| name | lean-manufacturing |
| description | Lean manufacturing — Toyota Production System (TPS), 8 wastes (DOWNTIME), value stream mapping (VSM), pull systems and kanban, one-piece flow vs. batch production, OEE (Overall Equipment Effectiveness), 5S methodology, SMED (Single Minute Exchange of Die), takt time calculation, standard work, poka-yoke (mistake-proofing), andon and jidoka, heijunka (production leveling), and lean metrics. |
| metadata | {"priority":7,"promptSignals":{"phrases":["lean manufacturing","Toyota Production System","value stream mapping","kanban","SMED","OEE"],"minScore":3}} |
Lean Manufacturing — Complete Skill
Toyota Production System (TPS) Foundations
Two Pillars
Jidoka (自働化 — Autonomation with human intelligence):
Equipment stops automatically when abnormal condition detected; separate human work from machine work
Andon cord: any worker can stop production line when defect found → swarm and fix root cause immediately
Never pass a defect downstream: detect at source, fix at source
Just-In-Time (JIT):
Produce only what is needed, when it is needed, in the amount needed
Pull system: downstream demand triggers upstream production; nothing produced speculatively
Required: reliable processes, short lead times, leveled production (heijunka), flexible workforce
The 8 Wastes (DOWNTIME)
D — Defects: rework, scrap, inspection (non-value-adding quality checks)
O — Overproduction: making more than customer needs; most severe waste (causes all others)
W — Waiting: idle time (machine, worker, material); demand exceeds supply locally
N — Non-utilized talent: employee skills/ideas not used; disengaged workforce
T — Transportation: moving material without adding value; layout-driven
I — Inventory: WIP, raw material, finished goods beyond flow requirements; hides problems
M — Motion: unnecessary human motion; ergonomic waste; searching for tools/parts
E — Extra-processing: doing more than customer requires; gold-plating; redundant operations
Priority: eliminate overproduction first (it drives inventory, transportation, waiting, defects)
Value Stream Mapping (VSM)
Mapping Methodology
VSM procedure:
- Select product family: group products with similar routings
- Draw current-state map: observe shop floor; don't use historical data; walk the flow
Symbols: process boxes, push arrows, pull/kanban loops, supplier, customer, data boxes
- Capture data per process: cycle time (C/T), changeover time (C/O), uptime %, batch size, WIP quantity
- Calculate current-state metrics:
- Total lead time (LT) = Σ(process times + WIP wait times) [days]
- Processing time (PT) = Σ(actual value-adding time) [minutes]
- Value-adding ratio = PT / LT × 100% [typical: 0.5–5%]
- Design future-state map: kaizen bursts showing improvements; apply lean principles
- Implement improvements; measure; iterate
VSM data box metrics:
C/T (cycle time): time from one part exit to next part exit
C/O (changeover time): time from last good piece to first good piece (next product)
Uptime: % of scheduled time equipment actually producing (not down for failure/setup)
FTQ (First Time Quality): % correct first pass; 100% − rework/scrap rate
Lead Time Calculation
Lead time (LT) formula:
LT = Σ(WIP_between_processes / demand_rate) + Σ(queue_time) + Σ(process_time)
WIP wait time = WIP_pieces / (production_rate downstream) [Little's Law: LT = N/λ where N = WIP, λ = throughput]
Example:
Process A C/T = 30 s; Process B C/T = 45 s; 200 pieces WIP between A and B; daily demand = 500 pcs
WIP wait time = 200 / (500/8h) = 200 / 62.5 = 3.2 hours
LT contribution from WIP = 3.2 h >> processing time → WIP dominates; reduce WIP
Takt Time
Definition and Calculation
Takt time:
Takt = Available production time / Customer demand rate [s/part or min/part]
Takt = T_avail / D [T_avail = net daily available time; D = daily customer demand]
Available time:
T_avail = (shifts × shift_length) − scheduled_breaks − planned_maintenance [net; not gross]
Example: 2 shifts × 8h = 16h; breaks = 2×60 min; T_avail = 14h = 50,400 s
Example:
Customer demand D = 840 parts/day; T_avail = 50,400 s
Takt = 50,400 / 840 = 60 s/part
Interpretation:
Cycle time = Takt: balanced; no overproduction, no waiting
C/T < Takt: operator has idle time → rebalance or reduce headcount
C/T > Takt: bottleneck; cannot meet demand → overtime, kaizen, or capacity addition
Constraint identification:
Operator with C/T > Takt = bottleneck → focus kaizen here first (Theory of Constraints alignment)
Pull Systems and Kanban
Kanban Card System
Kanban: visual signal authorizing production or movement of one container of parts
Two-bin kanban: part consumed from bin 1 → empty bin triggers replenishment → refilled bin returns
Card kanban: card attached to container; when container sent to production, card goes to upstream → authorization to produce
Number of kanban (kanban size calculation):
N_kanban = (D × L × (1 + SS)) / Q
Where: D = daily demand; L = lead time (days); SS = safety stock factor (0.1–0.2); Q = container quantity
Example:
D = 500 parts/day; L = 0.5 day; SS = 0.1; Q = 100 parts/container
N_kanban = (500 × 0.5 × 1.1) / 100 = 275 / 100 = 2.75 → 3 kanban cards
Electronic kanban: RFID, barcode, or ERP-triggered replenishment; same logic; no physical cards
OEE (Overall Equipment Effectiveness)
OEE Calculation
OEE formula:
OEE = Availability × Performance × Quality [expressed as % or decimal]
Availability: A = (Planned production time − Unplanned downtime) / Planned production time
Performance: P = (Actual output / Theoretical max output) = (Actual cycle time / Ideal cycle time)
Quality: Q = (Good parts / Total parts started) = First Pass Yield
Example:
Planned: 8h = 480 min; downtime = 60 min; A = 420/480 = 87.5%
Ideal cycle time = 1 min/part; actual avg = 1.2 min/part; P = 1.0/1.2 = 83.3%
Total parts = 350; defects = 10; Q = 340/350 = 97.1%
OEE = 0.875 × 0.833 × 0.971 = 70.8%
World-class OEE targets:
World class: OEE ≥ 85% (single machine); realistic improvement from 65% → 85% = 31% capacity gain without capex
Industry average: OEE = 60–65%
Six Big Losses (basis of OEE losses):
Availability losses: equipment failure, setup/changeover
Performance losses: minor stoppages, speed loss
Quality losses: startup defects, steady-state defects
SMED (Single Minute Exchange of Die)
Changeover Reduction
SMED goal: reduce changeover time to < 9 minutes (single-minute = one-digit minutes)
Actual target: reduce by 50% or to match requirements of flexible production
SMED 4-step process:
- Observe and record current changeover (video); include all activities
- Separate internal from external activities:
Internal: must be done with machine stopped (die removal, adjustment)
External: can be done while machine running (staging, preparation, transport)
- Convert internal to external: pre-set dies outside; pre-heat, pre-position, pre-verify
- Streamline remaining internal activities: standardize bolts, use quick clamps, eliminate adjustments (shim + set screw → fine-tune online)
Typical results: 50–90% reduction in changeover time achievable
Original 4-hour die change on stamping press → 19 minutes after SMED (classic Toyota example)
5S Methodology
5S: foundation for all lean improvement (before kaizen is possible):
Sort (Seiri 整理): remove unneeded items; red tag questionable items; clear workspace
Set in Order (Seiton 整頓): place for everything; shadow boards; clearly labeled storage
Shine (Seiso 清掃): clean to inspect; cleaning reveals abnormalities (leaks, wear, loose bolts)
Standardize (Seiketsu 清潔): create standards; visual management; checklists; routines
Sustain (Shitsuke 躾): audit system; leadership participation; make it culture
Visual management: color coding, floor tape, shadow boards, andon lights — communicate status instantly without searching
Heijunka (Production Leveling)
Heijunka: level production volume AND mix over scheduling period
Without heijunka: batches create demand spikes → WIP → demand amplification (bullwhip effect)
Heijunka box: scheduling tool; slots for time intervals; cards placed for each model/product → mixed production
Pitch: smallest unit of production scheduling time
Pitch = Takt × container_quantity [e.g., Takt = 60 s; container = 10 pcs → Pitch = 600 s = 10 min]
Every 10 min, pacemaker starts one container; production leveled in 10-min increments
Key Lean Metrics
| Metric | Formula | Target |
|---|
| OEE | A × P × Q | ≥ 85% |
| Value-adding ratio | PT/LT | Improve toward 20%+ |
| Takt time | T_avail/demand | Equals C/T at bottleneck |
| FTQ (First Time Quality) | Good/total | ≥ 99% |
| Changeover time | SMED result | < 10 min (SMED goal) |
| IPEI (Inventory Process Efficiency Index) | Value-added hours/total lead time | Improve continuously |
Standards and References
| Standard | Scope |
|---|
| SAE J4001 | Implementation of lean enterprise operations (automotive) |
| AS9100D | Aerospace quality management (incorporates lean principles) |
| ISO 9001:2015 | Process approach = lean systems thinking |
| Womack & Jones "Lean Thinking" (1996) | Foundational lean reference |
| Liker "The Toyota Way" (2004) | TPS principles; 14 principles |
| Shingo "Revolution in Manufacturing" (1985) | SMED methodology original |
Output
Provide: current state assessment (process mapped: yes/no; LT [days]; PT [hours]; value-adding ratio [%]; OEE [%]; 8 wastes identified: list top 3 by impact), takt time (T_avail [s]; demand D [parts/day]; Takt = T_avail/D [s/part]; compare to C/T at each station; bottleneck identified), kanban calculation (D; L; SS; Q; N_kanban = D×L×(1+SS)/Q; floor stock [#] per kanban×Q), OEE analysis (A [%]; P [%]; Q [%]; OEE [%]; six big losses itemized; priority loss for kaizen), SMED opportunity (current changeover time [min]; internal % vs. external %; quick-win conversions; target time after SMED [min]; savings in capacity [parts/day]), 5S audit (each S scored 1–5; weakest area; corrective actions), heijunka (production leveling plan: mix per pitch; pitch = Takt×container_qty [min]; pacemaker process identified), future state targets (LT reduction [%]; OEE target [%]; WIP reduction [%]; changeover time target [min]; implementation timeline), and applicable standard (SAE J4001 if automotive; AS9100D if aerospace; ISO 9001 as baseline).
