Among the statistical methods, there are junior and senior points, in SPC generally only involves primary statistical techniques, that is, involving only one factor or independent of each other multiple factors, such as control charts, cause and effect diagrams and so on.
2.1 Keep the process under control;
2.2 ? Ensure that the output of the process is consistent and compliant;
2.3 ? One of the preventive measures by turning after-the-fact inspection into before-the-fact prevention.
3.1 It is used for the control of manufacturing process/process output;
3.2 It helps to identify the causes of process fluctuations in order to prepare relevant countermeasures.
4.1 Variation
? Meaning: In the manufacturing process, the measurement of the output of the process is measured one by one, and the fluctuation of the measurement value is found, and the fluctuation of the quality characteristics of the individual product is called variance.
The causes of variation: *** two kinds, one is the common cause, is accidental, random causes. It is characterized by a) exists in any process; b) difficult to control with existing technology, such as tool wear caused by dimensional changes, the need for specialized tool wear compensation technology; c) the impact on the process is slight and uncertain, in the actual process, about 85% of more than belong to this situation, such as not, it is necessary to take measures in order to correct.
The second is a special cause, which is a systematic and non-random cause. It is characterized by a) sometimes appear, sometimes do not appear; b) does not necessarily exist in each process; c) a large impact on the process; d) can be controlled using existing technology.
In the actual process, about 15% belong to this situation.
4.2 Data
1) Metrological data, such as vernier calipers or micrometers measuring large data, where the data is continuous.
2) Counting-type data, such as a notification gauge measuring a threaded hole, where the data is intermittent.
5.1 Meaning
A control chart, one of the most commonly used tools in SPC, is a statistically designed chart that measures, records, and evaluates the values of a process quality characteristic to monitor whether the process is under control. Control charts are proposed by Hughart, also known as Hughart diagrams.
The graph is a right-angled coordinate plot with the horizontal coordinate being the time series or sequence of sample numbers and the vertical coordinate being the value of the sample statistic.
5.2 Control Chart Uses
a) It is used to help determine if there is a special cause for the process. If there is, the cause is analyzed to find the cause and solve the problem, and this process is repeated until the process is under control.
b) When the process is under control, the control chart is used to maintain it, and when there is a special cause, we can find it on the control chart.
c) It reduces over-control or under-control.
Process capability index Cp or Cpk
Cp=(USL-LSL)/6σ=T/6σ
where T is the tolerance, USL is the upper specification limit, LSL is the lower specification limit, and σ is the standard deviation of the process characteristic value in the controlled state;
Use M to denote the center of the specification, M=(USL-LSL)/2
< p> Denote the center of distribution by u, the process mean.With an offset, M and u are not equal, and the process capability is denoted by Cpk
Cpk=(1-k)Cp=(T-2 |M-u |)/(6σ)
Where k=2 |M-u |/T, which represents the relative offset of the mean from the center of the specification, and is also known as the offset coefficient.
Process performance index Pp or Ppk
Process performance is from the perspective of the total process fluctuations in the process to reflect the actual ability to process, it does not need to consider whether the process is under control, and does not require that the quality characteristics of the process output must be obeyed by a certain distribution of positive too.
The process is inconvenient and bi-directionally toleranced:
Pp=(USL-LSL)/6σ=｜T｜/6σ
The process is biased and bi-directionally toleranced,
Pp=min(USL-u)/3σ, (u-LSL)/3σ
When using unilateral bias:
Ppu=(USL-u) /3σ=(USL-X mean)/3σ
Ppl=(u-LSL)/3σ=(X mean-LSL)/3σ
Ppk=min(Ppu, Ppl)