Predictive maintenance estimates future deterioration or failure from patterns and models. A filter being changed because it reaches a condition threshold is a simpler condition-based decision, without requiring a prediction system.
The methods can sit together. Keep scheduled safety tasks while using inspection findings to decide whether extra cleaning, adjustment or replacement of a part is needed.
Choose an asset and a useful signal
Start with the fault you want to detect. Increasing vibration, filter pressure difference, unexpected temperature or visible wear may offer warning, but the chosen check must suit the equipment and way it deteriorates.
A changed reading is not always a fault. Equipment working harder may produce different temperatures or vibration. Record what it was doing and compare similar conditions before authorising repairs from one unexplained figure.
There must be time to respond to a warning. Where failure can happen suddenly without a detectable sign, other planned tasks, standby arrangements or replacement decisions may be needed.
Build a complete response process
- Identify the item, its usual operating state and previous history.
- Agree how it will be checked and who has the necessary competence.
- Use instructions and technical advice to define results needing attention.
- Choose who reviews a warning and arranges the next step.
- Keep the repair result and confirm whether condition became acceptable.
Give warnings an owner. Agree what happens when a reading is absent, a sensor fails or access prevents a check. Do not mark these gaps as satisfactory results.
Retain the tasks that still apply
Normal condition readings do not automatically remove legal inspections, manufacturer requirements or other duties. Get any proposed timetable change assessed against the equipment’s requirements and risks.
Read HSE's maintenance guidance for safe planning and competence. The maintenance procedures guidance is written for major-hazard plant, but illustrates why the chosen method needs to fit each item and operating context.
What makes the approach worthwhile?
Consider the effort of checking against the consequence of failure and your ability to act. A reliable periodic inspection may give enough warning; occasional checks may be insufficient where deterioration progresses quickly.
For industrial premises, production patterns and load affect the readings. For offices, condition findings can guide plant maintenance without treating every item as critical.
Common questions
Does condition-based maintenance require sensors?
No. Visual inspections or portable instruments can provide useful evidence. Choose according to the item, how it deteriorates and the warning time required.
Can it replace all planned maintenance?
It cannot replace everything. Keep the scheduled work and inspections that apply, using condition evidence only where it supports a justified change.
What should happen when a threshold is exceeded?
Follow the agreed response. The next step may be verifying a reading, investigating or commissioning work; simply clearing the warning does not explain the cause.
How do we know it is helping?
Look at faults detected, time to act, work finished and problems that return. More readings alone do not show that the maintenance decisions improved.
Related services and next steps
See Condition-Based Maintenance, Asset Maintenance, Mechanical and Planned Replacement Programmes for related work options.
Browse services, maintenance, projects, sectors and locations. To Discuss condition-based maintenance, describe the equipment, previous faults and decision you want the checks to inform.
