A, conventional control chart

For continuous data monitoring, statistical process control (SPC) is generally adopted in the control stage, and the commonly used method is control chart. Control chart is the most effective method to monitor the process performance, and its purpose is to monitor the process and quickly find abnormal fluctuations in the process. ..... People have the following understanding of the fluctuation of quality:

1, a process has many fluctuations at the same time, and it is inevitable;

2. Every kind of fluctuation is random: sometimes it is hidden, sometimes it is small, sometimes it is negative, so it is inevitable.

The predicted trend affects the output quality characteristics of the process;

3. It is normal for the quality characteristics to fluctuate, and "no fluctuation" is caused by the false phenomenon or low resolution of measuring instruments;

It is impossible to eliminate the fluctuation, but it is possible to reduce it.

The management and operation of any process is to limit the fluctuation within the allowable range, and to minimize the fluctuation beyond the allowable range or report it in time; Late reporting will cause losses, which is dereliction of duty.

In order to further understand the source of fluctuation and reduce it, people divide the causes of process fluctuation into two categories:

Common reasons and special reasons. Among them, accidental reasons, referred to as accidental reasons; Abnormal cause, also known as ascertainable cause, is called abnormal cause for short. The national standard "Conventional Control Chart" (GB/T 409 1-200 1) explains this in detail.

Second, the elements of control chart

Data points (in chronological order) can be continuous or discrete data. Each data point can represent a single measurement (such as the time taken to complete each contract) or a subgroup statistic of the sample (such as the average value of a group of products or the defective rate of each batch of goods).

The median line, CL) usually refers to the average value of the sample (however, the trend chart sometimes represents the median value instead of the average value).

Control limit [control limit, upper control limit (UCL) or lower control limit (LCL)] indicates that the data exceeds the limit of normal fluctuation. According to statistical convention, they are plotted at about 3 times the standard deviation of the average.