigating the cause (misalignment, lubrication, currents).
Over-lubricating: too much grease also does damage.
Installing sensors and not defining who reviews the alerts.
Ignoring the data the drive already provides.
Not storing measurements with the equipment, the date and the operating condition.

Checklist

Inventory of motors and drives with criticality.
Technique and frequency defined by failure mode.
Baseline recorded with speed and load.
Alarm criteria by level and by trend.
Thermography of terminals and the drive in the periodic route.
Insulation tests planned with the equipment de-energized and locked out.
Review of drive fans, filters and capacitors according to hours.
Work orders tied to each finding.
Report of results and of program effectiveness.

How we accompany you

We carry out electrical tests and thermography on motors and panels, and the commissioning tests that leave the baseline, in Colombia or elsewhere on request. We also support the design of the instrumentation and communication to bring the data to the supervision system from engineering design. We quote according to your equipment fleet.

References

ISO 20816 series, Mechanical vibration: measurement and evaluation of machine vibration (supersedes ISO 10816).
NFPA 70B, Recommended Practice for Electrical Equipment Maintenance.
IEEE 43, Recommended Practice for Testing Insulation Resistance of Electric Machinery.
Technical literature on motor current signature analysis and bearing currents in adjustable speed drives.

Frequently asked questions

Which technique detects a bearing failure earliest?

Vibration analysis, with a frequency spectrum, is the most used technique because it reveals bearing defects long before the overall vibration level or temperature rise. Ultrasound and current analysis complement it in particular cases.

Do all motors deserve online monitoring?

No. Continuous monitoring is justified on critical equipment, hard-to-reach equipment or equipment with a high downtime cost. For most, a periodic measurement route is enough; small, cheap motors are usually handled by replacement.

Does a drive need predictive maintenance?

Yes, although it has no rotating parts except the fans. The wear points are the fans, the DC bus capacitors, the connections and the air filters. Many drives report operating hours, temperature and counters that are useful for planning.

Does thermography replace vibration?

No, they measure different things. Thermography detects loose connections, phase unbalance, overheating and blocked ventilation; vibration detects mechanical problems such as misalignment, unbalance, looseness and bearings. They are used together.