Random thoughts today: Double exhaustion from the rainy weather and work
Today, the sky is covered by dark clouds. Fine raindrops are like broken beads on a string, floating down in profusion, casting a wet veil of sorrow over the whole world. Such weather seems to carry an inherent oppressive aura, making people's moods involuntarily turn heavy. And I, in this gloomy and rainy atmosphere, feel even more exhausted both physically and mentally.
However, even when physically and mentally exhausted, the work tasks still lie ahead. Today, the main focus is on the selection of control charts, which is a crucial part of quality management. Below, I will elaborate on the methods for selecting control charts under different circumstances.
Clarify characteristics and data types
Before selecting a control chart, the primary task is to determine the characteristics for which the control chart is to be developed. This is like finding a navigational beacon in the vast ocean. We need to distinguish whether the data is measurement data or attribute data. Measurement data can usually take continuous values, such as length, weight, temperature, etc. They can accurately reflect the specific characteristics of products or processes. In contrast, attribute data is discrete and can only take integer values, such as the number of non - conforming products, the number of defects, etc. Only by accurately distinguishing the data type can we lay the foundation for the subsequent selection of an appropriate control chart.
Selection of control charts for attribute data
Pay attention to the situation of the non-conforming product rate
When we are concerned with the nonconforming product rate, that is, the percentage of "bad" parts, whether the sample size is constant or not will affect the selection of control charts. If the sample size is constant, we have two options, namely the p-chart and the np-chart. The p-chart is a control chart with the nonconforming product rate as the statistic, which can intuitively show the changes in the nonconforming product rate in different samples. The np-chart, on the other hand, uses the actual number of nonconforming products as the statistic and is more suitable for scenarios where people are more concerned about the specific number of nonconforming products. When the sample size is not constant, the p-chart becomes the first choice. Because the p-chart is not affected by the change in sample size and can stably monitor the fluctuations in the nonconforming product rate.
Pay attention to the situation of the number of non - conforming defects
If what we are concerned about is the number of non - conforming defects (defective products), that is, the number of non - conforming defects per unit part, we also need to consider the factor of sample size. When the sample size is constant, the u - chart and the c - chart can be selected. The u - chart uses the number of defects per unit product as the statistic, which can clearly reflect the average level of defects on each unit product. The c - chart uses the total number of defects in the sample as the statistic and is suitable for situations where one is more sensitive to changes in the total number of defects. When the sample size is not constant, the u - chart shows its advantage. It can effectively monitor the changes in the number of defects per unit product when the sample size changes.
Selection of control charts for measurement data
Situations where there is little difference in parameter values or the values are set in a destructive way
For measurement data, if the parameter values have very little variability, or the measurement is destructive, or it is impossible to obtain more values, the I - MR control chart is the best choice. The I - MR control chart, that is, the individual - moving range control chart. It only requires a single data point for analysis and is very suitable for situations where it is difficult to obtain multiple data. For example, in some scenarios where products are subjected to destructive tests, each test will damage the product and multiple measurements cannot be made. In this case, the I - MR control chart can play a good role in helping us monitor the stability of the process.
Situations with different sample sizes
The size of the sample capacity also affects the selection of control charts. If the sample capacity is greater than or equal to 10, the X - S chart is more appropriate. The X - S chart is a control chart with the sample mean and sample standard deviation as statistics. It can make full use of the information in the sample and monitor process changes more accurately. When the sample capacity is less than 10, the X - R control chart is more suitable. The X - R control chart uses the sample mean and sample range as statistics. It is relatively simple and easy to understand, and can quickly and effectively monitor the process for cases with a small sample capacity.
Special cases of manual calculation
Under normal circumstances, unless the manual calculation method is adopted, the median chart will be selected at this time. The median chart is a control chart with the sample median as the statistic. Its calculation is relatively simple and does not require complex mathematical operations. For some scenarios where there are no professional calculation tools and only manual calculation can be relied on, it is a practical choice.
On these rainy days, although I'm tired, through in - depth research on the selection method of control charts, I've made a little progress in my work. I hope this continuous rain will stop soon and let the sunshine drive away the physical and mental fatigue.