Compressor reliability is the backbone of any refrigeration installation. When it fails, production stops, product can be lost, the cold chain breaks and energy costs can spike. This guide explains predictive maintenance of a compressor (reciprocating, screw, scroll, hermetic and semi-hermetic) for maintenance leads and plant engineers—what to measure, how to read the signals, and how to turn evidence into effective work orders. With a disciplined approach, predictive maintenance of a compressor cuts downtime, improves efficiency and safeguards the cold chain.
See also: Types of industrial maintenance · Condition based maintenance (CBM) · Industry 4.0 predictive maintenance plan.
How does a compressor work in a refrigeration system?
In a vapour-compression cycle the compressor raises the refrigerant pressure and temperature, sending superheated gas to the condenser. After heat rejection and subcooling, an expansion device drops the pressure; liquid/vapour enters the evaporator, absorbs heat and returns as low-pressure vapour to the compressor. Monitoring the health of the compressor—and the process variables that describe this cycle—is central to a robust predictive maintenance of a compressor programme.
What to monitor on a compressor—and why
A solid plan rests on three pillars: mechanical condition, thermodynamic process and lubricant condition. Add targeted technologies for leaks and hot-spots, and—where relevant—electrical signatures.
- Predictive maintenance vibration analysis (rotor/bearings, misalignment, looseness, screw gearmesh). Use machine-level vibration measurement and evaluation to set alarms and judge severity; rely on structured diagnostic procedures to ensure consistent routes and interpretation. See ISO 20816-1 for measurement/evaluation and the ISO 13373 series for vibration diagnostics practice.
- Process variables (suction/discharge pressure and temperature, superheat/subcooling, oil temperature). Trends reveal efficiency losses, liquid slugging or valve damage. Their selection and control logic should sit inside a condition-monitoring programme as outlined in ISO 17359.
- Lubricant condition (viscosity, water/acid number, particle count, wear metals). The oil “tells the story” of internal wear and thermal/chemical degradation.
- Ultrasound / airborne acoustics. For early refrigerant leak location at joints and valves, abnormal friction and electrical effects (arcing/corona) in panels; governed by ISO 29821.
- Infrared thermography. Heads, casings, discharge lines, electrical panels and power terminations; interpret images within a condition-monitoring programme per ISO 18434-2.
- Motor electrical signature (current/voltage). Phase imbalance, cracked rotor bars, eccentricity, VFD harmonics. Energy tracking (kWh/tonne of refrigeration; COP/EER trend) quantifies performance drift.
Practical start-point. Begin predictive maintenance of a compressor with vibration + process (P/T, superheat, subcooling) + oil. Add ultrasound for leaks and thermography for electrical/mechanical hot-spots as the programme matures.
Predictive maintenance vibration analysis
Predictive maintenance vibration analysis is the compressor’s universal sensor: from a single discipline you can watch unbalance, misalignment, looseness, bearing faults and compressor-specific phenomena (pulsation in reciprocating units; gearmesh in screws).
What to capture in practice
- Global magnitudes (RMS) for surveillance and trending.
- FFT spectrum for 1×RPM (unbalance), 2×RPM (misalignment), sidebands from looseness, and bearing fault frequencies (BPFI/BPFO/BSF/FTF).
- Envelope for bearing faults and brinelling.
- Phase and orders for assembly issues or resonances.
Turning signals into decisions
- Classify machines and set thresholds using ISO 20816-1; apply ISO 13373 for diagnostic procedure (measurement points, repeatability, filters). Document severity (alert/action) and the linked response (inspection, deeper analysis, intervention).
Noise and ultrasound for predictive maintenance
Here “noise” covers ultrasound (contact/airborne) and acoustics.
- Refrigerant leaks: ultrasound pinpoints micro-leaks at joints, valves and separators—even in noisy environments.
- Friction and electrical discharges: dry bearings, cavitation, and arcing/corona in electrical cabinets generate characteristic ultrasonic signatures.
- Procedures: fixed routes, gain, distance, recording, and confirmation (e.g., tracer spray). ISO 29821 provides guidelines, procedures and validation for ultrasound in condition-monitoring programmes.
Lubricant analysis (condition monitoring)
Parameters to trend
- Viscosity (ageing/contamination).
- TAN/acid number (oxidation; reactions with refrigerant).
- Water (ppm) and particles (system cleanliness).
- Wear metals (Fe, Cu, Al, Sn by ICP) to localise sources (bearings/journals/gears).
Good practice
- Use a repeatable sampling procedure (point, purge, temperature) with traceability; compare trends, not isolated readings.
- Relate results to process events (e.g., liquid slugging drives dilution and additive breakdown).
- Integrate lab alarms with CMMS/GMAO so findings become prioritised work orders (see our Industry 4.0 predictive maintenance plan).
- ISO 17359 frames lubricant physico-chemical analysis inside CBM/PdM programmes, ensuring oil becomes “just another sensor” within predictive maintenance of a compressor.
Oil analysis in predictive maintenance
Treat oil as a condition sensor and read it in context:
- Thermal ageing: TAN rises + viscosity changes.
- Moisture contamination: risk of sludge/acids—often correlates with high discharge temperatures.
- Wear metals trending: sustained, significant increases = degradation alarm; validate with vibration/ultrasound.
Set sampling frequencies by criticality and hours of service, following the programme structure in ISO 17359 (procedures, records, continual review). This way, predictive maintenance of a compressor unifies chemistry, process and mechanics.
Bearing maintenance
Bearings are the most common mechanical root cause in compressors (reciprocating, screw, scroll). Embedding bearing maintenance inside the predictive maintenance of a compressor lets you catch lubrication issues, contamination, misalignment/soft-foot and VFD-related electrical erosion before they escalate.
What to watch
- Lubrication starvation/dilution, moisture/particle contamination, mounting or preload errors, electrical erosion. (Use ISO 15243 to code confirmed failures.)
- How to monitor: vibration RMS/FFT/envelope at DE/ND housings with machine-level limits (ISO 20816-1) and diagnosis per ISO 13373; lubricant analysis (condition monitoring) trending viscosity, TAN, water, particle count and wear metals (ISO 17359; cleanliness by ISO 4406); spot IR on housings/connectors; ultrasound for friction cues and leaks.
- Prevent before detect: correct fits/tolerances and alignment (ISO 492), review pipe strain and validate L10 life with ISO 281.
- Turn signals into action: combine vibration + oil + temperature deltas; when thresholds are breached, open a CMMS work order and record the confirmed defect using ISO 15243 to keep refining rules.
Industrial electric motor predictive maintenance
The compressor’s motor deserves a mini PdM plan of its own inside the wider predictive maintenance of a compressor:
-
- Electrical signature (current/voltage): phase imbalance, cracked rotor bars (induction motors), eccentricities, and VFD harmonics.
- IR thermography: connections, contactors, drives and terminals; hot joints imply high resistance or unbalanced loads—interpret thermograms with ISO 18434-2. ISO
- Insulation (periodic tests aligned with preventive tasks) and earthing integrity.
- Energy: track specific consumption (kWh per tonne of refrigeration) at comparable load; if it rises with a stable process, suspect mechanical or performance degradation.
Tip. Correlate electrical findings with vibration and process to distinguish mechanical vs. electrical vs. control causes.
Implementation flow and working with CMMS
- Inventory and criticality of compressors (production, safety, environment impact).
- Measurement plan with repeatable routes (vibration, ultrasound), process variables and oil schedule.
- Thresholds:
- Vibration per ISO 20816-1; diagnostics per ISO 13373. ISO+1
- Process: target bands for superheat/subcooling and discharge temperature.
- Oil: internal limits and trend-based criteria.
- Multiprotocol platform (OPC UA/Modbus/MQTT) + CMMS: every alert creates a work order with priority, task and verification.
- Closed loop: record the confirmed defect (e.g., inner-race bearing) to refine rules and response times.
Compressor-specific KPIs
- MTBF/MTTR for the compressor-motor set.
- Leak rate (kg/year) and number of tightness incidents.
- COP/EER trend at comparable load.
- Discharge temperature, superheat and subcooling within target windows.
- Asset Health Index (AHI) combining vibration, oil and energy.
- Unplanned stops attributable to the compressor.
Set explicit targets (e.g., +25% MTBF and −30% leak rate in 12 months) and review quarterly. This keeps predictive maintenance of a compressor demonstrably valuable and high on the priority list.
Conclusion: a practical path to protect the refrigeration “heart”
Predictive maintenance of a compressor prevents stoppages, avoids losses and improves efficiency. A robust minimum viable scope combines vibration (per ISO 20816-1/ISO 13373), well-instrumented process (framed by ISO 17359), oil trending, ultrasound for leaks and thermography on power/casing (ISO 18434-2). All of it must operate under safety and environmental procedures consistent with ISO 5149-4 / EN 378-4 and the EU F-gas Regulation now in force.
Compressors are only one possible application. Discover other examples of predictive maintenance across sectors that highlight its real impact
Will you prepare for change? Discover how to implement predictive maintenance in your industry.
