SPC Statistical Process Control-Control Diagram

Continuous Data and discrete data

Continuous: continuous numerical values that can be meaningful and infinitely divided using measurements. (Time, length)
Discrete: class information, which can be counted but cannot be meaningful. (Qualified/unqualified)

A Control Chart consists of a central line, a control upper limit (UCL), and a control lower limit (LCL. Note the differences between the control limit and the specification limit. The control limit (UCL, LCL) is calculated based on the average value and calculated based on the three standard deviations of the Process center value +. That is, the control limit is calculated based on the sample data and is the internal feature of the process. The control limit is determined by the process capability. The usl (LSL) is determined by the execution standard and is an external feature of the process. Most specifications are about individual values, which are determined by the customer's requirements.

The fluctuations shown in the control chart are divided into fluctuations caused by general causes and those caused by special causes. The fluctuations caused by special causes are more likely to be discovered, for example, if the limit is exceeded or the lower limit is exceeded, or the control chart is usually used to determine the seven-point rule. The role of a control chart is to discover these exceptions and take corrective actions to analyze the root cause.

Use and selection of Control Charts
 

• Continuous grouped data: xbar-r control chart and xbar-s control chart.
• Continuous single value data: I-MR control chart.
• Number of nonconforming products that meet the two-item distribution discretely: (NP control diagram for equal group sample capacity, P Control Diagram for unequal distribution)
• Number of discrete defects that conform to Poisson distribution: (c control diagram for equal group sample capacity, and U control diagram for unequal use)

Xbar-r control diagram

The xbar (average value control chart) reflects the trend of variable X over time and the variability between group samples. Note that each vertex in the control chart is the average value of each group, and the center line of the control chart is the average value of the group. The R (range control chart) range control chart monitors the changes within the group sample over time. The center line of the graph represents the very short average of long grouped samples, or R. The r control chart is only suitable for scenarios with small sample size.

Xbar-s Control Chart

The xbar (average value control chart) reflects the trend of variable X over time and the variability between group samples. This is the same as the xbar-r control chart. The S control chart is the value standard deviation, and the standard deviation control chart monitors the changes within the group sample over time. TheThe center line of the chart represents the average value of the long-term standard deviation of the group sample. The standard deviation chart can be applied to any scenario where the group sample size (that is, n) is greater than 2. (To check whether the verification process is stable, 10 data values are sampled every day for a total of 10 days .)

I-MR (individuals and moving range) Control Chart

It mainly reflects the change of continuous single-value data over time. The application scope has fewer feature values in the process, and only one data can be obtained each time. The I-MR diagram is more susceptible to interference than the X Bar-r diagram due to the use of individual values. For example, we want to record whether the time of the vehicle to and from is controlled, we can record a series of continuous data values for I-MR control chart analysis.

P and NP graphs (number of unqualified discrete data)

The P chart is a count-type control chart that draws the nonconforming product rate for each sample. Each group sample can have the same or different sample sizes. This graph has the strongest versatility and is the most widely used in counting control charts. The p Diagram usually requires a large sample size. The better the quality, the larger the group sample is needed to detect a process out of control. (Records the number of welding points and the number of poor solder joints on a daily basis. The number of welding points can be different for group samples on a daily basis .) NP: A count-type control chart that draws the number of nonconforming items in each group sample. Each group of samples must have the same sample size or each sample size is similar enough to be considered equal.

Diagram C and diagram U (number of defects in discrete data)

Figure C is a count-type control chart that draws the number of defects (non-conformances) in each sample ). When all samples have the same sample size, graph C is a very practical choice. U chart: A count-type control chart that draws the average number of defects per unit in each sample, that is, the number of defects per unit when the number of samples changes. Pay attention to the difference between the number of unqualified items and the number of defects. The number of nonconforming items is for the samples themselves (either qualified or unqualified); the number of defects is for the samples, describes the conformity of the sample, rather than the sample itself.