Hướng dẫn chi tiết sắp ra mắt
Chúng tôi đang chuẩn bị hướng dẫn giáo dục toàn diện cho OEE (Overall Equipment Effectiveness) Calculator. Quay lại sớm để xem giải thích từng bước, công thức, ví dụ thực tế và mẹo từ chuyên gia.
Overall Equipment Effectiveness (OEE) is the gold standard KPI for measuring manufacturing productivity. Developed within the Total Productive Maintenance (TPM) framework pioneered by Seiichi Nakajima at the Japan Institute of Plant Maintenance in the 1960s and 1970s, OEE quantifies what percentage of planned production time is truly productive — generating good-quality output at the designed rate without unplanned interruptions. OEE is expressed as the product of three independent factors: Availability (what fraction of planned time the machine was actually running), Performance (how fast it was running relative to its designed speed), and Quality (what fraction of output met specifications without defect). Each factor can range from 0% to 100%, and their product — OEE — represents the overall utilization of the machine's full productive potential. A machine running 90% of the time, at 95% of ideal speed, producing 99% good parts achieves an OEE of 84.6%. World-class OEE is defined as 85% or above — a benchmark that most manufacturers aspire to but few achieve consistently. Industry surveys suggest average manufacturing OEE runs between 55–65%. The gap between current OEE and world-class represents the maximum addressable improvement opportunity — the hidden factory of lost capacity that exists within every production environment. The power of OEE lies not just in the composite score but in its decomposition into the three factors. An Availability problem points to unplanned downtime and equipment reliability issues (MTBF, MTTR). A Performance problem points to speed losses, micro-stoppages, and idling. A Quality problem points to defects, rework, and startup scrap. This diagnostic clarity makes OEE a starting point for improvement rather than just a measurement — it directs maintenance, engineering, and quality teams toward the specific loss category with the greatest improvement potential.
OEE = Availability × Performance × Quality Availability: Availability = Run Time ÷ Planned Production Time Run Time = Planned Production Time − Downtime (unplanned stops, changeovers) Performance: Performance = (Ideal Cycle Time × Total Parts) ÷ Run Time Or equivalently: Actual Rate ÷ Ideal Rate Quality: Quality = Good Parts ÷ Total Parts Good Parts = Total Parts − Defective Parts (rejects + rework) OEE % = Availability × Performance × Quality × 100 Worked Example — Injection Moulding Press: Planned time: 480 min/shift | Downtime: 47 min | Ideal cycle time: 1.0 min/part Actual parts produced: 392 | Defective parts: 8 Availability = (480 − 47) ÷ 480 = 433 ÷ 480 = 90.2% Performance = (1.0 × 392) ÷ 433 = 392 ÷ 433 = 90.5% Quality = (392 − 8) ÷ 392 = 384 ÷ 392 = 98.0% OEE = 90.2% × 90.5% × 98.0% = 80.1%
- 1Define Planned Production Time — this is the shift time minus planned stops such as scheduled maintenance, breaks, and meetings. Planned stops are not counted as OEE losses; only unplanned losses reduce OEE.
- 2Record all Downtime events — every unplanned stop that exceeds a threshold (typically 5 minutes) is logged with its cause category: equipment failure, material shortage, operator unavailability, tooling change, etc. Sum all downtime to calculate Run Time.
- 3Calculate Availability: Run Time ÷ Planned Production Time. This captures the first category of OEE losses: Unplanned Downtime and Setup/Adjustment time.
- 4Determine the Ideal Cycle Time — the fastest theoretically possible cycle time under optimal conditions, typically derived from the machine's nameplate speed rating or design specification. Multiply by total parts produced and divide by Run Time to get Performance.
- 5Calculate Quality: Good Parts ÷ Total Parts Produced. Count all defective parts including scrapped parts, reworked parts (counted as defective even if they are recovered), and startup scrap during warmup.
- 6Multiply Availability × Performance × Quality to get OEE. Benchmark the result against world-class (85%), industry average (60–65%), and your own historical trend.
- 7Decompose losses into the Six Big Losses categories — Equipment Failure, Setup/Adjustment (Availability losses), Idling/Minor Stoppages, Reduced Speed (Performance losses), Process Defects, Reduced Yield (Quality losses) — and prioritize the largest loss category for improvement initiatives.
This line exceeds the 85% world-class benchmark. The primary improvement opportunity is Performance (94.2%) — investigating minor stoppages and speed reductions could push OEE above 90%.
Quality at 90% is the biggest drag on OEE — a 10% reject rate is unacceptable in pharma. The hidden cost: 820 rejected tablets per shift means wasted API, re-inspection labor, and batch documentation rework. Quality improvement is the top priority.
Low Availability (75%) driven by frequent product changeovers is the primary loss. SMED (Single-Minute Exchange of Die) methodology targets changeover time reduction — cutting changeover by 30 min would lift OEE to 79.3%.
Many manufacturers are surprised to find their OEE is 55–65%. This means 35–45% of planned production time is wasted. The hidden factory within this gap could produce millions of dollars in additional output with no capital investment.
Automotive Tier 1 suppliers report OEE by line to OEM customers as part of production capacity commitments and quality system audits — OEE below 75% triggers supplier improvement plans under IATF 16949 quality management requirements.
Pharmaceutical manufacturers use OEE as part of continuous process verification (CPV) under FDA 21 CFR Part 211, tracking equipment performance over time to validate that manufacturing processes remain in control and product quality is consistently achieved.
Consumer packaged goods (CPG) companies like Procter & Gamble and Unilever use OEE dashboards in manufacturing execution systems (MES) to drive daily production meetings and allocate engineering resources to the highest-impact improvement opportunities.
Industrial IoT (IIoT) platforms from Siemens, Rockwell, and GE Digital calculate OEE in real time from sensor data — machine uptime, cycle times, and quality sensor readings — feeding live OEE scores to plant managers' dashboards without manual data collection.
Bottleneck machines should be the primary focus of OEE measurement and
Bottleneck machines should be the primary focus of OEE measurement and improvement — Goldratt's Theory of Constraints teaches that throughput of an entire factory is determined by its bottleneck. Improving OEE on a non-bottleneck machine produces more WIP inventory but no additional factory output; improving OEE on the bottleneck directly increases revenue-generating throughput.
Multi-machine or production cell OEE requires deciding whether to measure OEE
Multi-machine or production cell OEE requires deciding whether to measure OEE at the individual machine level or the cell/line level. Cell OEE (also called Line OEE) multiplies individual machine OEEs along the production sequence and is always lower than any individual machine's OEE — one machine at 85% OEE in a 5-machine line limits cell OEE to below 85% regardless of other machines' performance.
High-mix, low-volume production (HMLV) environments with frequent product
High-mix, low-volume production (HMLV) environments with frequent product changeovers face a structural OEE challenge because Setup/Adjustment losses are unavoidably high. In HMLV, SMED (reducing changeover time from hours to minutes) is the highest-leverage OEE improvement methodology, and world-class benchmarks are adjusted to 75–80% rather than 85% to reflect the changeover-intensive environment.
| Industry | World-Class OEE | Industry Average OEE | Typical Biggest Loss Factor |
|---|---|---|---|
| Automotive assembly | 90%+ | 70–80% | Availability (changeovers, breakdowns) |
| Semiconductor fab | 88%+ | 65–75% | Performance (yield loss, recipe time) |
| Pharmaceutical | 85%+ | 60–70% | Quality (validation, reject rates) |
| Food & beverage | 85%+ | 55–65% | Availability (CIP cleaning, changeovers) |
| Printing/packaging | 80%+ | 60–70% | Performance (speed loss, minor stoppages) |
| Metal fabrication | 80%+ | 55–65% | Availability (setup, tooling changes) |
| Electronics assembly | 85%+ | 65–75% | Quality (solder defects, inspection fails) |
What is a good OEE score?
World-class OEE is defined as 85% or above — a benchmark established through TPM research at the Japan Institute of Plant Maintenance. This means 85% of all planned production time is being used to make good-quality parts at full speed. Industry surveys typically show average manufacturing OEE between 55–65%. An OEE of 65% is considered acceptable for many industries; 75%+ is good; 85%+ is world-class. Less than 50% indicates significant improvement opportunities requiring immediate attention.
What are the Six Big Losses in OEE?
The Six Big Losses framework categorizes all OEE losses: (1) Equipment Failure — unplanned breakdowns reducing Availability; (2) Setup and Adjustments — changeover time reducing Availability; (3) Idling and Minor Stoppages — brief stops under 5 minutes reducing Performance; (4) Reduced Speed — running below ideal rate reducing Performance; (5) Process Defects — defects and rework reducing Quality; (6) Reduced Yield — startup scrap until stable production reducing Quality. Each loss type has different root causes and different corrective tools.
What is the difference between OEE and TEEP?
OEE measures productivity against Planned Production Time — time the machine was scheduled to run. TEEP (Total Effective Equipment Performance) measures productivity against all calendar time (24/7/365), including planned downtime, weekends, and holidays. TEEP = OEE × Utilization, where Utilization = Planned Production Time ÷ Total Calendar Time. TEEP is used for capacity planning decisions; OEE is used for operational improvement. A machine running one shift per day might have OEE of 85% but TEEP of only 28% — highlighting available capacity on unscheduled shifts.
How do I calculate Ideal Cycle Time?
Ideal Cycle Time is the fastest theoretically possible time to produce one part — the machine's nameplate or design speed under optimal conditions. Sources include: equipment manufacturer's specification sheets, engineering time studies under optimal conditions, or the best observed cycle time over a sustained production run (best-of-best approach). Ideal Cycle Time should be the mathematical best case, not the average actual cycle time. Using actual average cycle time as the 'ideal' eliminates Performance losses from the OEE calculation and overstates true OEE.
Should planned maintenance be included in OEE downtime?
No — planned maintenance, scheduled changeovers, team meetings, and operator breaks are not included as OEE losses. OEE is calculated against Planned Production Time — the time the machine is scheduled to produce. Planned stops reduce the total production capacity but are not counted against OEE. However, if planned maintenance runs over schedule into production time, the overtime becomes unplanned downtime and does count against OEE Availability.
How is OEE used in lean manufacturing?
OEE is a cornerstone metric in lean manufacturing and Total Productive Maintenance (TPM). Lean teams use OEE as a baseline measurement before kaizen improvement events, then track OEE after improvements to validate results. OEE is posted on production boards and reviewed in daily production meetings (obeya rooms). The Six Big Losses framework connects OEE directly to lean improvement tools: SMED for Setup losses, TPM preventive maintenance for Equipment Failure, poka-yoke for Quality losses, and 5S for Minor Stoppages.
Can OEE be applied to non-manufacturing processes?
Yes — OEE principles have been extended to service operations and knowledge work. In a call center, Availability is agent uptime, Performance is calls handled vs. ideal handle time, and Quality is first-call resolution rate. In software development, OEE analogues measure sprint capacity utilization, velocity relative to potential, and rework-free deliveries. Healthcare uses OEE-style metrics for operating room utilization. The core framework adapts wherever there is a constrained resource with planned capacity and measurable output.
Mẹo Chuyên Nghiệp
Track OEE losses by shift, machine, and loss category, and share the results with operators on production floor displays updated in real-time or at least each shift. Research consistently shows that simply measuring and displaying OEE to operators — without any other intervention — improves OEE by 5–10% within 30 days through operator awareness and self-correction of avoidable losses.
Bạn có biết?
Toyota's assembly plants in Japan consistently achieve OEE above 90%, which is considered the practical upper limit for high-complexity automotive assembly. The Toyota Production System achieves this through relentless kaizen (continuous improvement), jidoka (automatic detection of abnormalities), and built-in quality rather than end-of-line inspection. By comparison, the average global automotive assembly plant runs at 70–75% OEE — meaning Toyota extracts 20–25% more output from the same equipment investment.