[中英文关键工艺参数的定义指南]关键的定义

Guidance in defining critical process parameters

关键参数的定义指南

The criticality of a process parameter is an assessment of the probability that it can be consistently and reproducibly controlled within the proven acceptable range (PAR) during routine manufacturing.

工艺参数的关键性是指在生产中对可接受范围的持续重复控制的可能性的评估，

This probability depends on

可能性主要取决于：

the robustness of the process parameter (the width of the PAR)

工艺参数的稳定性（可接受范围的宽度）

 the ability to technically control the parameter (technical limitations)

参数控制的技术能力（技术限制）

 the uncertainty of the measurement of the parameter (reflected by the standard deviation σ)

参数测量的不确定性（用标准偏差来反映） 

The robustness of a process parameter is reflected by the width of the proven acceptable range. The wider the range within which a parameter could be varied during process qualification without impacting product quality, the more robust it is.

可接受的参数范围反映工艺参数的稳定性。在参数确认中不影响产品质量的可改变的参数的范围越宽，工艺越稳定。

The ability to technically control a parameter is a function of a combination of process properties and equipment capabilities. For example, a highly exothermic reaction might be well controlled in a small stainless steel vessel, but could prove impossible to control in a larger glass-lined vessel.

参数的技术控制能力是工艺特性和设备能力的综合功能。例如，一个高放热的反应可能在一个小的不锈钢反应釜内能够很好的被控制，但是在一个大的塘玻璃反应釜内证明是不可能控制的。 The uncertainty of a measurement (σ) is the combined uncertainty of the calipation of the probe, the variance of the probe itself, and the variance of the signal transmission from the probe to the distributed control system (DCS). If the measurement is normally distributed, 3.4 out of a million data points will be outside a range of +/- 4.5 standard deviations. This is the basis of the ‘six sigma’ theory. The difference of 1.5 between 4.5 and 6 is an empirical value chosen based on the observation that the mean of real processes tends to drift by this value over time. Based on the ‘6σ’-concept we define a parameter as ‘critical’ when the proven acceptable range is narrower than ± six standard deviations.

测量的不确定性是探头或检测器的校验，监测器本身的变化和从探头到DCS的信号传输的变化的不确定因素的综合。如果测量结果是正态分布的，3.4PPM（百万分之一）将在+/-4.5标准偏差的范围之外。这是六西格玛理论的基础。4.5和6之间1.5的差异是一个经验值基于观察实际工艺趋势相对时间数值的漂移。基于六西格玛概念当一个参数的可接受的范围比+/-六个标准偏差窄时，我们定义这个参数为关键参数。

figure 1: critical parameter

setpointfigure 2: non-critical parametersetpoint

Whereas a proven acceptable range depends only on the process, the variance of a control system is equipment-specific. The variance of a control system for the same process parameter might be different in another operational environment, which could result in a different criticality assessment (e.g. a critical process parameter in one plant may be judged to be non-critical in another plant which has a control system with a lower variance).

如果一个可接受的参数范围仅仅依赖于工艺，控制系统的变异是设备属性。对同样的工艺参数的控制系统的变异可能在另一个操作环境下不同，这个可能导致一个不同的关键性评估（比如在一个工厂的一个关键参数在另一个有较低变异的控制系统的工厂就可能被判定为不关键）

The following table can be used for guidance to assess the criticality of a process parameter 以下列表在指南中用于评定一个工艺参数的关键性

1) The pH is typically much more difficult to be controlled than to be measured. Unless a system is in place which significantly reduces the control uncertainties (local pH, re-adjustment, etc.) a PAR of +/-1.0 pH units is recommended.

pH是典型的控制比检测要难得多。如果一个适当的系统能够显著的降低控制的不确定性（在线pH，重新校正等）推荐使用+/-1.0pH的可接受的范围。

2) Taking into account the difficulties of controlling a pressure under changing process conditions as observed during e.g. drying, distillation operation and reactions under pressure, a PAR of +/-20% is recommended.

考虑到在干燥、蒸馏操作和压力下反应中观察到的工艺条件改变时压力控制的难度。推荐使用+/-20%的可接受的参数范围

Examples:

例子：

1）A reaction temperature is likely to be not critical if the proven acceptable range (PAR) for a set point of 50°C is determined as 42°- 58°C for a calipated temperature range of 200°C (e.g. -20°C to +180°C).

如果一个可接受参数范围是温度范围是42~58°C，设定点是50°C，而校验过的温度为200°C（比如-20°C~180°C）,那么一个反应温度或许就是不关键的。

2）A pH has to be adjusted to 5.0 (PAR: 4.4 – 5.6) by charging acid. The final pH is relevant; an overcharge of acid can be corrected by charging base. This parameter is well controlled and not critical.

一个pH必须通过加酸调节至5.0（范围：4.4~5.6），最终pH是相应的。加过量的酸可以通过加碱来校正。那么这个参数可以被很好的控制并不关键。

3）During a reaction an acid is released and the pH has to be controlled between 6.4 and 7.6 by charging base. This parameter is likely to be critical as the local pH might differ significantly from the value measured and as the charge of base has to be controlled.

在一个反应中酸已经加了，pH必须通过加碱控制在6.4~7.6，因为现场的pH与检测的值可能有显著不同，碱的加入必须控制，所以这个参数可能是关键的。

4）A distillation is to be carried out under a pressure of

一个蒸馏在压力小于10KPa的条件下进行。这个设定点应该定义为8KPa。这个参数可能就是不关键的。

5）A hydrogenation has to be carried out at 0.8 Mpa, it can be run between 0.6 and 1 MPa; below 0.6MPa the reaction stops and more than 1MPa could not be tested. This parameter is very likely to be not critical.

一个氢化反应必须在0.8Mpa下进行。它可能在在0.6~1MPa下进行。低于0.6MPa反应停止，大于1MPa没有测试过。那么这个参数非常可能就是不关键的。

6）A reaction has to be carried out at a pressure of 5 bar; at 4 bar and at 6 bar different side products form. The edge of failure has to be determined and the parameter is to be regarded as critical.

一个反应必须在0.5MPa下进行，0.4MPa和0.6MPa是不同的两边产品形式。失败的边缘必须被定义，这个参数考虑是关键的。

6个西格玛＝3.4失误/百万机会―意味着卓越的管理，强大的竞争力和忠诚的客户 5个西格玛＝230失误/百万机会－优秀的管理、很强的竞争力和比较忠诚的客户 4个西格玛＝6,210失误/百万机会－意味着较好的管理和运营能力，满意的客户 3个西格玛＝66,800失误/百万机会－意味着平平常常的管理，缺乏竞争力

2个西格玛＝308,000失误/百万机会－意味着企业资源每天都有三分之一的浪费 1个西格玛＝690,000失误/百万机会－每天有三分之二的事情做错的企业无法生存