OEE (Hiệu suất thiết bị tổng thể): Tính toán, Hiểu và Cải thiện

OEE is calculated as the product of three factors: Availability x Performance x Quality. Each factor is expressed as a percentage, and the final OEE score is also a percentage. A world-class OEE score is 85%, but most plants operate between 40-60%. Understanding why your score is what it is matters far more than the number itself.

Availability measures the percentage of planned production time that the equipment is actually running. It accounts for unplanned downtime (breakdowns, material shortages) and planned downtime (changeovers, cleaning). The formula is: Availability = Run Time / Planned Production Time. If your machine is scheduled for 480 minutes but only runs for 400, your availability is 83.3%.

Performance measures how fast the equipment runs relative to its designed speed. It captures speed losses like slow cycles, minor stops, and idling. The formula is: Performance = (Ideal Cycle Time x Total Count) / Run Time. Quality measures the proportion of good parts out of total parts produced: Quality = Good Count / Total Count. Multiply all three and you have your OEE.

OEE losses fall into six categories, often called the Six Big Losses. Availability losses include equipment breakdowns and setup/adjustment time. Performance losses include idling and minor stops as well as reduced speed. Quality losses include process defects and reduced yield (startup losses). Each category requires different countermeasures.

Breakdowns are the most visible loss but rarely the largest. In many plants, minor stops and speed losses account for more lost production than dramatic failures. The insidious nature of these micro-losses is that they become normalized -- operators stop noticing 3-second hesitations or 10% speed reductions because they happen so often.

Tracking losses by category is essential for prioritization. Use Pareto analysis to identify which of the six losses consumes the most capacity. Then apply targeted countermeasures: autonomous maintenance for breakdowns, SMED for changeovers, root cause analysis for chronic minor stops, and standard work for speed optimization.

Starting OEE measurement does not require expensive software. A simple spreadsheet with columns for planned time, downtime, total count, defect count, and ideal cycle time is enough to begin. The most important thing is consistency: define your terms clearly, train your operators, and collect data at the same granularity every shift.

Define your planned production time carefully. Should planned breaks be included or excluded? What about scheduled maintenance? There is no single right answer, but you must be consistent and transparent. Most organizations exclude planned breaks and scheduled maintenance from planned production time, which gives a cleaner picture of equipment losses.

Ideal cycle time is often the most debated input. Use the design speed of the equipment, not the current actual speed. If a machine was designed to produce 100 parts per hour but currently runs at 80, the ideal cycle time should be based on 100. This ensures your OEE score accurately reflects the performance gap.

Quick wins in OEE improvement often come from addressing changeover times. SMED methodology can reduce changeover durations by 50-90% by converting internal setup tasks (done while the machine is stopped) to external tasks (done while the machine is running). Even a 15-minute reduction in changeover time, multiplied across dozens of changeovers per week, adds up to significant capacity recovery.

Autonomous maintenance empowers operators to perform basic care tasks -- cleaning, inspection, lubrication, tightening -- that prevent breakdowns before they occur. When operators own the condition of their equipment, mean time between failures increases dramatically. The key is structured training and clear standards for what operators should check and when.

For chronic speed and quality losses, root cause analysis is essential. Tools like 5-Why analysis, fishbone diagrams, and designed experiments help you move beyond symptoms to the underlying mechanisms causing losses. Often, the root cause is not where you expect it: a quality defect might trace back to a upstream material variation, not the machine itself.

World-class OEE is generally defined as 85% or higher, with availability at 90%, performance at 95%, and quality at 99.9%. These numbers serve as aspirational targets, but context matters enormously. A highly flexible job shop with frequent changeovers will have a different achievable OEE than a dedicated high-volume line.

Do not chase OEE as an end in itself. The purpose of OEE is to make losses visible and guide improvement priorities, not to generate a score for a management dashboard. An OEE of 60% that is thoroughly understood and actively improving is more valuable than an OEE of 80% that nobody can explain.

Use OEE trends rather than absolute values for performance management. A line that improves from 55% to 65% in six months has made tremendous progress, even though it is still below world-class. Celebrate the direction and velocity of improvement, and use the detailed loss analysis to keep setting meaningful targets.