Total productive maintenance (TPM): eight pillars and a practical example

Total productive maintenance, or TPM, is an approach to preventing equipment losses through shared work across operations, maintenance, engineering and other functions. It connects daily care, technical maintenance and improvement over the equipment's life.

The conveyor stops again. Production calls maintenance, someone gets it running, and the shift catches up as best it can. The ticket closes. Two days later, the same fault returns.

The repeated callouts leave little time to deal with what keeps bringing the fault back.

The Lean Enterprise Institute's definition stresses participation beyond the maintenance team and attention to losses such as stops, slow running and defects. Fixing breakdowns is part of the work. TPM extends into the conditions that make them recur.

The eight pillars of TPM

JIPM's TPM checklist names eight pillars. The work spans operations, maintenance, engineering and support teams.

Pillar	Work it covers
Focused improvement	Investigate and reduce specific recurring losses
Autonomous maintenance	Trained operators carry out agreed basic care and report abnormal conditions
Planned maintenance	Prepare technical work, skills, parts and access
Quality maintenance	Maintain equipment conditions needed for conforming output
Education and training	Build the skills needed to operate, maintain and improve equipment
Safety, health and environment	Prevent harm through the relevant operating arrangements
Early management	Apply operating experience when developing equipment and processes
Office improvement	Improve the administrative and support work that production depends on

5S can support visible, orderly equipment care. It does not replace the other work. The conveyor example below follows several of these responsibilities through one recurring fault.

Autonomous maintenance vs planned maintenance

Autonomous maintenance means operators take responsibility for agreed equipment-care tasks they are trained and authorised to perform. They know the normal condition, notice a change and have a clear way to get help.

Planned maintenance is organised technical work based on equipment needs, history and condition. The plan provides the right skills, parts and access before the work is due. It can include time-based and condition-based tasks.

Engineering contributes when restoring the current condition is not enough, or when the design makes the same problem likely to return. JIPM's TPM material describes distinct capabilities for operators, maintenance staff and production engineers, including feeding experience into equipment design.

Work	Typical responsibility in our example	Evidence worth keeping
Routine observation and agreed basic care	Trained operator, supported by the shift lead	What changed, where and when; completed care task or raised abnormality
Technical inspection and restoration	Qualified maintenance staff, with planned access	Finding, work performed and condition checked before return to service
Recurrence and design review	Engineering with maintenance and operators	Proposed change, approval, trial conditions and later result
Time, support and follow-up	Production and maintenance leaders	Available window, named response owner and unresolved work

Operators need time for the agreed care tasks and technical help when they find a problem.

TPM example: a recurring conveyor fault

In our fictional packing area, the conveyor has repeated sensor-related stops. The operator can observe the fault indication from the normal working position and report it. Intrusive inspection and repair stay with qualified staff under the site's approved access and isolation arrangements.

Monday's operator records the fault code, location and operating conditions. The entry includes that the fault returned after an earlier restart. It does not assume the sensor itself is defective. A gemba walk can help the team observe the operating condition before choosing an explanation.

On Tuesday, maintenance and production agree a Wednesday access window, the assigned technician and the checks to perform. The planning matters because “maintenance will look at it” can otherwise remain open across several shifts.

During Wednesday's inspection, the technician finds a loose sensor mounting, restores it to the approved condition and completes the defined return-to-service checks. The finding goes into the same equipment history as Monday's observation.

That repair addresses the condition found. Thursday's review asks why the mounting became loose and whether the existing care and maintenance tasks would detect it early enough.

The operator's observation gave the technician a more specific starting point. The technical finding then gave engineering something to investigate. If each team kept a separate unlinked record, the next shift could start again with “conveyor stopped.”

How to make autonomous maintenance usable

“Check the machine” is too broad. An agreed task needs a location, a normal condition, a frequency and a response when something is outside that condition. It also needs enough time in the shift to be done properly.

For this conveyor, the observation task could identify the relevant fault display, the record to update and the person to call. Any cleaning, inspection or adjustment task needs its own approved method and training.

The standard work and SOP guide explains how repeatable work and its instructions fit together. Keep specialist work with the people authorised to do it; the word “autonomous” does not grant permission to repair unfamiliar equipment.

The jidoka and andon example follows how a call for help connects to a hold and a response. Leaders also need to close that response. If operators report the same abnormality for weeks without help, another reminder to fill in the sheet is unlikely to improve the equipment.

Vrolen's operational excellence guide covers that wider link between a reporting routine, time and decision authority. A fault report needs to reach someone with time and authority to act.

TPM measures: recurrence, MTBF, MTTR and OEE

For the conveyor, keep the original stop records and the restoration evidence. Engineering can then compare the recurrence pattern before and after an approved change under comparable operating conditions.

Record exposure too. A quiet week with little running is different from a full week without the fault. Product mix, load and other changes may affect the comparison. An empty fault log is only useful if people are still recording the events.

MTBF and MTTR separate observed time between failures from time to restore operation. OEE places equipment losses alongside speed and quality. Use the records underneath those measures to choose the work; an improving headline percentage cannot explain the cause on its own.

Start with the recurring loss and the missing responsibility

For this conveyor, the immediate repair is complete. The next actions are to finish the recurrence review, agree any change, check the care instructions and evaluate the result after enough relevant operation.

A TPM programme is broader than one conveyor. Starting with one recurring loss makes the responsibilities concrete: what the operator can do, what maintenance must plan, what engineering needs to change and what leaders must make possible.

Sources and example notes

• JIPM. TPM Excellence self-checklist, section2-1. The eight-pillar structure.
• Lean Enterprise Institute. Total productive maintenance. Participation, equipment losses and lifecycle care.
• Japan Institute of Plant Maintenance. Basics of TPM, printed page 57. Distinct skills and equipment improvement over its life.

The equipment, fault, finding and work plan are fictional. Tasks, frequencies, access and authorisations depend on the actual equipment and site. The example ends with follow-up planned, before a result has been measured.