Quality control system based on multivariate exponential weighted moving average control chart

Abstract

The invention provides a quality control system based on a multivariate exponential weighted moving average control chart, and belongs to the technical field of quality management and statistical process control. The system comprises a quality data acquisition module, a data preprocessing module, a statistical process control module, an anomaly judgment module and a quality evaluation and improvement module, and is used for monitoring and controlling the production process of a product with a plurality of quality characteristics in real time. The method comprises the following steps of: acquiring quality data of a plurality of quality characteristics in a production process by a system, preprocessing the quality data, and constructing a multivariate quality characteristic data set; estimating a mean vector and a covariance matrix in a controlled state based on the historical quality data, performing weighted updating on the multivariate quality data by adopting a multivariate exponential weighted moving average method, calculating a corresponding MEWMA statistic and generating an MEWMA control chart; and comparing the MEWMA statistical magnitude with a preset control limit to judge whether the production process is in an out-of-control state or not, and outputting early warning information and a corresponding quality evaluation result and improvement suggestion when abnormality is detected. The method can comprehensively analyze related information among multivariate quality characteristics, improves the sensitivity and accuracy of anomaly detection in the production process, and effectively improves the quality control level of the production process in the manufacturing industry.

Description

Technical Field

[0001] This invention relates to the field of quality control and statistical process control technology, specifically to the analysis and management of multivariate quality data, and in particular to a quality control system based on a multivariate index-weighted moving average control chart, which is suitable for quality monitoring and anomaly detection in the production process of products with multiple quality characteristics. Background Technology

[0002] Quality control is a crucial component of quality management in manufacturing. It involves collecting and analyzing quality data during the production process to monitor and control the product quality formation process. Traditional quality control methods typically employ Statistical Process Control (SPC) techniques to monitor quality characteristics during production to determine whether the process is under control.

[0003] In existing statistical process control methods, univariate control charts are widely used for quality monitoring in production processes. However, in actual manufacturing processes, product quality is often determined by multiple interrelated quality characteristics. Univariate control charts are difficult to reflect the correlation between multiple quality characteristics, which can easily lead to delayed or misjudged anomalies, thereby reducing the effectiveness of quality control.

[0004] To address the issue of monitoring multiple quality characteristics, existing technologies have proposed multivariate statistical process control methods. However, these methods still have the following shortcomings in practical applications: Firstly, they lack sensitivity to minor changes in the production process, making it difficult to identify potential quality risks in a timely manner. Secondly, some multivariate control methods are computationally complex and have low visualization capabilities, which hinders production managers from making intuitive judgments and quick decisions regarding the production process status.

[0005] Exponentially weighted moving average (MEWMA) control charts have been increasingly adopted in quality control due to their high sensitivity to small process deviations. Furthermore, multivariate exponentially weighted moving average control charts can comprehensively consider multiple quality characteristics and their correlations, showing promising application prospects in multivariate quality monitoring. However, existing technologies lack systematic implementation schemes for multivariate exponentially weighted moving average control charts, particularly lacking quality control systems that integrate multivariate quality data acquisition, statistical analysis, control chart construction, anomaly detection, and quality improvement.

[0006] Therefore, there is an urgent need for a quality control system based on a multivariate index-weighted moving average control chart to effectively monitor multivariate quality characteristics, improve the timeliness and accuracy of anomaly detection in the production process, and thus enhance the quality control level of the manufacturing process. Summary of the Invention

[0007] To address the problems of lagging anomaly identification and insufficient consideration of multivariate correlation in the monitoring of multiple quality characteristics in existing quality control systems, this invention proposes a quality control system based on a multivariate index-weighted moving average control chart.

[0008] This invention constructs a statistical process control module with a multivariate index-weighted moving average statistical model as its core, enabling real-time monitoring and anomaly detection of multidimensional quality data. It also integrates multivariate quality data acquisition, statistical analysis, process control, and quality improvement functions into a unified system platform, thereby improving the efficiency and accuracy of quality control in multi-quality characteristic production processes.

[0009] To achieve the above objectives, the present invention provides the following technical solution: A quality control system based on a multivariate index-weighted moving average control chart is characterized by comprising a quality data acquisition module, a data preprocessing module, a statistical process control module, an anomaly detection module, and a quality assessment and improvement module.

[0010] Specifically, the statistical process control module constructs a MEWMA control chart based on the multivariate exponentially weighted moving average method for real-time monitoring and anomaly detection of multivariate quality data collected during production. After quality data collection is complete, the quality control system automatically initiates the statistical analysis process, calculates the corresponding exponentially weighted moving average statistic based on the collected quality data, and plots the results on the control chart. The horizontal axis of the control chart represents the sample number or time series, and the vertical axis represents the EWMA or MEWMA statistic. Simultaneously, the center line, upper control limit, and lower control limit are determined based on statistical theory.

[0011] The control chart can intuitively reflect the changing trends of quality characteristics during the production process and can be used to identify potential abnormal fluctuations, thereby helping production managers to grasp the operating status of the production process in a timely manner.

[0012] Specifically, the steps for constructing a MEWMA control chart include: (1) Parameter estimation: Based on historical quality data, estimate the mean vector of the production process under controlled conditions. Covariance Matrix .

[0013] (2) Selecting the smoothing constant : A value between 0 and 1 (commonly 0.1-0.3) is used to adjust the sensitivity of the statistic to new observation data.

[0014] (3) Initialize the EWMA vector: set the initial value to (If the data has been centralized) ).

[0015] (4) Update the EWMA vector: for each new quality observation The MEWMA vector is updated using an exponentially weighted method, and its calculation formula is as follows: (5) Calculate the control statistic: using the steady-state covariance matrix ,calculate Statistic: (6) Determine the control limits: Under steady state, Approximately follows the degree of freedom of The chi-square distribution of (number of variables) is used to determine the upper control limit based on the preset significance level. The upper control limit is set as follows: in significance level (e.g.) (Corresponding to the 99% control limit).

[0016] (7) Monitoring process: When the MEWMA statistic exceeds the control limit, The judgment process got out of control.

[0017] Specifically, when the user selects only a single quality characteristic, the statistical process control module generates a univariate EWMA control chart; when the user selects multiple quality characteristics, a multivariate exponentially weighted moving average control chart is generated. For the univariate case, the system calculates the EWMA statistic and corresponding control limits based on the smoothing coefficient and process variance; when the calculated lower control limit is less than zero, it is set to zero. For the multivariate case, the system approximates the upper control limit based on the chi-square distribution; when the dimension of the quality characteristic exceeds a preset range, linear extrapolation is used to determine the corresponding control limits.

[0018] Calculate the univariate EWMA vector: in, Let represent the EWMA vector of the t-th sample. This represents the observation value of the t-th sample. Represents the smoothing coefficient (default 0.2). This represents the EWMA vector of the previous sample.

[0019] Calculate the variance: in, It is the original process variance.

[0020] Calculate the standard deviation (subscript) (Used to uniquely identify a specific quality characteristic and avoid confusion between multiple characteristics) Calculate control limits: in, This represents the upper limit of control at time t. This represents the lower control limit at time t. Represents the center line.

[0021] When the user selects multiple quality characteristics, a MEWMA control chart is generated. If the user selects a custom chart, the user-provided control limits are used; otherwise, the upper control limit is calculated using a chi-square distribution approximation, and the corresponding degrees of freedom are obtained from a predefined chi-square distribution approximation table. The value of ). When the degrees of freedom Exceeding the table range (usually) When >10), based on The value is then extrapolated linearly. The linear extrapolation formula is: Calculate the deviation vector: in, It is the MEWMA statistic vector at the current moment (dimension 1). (i.e., the number of variables) It is the mean vector of the process when it is in a steady state (the dimension is also 1). ), that is, the initial value.

[0022] calculate Statistic: in, Represents time t Statistic, Represents a A dimensional sample mean vector, Represents a The target mean vector of dimension, represent The covariance matrix of dimension , This represents the scaling factor, which can be omitted or set to 1.

[0023] Specifically, the system includes a data storage module for storing quality data, user information, and related data generated during system operation. This data storage module supports historical tracing, statistical analysis, and result querying of quality data.

[0024] The present invention proposes a quality control system based on a multivariate exponential weighted moving average control chart, which has the following advantages compared with the prior art: 1. By introducing a multivariate index-weighted moving average control chart, the correlation between multiple quality characteristics is comprehensively considered, thereby improving the sensitivity and accuracy of anomaly detection; 2. Enable real-time monitoring and early warning of anomalies in the production process, reducing the lag in the discovery of quality problems; 3. Integrate quality data collection, statistical process control, quality assessment and quality improvement functions into a unified system platform to improve the systematization and automation of quality management; 4. It is applicable to quality control scenarios in manufacturing processes with multiple quality characteristics, and has strong engineering application value. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the system flow of a quality control system based on a multivariate index-weighted moving average control chart according to the present invention. Figure 2 This is a system overall functional module architecture diagram of a quality control system based on a multivariate index-weighted moving average control chart according to the present invention; Figure 3 This is a schematic diagram of the data entity relationships in a quality control system based on a multivariate index-weighted moving average control chart, as described in this invention. Figure 4 This invention relates to a bimetallic thermostat MEWMA control chart for a quality control system based on a multivariate exponential weighted moving average control chart. Figure 5 This is an interface diagram of the control chart module and the anomaly identification result module of a quality control system based on a multivariate index-weighted moving average control chart according to the present invention. Figure 6 This is a comprehensive analysis report interface diagram of a quality control system based on a multivariate index-weighted moving average control chart, according to the present invention. Detailed Implementation

[0026] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0027] It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0028] In this invention, unless otherwise stated, the directional terms such as "up" and "down" generally refer to the directions shown in the accompanying drawings, or to the vertical, perpendicular, or gravitational direction; similarly, for ease of understanding and description, "left" and "right" generally refer to the left and right shown in the accompanying drawings; "inner" and "outer" refer to the inner and outer contours of each component itself, but the above directional terms are not intended to limit this invention.

[0029] Reference Appendix Figure 1 This embodiment provides a quality control system based on a multivariate index-weighted moving average control chart, including a user management module, a quality data acquisition module, a data preprocessing module, a statistical process control module, an anomaly detection module, and a quality assessment and improvement module. During system operation, users first register and log in through the user management module, gaining system operation privileges upon successful login. Users can import historical quality data through the quality data acquisition module or upload quality data collected in real-time during production. The quality data includes measurement data corresponding to at least one quality characteristic; when the quality data contains multiple quality characteristics, a multivariate quality characteristic dataset is formed.

[0030] The data preprocessing module is used to organize and standardize the collected quality data, including but not limited to data format verification, preliminary screening of outliers, and data structure unification, so that it can be called by the subsequent statistical analysis module.

[0031] The statistical process control module analyzes the preprocessed quality data based on a multivariate exponentially weighted moving average method. The system automatically determines whether to use a univariate exponentially weighted moving average control chart or a multivariate exponentially weighted moving average control chart based on the number of quality characteristics selected by the user.

[0032] When only a single quality characteristic is selected, the system generates a univariate EWMA control chart; when multiple quality characteristics are selected, the system generates a MEWMA control chart. The statistical process control module updates the quality data exponentially according to a preset smoothing coefficient, calculates the corresponding EWMA or MEWMA statistics, and plots the statistics on the control chart in time series or production batch order.

[0033] The control chart sets a center line and corresponding upper and lower control limits, which are determined based on historical quality data and a statistical distribution model.

[0034] The anomaly detection module compares the statistics output by the statistical process control module with the control limits. When a statistic exceeds the control limits, the system automatically determines that the production process is in an abnormal state and triggers an early warning mechanism.

[0035] After an anomaly warning is triggered, the quality assessment and improvement module analyzes the quality data corresponding to the time of the anomaly, generates a quality assessment result, and outputs quality improvement suggestions related to the abnormal quality characteristics to assist production managers in making process adjustments and quality optimizations.

[0036] Reference Appendix Figure 2 The quality control system, through the coordinated operation of its various functional modules, enables the collection, analysis, anomaly detection, and output of quality improvement suggestions for diverse quality data. Figure 2 The overall functional module architecture of the system is shown in detail.

[0037] Reference Appendix Figure 3 The system uses a data entity relationship model to manage user information, quality data, statistical analysis results, and anomaly records in a unified manner, in order to support historical traceability of quality data, query of statistical results, and evaluation of the effectiveness of quality improvement.

[0038] In the statistical process control module, the system continuously monitors the production process based on the MEWMA control chart. When a statistical quantity is detected to exceed the control limit, an anomaly warning is automatically triggered, and the corresponding quality assessment results and improvement suggestions are output.

[0039] Reference Appendix Figure 4 Using product data from bimetallic thermostats, the quality control system was tested. In the statistical process control module, the bimetallic thermostat product was selected, and five quality characteristics were analyzed to generate a MEWMA control chart. Figure 4 The results show that all selected features are within control limits, there are no out-of-control points, and the production process is stable.

[0040] Reference Appendix Figure 5 In the statistical process control module of the quality control system, the generated MEWMA control chart has an anomaly identification result displayed on the right. The system determines whether there are out-of-control points in the sample data based on the anomaly criteria. If an out-of-control point appears in the control chart, it will be displayed in this section, along with the corresponding anomaly cause analysis. (See attached diagram.) Figure 5 As shown, one anomaly was detected: T 2 The statistics showed a continuous increase of 9 points from point 7 to 15, and an analysis of the reasons for the anomaly was given: it may be due to fluctuations in the production process, equipment calibration problems, or changes in raw material batches.

[0041] Reference Appendix Figure 6 The quality control system generates a comprehensive analysis report based on abnormal situations, including some basic information of the control chart, key findings and preliminary interpretations, and follow-up suggestions for user reference, so that users can further control and improve the product production process, achieve quality control of the product, and improve the overall product quality.

[0042] This invention provides a quality control system that integrates multi-dimensional quality data management and statistical process control. In the statistical process control module, MEWMA control charts are primarily used for data visualization and analysis, identifying outliers to pinpoint key issues in the production process. This system overcomes the limitations of single indicators, comprehensively considering multiple quality characteristics and process parameters to reduce production defects and resource waste, effectively improving the quality level of manufactured products.

[0043] It should be emphasized that although the technical solutions of this invention have been described in detail through multiple embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions in the embodiments without departing from the spirit of this invention. These modifications or substitutions will not change the core technical content and innovative points of this invention. Therefore, the scope of protection of this invention should be determined by the claims.

Claims

1. A quality control system based on a multivariate exponential weighted moving average control chart, characterized in that, It includes a quality data acquisition module, a data preprocessing module, a statistical process control module, an anomaly detection module, and a quality assessment and improvement module; The quality data acquisition module is used to collect quality data of multiple quality characteristics during the production process; The data preprocessing module is used to organize the collected quality data to form a multivariate quality characteristic dataset. The statistical process control module is used to update the multivariate quality data based on the multivariate quality characteristic dataset using the multivariate index weighted moving average method, calculate the MEWMA statistic, and construct the corresponding MEWMA control chart. The anomaly detection module is used to compare the MEWMA statistic with a preset control limit to determine whether an anomaly has occurred in the production process. The quality assessment and improvement module is used to output corresponding quality assessment results and quality improvement information when an abnormality is determined to occur in the production process.

2. The quality control system based on a multivariate exponential weighted moving average control chart according to claim 1, characterized in that, The statistical process control module estimates the mean vector and covariance matrix under controlled conditions based on historical quality data, which are used for the calculation of the MEWMA statistic.

3. A quality control system based on a multivariate exponential weighted moving average control chart according to claim 1, characterized in that, When the quality data contains only a single quality characteristic, the statistical process control module generates a univariate exponentially weighted moving average control chart; when the quality data contains multiple quality characteristics, the statistical process control module generates a multivariate exponentially weighted moving average control chart.

4. A quality control system based on a multivariate index-weighted moving average control chart according to claim 1, characterized in that, The control limit is determined based on a multivariate statistical distribution model, and the anomaly detection module determines that the production process is out of control when the MEWMA statistic exceeds the control limit.

5. A quality control system based on a multivariate exponential weighted moving average control chart according to claim 1, characterized in that, After determining that an abnormality has occurred in the production process, the quality assessment and improvement module generates an early warning message and outputs the quality analysis results and improvement suggestions corresponding to the abnormality.

6. A quality control system based on a multivariate exponential weighted moving average control chart according to claim 1, characterized in that, The statistical process control module updates the multivariate quality data with an index-weighted method based on a preset smoothing coefficient.

7. A quality control system based on a multivariate exponential weighted moving average control chart according to claim 1, characterized in that, The quality control system is used for product quality monitoring during the manufacturing process.

Images