六西格玛 is a quality management methodology aiming for near-perfection. A process operating at a “six sigma” level has a yield of 99.9997%, meaning it produces only 3.4 Defects Per Million Opportunities (DPMO). This standard corresponds to a process where the nearest specification limit is six standard deviations away from the process mean, accounting for a 1.5 sigma shift in the mean over time (refer to the specific details on the debated 1.5 sigma shift as an empirical observation).
Six Sigma Quality Standard (3.4 DPMO)
- Bill Smith
- Mikel Harry
The core idea of Six Sigma is that a process should be so well-controlled that its output consistently meets customer requirements. Statistically, this is defined by the 过程能力. A process is considered capable at a six sigma level if the distance from the process mean to the nearest specification limit is at least six times the process standard deviation ([latex]\sigma[/latex]).
However, the widely cited 3.4 DPMO figure is not derived from a pure six sigma distribution. It incorporates an empirical observation that process means can drift over time. Motorola engineers observed a long-term drift of about 1.5 sigma. Therefore, the calculation for 3.4 DPMO is based on the probability of a defect when the specification limit is [latex]6\sigma – 1.5\sigma = 4.5\sigma[/latex] away from the mean. This 1.5 sigma shift is a crucial, and sometimes controversial, component of the Six Sigma methodology, distinguishing it from a purely theoretical statistical 过程控制 model. It makes the standard more attainable in real-world applications where perfect process centering is difficult to maintain over long periods.
Important note: while this target is a very “near perfection” target, one should not forget that in industries where safety is crucial AND that produces products or services with high volumes, the defect (DPMO to improve) CAN arrive, thus needing potential combined and counter mesures to lower even more or suppress completely the consequences.
类型
中断
使用方法
前体
- Statistical Process Control (统计过程控制) by Walter A. Shewhart
- Total Quality Management (TQM) principles
- the concept of the normal distribution by Carl Friedrich Gauss
- process capability studies
- quality circles movement in Japan
应用
专利:
迎接新挑战
机械工程师、项目、工艺工程师或研发经理
可在短时间内接受新的挑战。
通过 LinkedIn 联系我
塑料金属电子集成、成本设计、GMP、人体工程学、中高容量设备和耗材、精益制造、受监管行业、CE 和 FDA、CAD、Solidworks、精益西格玛黑带、医疗 ISO 13485
相关发明、创新和技术原理