Statistical process control method and apparatus

Abstract

The invention relates to a statistical process control method and a statistical process control device applying the method. The method comprises the following steps: acquiring initial data to be analyzed during production, dividing the initial data into non-continuous distribution data, autocorrelation data, nonnormal distribution data and normal distribution data, determining the control limit according to different data types respectively, and monitoring and / or analyzing the production process according to the determined control limit. The statistical process control method simplifies classification of control charts through data classification, solves the problem of calculating the control limit of limited data points subjected to normal distribution and nonnormal distribution, realizes monitoring of a complex system with a large quantity of parameters which can not be realized in the prior art, saves the data processing time, and improves the work efficiency.

Description

technical field [0001] The invention relates to a statistical process control method and device. Background technique [0002] In the production department of an enterprise, a large number of products are produced or processed every day, and the performance and quality of the product are related to the lifeline of the enterprise. In order to ensure product quality, it is necessary to monitor, detect and analyze the production process in time. Statistical Process Control (SPC) is a tool for statistics and control of the production process by means of statistics. In the production process, statistical process control tools are widely used to collect and analyze the parameters in the production process, and provide them to engineers in the form of statistical signals, such as calculating the control boundary of the control chart for the sample data collected in the production process , make a control chart, histogram for analysis; or use the control circle in the control char...

AI Technical Summary

Problems solved by technology

This is because on the one hand, if the control boundary is set too close to the historical average value, false alarms will increase; on the other hand, if the control boundary is set far from the historical average value, there may be system instability without alarm Case

[0005] The Chinese patent application with the application number 200480037968.6 mentions that "extract time series from a multidimensional database, and use Statistical Process Control (SPC) techniques from the extracted time series to automatically calculate one or more values ​​for specified metrics in specified dimensions. control limits” and monitor data crossings according to the “one or more control limits”, but the technical solution described in this patent application is complex in data processing, and the calculation process is based on the traditional standardized calculation process. There are requirements for both nature and nature, that is, the data are required to obey continuous normal distribution, unimodality and no autocorrelation, and the number of data used to calculate the control boundary must be greater than thirty and the data does not contain Outlier, under the premise that the data meets the above conditions, the control boundary is three times the standard deviation of the monitored data

However, in the actual production process, many production process parameters do not meet the above conditions, so the results obtained by applying the traditional standard statistical process control to the calculation of the control circle are inaccurate

Similar to this, most of the existing software systems use traditional standard statistical process control for analysis, which cannot meet the actual production needs

Images