Condition-based maintenance (CBM) is a maintenance strategy in which equipment is serviced only when measurements of its actual condition show that a fault is developing. Instead of following a fixed calendar or waiting for a breakdown, CBM uses data such as vibration, temperature, oil quality or motor current to decide when maintenance is needed.
Key facts:
How condition-based maintenance works
- Measure. Equipment condition is measured, either continuously by a permanent monitoring system or periodically with portable instruments.
- Compare. Readings are compared against a healthy baseline and set alarm thresholds.
- Diagnose. When a reading moves outside normal limits, the data is analysed to identify the fault and how serious it is.
- Plan. Maintenance is scheduled for a convenient time, with parts and people ready, before the fault causes a failure.
- Act and review. The repair is carried out and the results feed back into thresholds and future planning.
Condition-based maintenance techniques
Most CBM programmes use one or more of five techniques. Each detects different faults, so the right mix depends on your equipment.
|
Technique |
What it detects |
|---|---|
| Vibration analysis | Imbalance, misalignment, looseness, bearing wear |
| Thermography | Overheating, poor connections, friction |
| Oil analysis | Wear particles, contamination, lubricant breakdown |
| Ultrasound | Early bearing wear, leaks, electrical discharge |
| Electrical signature analysis (ESA / MCSA) | Bearing, alignment and rotor faults, winding faults, pump and process problems, supply quality |
Vibration analysis
Vibration analysis measures how a machine shakes as it runs. Changes in the vibration pattern reveal mechanical faults such as imbalance, misalignment and bearing wear. It is the most established CBM technique, but it needs a sensor or a technician at every machine.
Thermography
Thermography uses an infrared camera to find components running hotter than they should, such as overloaded electrical connections or bearings short of lubricant. It is quick to carry out but usually detects faults at a later stage, once heat has built up.
Oil analysis
Oil analysis tests lubricant samples for metal wear particles, contamination and breakdown of the oil itself. It is well suited to gearboxes, engines and hydraulic systems, but results depend on laboratory turnaround.
Ultrasound
Ultrasonic testing picks up high-frequency sound that people cannot hear, revealing early-stage bearing defects, compressed air and steam leaks, and electrical discharge.
Electrical signature analysis and motor current signature analysis
Electrical signature analysis (ESA), including motor current signature analysis (MCSA), measures the voltage and current supplying an electric motor. Because the motor acts as a sensor for everything it drives, faults in bearings, couplings, pumps and fans create small but recognisable patterns in that supply. Measurements are taken in the motor starter panel, so no sensors are fitted to the machine. This makes ESA well suited to machines that are hard to reach, hazardous or submerged. Read more about electrical signature analysis.
The P-F curve: why early detection matters
The P-F curve shows how a machine’s condition declines over time. Point P is when a developing fault first becomes detectable; point F is functional failure. The time between them, the P-F interval, is your window to plan a repair.
The earlier a technique detects a fault, the wider that window. Temperature rises and audible noise usually appear late, close to failure, while vibration, ultrasound and electrical analysis can detect faults much earlier. At one power station, Faraday Predictive’s monitoring flagged a fan bearing fault 18 months before the existing alarm system.
CBM vs preventive vs predictive maintenance
Condition-based maintenance differs from other strategies in what triggers the work.
| Reactive | Preventive (time-based) | |
|---|---|---|
| Work is triggered by | A breakdown | A fixed schedule or running hours |
| Data needed | None | Manufacturer intervals |
| Planning window | None | Long, but not linked to condition |
| Main risk | Costly unplanned downtime | Unnecessary work, and failures between services |
| Best for | Cheap, non-critical items | Simple assets with predictable wear |
Time-based maintenance (TBM) is another name for preventive maintenance. In practice, CBM and predictive maintenance overlap: predictive maintenance is condition-based maintenance that also forecasts how condition will change. Modern monitoring systems often do both, for example by showing the condition of a machine now and predicted one and three months ahead.
7 benefits of condition-based maintenance
1. Reduced unplanned downtime
Unplanned downtime now costs the world’s 500 largest companies around $1.4 trillion a year, 11% of their revenues, according to Siemens’ True Cost of Downtime 2024 report. Condition-based maintenance gives early warning of developing faults, so repairs can be planned for a convenient time with parts and people ready. McKinsey found analytics-based maintenance can cut machine downtime by 30 to 50%.
2. Lower maintenance costs
Time-based maintenance services equipment whether it needs it or not, and reactive maintenance pays the premium of emergency repairs. CBM avoids both by carrying out work only when the condition data shows it is needed. The US Department of Energy’s Operations and Maintenance Best Practices Guide estimates predictive, condition-driven maintenance saves 8 to 12% compared with preventive maintenance, and 30 to 40% compared with reactive maintenance.
3. Extended equipment lifespan
Catching faults early stops minor problems, such as slight misalignment or early bearing wear, from causing secondary damage to the rest of the machine. Fewer catastrophic failures and less unnecessary dismantling both reduce wear over time. McKinsey research found that analytics-based maintenance can extend machine life by 20 to 40%.
4. Improved safety
Equipment failures can put people at risk through overheating, fires, leaks or parts breaking loose. Condition monitoring identifies these hazards while they are still developing, so they can be dealt with under controlled conditions. Techniques that measure from the switchroom also mean fewer visits to hazardous or hard-to-reach machinery.
5. Better resource allocation
With CBM, maintenance teams focus their time on the machines that actually need attention, rather than working through routine tasks on healthy equipment. Planned work also means spare parts can be ordered in advance, reducing the need to hold large stocks or pay for urgent deliveries.
6. Higher operational efficiency
Machines in good condition run more smoothly and use less energy. Many developing faults, such as misalignment or a worn pump impeller, waste energy long before they cause a failure. Electrical monitoring can estimate how much energy each fault is wasting, giving a further, measurable reason to act early.
7. Stronger return on investment
The cost of monitoring is quickly outweighed by the failures it prevents. At one oil and gas facility, half an hour of testing with a portable condition monitoring device found a bearing and seal problem on a pump. It was fixed at low cost, avoiding a repeat of an incident conservatively estimated at £300k to £500k.
Condition-based maintenance in practice
- Borehole pump failure predicted. A water utility monitored six pumps, including two hard-to-reach borehole pumps. Monitoring flagged an impeller problem and predicted when the pump would need replacing. Left running as a test, it failed within a couple of days of the predicted date.
- £300k+ incident avoided in half an hour. After a lube oil pump failure caused an unplanned shutdown and flaring, an oil and gas facility tested its other pumps. A bearing and rubbing problem linked to a failing seal was found and fixed at low cost, avoiding a repeat incident conservatively estimated at £300k to £500k.
- Hidden corrosion repaired for a tenth of the cost. On an LNG tanker, monitoring found internal corrosion in a seawater pump that handheld vibration checks had missed. It was repaired for a few hundred dollars, less than a tenth of the cost of a replacement.
See more condition-based maintenance case studies.
Challenges of condition-based maintenance and how to overcome them
- Upfront cost. Fitting sensors to every machine is expensive. Start with your most critical assets, or use methods that measure from the switchroom, so one installation covers the motor and everything it drives.
- Access to machines. Submerged, sealed or hazardous equipment is hard to fit sensors to. Electrical signature analysis avoids this, because measurements are taken at the motor starter.
- Data and skills. Raw condition data needs interpreting. Choose systems that diagnose faults automatically and recommend actions, so you are not reliant on a specialist analyst.
- False alarms. Poorly set thresholds erode trust and savings. McKinsey found a 10% false-positive rate wiped out the savings of one maintenance programme. Establish good baselines and review thresholds regularly.
- Gaps between checks. Periodic checks can miss faults that develop quickly. Use continuous monitoring on critical assets and portable checks on the rest.
How to implement condition-based maintenance
- Rank your assets by criticality. Consider the cost of failure, safety risk, production impact and lead time for spares.
- Choose the right techniques. Match techniques to the faults and assets that matter most, and consider how easy each machine is to access.
- Set baselines. Record healthy readings for each machine so changes can be detected.
- Define thresholds and workflows. Decide what each alert means and who acts on it, ideally linked to your maintenance management system.
- Pilot, then scale. Prove results on a few machines first. An on-site diagnostic survey or a P100 portable kit is a low-risk way to start, before moving critical assets to continuous monitoring with Inpod.
Is condition-based maintenance worth it?
For critical, high-value machinery, yes. Total maintenance cost is lowest when work is done only when needed: reactive maintenance carries the cost of breakdowns and lost production, while preventive maintenance spends money servicing healthy equipment. CBM sits at the lowest point between the two. For cheap, non-critical assets, simple preventive maintenance or even run-to-failure may still make more sense, so most sites use a mix.
To see what condition-based maintenance could find on your plant, talk to Faraday Predictive or book an on-site diagnostic survey.
FAQs
Condition-based maintenance means fixing equipment when it shows signs of wear, rather than on a fixed schedule or after it breaks. Measurements such as vibration, temperature or motor current reveal developing faults, so repairs are planned only when they are actually needed.
A common example is a pump whose motor current shows a developing bearing fault. Instead of waiting for a breakdown, the bearing is replaced during a planned stop. At one oil and gas site, finding this kind of fault early avoided an incident estimated at £300k to £500k.
Time-based maintenance (TBM) services equipment at fixed intervals, whether it needs it or not. Condition-based maintenance (CBM) services equipment only when measurements show its condition is declining. CBM avoids unnecessary work on healthy machines and catches faults that develop between scheduled services.
Condition-based maintenance acts when measured condition crosses a threshold, while predictive maintenance also forecasts when a failure is likely to happen. Predictive maintenance is a more advanced form of CBM. Many modern monitoring systems do both, showing condition now and predicted weeks or months ahead.
The five main techniques are vibration analysis, thermography, oil analysis, ultrasound and electrical signature analysis. Electrical signature analysis, including motor current signature analysis, stands out because it monitors the motor and its driven machine from the switchroom, with no sensors fitted to the machine.
Costs depend on the technique and how many machines you monitor. Portable checks need no installation, and permanent systems need hardware per machine or panel. As examples, Faraday Predictive’s on-site diagnostic survey costs £1045 per day plus expenses, and its Inpod permanent monitoring system starts from £3,250.
For critical, high-value machinery, it usually is. The US Department of Energy estimates condition-driven, predictive maintenance saves 8 to 12% over preventive maintenance and 30 to 40% over reactive maintenance. For cheap, non-critical assets, simple preventive maintenance may still be more cost-effective.
Condition-based maintenance is used wherever unplanned downtime is costly, including oil and gas, water utilities, power generation, chemicals and process, pharmaceutical, manufacturing, marine and transport. It is most valuable on critical rotating machinery such as motors, pumps, fans and compressors.
Start by ranking assets by how critical they are, then choose techniques that suit them and record healthy baseline readings. Pilot on a few machines before scaling up. A one-day on-site survey is a low-cost way to see what condition monitoring finds on your own equipment.