What Is a Control Chart?

Control charts are charts with control limits used to analyze and judge whether the process is in a stable state. They are functional charts that distinguish between normal fluctuations and abnormal fluctuations, and are often important statistical tools in field quality management. [1] Routine control charts include measurement value control charts (including single value control charts, average and range control charts, median and range control charts) and count value control charts (including non-conforming product number control charts, unqualified Product rate control chart, defect number control chart, unit defect number control chart, etc.). [2]

The world's first control chart is the academic leader of the process control group of the Bell Telephone Laboratory Quality Research Group
Control Chart (Control Chart)
One of the purposes of using control charts is to analyze and judge whether an abnormality occurs in the production process by observing the distribution of product quality characteristic values on the control chart. Once abnormalities are found, necessary measures should be taken to eliminate them in time to restore stability status. Control charts can also be used to bring the production process to a state of statistical control. The distribution of product quality characteristics is a statistical distribution. Therefore, drawing control charts requires the application of relevant theories and knowledge of probability theory.
According to the use of control charts, control charts can be divided into: control charts for analysis and control charts.
According to the type of statistical data, control charts can be divided into: metering control charts and counting control charts (including piece count control charts and point control charts). They are suitable for different production processes. Each category can be subdivided into specific control charts, which initially consist of seven basic charts. [2]
1.Identify key processes
The formation of a product quality requires many processes (processes), some of which play a vital role in the quality of the product. Such a process is called a key process. The SPC control chart should be used for the key process first, not all of them. Procedure. Therefore, the first step in implementing SPC is to identify key processes.
Then, analyze and study the key processes to identify the structure of the processes (inputs, outputs, resources, activities, etc.).
2. Determine process key variables (characteristics)
Analyze key processes (cause and effect diagrams, permutations, etc.) to find the variables (characteristics) that have the greatest impact on product quality.
3. Develop process control plans and specifications
This step is often the most difficult and time-consuming, and some experimental methods can be used to refer to the relevant standards.
4. Collection and arrangement of process data
5. Initial analysis of process controlled state
Analyze control charts to analyze whether the process is controlled and stable. If special reasons are found for uncontrolled or deteriorating, measures should be taken.
Note: At this time, the distribution center (X) of the process, the mean difference , and the chart limit may be unknown.
6. Process capability analysis
Only when the process is controlled and stable is it necessary to analyze the process capability. When insufficient process capability is found, measures should be taken.
7.Control chart monitoring
Only when the process is controlled and stable and the process capacity is sufficient can the control chart for monitoring be used to enter the SPC implementation phase.
8.Monitoring, diagnosis and improvement
In the monitoring process, when abnormalities are found, the reasons should be analyzed in time to take measures to restore the process to normal. For the controlled and stable process, we must also continuously improve, reduce common causes of deterioration, improve quality and reduce costs.
In the production process, the product quality is deteriorated due to the influence of random factors and system factors; the former is caused by the superposition of a large number of small accidental factors, and the latter is caused by identifiable and obvious causes. Appropriate measures can be found and eliminated. When a production process is only affected by random factors, so that the average value and variation of the quality characteristics of the product basically remain stable, it is said to be in a controlled state. At this time, the quality characteristics of the product are subject to determination
When you want to predict the range of change in process output;
When you judge whether a process is stable (in a statistically controlled state);
When you analyze whether the source of process variation is random or non-random;
When you decide how to complete a quality improvement project-prevent special problems or make fundamental changes to the process;
When you want to control the current process, you can detect and take remedial action when problems arise.
How does the control chart implement the precautionary principle? This can be seen from the following two points:
(1) Application of control charts to continuous monitoring of the production process. When abnormal factors emerge, they can be found in time even before the non-conforming product is caused. Measures are taken to eliminate the non-conforming product before this trend. To preventive effects.
(2) At the scene, it is more often that the control chart shows an abnormality, indicating that the cause of the abnormality has occurred. At this time, it is necessary to implement "detect the cause of the abnormality, take measures to ensure elimination, no longer appear, and include the standard." Otherwise, control The picture is nothing, so it's better not to do it. Each implementation (ie, after such a cycle) eliminates an abnormal factor, so that it no longer appears, thus playing a preventive role.
To get the exact value of the population, you need to collect the value of each sample of the population. This for a
Commonly used measurement value control charts are: average and range control chart (-R chart) median and range control chart (-R chart) and so on. Among them, the -R chart is most used. It has a strong control ability on the processing process and is the most practical and effective tool for controlling product quality. [2]
Steady state is the goal pursued by the production process. So how to use the control chart to determine whether the process is in a steady state? To this end, criteria for determining steady state need to be developed.
Criteria for stability: In the case of randomly arranged ideas, the process is considered to be stable when one of the following points is met:
(1) 25 consecutive ideas are within the control limit;
(2) Up to 1 pip in 35 consecutive pips falls outside the control limit;
(3) A maximum of 2 consecutive 100 ideas fall outside the control limit. When discussing the principle of control charts, you already know that an idea is out of bounds to determine anomalies. This is the most basic criterion for determining anomalies. In order to increase the confidence of users of control charts, it is necessary to observe whether the ideas are randomly arranged even within the control limits. If the arrangement of the points within the bounds is not random, the judgment is abnormal.
Criteria for judging anomalies: Abnormal factors are considered to exist in the process if one of the following points is met:
(1) The idea is outside or just above the control limit;
(2) The arrangement of ideas within the control limit is not random;
(3) Chain: continuous chain, 9 consecutive points are arranged below or above the center line; discontinuous chain, most of the points are on one side
(4) Most points are repeatedly close to the control limit (between 2 and 3 times
The following issues need to be considered when applying control charts:
(1) Where is the control chart used? In principle, control charts can be applied to any process where quality control is required. However, it is also required here that: for the identified control object-the quality index should be quantifiable so that the measurement value control chart can be applied. If there is only a qualitative description and not a quantifiable one, then only count value control charts can be applied. The process being controlled must be repetitive, i.e.
(Shewhart control chart)
There are different types of control charts, and the following are commonly used:
Applicable to the mean number control chart and the range R control chart that follow the measurement characteristics of the normal distribution. These two charts must be used together and are generally called -R control charts. Of which if used
1. The data of the control chart has a chronological order and must not be reversed, that is, it should be arranged and drawn in accordance with the order of acquisition (production), that is, a series of data is a characteristic that contains time series.
2. A series of ideas on the control chart must have fluctuations. This is caused by variation. The causes of variation are divided into two categories: one for opportunity, one for non-opportunity, and large fluctuations caused by non-opportunity. Larger should be avoided, that is, the greater the fluctuation, the more unstable the quality, and the wider the upper and lower limits of the control will be. At this time, the individual value distribution range of the points on the chart is also larger. If compared with the specifications, it is easier to escape. Defective products outside the specifications.
3. The control chart must contain statistical limits, that is, upper and lower control limits. The upper and lower control limits without statistics do not conform to the control chart principle. The upper limit of ± 3 is usually the abscissa is time (ie group), and the ordinate is the scale of quality.
Generally, the vertical axis of the control chart is set to the quality characteristics of the product, and the process change data is used as the scale.The horizontal axis is the group code or number or year, month, and day of the product to be tested, in terms of time or manufacturing order, in order. The points are drawn on the graph.
There are three straight horizontal lines on the control chart, the central one is the Central Line (CL), which is generally drawn with a solid blue line; the upper one is called the Upper Control Limit (UCL); The lower one is called the lower control limit (LCL). The drawing of the upper and lower control limits is generally represented by a red dashed line to indicate the acceptable range of variation; as for the dot line bars of actual product quality characteristics, most of them are drawn by solid black lines.
1. The production line worker or team leader first finds that the SPC control is abnormal; self-inspection, whether the operation is strictly in accordance with the operating standard (SOP or WI), cross-check by adjacent operators; if the situation is serious, or the cause cannot be found, the quality engineer and process must be notified immediately engineer.
2. After the quality engineers and process engineers have performed on-site analysis, can they find the cause of the abnormality in a short period of time (0.5 ~ 1 hour), and use 4M1E to analyze the process; if the root cause is still not found, and the situation is serious (for example, the P defect rate is large) (Exceeded the standard), report to the superior to decide whether to stop the line; the quality engineer convenes the relevant departments to meet to discuss the root cause (process, design, material or other).
3 After the cause of the abnormality of SPC was found and corrective and preventive measures were found and implemented, the SPC control chart shifted to the opposite direction of the abnormal control, indicating that the countermeasures were effective; normal production was resumed. This process must be closely monitored.
CPK is an important parameter reflecting the process capability:
What is CPK:
CPK: Abbreviation for ComplexProcess Capability index, is an indicator used by modern enterprises to indicate process capability. Only with strong process capability can it be possible to produce products of high quality and reliability. The process capability index is a method to indicate the level of the process, and its actual function is to reflect the level of process qualification. The research of process capability is to confirm the degree to which these characteristics meet the specifications to ensure that the yield of the finished product is above the required level, which can be used as a basis for continuous improvement of the process. The specifications are divided into unilateral specifications and bilateral specifications according to the upper and lower limits. A specification with only the upper specification limit and the specification center or only the lower specification limit and the specification center is called a unilateral specification. There are specifications with upper and lower limits and center values, and specifications with upper and lower limits and center values are called bilateral specifications. After our products pass the GageR & R test, we can start the test of Cpk value. A larger CPK value indicates better quality. Indicator description:
If CPK 1.33, it indicates that the process capability is good and needs to be maintained;
If 1.33CPK1, it shows that the process capability is average and needs to be improved.
If CPK1, it indicates that the process capability is poor, and improvement is urgently needed.

IN OTHER LANGUAGES

Was this article helpful? Thanks for the feedback Thanks for the feedback

How can we help? How can we help?