A System for On-Line Monitoring of Mean and Standard Deviation

Abstract

A system for monitoring a mean value and a standard deviation in an on-line way comprises an original data acquisition device, a data grouping processing unit, a plotting processing unit and a front end displayer all connected in sequence. The system adopts a central limit theorem, the statistical inference and a box plot tool to realize that the content which needs to be monitored by a traditional Xbar-S / Xbar-R control chart on two charts can be monitored by the system on one chart, and make up for the defect that a pre-control chart cannot monitor wave (S or R). As the box plot not only can identify an abnormal value quickly, but also can enable an IQR (Inter Quartile Range) of the box plot to be less affected or not to be affected by the abnormal value, the box plot is more suitable for the quality control of tobacco production at the aspect of identifying the abnormal value and expressing the procedure wave. In addition, the system has the advantage that a confidence coefficient combination can be configured flexibly to meet the requirements of different control stringencies. The traditional control chart is not endowed with the advantage.

Description

technical field [0001] The system can be applied to online real-time monitoring and analysis of single variable (or single quality characteristic) in the cigarette production process, and can also be applied to other similar production processes outside the tobacco industry. Background technique [0002] At present, domestic cigarette production is a way of coexistence of large-scale production, multi-variety and small-batch production, and it is becoming more and more popular. This change makes the theoretical basis of normality and large samples on which the traditional control chart is based deviates. Moreover, under the condition that cigarette production is becoming more and more personalized, how to implement the essence of classical quality control technology has important and far-reaching theoretical significance for improving the stability of cigarette quality and improving the core processing capabilities of enterprises. [0003] Due to the limited number of batche...

AI Technical Summary

Problems solved by technology

[0003] Due to the limited number of batches of multi-variety and small-batch cigarette brands produced under the same or similar production conditions, if the traditional SPC method is directly used, it is sometimes difficult to limit the monitoring object to processing quality characteristics or equipment process parameters. To ensure the sample size required for statistics, sometimes it is difficult to find the cause of abnormalities, and sometimes variables or quality characteristics do not conform to the normal distribution

Moreover, the cigarette manufacturers in Shanghai, Hunan, Hubei, Jiangsu and other provinces, as well as our own practice of applying the traditional SPC method, have found that it is not suitable for quality control in the current cigarette production environment.

Therefore, cigarette manufacturers in Hunan, Hubei, Henan, Zhejiang and other provinces began to introduce the pre-control map (commonly known as the rainbow map) into the cigarette production process, and achieved some success. However, the pre-control map technology only monitors and analyzes the single variable (or single-mass properties), the fluctuations of which are not monitored

Images