Plant maintenance is the work required to keep manufacturing equipment running safely and reliably. It covers four approaches: breakdown, preventive, condition-based and predictive maintenance. Plants can combine these approaches based on equipment criticality, failure consequences and operating conditions. Performance is commonly tracked using MTBF, MTTR, PM compliance rate and breakdown frequency.
- No single strategy fits a whole plant. The judgement is which equipment goes into which category, not which approach is best overall.
- Downtime cost sits across five departments. Production, HR, logistics, quality and commercial each absorb part of it. Nobody adds it up.
- A few measures make maintenance performance visible. MTBF, MTTR, PM compliance rate and breakdown frequency are the common starting set.
- A utility failure can affect more than one line. Utilities often get less attention than production equipment, yet a single failure can stop several areas at once.
- Technology does not fix a broken process. AI and condition monitoring need maintenance records, working schedules and people available to act on alerts.
- Skilled technicians are often the constraint. Equipment is usually easier to buy than to staff. Finding and retaining people who can maintain it is a common difficulty.
Every plant has a maintenance strategy. Most were never chosen.
Over time, maintenance practices can accumulate without a clear strategy behind them. A schedule may cover some equipment, while other machines rely on breakdown response, stalled monitoring pilots or unreviewed OEM contracts.
That is not a strategy. It is a set of habits.
This guide covers the four maintenance approaches, how to decide which equipment gets which, what to measure, and where plants typically get stuck.
What this guide covers
The four maintenance approaches
Every maintenance activity falls into one of four categories. The difference is when the work happens.
| Approach | Work happens | Cost pattern | Best for |
|---|---|---|---|
| Breakdown | After failure | Low until it fails | Low-value, quickly replaced equipment |
| Preventive | On a fixed schedule | Steady and predictable | Critical production equipment |
| Condition-based | When equipment condition indicates maintenance is needed | Medium, needs monitoring setup | High-value rotating equipment |
| Predictive | When data indicates a developing failure risk | High setup, lower running | Critical assets with failure history |
Breakdown maintenance
Repair after failure. It comes in two forms — run-to-failure, which is a deliberate choice on cheap equipment, and emergency maintenance, which is an unplanned response to a critical stoppage.
Read the full guide to breakdown maintenance and when running to failure makes sense.
Preventive maintenance
Scheduled servicing before failure, based on running hours, production cycles or calendar time. Covers lubrication, inspection, part replacement, alignment and calibration.
Read the full guide to preventive maintenance in Indian manufacturing plants.
Condition-based maintenance
Work triggered by measured equipment condition rather than elapsed time. Vibration analysis, thermal imaging and oil analysis are the common methods. It solves the main weakness of preventive maintenance, which is that a calendar does not know how the machine is actually running.
Predictive maintenance
Condition data and historical trends are analysed to estimate developing failure risks. Continuous monitoring may be used where the equipment and application justify it. Requires maintenance history to learn from and people available to act on the alerts.
Read more on AI in plant maintenance and where implementations fail.
Choosing the right approach for each machine
This is the decision that matters. Three questions answer it.
1. What happens when it stops?
A press that halts the whole line ranks higher than a standby compressor. A machine with three identical siblings ranks lower than a single-source one.
2. How long does the repair take?
Equipment with long spare part lead times needs preventive attention even if the machine itself is not expensive. Waiting six weeks for an imported component is the real cost.
3. Does failure give warning?
Rotating equipment usually degrades gradually and shows up in vibration data. Electronic components often fail suddenly. The first suits condition monitoring. The second does not.
Rank machines using those three questions. Use the result to decide where preventive, condition-based, predictive or breakdown maintenance is appropriate.
What downtime actually costs
A breakdown costs much more than the repair bill. Lost output, idle workers, rush spares and late deliveries often cost more than the part that failed.
These costs sit in different departments, so they are rarely added up. To see the full list and work out the cost for your own plant, read our guide to breakdown maintenance.
A plant maintenance audit can also bring these numbers together.
What to measure
Track these measures regularly and judge the trend over time.
MTBF — Mean Time Between Failures
Average running time between breakdowns. Should rise as maintenance improves. The slowest metric to move because it needs enough running time to measure.
MTTR — Mean Time To Repair
Time from failure to production restart. Driven by spares availability, documented procedures and technician availability rather than by the maintenance schedule.
PM compliance rate
Percentage of scheduled tasks completed on time. A falling compliance rate can show that planned maintenance is not being completed as scheduled. Review the causes before changing the maintenance programme.
Breakdown frequency
Unplanned stoppages per month. Track the trend alongside the causes, equipment involved and maintenance actions.
OEE is worth adding where production performance is the concern rather than equipment reliability. It combines availability, performance and quality into one figure.
Utilities and support systems
Plant maintenance discussions often focus on production equipment. Utilities also need planned inspection and maintenance because failures can affect multiple production areas.
When a production machine fails, one line stops. A failure in a shared compressed-air system can affect multiple pneumatic machines.
Boilers, compressors, chillers, DG sets, water and effluent treatment, fire systems and electrical distribution all sit in this category. Several carry statutory inspection obligations.
Read the full guide to utility maintenance in manufacturing plants.
Where technology helps and where it does not
Condition monitoring and AI-based predictive systems work. The question is whether your plant is ready for them.
Three things need to be in place first.
Maintenance records
What failed, when, why and what was done. Predictive models learn from failure history. Without records there is nothing to train on.
Someone to act on alerts
The system is only useful when someone has the time, authority and process to respond to an alert.
A working preventive schedule
Predictive maintenance works alongside basic maintenance practices. If planned maintenance is not being executed, adding sensors will not solve the underlying process gap.
Start with the right level of monitoring. Vibration analysis and thermal imaging can support condition-based maintenance without requiring a full AI system. Plants can expand into predictive systems as data, processes and skills mature.
The workforce constraint
A maintenance schedule needs people to execute it. A condition monitoring alert needs a technician to act on it.
Reading a vibration trend, hearing early bearing noise, judging whether a belt has another month in it — these skills take years to build.
Finding and retaining people who can maintain modern equipment can be difficult, particularly across night shifts and weekends.
This is why plants use technical manpower outsourcing rather than carrying specialist headcount that is not needed full time.
Where plants get stuck
Trying to cover everything at once
Trying to schedule every machine at once can spread maintenance resources too thin. Start with critical equipment and expand once the process is working.
Production always wins the argument
When the line is running, maintenance windows get cancelled. Without management backing, planned maintenance windows can be repeatedly cancelled.
Findings are recorded but never closed
Audit observations without a named owner and a closure date reappear as fresh findings in the next audit. See our plant safety audit checklist for how closure tracking should work.
Repeat failures treated as bad luck
Repeated bearing failures should trigger a root-cause review. Misalignment, contamination, lubrication and installation practices are examples of factors worth checking.
Tasks written too vaguely
"Check the gearbox" gets interpreted differently by every technician. Specify what to check and against what standard, or it will be skipped.
Frequently asked questions
What is plant maintenance?
Plant maintenance is the work required to keep manufacturing equipment running safely and reliably. It covers breakdown, preventive, condition-based and predictive approaches.
Which maintenance strategy is best?
No single approach suits a whole plant. Most plants apply preventive schedules to critical equipment, condition monitoring to high-value rotating assets, and breakdown maintenance to low-value equipment.
How do I decide which equipment gets a preventive schedule?
Rank machines on what happens when they stop, how long repair takes including spare part lead times, and whether failure gives warning. Use that ranking to set maintenance priorities.
What KPIs measure maintenance performance?
MTBF, MTTR, PM compliance rate and breakdown frequency. OEE is worth adding where production performance rather than equipment reliability is the concern.
Why does maintenance spending get cut first?
Because the maintenance budget is visible and the cost of breakdowns is spread across production, HR, logistics, quality and commercial. Nobody adds the second number up.
Should plant maintenance be outsourced?
It depends on skill availability and how much your workload varies. Plants often outsource where specialist skills are needed occasionally, where shift cover is difficult, or where recruitment and compliance overhead outweighs the benefit of in-house headcount.
How long before a new maintenance programme shows results?
Breakdown frequency and MTBF should be reviewed as trends after the programme has been implemented consistently. MTBF needs enough running time and failure data to produce a meaningful trend.
Build a maintenance programme that holds
SGK India sets up and runs maintenance programmes for manufacturing plants across India.
Equipment criticality ranking, schedule design, execution, spares planning, breakdown response and monthly KPI reporting on MTBF and MTTR.
Explore our plant maintenance services