Production data statistical analysis method based on cloud edge collaboration, medium and terminal

Abstract

The invention is suitable for the technical field of electric power Internet of Things, and relates to a production data statistical analysis method based on cloud edge collaboration, a medium and a terminal, and the method comprises the steps: inputting production basic data at a cloud end; a production data set X is automatically collected from an MES system at regular time and marked to obtain a corresponding label data set Y and a data set D = (X, Y); the edge agent downloads the data set D from the cloud platform, executes an analysis task, and calculates and generates a baseline and a control chart for analysis; the edge agent calculates an area A, an area B and an area C, automatically judges the difference of the remaining data set D, calculates upper and lower limit information according to a base line, locally decides whether to report to the master station or not, generates and eliminates invalid data, and reports early warning data meeting the difference judgment condition and sample data to the cloud master station; and generating a control chart for production by using a weighting algorithm, and issuing new sample data to an edge agent to carry out difference judgment. The method is simple in process and convenient to operate, and improves the accuracy and efficiency of statistical analysis of production data.

Description

Technical Field

[0001] The present invention belongs to the technical field of the power Internet of Things, and particularly relates to a production data statistical analysis method, medium and terminal based on cloud-edge collaboration. Background Art

[0002] With the development of digital production, the statistical analysis and control of production data become more and more important. Statistical analysis of production data can help enterprises understand various indicators and problems in the production process, and can also reflect the production trend of the enterprise, providing a basis for the long-term planning of the enterprise. Traditional manual or using third-party software for statistical analysis of production data is time-consuming and laborious, and the statistical analysis accuracy and precision of common production data statistical methods such as mean, median, standard deviation, variance, range, etc. are difficult to be effectively guaranteed. Only by preprocessing the data near the object or data source side for massive production data can the network bandwidth pressure and the backend computing and storage pressure be effectively reduced, the overall analysis efficiency be improved, and the requirements such as real-time response of the service be met. Although cloud computing uses its super computing power to process and statistically analyze the data collected and uploaded from the device layer to the cloud platform in processing big data, the increase in the number of devices leads to a sharp increase in the amount of data generated by the devices. The growth rate of the network bandwidth far lags behind the growth rate of the data, making the network bandwidth a bottleneck, and the increasingly complex network environment makes the network latency problem more obvious, ultimately resulting in low statistical analysis accuracy.

[0003] The patent with the publication number CN108595896B provides an analysis method for material data in automotive panel stamping simulation. The method steps are as follows: (1) Determine the typical thickness and select the large production data of this thickness; the yield strength value for the simulation analysis is the average yield strength + A * σ, where the average yield strength is the arithmetic average of the large production data, the value range of A is 0.8 to 0.9, and σ is the standard deviation of the large production data; (2) Select all the data of the materials with the yield strength of the material in the range of the average + 0.68σ to 1.04σ; the tensile strength value and n value for the simulation analysis are both the average values of this batch of data, and the r value in a certain direction is the average value of the r in this direction of this batch of data; (3) Randomly select several batches of materials, select the materials with the yield strength above the average value among these materials, calculate the average values of the r in the other two directions of the selected materials, and the calculation result is the r value in the other two directions for the simulation analysis. In this patent, statistical analysis is carried out through the mean and standard deviation, and the accuracy cannot be guaranteed, and there are the same drawbacks as the existing technology.

[0004] Therefore, how to improve the accuracy and efficiency of production data statistical analysis is an urgent problem to be solved by those skilled in the art. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the object of the present invention is to provide a production data statistical analysis method based on cloud-edge collaboration to solve the problems of low accuracy and efficiency in the statistical analysis of production data in the prior art; in addition, the present invention also provides a production data statistical analysis medium and terminal based on cloud-edge collaboration.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides a production data statistical analysis method based on cloud-edge collaboration, including the following steps:

[0008] S10. Enter production basic data in the cloud;

[0009] S20. Regularly collect the production data set X from the MES system automatically, and perform marking to obtain the corresponding label data set Y and the data set D=(X, Y);

[0010] S30. The edge agent downloads the data set D from the cloud platform and executes the analysis task, calculates and generates a baseline and a control chart for analysis;

[0011] S40. The edge agent calculates regions A, B, and C according to the calculated center line, upper control limit, and lower control limit, automatically performs outlier judgment on the remaining data set D, and makes an in-situ decision on whether to report to the master station and generate and eliminate invalid data according to the baseline. When the number of data samples reaches a certain amount, the warning data and sample data that meet the outlier judgment conditions are reported to the cloud master station;

[0012] S50. The cloud master station generates a production control chart using the weighted algorithm for the sample data reported from the edge side, sends the new sample data to the edge agent for outlier judgment, and analyzes and processes the reported warning data and makes a decision on whether to report to OA or send an email notification.

[0013] Further, in the step S10, the production basic data includes product information, production line information, and process information.

[0014] Further, in the step S20, the production data set is defined as X={x 1 ,x 2 ,…,x N}, the label data set is Y={y 1 ,y 2 ,…,y N}, the data set D=(X, Y), the product test items and sub-items that need to be analyzed are set, as well as the outlier judgment criteria required, and the data set D=(X, Y) and the production basic data are uploaded to the cloud platform.

[0015] Further, the specific steps of step S30 are as follows:

[0016] S301. Calculate the standard deviation σ = sqrt(((x 1 - x)^2+(x 2 - x)^2+……(x n - x)^2) / (n - 1)), and the mean Xbar = (X 1 + X 2 +……X n ) / n;

[0017] S302. Calculate Ppk. The calculation formula of Ppk is as follows:

[0018] Ppk=(1 - k)*Pp=(USL - LSL - 2|M - μ|) / (3σ);

[0019] k = |M - μ| / (T / 2)=ε / (T / 2);

[0020] ε = |M - μ|;

[0021] Pp=(USL - LSL) / 6σ=T / 6σ;

[0022] T=(USL - LSL);

[0023] Among them, σ is the standard deviation, μ is the mean, M is the target value, USL is the upper specification limit, and LSL is the lower specification limit;

[0024] S303. Through the calculated center line, upper control limit, and lower control limit, perform stability judgment to obtain baseline data. The stability judgment criteria are as follows:

[0025] All 25 consecutive data points are within the control limits;

[0026] Among 35 consecutive data points, at most only one data point exceeds the control limit;

[0027] Among 100 consecutive data points, at most only two points exceed the control limit;

[0028] Generate a control chart for analysis through the baseline data and the Ppk.

[0029] Further, the out-of-control criteria are as follows:

[0030] Criterion 1 is that the K point falls outside area A; Criterion 2 is that consecutive K points fall on the same side of the center line; Criterion 3 is that consecutive K points increase or decrease; Criterion 4 is that the adjacent points of consecutive K points alternate up and down; Criterion 5 is that among consecutive K + 1 points, K points fall outside area B on the same side of the center line; Criterion 6 is that among consecutive K + 1 points, K points fall outside area C on the same side of the center line; Criterion 7 is that consecutive K points fall within area C; Criterion 8 is that consecutive K points fall on both sides of the center line, but none fall within area C.

[0031] Further, the control chart for production in step S50 is generated using the following weighted algorithm:

[0032] S501. Assume that the overall data X ~ N(μ, σ), then the distribution of X follows N(μ, σ / √n), and the distribution of R follows (k 2 σ, k 3 σ), where k 2 and k 3 are coefficients varying with n. Assume that the sample group number i ∈ [1, m], the jth sample observation value in the ith group is x ij , j ∈ [1, n], the sample mean of the ith group is The total sample mean is The sample range of the ith group is R i , and the average value of the ranges is Then:

[0033]

[0034] R i = x imax - x imin ;

[0035] S502. The definitions of the center line UCL, upper control limit CL, and lower control limit LCL of the chart are:

[0036]

[0037] Among them,

[0038] The definitions of the center line UCL, upper control limit CL, and lower control limit LCL of the R chart are:

[0039]

[0040] Among them,

[0041] S504. Weight the control chart. Assume that a weight value is given to each observation value, and the sum of all weight values is 1. If p k is the weight value of the kth observation value, qk If it is the k-th observed value, then the weighted i-th observed value is expressed as p k q k , according to the model prediction, the estimated value ^q of the (n + 1)-th point n+1 is represented by the linear combination of the weighted observed values of the first n points and the weighted predicted value of the first point:

[0042] ^q n+1 = λq n + λ(1 - λ)q n-1 +... + λ(1 - λ) z q n-z +... + λ(1 - λ) n-1 q 1 +(1 - λ) n ^q 1 ;

[0043] Among them, the weight of each observed value is:

[0044]

[0045] The weighted mean and weighted variance are respectively:

[0046]

[0047]

[0048] Then the weighted control limit is:

[0049]

[0050] The weighted R control limit is:

[0051]

[0052] In a second aspect, the present invention also provides a computer-readable storage medium, where the storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned method is implemented.

[0053] In a third aspect, the present invention also provides an electronic terminal, including: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the terminal executes the above-mentioned method.

[0054] Compared with the prior art, the production data statistical analysis method, medium and terminal based on cloud-edge collaboration provided by the present invention have at least the following beneficial effects:

[0055] Traditional manual or using third - party software for statistical analysis of production data is time - consuming and laborious, and it is difficult to effectively guarantee the accuracy and precision of common production data statistical methods such as mean, median, standard deviation, variance, range, etc. The process of the present invention is simple and the operation is convenient. Considering the computing power at the edge, by filtering redundant data at the edge and performing basic calculation and statistical analysis, the edge statistical analysis speed is improved. At the same time, data is collected at the edge and a weighted algorithm is continued to be used in the cloud to improve the statistical analysis speed. The cloud - edge collaboration solution can not only meet the requirements of real - time and security in specific scenarios, but also transmit high - quality statistical analysis data and analysis results to the cloud as needed, greatly reducing the redundant video data at the edge and reducing the storage energy consumption at the edge. The present invention adopts a weighted control algorithm, which improves the effective control rate of production data consistency, thereby improving product quality and production efficiency, and improving the accuracy and efficiency of data analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the solution of the present invention, the following will give a simple introduction to the drawings required for description in the embodiments. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0057] Figure 1 It is a flowchart of a method for statistical analysis of production data with cloud - edge collaboration provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0058] To facilitate the understanding of the present invention, the following will describe the present invention more comprehensively with reference to the relevant drawings. The preferred embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure content of the present invention more thorough and comprehensive.

[0059] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0060] The present invention provides a method for statistical analysis of production data based on cloud - edge collaboration, which is applied to the statistical analysis and control process of production data. The method for statistical analysis of production data based on cloud - edge collaboration includes the following steps:

[0061] S10. Enter the production basic data in the cloud; S20. Automatically collect the production data set X from the MES system at regular intervals, and perform marking to obtain the corresponding label data set Y and data set D=(X, Y); S30. The edge agent downloads the data set D from the cloud platform and executes the analysis task, calculates and generates the baseline and the control chart for analysis; S40. The edge agent calculates regions A, B, and C according to the calculated center line, upper control limit, and lower control limit, automatically performs outlier detection on the remaining data set D, and makes an on-site decision on whether to report to the master station and generate and eliminate invalid data according to the baseline calculation of the upper and lower limit information. When the number of data samples reaches a certain amount, report the warning data and sample data that meet the outlier detection conditions to the cloud master station; S50. The cloud master station generates a production control chart using the weighted algorithm for the sample data reported from the edge side, sends the new sample data to the edge agent for outlier detection, and at the same time analyzes and processes the reported warning data and makes a decision on whether to report to the OA or send an email notification.

[0062] The process of the present invention is simple and convenient to operate, improving the accuracy and efficiency of production data statistical analysis.

[0063] In order to enable those skilled in the art of this technology to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings.

[0064] The present invention provides a method for statistical analysis of production data based on cloud-edge collaboration, which is applied to the statistical analysis and control process of production data. In combination with Figure 1 to the drawings, in this embodiment, the method for statistical analysis of production data based on cloud-edge collaboration includes the following steps:

[0065] S10. Enter the production basic data in the cloud.

[0066] Specifically, the production basic data includes product information, production line information, and process information.

[0067] S20. Automatically collect the production data set X from the MES system at regular intervals, and perform marking to obtain the corresponding label data set Y and data set D=(X, Y).

[0068] Specifically, automatically collect the production data set X={x 1 ,x 2 ,…,x N} from the MES system every day at regular intervals, and perform marking on it to obtain the corresponding label data set Y={y 1 ,y 2 ,…,y N} and data set D=(X, Y), set the product test items and sub-items that need to be analyzed, as well as the required outlier detection criteria (8 major outlier detections), and upload the data set D=(X, Y) and production information to the cloud platform.

[0069] S30. The edge proxy downloads the dataset D from the cloud platform and executes the analysis task, calculates and generates the baseline and the control chart for analysis.

[0070] Specifically, the specific steps of step S30 are as follows:

[0071] S301. Calculate the standard deviation σ = sqrt((x 1 - x)2+(x 2 - x)2+……(x n - x)2) / (n - 1)), and the mean Xbar = (X 1 + X 2 +……X n ) / n;

[0072] S302. Calculate Ppk, and the calculation formula of Ppk is as follows:

[0073] Ppk = (1 - k)*Pp = (USL - LSL - 2|M - μ|) / (3σ);

[0074] k = |M - μ| / (T / 2) = ε / (T / 2);

[0075] ε = |M - μ|;

[0076] Pp = (USL - LSL) / 6σ = T / 6σ;

[0077] T = (USL - LSL);

[0078] Where, σ is the standard deviation, μ is the mean, M is the target value, USL is the upper specification limit, and LSL is the lower specification limit;

[0079] S303. Through the calculated center line, upper control limit and lower control limit, perform stability judgment to obtain the baseline data. The stability judgment criteria are as follows:

[0080] All 25 consecutive data points are within the control limits;

[0081] Among 35 consecutive data points, at most only one data point exceeds the control limit;

[0082] Among 100 consecutive data points, at most only two points exceed the control limit;

[0083] Generate the control chart for analysis through the said baseline data and the said Ppk.

[0084] S40. The edge proxy calculates three areas, namely A, B, and C, based on the calculated center line, upper control limit, and lower control limit. It automatically performs outlier detection on the remaining dataset D, makes an on-site decision on whether to report to the master station according to the upper and lower limit information calculated based on the baseline, and generates and eliminates invalid data. When the number of data samples reaches a certain amount, it reports the warning data and sample data that meet the outlier detection conditions to the cloud master station.

[0085] Specifically, the outlier detection criteria are as follows:

[0086] Criterion 1: The K point falls outside the A area;

[0087] Criterion 2: K consecutive points fall on the same side of the center line;

[0088] Criterion 3: K consecutive points are increasing or decreasing;

[0089] Criterion 4: The adjacent points of K consecutive points alternate up and down;

[0090] Criterion 5: Among K + 1 consecutive points, K points fall outside the B area on the same side of the center line;

[0091] Criterion 6: Among K + 1 consecutive points, K points fall outside the C area on the same side of the center line;

[0092] Criterion 7: K consecutive points fall within the C area;

[0093] Criterion 8: K consecutive points fall on both sides of the center line, but none of them fall within the C area.

[0094] S50. The cloud master station generates a production control chart from the sample data reported by the edge side using a weighted algorithm, sends the new sample data to the edge proxy for outlier detection, and analyzes and processes the reported warning data and makes a decision on whether to report to OA or send an email notification.

[0095] Specifically, the following weighted algorithm is used to generate the production control chart in step S50:

[0096] S501. Assume that the overall data X ~ N(μ, σ), then the distribution of X follows N(μ, σ / n), and the distribution of R follows (k 2 σ, k 3 σ), where k 2 , k 3 are coefficients that change with n. Assume that the number of sample groups i ∈ [1, m], the jth sample observation value of the ith group is x ij , j ∈ [1, n], the sample mean of the ith group is The total sample mean is The sample range of the ith group is R i , and the average range is Then:

[0097]

[0098]

[0099] R i = x imax -x imin ;

[0100] S502, The definitions of the center line UCL, upper control limit CL, and lower control limit LCL of the graph are as follows:

[0101]

[0102] Among them,

[0103] The definitions of the center line UCL, upper control limit CL, and lower control limit LCL of the S503, R graph are as follows:

[0104]

[0105] Among them,

[0106] S504, For The control chart is weighted. Let each observed value be given a weight value, and the sum of all weight values is 1. If p k is the weight value of the kth observed value, q k is the kth observed value, then the weighted ith observed value is expressed as p k q k , According to the model prediction, the estimated value ^q n+1 of the (n + 1)th point is represented by the linear combination of the weighted observed values of the first n points and the weighted predicted value of the first point:

[0107] ^q n+1 = λq n + λ(1 - λ)q n-1 +... + λ(1 - λ) z q n-z +... + λ(1 - λ) n-1 q 1 +(1 - λ) n ^q 1 ;

[0108] Among them, the weight of each observed value is:

[0109]

[0110] Also because:

[0111] λ + λ(1 - λ)+... + λ(1 - λ) z+...+ λ(1 - λ) n-1 +(1 - λ) n = 1;

[0112] It is explained that for any given λ ∈ [0, 1], a set of weights p k , k ∈ [1, n] can always be found to make the above hold;

[0113] The weighted mean and weighted variance are respectively:

[0114]

[0115] Then after weighting The control limits are:

[0116]

[0117] The weighted R control limits are:

[0118]

[0119] Furthermore, in this embodiment, the platform includes an edge management cloud platform running in the cloud and an edge steward running at the edge. The edge management cloud platform in the cloud is designed with microservices, and the edge steward is designed modularly. Through the edge management cloud platform with a Web interface operation in the cloud, all edge devices and the applications running therein are deployed, managed, updated, and monitored. The edge side connects to the edge management cloud platform through the edge steward, receives and processes the deployment, management, and update commands sent by the edge management cloud platform, and reports monitoring data, supporting the online upgrade of the operating system, kernel, patches, and the edge steward. The edge management platform separates the front end and the back end, enabling front-end personnel to focus on page development and back-end developers to focus on business logic. The back-end service is written in Java, the data is stored in MySQL, and Redis is used for data caching to accelerate data access speed. The cloud services are microserviced for easy subsequent expansion. The basic components at the edge are modularized and divided into an Edge Hub forwarding module, an EdgeAgent edge proxy module, an Edge Monitor monitoring module, a log module, a plug-and-play identity authentication module, a security proxy module, a container-internal APP monitoring service, an Edge Daemon deployment module, a local operation and maintenance program, a device management module, a blockchain docking service. The modules at the edge communicate with each other through Beehive and MQTT.

[0120] Furthermore, in this embodiment, the specific operation process of the cloud interface is as follows:

[0121] Basic data maintenance: Enter basic data; Menu: Product management, production line management, process management; Prerequisite: None.

[0122] MES Synchronization Data Setting: Set the data that needs to be collected from the MES system; Menu: MES Data Synchronization; Prerequisite: Product, production line, and process information have been well maintained.

[0123] Analysis Item Setting: Set the product test items and sub-items that need to be analyzed, as well as the out-of-control criteria required; Menu: Analysis Item Setting; Prerequisite: The corresponding product test items and sub-items have been collected from the MES system.

[0124] Baseline Update and Analysis Chart Viewing: Use the data collected from MES to automatically / manually trigger the calculation of control upper and lower limit information, and the generated baseline information and analysis control charts can be viewed; Menu: Analysis Control Chart; Prerequisite: The analysis chart data can only be seen after the baseline for the corresponding product test items and sub-items has been generated.

[0125] Automatic Alarm Judgment: Automatically judge the out-of-control of daily production data according to the set warning types; Menu: None (no manual operation required); Prerequisite: The analysis items have been set.

[0126] Control Chart Viewing: View the production process control charts xbar-s chart, p chart, and the production process control data of each day can be exported; Menu: Production Process Control; Prerequisite: The analysis items have been set, the MES data of the corresponding product test items and sub-items has been collected, and the alarm judgment has been performed.

[0127] Generation of Abnormality Viewing and Handling: View the abnormal information generated during the out-of-control judgment process; Menu: Production Abnormality Query; Prerequisite: The analysis items have been set, the MES data of the corresponding product test items and sub-items has been collected, and alarm information has been generated.

[0128] Process Capability Analysis: Used to calculate the order CPK; Menu: Process Capability Analysis.

[0129] Furthermore, in this embodiment, the system permissions are split into a four-layer structure: Web API -> Menu Page -> User Group -> User 4 layers; The interface API access and menu binding are solidified and hidden from the user without user operation; Multiple menus are combined into user groups to facilitate the assignment of access permissions to users; It is possible to specify which user group a user belongs to (a user can belong to multiple groups), and the user can then have all the menu access permissions under this user group.

[0130] The embodiment of the present invention also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements any one of the methods in this embodiment.

[0131] An embodiment of the present invention further provides an electronic terminal, including: a processor and a memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory so that the terminal executes any one of the methods in this embodiment.

[0132] For the computer-readable storage medium in this embodiment, those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by hardware related to computer programs. The foregoing computer program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps including the above method embodiments; and the foregoing storage medium includes: various media such as ROM, RAM, magnetic disk or optical disc that can store program codes.

[0133] The electronic terminal provided in this embodiment includes a processor, a memory, a transceiver, and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store a computer program, the communication interface is used for communication, and the processor and the transceiver are used to run the computer program so that the electronic terminal executes each step of the above method.

[0134] Compared with the prior art, for the method, medium and terminal for statistical analysis of production data based on cloud-edge collaboration described in the above embodiments, traditional manual or using third-party software to perform statistical analysis on production data is time-consuming and laborious, and it is difficult to effectively guarantee the statistical analysis accuracy and precision of common production data statistical methods such as mean, median, standard deviation, variance, range, etc. The process of the present invention is simple and the operation is convenient. Considering the computing power of the edge side, by filtering redundant data and performing basic calculation and statistical analysis on the edge side, the edge statistical analysis speed is improved. At the same time, data is collected on the edge side, and the weighted algorithm is continued to be used in the cloud to improve the statistical analysis speed. The cloud-edge collaboration solution can not only meet the requirements of real-time and security in specific scenarios, but also transmit high-quality statistical analysis data and analysis results to the cloud as needed, greatly reducing the redundant video data at the edge side and reducing the storage energy consumption at the edge side. The present invention adopts a weighted control algorithm to improve the effective control rate of production data consistency, thereby improving product quality and production efficiency and improving the accuracy and efficiency of data analysis.

[0135] Obviously, the embodiments described above are only the preferred embodiments of the present invention, rather than all embodiments. The preferred embodiments of the present invention are shown in the accompanying drawings, but they do not limit the patent scope of the present invention. The present invention can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosed content of the present invention more thorough and comprehensive. Although the present invention has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements on some of the technical features. Any equivalent structure that makes use of the content of the specification and drawings of the present invention, directly or indirectly applied in other related technical fields, is similarly within the scope of the patent protection of the present invention.

Claims

1. A production data statistical analysis method based on cloud-edge collaboration, characterized in that: The following steps are involved: S10. Input basic production data in the cloud; S20, automatically collecting the production data set X from the MES system at a fixed time, and marking it to obtain the corresponding label data set Y and data set D = (X, Y); S30, the edge agent downloads the data set D from the cloud platform and performs the analysis task, calculates and generates the baseline and control chart for analysis; S40, the edge agent calculates the three areas A, B, and C based on the calculated center line, control upper limit, and control lower limit, and automatically judges the remaining data set D for abnormality, and decides on-site whether to report to the main station and generate and eliminate invalid data based on the upper and lower limit information of the baseline calculation. When the data samples reach a certain number, the warning data and sample data that meet the abnormality judgment conditions will be reported to the cloud main station; S50. The cloud master station generates a production control chart using a weighted algorithm for the sample data reported by the edge side, and sends the new sample data to the edge agent for abnormality judgment. At the same time, it analyzes and processes the reported warning data and decides whether to report it to OA or send an email notification.

2. According to the method of production data statistical analysis based on cloud-edge collaboration in claim 1, it is characterized in that: In the step S10, the basic production data includes product information, production line information and process information.

3. According to the method of production data statistical analysis based on cloud-edge collaboration in claim 1, it is characterized in that: In step S20, the production data set is defined as X = {x1, x2, ..., x N }, the label data set is Y = {y1, y2, ..., y N }, the data set D = (X, Y), set the product test items and sub-items that need to be analyzed, and the required judgment criteria, and upload the data set D = (X, Y) and the basic production data to the cloud platform.

4. According to the method of production data statistical analysis based on cloud-edge collaboration in claim 1, it is characterized in that: The specific steps of step S30 are as follows: S301, calculate the standard deviation σ=sqrt((x1-x)) through the collected data set D 2 +(x2-x) 2 +……(x n -x) 2 ) / (n-1)), and the mean Xbar=(X1+X2+……X n ) / n; S302, calculate Ppk, the Ppk calculation formula is as follows: Ppk=(1-k)*Pp=(USL-LSL-2|M-μ|) / (3σ); k=|M-μ| / (T / 2)=ε / (T / 2); ε=|M-μ|; Pp=(USL-LSL) / 6σ=T / 6σ; T = (USL-LSL); Among them, σ is the standard deviation, μ is the mean, M is the target value, USL is the upper specification limit, and LSL is the lower specification limit; S303, using the calculated center line, control upper limit and control lower limit, a stability determination is performed to obtain baseline data, and the stability determination criteria are as follows: 25 consecutive data points are within the control limits; Among 35 consecutive data points, at most one data point exceeds the control limit; Among 100 consecutive data points, at most two points are outside the control limits; A control chart for analysis is generated using the baseline data and the Ppk.

5. According to the method of production data statistical analysis based on cloud-edge collaboration in claim 3, it is characterized in that: The discrimination criteria are as follows: Criterion one is that point K falls outside zone A; Criterion two is that consecutive points K fall on the same side of the center line; Criterion three is that consecutive points K increase or decrease; Criterion four is that consecutive points K alternate up and down; Criterion five is that one of the consecutive K+1 points falls outside zone B on the same side of the center line; Criterion six is ​​that one of the consecutive K+1 points falls outside zone C on the same side of the center line; Criterion seven is that consecutive points K fall within zone C; Criterion eight is that consecutive points K fall on both sides of the center line, but no point is within zone C.

6. According to the method of production data statistical analysis based on cloud-edge collaboration in claim 1, it is characterized in that: The following weighted algorithm is used to generate the production control chart in step S50: S501. Assume that the overall data is X~N(μ,σ), then the distribution of X follows N(μ,σ / n), and the distribution of R follows (k2σ,k3σ), where k2 and k3 are coefficients that vary with n. Assume that the number of sample groups is i∈[1,m], and the observed value of the jth sample in the i-th group is x ij , j∈[1,n], the mean of the i-th group of samples is The total sample mean is The range of the i-th group of samples is R i , the average value of the range is but: R i =x imax -x imin ; S502, The center line UCL, upper control limit CL, and lower control limit LCL of the chart are defined as: in, S503, the center line UCL, upper control limit CL, and lower control limit LCL of the R chart are defined as: in, S504, yes The control chart is weighted, and each observation is weighted, and the sum of all weighted values ​​is 1. If p k is the weight of the kth observation, q k is the kth observation value, then the weighted i-th observation value is expressed as p k q k , according to the model prediction, the estimated value of the n+1th point ^q n+1 It is expressed as a linear combination of the weighted observations of the first n points and the weighted predicted value of the first point: ^q n+1 =λq n +λ(1-λ)q n-1 +...+λ(1-λ) z q n-z +...+λ(1-λ) n-1 q1+(1-λ) n ^q1; The weight of each observation is: The weighted mean and weighted variance are: The weighted The control limits are: The weighted R control limits are:

7. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

8. An electronic terminal, characterized in that: include: Processor and memory; The memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory, so that the terminal executes the method according to any one of claims 1 to 6.

Images