TPMTotal Productive Maintenance

TPM is a structured production-system framework — not a maintenance department initiative. The word "Total" carries three meanings simultaneously: total participation (every employee from operator to plant manager), total equipment effectiveness (OEE as the unifying metric) and total life-cycle (from design through commissioning, operation, and end-of-life).

Nakajima's central insight was that maintenance technicians cannot prevent breakdowns alone — operators, who are physically with the equipment all shift, must own basic cleaning, lubrication, inspection and tightening (the four CLIT activities). Maintenance technicians are then freed to do skilled planned + predictive work instead of constantly reacting. The result is the operational shift from reactive → preventive → predictive → proactive maintenance.

TPM defines six recurring equipment loss categories that decompose directly into the three OEE components. Eliminating them is the entire operational goal of the framework.

OEE = Availability × Performance × Quality, with each multiplier directly attacked by 2 of the 6 Big Losses. World-class TPM plants reach 85%+ OEE sustainably (the JIPM TPM Excellence threshold); typical untreated plants run 40-60%. The gap is almost entirely the 6 Big Losses.

Regulated environments (pharma, medical devices, supplements, food, cosmetics, radiopharm) impose specific equipment requirements that TPM aligns with cleanly — but with adaptations.

1. 21 CFR 211.67 + EU GMP Ch.3 §3.34 require written equipment-cleaning + maintenance procedures with execution records — TPM's planned-maintenance pillar produces exactly this artefact when properly instrumented.
2. 21 CFR 820.70(g) + 820.72 require documented maintenance + calibration for production + inspection equipment — TPM's maintenance backbone covers this if records are Part-11 / Annex-11 compliant.
3. ICH Q9(R1) Quality Risk Management informs maintenance prioritisation — failure-mode + risk-criticality drive PM frequency; not every asset gets the same treatment.
4. Operator-performed autonomous maintenance must stay within trained-task boundaries — pillar 1 in regulated plants must explicitly delineate what operators CAN do (clean external surfaces, check lubrication levels, log condition observations) from what requires qualified technicians (calibration, hot work, GMP-impact maintenance).
5. Maintenance work orders that touch product-contact surfaces or change a validated state must trigger requalification per Annex 15 / cleaning validation — TPM execution cannot bypass change control.
6. Equipment criticality classification (direct-impact, indirect-impact, no-impact) per ISPE Baseline Guide drives TPM PM intensity + the required documentation rigour.

• OEE (Availability × Performance × Quality) per line, per shift, per product — the headline TPM metric; benchmarks: 40-60% untreated, 60-75% improving, 75-85% strong, 85%+ JIPM TPM Excellence level.
• MTBF (Mean Time Between Failures, ISO 22400-2 §6.5) — direct measure of breakdown elimination; trend upward as TPM matures.
• MTTR (Mean Time To Repair, ISO 22400-2 §6.7) — direct measure of maintenance organisation responsiveness; trend downward as planned + spare-parts strategy matures.
• PM compliance rate — fraction of scheduled PM completed on time; below 90% is a leading indicator of regression to reactive maintenance.
• Autonomous maintenance step-up — fraction of operators certified at each AM step (1: initial cleaning, 2: counter-measures to contamination sources, 3: cleaning + lubrication standards, 4: general inspection, 5: autonomous inspection, 6: workplace organisation, 7: full autonomous maintenance).
• Equipment loss tree (cumulative percentage by loss category) — focused-improvement target list; refresh quarterly + drive kaizen events against the top-3 losses.
• Near-miss + incident rate (SHE pillar) — leading safety indicator; trends with TPM maturity.
• Cost of maintenance per unit of output — the executive-visible ROI; falling cost-per-unit with rising OEE is the unambiguous TPM signal.

Mistake 1 — TPM as a maintenance-department initiative

If maintenance owns TPM alone, operators stay disengaged and the autonomous-maintenance pillar — the foundation — never lands. TPM is a production-system initiative; operations must own it equally.

Mistake 2 — skipping 5S

Autonomous maintenance in a cluttered, unmarked, dirty workplace is impossible. 5S is the visible foundation; without it, operators cannot tell normal from abnormal and CLIT becomes performative.

Mistake 3 — over-engineering the AM standards

200-step autonomous-maintenance checklists drive immediate operator rejection. Start with 5-10 critical checks per shift, prove the value, then expand. AM is a 7-step maturity model — respect the steps.

Mistake 4 — measuring OEE without acting on the loss tree

OEE that nobody investigates is a vanity number. The loss tree must drive the focused-improvement (kobetsu kaizen) backlog quarterly + kaizen events must close losses, not just describe them.

Mistake 5 — ignoring early-equipment management

New equipment that arrives without TPM input takes 6-18 months to reach stable OEE. Maintenance-prevention thinking at design + commissioning compresses this to weeks — and avoids inheriting the same chronic losses as the previous generation.

Mistake 6 — bypassing change control

In regulated plants, an operator-adjusted lubrication interval or a kaizen-modified guard is a change-control event. "It's just TPM" is not a defence in front of an FDA / Notified Body inspector. Every TPM-driven change must run the QMS gate.

V5 instruments TPM end-to-end: maintenance work orders, OEE telemetry, autonomous-maintenance tasks rendered on the kiosk, loss-tree analytics + the regulated-overlay audit trail — all wired into the QMS so TPM execution stays inside change control.

• Maintenance work orders — PM + corrective + predictive WOs as first-class records with skills-required + parts-required + downtime-impact; supervisor scheduler view by line + by skill.
• Autonomous maintenance kiosk tile — operator-owned CLIT tasks render per shift with one-tap completion + condition-observation capture + photo evidence; training-record hard-block on unqualified tasks.
• OEE + loss-tree dashboard — live OEE per line / shift / product + loss-tree decomposition into the 6 Big Losses; weekly kaizen-event candidate list auto-populated from top-3 losses.
• MTBF + MTTR streams — ISO 22400-2 compliant computation from the same event ledger as OEE; trend per asset + per failure mode.
• Predictive-maintenance hooks — historian-driven vibration / temperature / current / pressure thresholds firing PdM WOs before breakdown; PdM hit-rate tracked as a leading indicator.
• Spare-parts strategy — criticality-classified spares inventory with min-max per location; PM work orders auto-reserve required parts; WMS replenishment triggers automatic.
• Change-control integration — any maintenance WO touching a validated state requires change-control approval before execution; the Part-11 audit trail captures the gate.
• Skill matrix + training-record linkage — every WO checks operator + technician training currency against task requirements; out-of-date training hard-blocks execution.
• Equipment-criticality classification — direct-impact / indirect-impact / no-impact registry drives PM intensity + documentation rigour automatically per ISPE Baseline Guide.
• Mobile-safe — operator + technician + supervisor TPM views work at iPhone (≤390 px CSS width) with no horizontal scroll; floor + maintenance shop see the same WO + OEE state.

TPM vs PM — what's the difference?

PM (Preventive Maintenance) is one pillar of TPM. TPM is the whole framework — eight pillars, operator ownership, OEE measurement, focused improvement, early-equipment management. A PM programme without operator ownership + OEE + kaizen is not TPM; it is just scheduled maintenance.

How long does a TPM implementation take?

JIPM TPM Excellence Award typically takes 3-5 years of disciplined implementation; meaningful OEE improvements (10-20 percentage points) often appear in the first 12-18 months. Plants that try to compress this routinely fail; plants that commit see compounding returns for a decade.

Is TPM compatible with pharma / FDA / EU GMP?

Yes — and the maintenance-record artefacts TPM produces are directly what 21 CFR 211.67, EU GMP Chapter 3 + 21 CFR 820.70(g) require. The adaptation is that operator-performed autonomous-maintenance tasks must stay within trained-task boundaries + any change to a validated state must run change control. TPM execution must be inside the QMS.

What's the right starting point for a TPM rollout?

Pick one model line (typically a chronic-loss leader); deploy 5S; train operators on AM step 1 (initial cleaning + initial inspection); install OEE measurement + the loss tree; run focused-improvement kaizen against the top loss every 4-6 weeks; layer planned maintenance under it; declare success + replicate to the next line. Model line in 6-12 months; plant rollout in 2-3 years.

What's the difference between MTBF and MTTR?

MTBF (Mean Time Between Failures, ISO 22400-2 §6.5) measures reliability — how long an asset runs between failures, in hours. MTTR (Mean Time To Repair, §6.7) measures responsiveness — how long it takes to restore a failed asset, in hours. Availability = MTBF ÷ (MTBF + MTTR). TPM drives MTBF up + MTTR down simultaneously.

How does V5 support TPM?

Maintenance WOs as first-class records with skills + parts; autonomous-maintenance kiosk tile with CLIT tasks + training hard-block; OEE + loss-tree dashboard with kaizen-candidate auto-population; MTBF + MTTR streams; predictive-maintenance hooks from historian thresholds; criticality-classified spares with auto-reserve; change-control integration for validated-state changes; skill matrix linkage; equipment-criticality registry; Part 11 + Annex 11 audit trail throughout; mobile-safe on iPhone for floor + shop.

What OEE target should I aim for?

Benchmark by industry + process type. Continuous chemicals + sterile pharma: 85%+ is achievable + the JIPM Excellence floor. Discrete assembly: 80-85%. Pharma packaging: 65-75% is strong. Food filling: 70-80%. The number matters less than the trend + the loss-tree decomposition driving improvement; honest OEE that improves quarter-over-quarter beats stagnant inflated OEE.