A digital production line product quality monitoring method and system

Abstract

The application discloses a kind of digital production line product quality monitoring method and system, including set equipment data and production process data, and construct database;Based on data acquisition result, the historical data of key quality characteristics of key working procedure is obtained, the test design of key quality characteristics of key working procedure is carried out, and transfer function is established;Real-time monitoring feedback result of transfer function, whether transfer function analysis result meets quality management requirement is judged, if yes, then it is transferred into reverse control based on transfer function, if no, then product is isolated, if there is risk, then risk early warning is carried out;Based on the result of judgment, the prediction result of product quality is shown.The application can realize early prediction of the source of quality problem by the key quality characteristics of key working procedure, improve product quality monitoring effect;Through real-time monitoring feedback result of transfer function, product quality monitoring accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of product quality monitoring, and in particular to a method and system for monitoring product quality of a digital production line. Background Art

[0002] Product quality monitoring is a crucial step in a company's production process. Without the correct quality management ideas and methods, product quality can fluctuate. However, strict quality control can lead to a decline in production efficiency; conversely, excessive efficiency control can lead to quality issues. Therefore, achieving product quality management while ensuring production efficiency has become a critical concern for manufacturing companies.

[0003] Currently, most manufacturing processes rely on parameter monitoring through SPC (Statistical Process Control). SPC is a process control tool that utilizes mathematical statistics to analyze and evaluate the production process. Based on feedback, it promptly identifies signs of systemic factors and takes measures to eliminate their influence, maintaining the process in a controlled state influenced only by random factors to achieve quality control. It states that when a process is influenced only by random factors, it is in a state of statistical control (referred to as the controlled state); when systematic factors are present, it is in a state of statistical out-of-control (referred to as the out-of-control state). Because process fluctuations exhibit statistical regularity, when a process is under control, its characteristics generally follow a stable random distribution; when out-of-control, the process distribution changes.

[0004] However, SPC doesn't consider the interplay between multiple processes. Consequently, using SPC to monitor product quality can result in situations where the SPC for each process is within specifications, yet the final product quality is substandard. This means that existing technologies suffer from poor product quality monitoring and inaccurate results. Summary of the Invention

[0005] The main purpose of the present invention is to provide a digital production line product quality monitoring method and system, aiming to solve the technical problems of poor product quality monitoring effect and inaccurate product quality monitoring results in the prior art.

[0006] To achieve the above-mentioned purpose, the present invention provides a digital production line product quality monitoring method, which includes the following steps: collecting equipment data and production process data, and constructing a database; based on the data collection results, obtaining the historical data of key quality characteristics of key processes, conducting experimental design on the key quality characteristics of key processes, and establishing a transfer function; monitoring the feedback results of the transfer function in real time, judging whether the transfer function analysis results meet the quality management requirements, and if so, switching to reverse control based on the transfer function; if not, isolating the product, and issuing a risk warning if there is a risk; based on the judgment result, displaying the prediction results of product quality.

[0007] Optionally, equipment data and production process data are collected, specifically based on the type of production equipment, selecting the corresponding preset data collection method to collect equipment data and production process data; wherein, if the production equipment has a data interface, select the first preset data collection method, and transmit the data to the gateway through the bus or industrial Ethernet network for data collection; if the production equipment does not have a data interface, select the second preset data collection method, and collect data through an external sensor.

[0008] Optionally, the equipment data includes at least equipment temperature, vibration, current, and stress-strain data; the production process data includes at least production part characteristics, and production data of each production unit or production equipment in the production process.

[0009] Optionally, the database is at least used to store collected data, and the database includes at least a real-time database and a historical database. The real-time database is also used to build a real-time quality monitoring system, and the historical database is also used to store collected data, query historical sample values ​​of real-time production data, and perform subsequent analysis and modeling; reverse control includes displaying the predicted results of product quality and the control of subsequent processes; displaying the predicted results of product quality is specifically displayed in a visual form.

[0010] Optionally, a key process is a process that has a direct impact and plays a decisive role on at least one or more of the product quality, performance, function, life, reliability, and cost, and is also a process that forms key quality characteristics.

[0011] Optionally, based on the data collection results, historical data of key quality characteristics of key processes are obtained. Specifically, the SOV source of variation analysis statistical model is used in combination with the embedded method to select key quality characteristics and then obtain the corresponding historical data.

[0012] Optionally, an experimental design is performed on the key quality characteristics of the key process, and a transfer function is established to obtain an experimental factor setting, which specifically includes the following steps: step a. determining the key quality characteristic Y; step b. preliminarily determining the influencing factor X of Y and the level of the influencing factor X based on the key quality characteristic Y, wherein the number of influencing factors X is multiple; step c. selecting an orthogonal experimental table according to the number of levels of each influencing factor X; step d. collecting data based on the experimental table experiment to verify whether the degree of influence of the influencing factor X and the size of its influence meet the preset standards; if so, executing step e; if not, repeating steps bc; step e. establishing a transfer function model through regression analysis to determine the relationship between the key quality characteristic Y and its influencing factor X.

[0013] Optionally, monitor the feedback results of the transfer function in real time to determine whether the transfer function analysis results meet the quality management requirements, specifically:

[0014] The following transfer function prediction model is established:

[0015]

[0016] Among them, the response variable y i represents the i-th key quality characteristic, explanatory variable x ip represents the pth influencing factor of the i-th response variable, β0, β1…, β p is the regression model coefficient, ε i is a random error term, which is used to represent the influence of random factors on the model, ε1,ε2,…,ε n iidN(0,σ 2 ) is the model assumption, that is, the error terms are independent of each other and normally distributed, with an expected value of 0 and a constant variance;

[0017] The input X of the transfer function prediction model is divided into two parts. One part is the influencing factors involved in the completed process, which is recorded as The value of is certain; the other part is the influencing factors involved in the unfinished process It is unknown and is obtained by using historical data distribution;

[0018] Then use Monte Carlo simulation to predict the distribution of predicted values ​​when the product is offline according to with y i The area overlap of the specification distribution is used to predict whether the quality of the product meets the quality management requirements.

[0019] Optional, when with yi If the area overlap is greater than the preset qualified threshold, it is predicted that the product quality meets the quality management requirements and the prediction results are displayed; if with y i If the area is between the preset risk threshold and the preset qualified threshold, it is predicted that the product quality has a risk of poor quality, the prediction result is displayed, and a risk warning is issued; if with y i If the area overlap is lower than the preset risk threshold, it is predicted that the product quality does not meet the quality management requirements, the prediction result is displayed, and the product is isolated.

[0020] Corresponding to the digital production line product quality monitoring method, the present invention provides a digital production line product quality monitoring system, which includes: a data acquisition module, used to collect equipment data and production process data, and build a database; an experimental design module, used to obtain historical data of key quality characteristics of key processes based on data acquisition results, perform experimental design on key quality characteristics of key processes, and establish a transfer function; a real-time monitoring module, used to monitor the feedback results of the transfer function in real time, judge whether the transfer function analysis results meet the quality management requirements, and if so, switch to reverse control based on the transfer function; if not, isolate the product, and if there is a risk, issue a risk warning; a display module, used to display the prediction results of product quality based on the judgment results.

[0021] The beneficial effects of the present invention are:

[0022] (1) By building a database as the basis for full production process monitoring and product quality tracking, the production data is integrated to reversely optimize the process parameters and achieve coordinated control of process quality; through the key process and key quality characteristics, the workshop can quantitatively evaluate the quality assurance capabilities of the key processes of key products, and predict the source of quality problems in advance, thereby avoiding further losses and improving the product quality monitoring effect; at the process level, through real-time monitoring of the feedback results of the transfer function, the production process monitoring, process quality prediction and early warning, process parameter reverse optimization and other functions are realized, and the accuracy of product quality monitoring results is improved to achieve continuous improvement of production and manufacturing processes and quality levels;

[0023] (2) According to the type of production equipment, the corresponding preset data collection method is selected to effectively improve the efficiency of data collection;

[0024] (3) By building a database, all production processes can be uniformly monitored and managed. The predicted results of product quality can be displayed in a visual form, which is vivid and intuitive. This allows on-site staff to not only understand the real-time production situation in the workshop, but also have an intuitive judgment of product quality, thereby improving productivity and product quality.

[0025] (4) Realize overall monitoring of the entire product process / manufacturing period through data modeling, including full connectivity control between different equipment, different processes, and different factory areas, fully consider the mutual influence between multiple processes, and improve the effectiveness and accuracy of product quality monitoring;

[0026] (5) Through DOE (DESIGN OF EXPERIMENT), the key quality characteristics of the key processes are designed and a transfer function is established. Combined with the feedback of the transfer function, the key quality characteristics of each process in the product manufacturing process are visually monitored, and the status of product quality is displayed in real time according to the production rhythm, thereby improving production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0028] Figure 1 This is a simplified flow chart of the product quality monitoring method for a digital production line according to the present invention;

[0029] Figure 2 This is a simplified schematic diagram of predicting product quality based on area overlap according to an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of the display form of the display module in one embodiment of the digital production line product quality monitoring system of the present invention;

[0031] Figure 4 The figure is a simplified schematic diagram of a workpiece process parameter monitoring diagram according to an embodiment of the present invention. DETAILED DESCRIPTION

[0032] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0033] like Figure 1As shown, a digital production line product quality monitoring method of the present invention includes the following steps: collecting equipment data and production process data, and building a database; based on the data collection results, obtaining the historical data of key quality characteristics of key processes, conducting experimental design on the key quality characteristics of key processes, and establishing a transfer function; monitoring the feedback results of the transfer function in real time, judging whether the transfer function analysis results meet the quality management requirements, and if so, switching to reverse control based on the transfer function; if not, isolating the product, and issuing a risk warning if there is a risk; based on the judgment result, displaying the prediction result of product quality.

[0034] The present invention constructs a database as the basis for full production process monitoring and product quality tracking, integrates production data to reversely optimize process parameters, and realizes coordinated control of process quality; through key processes and key quality characteristics, the workshop can conduct quantitative evaluation of the quality assurance ability of key processes of key products, and predict the source of quality problems in advance, thereby avoiding further losses and improving the product quality monitoring effect; at the process level, through real-time monitoring of the feedback results of the transfer function, the production process monitoring, process quality prediction and early warning, process parameter reverse optimization and other functions are realized, and the accuracy of product quality monitoring results is improved, so as to achieve continuous improvement of production and manufacturing processes and quality levels.

[0035] In this embodiment, the equipment data and the production process data are collected by selecting a corresponding preset data collection method based on the type of production equipment to collect the equipment data and the production process data.

[0036] Preferably, if the production equipment has a data interface, such as a machining center, a grinder, a PLC controller, a robot, an instrument, etc., the first preset data acquisition method is selected to transmit the data to the gateway through a bus or an industrial Ethernet network for data acquisition.

[0037] If the production equipment does not have a data interface, the second preset data collection method is selected, and data is collected through external sensors. Data is transmitted by wired communication or wireless communication to improve communication capabilities. After the data is transmitted to the network, on-site data analysis and storage are completed through edge computing.

[0038] Complete manufacturing equipment should have comprehensive archival information, including product number, description, status, and timestamp. Devices and data can be interconnected with other devices, equipment, and the execution layer through defined communication interfaces. This invention selects a corresponding pre-set data collection method based on the type of production equipment, effectively improving data collection efficiency.

[0039] In the embodiment, the equipment data at least includes equipment temperature, vibration, current, stress-strain data; and the production process data at least includes production part features, production data of each production unit or production equipment in the production process.

[0040] In the embodiment, the database is used at least for storing the collected data, and the database at least includes a real-time database and a historical database; the real-time database is further used for constructing a quality real-time monitoring system; the historical database is further used for storing the collected data, querying historical sample values of real-time production data, and subsequent analysis and modeling; the reverse control includes displaying a prediction result of product quality and controlling a subsequent process; and the prediction result of product quality is displayed in a visual form.

[0041] The application can realize unified monitoring and management of each production process by constructing a database; and the prediction result of product quality is displayed in a visual form, which is intuitive and can help on-site staff to understand real-time production conditions in the workshop and make intuitive judgments on product quality, thereby improving productivity and product quality.

[0042] In the embodiment, the key process is a process that has a direct impact on and plays a decisive role in one or more aspects of product quality, performance, function, life, reliability and cost, and is also a process in which key quality characteristics are formed.

[0043] In the embodiment, based on the data collection result, the historical data of the key quality characteristics of the key process are obtained, and the SOV variation source analysis statistical model is used to select the key quality characteristics in combination with an embedded method, and then the corresponding historical data are obtained.

[0044] In the embodiment, the embedded method refers to using a machine learning algorithm and model to train to obtain weight coefficients of each parameter, selecting features according to the coefficients from large to small, and determining the advantages and disadvantages of the parameters in the training process, thereby screening the key quality characteristics. The machine learning algorithm and model include but are not limited to a neural network model, a random forest model, an XGBoost algorithm, etc.

[0045] In the embodiment, the key quality characteristics of the key process are designed by a design of experiment (DOE), and a transfer function is established to obtain experimental factor settings, which specifically include the following steps: step a. determining a key quality characteristic Y; step b. preliminarily determining an influence factor X of the key quality characteristic Y and a level of the influence factor X according to the key quality characteristic Y, wherein the number of the influence factors X is multiple;

[0046] Step c. selecting an orthogonal test table according to the number of levels of each influence factor X;

[0047] Step d. Collect data based on the test table test to verify whether the degree of influence of the influencing factor X and its impact size meet the preset standards. If so, execute step e; if not, repeat steps bc;

[0048] Step e. Establish a transfer function model through regression analysis to determine the relationship between the critical quality characteristic Y and its influencing factor X. That is, step b is a preliminary screening of influencing factor X, and step d is a verification of the screening results of step b, determining the degree of influence of influencing factor X screened in step b on the critical quality characteristic Y and its magnitude, etc., identifying the main influencing factors and labeling them. If the main influencing factors cannot be determined in step d, it means that the influencing factors X screened in step b have no significant effect on the critical quality characteristic Y. Therefore, it is necessary to repeat steps bc and re-determine the influencing factor X until the main influencing factors can be determined in step d.

[0049] Preferably, the method for determining the critical quality characteristic Y includes but is not limited to being specified by the customer, determined based on the relevant experience of the product engineer / process engineer, and determined after analyzing data of similar products or problems found in the production process and on-site use (such as paretochart).

[0050] Preferably, the method for determining the influencing factor X includes but is not limited to brainstorming, fishbone diagram, and expert opinion; the method for determining the level of the influencing factor X includes but is not limited to determining based on engineer experience and determining based on analysis of historical production data.

[0051] Preferably, the identification method in step d is the variance contribution rate in variance analysis, which ranks the relative importance of the influencing factor X.

[0052] DOE is a method for studying and processing the relationship between multiple factors and response variables. By collecting a large amount of historical production process parameter data, rationally selecting experimental conditions, and analyzing the experimental results to obtain a transfer function, the method simultaneously optimizes multiple influencing factors X. While each key quality characteristic Y may not be optimal, the combined settings are optimal, ultimately resulting in an optimal set of experimental factor settings.

[0053] The present invention realizes overall monitoring of the entire product process / the entire manufacturing period through data modeling, including fully connected management and control between different equipment, different processes, and different factory areas, fully considering the mutual influence between multiple processes, and improving the effect and accuracy of product quality monitoring.

[0054] In this embodiment, real-time monitoring of the feedback results of the transfer function and determining whether the transfer function analysis results meet the quality management requirements specifically include the following steps:

[0055] Since a product's key quality characteristic Y usually does not have only one, and many key quality characteristics Y must meet quality management requirements at the same time, it is based on multiple transfer functions: Y = Xβ + ε, ε → N n (O,σ 2 I n ), the response variable vector Y represents the key quality characteristics that meet the quality management requirements Y=(y1,y2,…y n ), variable X represents the factor that affects the key quality characteristic Y (X=(x1,x2,…,x p );

[0056] And the transfer function is as follows: y i =β0+β1x i1 +β2x i2 +…+β p x ip +ε i ,ε1,ε2,…,ε n iidN(0,σ 2 ), where the response variable y i represents the i-th key quality characteristic, explanatory variable x ip represents the pth influencing factor of the i-th response variable, β0, β1…, β p is the regression model coefficient, ε i is a random error term used to represent the impact of random factors on the model. iid stands for independently and identically distributed in English and in Chinese for independent and identically distributed. In probability and statistics theory, if a sequence of variables or other random variables have the same probability distribution and are independent of each other, then these random variables are independent and identically distributed. n iidN(0,σ 2 ) is the model assumption, that is, the error terms are independent of each other and normally distributed, with an expected value of 0 and a constant variance.

[0057] The input X of the transfer function prediction model is divided into two parts. One part is the influencing factors involved in the completed process, which is recorded as The value of is certain; the other part is the influencing factors involved in the unfinished process It is unknown and is obtained by using historical data distribution; Monte Carlo simulation is then used to predict the distribution of predicted values ​​when the product is offline. according to with y iThe area overlap of the specification distribution is used to predict whether the quality of the product meets the quality management requirements.

[0058] The present invention uses DOE (Design of Experiment) to conduct experimental design on the key quality characteristics of key processes, establish a transfer function, and obtain experimental factor settings. Combined with the feedback of the transfer function, the key quality characteristics of each process in the product manufacturing process are visually monitored, and the status of product quality is displayed in real time according to the production rhythm, thereby improving production efficiency.

[0059] In this embodiment, when with y i If the area overlap is greater than the preset qualified threshold, it is predicted that the product quality meets the quality management requirements and the prediction results are displayed; if with y i If the area overlap is between the preset risk threshold and the preset qualified threshold, it is predicted that the product quality has a risk of poor quality, the prediction result is displayed, and a risk warning is issued; if with y i If the area overlap is lower than the preset risk threshold, it is predicted that the product quality does not meet the quality management requirements, the prediction result is displayed, and the product is isolated.

[0060] Preferably, the preset qualified threshold is 70%, and the preset risk threshold is 50%. with y i If the area overlap is greater than 70%, it is predicted that the product quality meets the quality management requirements and the prediction results are displayed; if with y i If the area overlap is between 50% and 70% (inclusive), it is predicted that the product quality has a risk of poor quality, the prediction result is displayed, and a risk warning is issued; if with y i If the area overlap is less than 50%, it is predicted that the product quality does not meet the quality management requirements, the prediction results are displayed, and the product is isolated.

[0061] like Figure 2 As shown, the upper part briefly shows x i with y i The left figure in the lower half shows the relationship between the predicted distribution and the specification distribution when the cross area is 46.25%. with y i If the area overlap is less than 50%, it is predicted that the product quality does not meet the quality management requirements, and the product is predicted to be defective and isolated; the right picture is The intersection area of ​​the predicted value distribution and the specification distribution is 85%, with y iIf the area overlap is greater than 70%, the product quality is predicted to meet the quality management requirements, and the product quality is predicted to be OK; the middle picture is The intersection area between the predicted value distribution and the specification distribution is 69%. with y i If the area overlap is between 50% and 70%, it is predicted that there is a risk of poor product quality and a risk warning is issued.

[0062] Corresponding to the digital production line product quality monitoring method, the present invention provides a digital production line product quality monitoring system, which includes: a data acquisition module, used to collect equipment data and production process data, and build a database; an experimental design module, used to obtain historical data of key quality characteristics of key processes based on data acquisition results, conduct experimental design on the key quality characteristics of key processes, and establish a transfer function to obtain experimental factor settings; a real-time monitoring module, used to monitor the feedback results of the transfer function in real time, judge whether the transfer function analysis results meet the quality management requirements, and if so, switch to reverse control based on the transfer function; if not, isolate the product, and if there is a risk, issue a risk warning; a display module, used to display the prediction results of product quality based on the judgment results.

[0063] like Figure 3 The figure shows a schematic diagram of the display form of the display module in an embodiment of the digital production line product quality monitoring system of the present invention. Specifically, in this embodiment, the statistical control concept based on the combination of source data and statistical analysis data is mainly adopted to represent the status of product quality in the production process as a block diagram. Figure 2 In the display, three different colors represent three quality prediction states. Dark gray represents the display result that the predicted product quality meets the quality management requirements. with y i The area overlap is greater than the preset qualified threshold; light gray represents the display result of the predicted product quality risk. with y i The area overlap is between the preset risk threshold and the preset qualified threshold; black represents the display result that the product quality is predicted to not meet the quality management requirements. with y i The area overlap is lower than the preset risk threshold.

[0064] It should be understood that the above colors are for example only and do not constitute an improper limitation on this application.

[0065] like Figure 4As shown, it is a brief schematic diagram of a workpiece process parameter monitoring diagram of an embodiment of the present invention. It can be seen that the present invention realizes unified monitoring and management of various production processes, which is vivid and intuitive, so that on-site staff can not only understand the real-time production situation of the workshop, but also display the predicted results of product quality in a visual form, which is vivid and intuitive. The staff can also have an intuitive judgment on the product quality, thereby improving productivity and product quality.

[0066] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. For similar or identical parts between the various embodiments, reference can be made to each other. For the apparatus embodiments, device embodiments, and storage medium embodiments, since they are generally similar to the method embodiments, their descriptions are relatively simple. For relevant parts, reference can be made to the descriptions of the method embodiments.

[0067] Furthermore, in this document, the terms "comprises," "comprising," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0068] While the foregoing description shows and describes preferred embodiments of the present invention, it should be understood that the present invention is not limited to the forms disclosed herein and should not be construed as excluding other embodiments. Rather, the present invention can be used in various other combinations, modifications, and environments, and can be modified within the scope of the present invention by the teachings herein or by techniques or knowledge in the relevant art. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the present invention are intended to be within the scope of the appended claims.

Claims

1. A digital production line product quality monitoring method, characterized in that: The following steps are involved: Collect equipment data and production process data and build a database; the database is used to store at least the collected data, and includes at least a real-time database and a historical database. The real-time database is also used to build a real-time quality monitoring system, and the historical database is also used to store collected data, query historical sample values ​​of real-time production data, and perform subsequent analysis and modeling; Based on the data collection results, obtain the historical data of the key quality characteristics of the key processes, conduct experimental design for the key quality characteristics of the key processes, and establish transfer functions. The key process is the process that has a direct impact and plays a decisive role on at least one or more of the product quality, performance, function, life, reliability, and cost, and is also the process where the key quality characteristics are formed; Based on the data collection results, historical data of key quality characteristics of key processes are obtained. Specifically, the SOV source of variation analysis statistical model is used, combined with the embedded method to select key quality characteristics, and then the corresponding historical data is obtained; Real-time monitoring of transfer function feedback results to determine whether the transfer function analysis results meet quality management requirements. If so, reverse control based on the transfer function is initiated. If not, the product is isolated and, if risks exist, a risk warning is issued. Reverse control includes displaying product quality prediction results and controlling subsequent processes. Monitor the feedback results of the transfer function in real time to determine whether the transfer function analysis results meet the quality management requirements, specifically: The following transfer function prediction model is established: Among them, the response variable y i represents the i-th key quality characteristic, explanatory variable x ip represents the pth influencing factor of the i-th response variable, β0, β1…, β p is the regression model coefficient, ε i is a random error term, which is used to represent the influence of random factors on the model, ε1,ε2,…,ε n iidN(0,σ 2 ) is the model assumption, that is, the error terms are independent of each other and normally distributed, with an expected value of 0 and a constant variance; The input X of the transfer function prediction model is divided into two parts. One part is the influencing factors involved in the completed process, which is recorded as The value of is certain; the other part is the influencing factors involved in the unfinished process It is unknown and is obtained by using historical data distribution; Then use Monte Carlo simulation to predict the distribution of predicted values ​​when the product is offline according to with y i The area overlap of the specification distribution can be used to predict whether the product quality meets the quality management requirements; Based on the judgment results, the prediction results of product quality are displayed; Conduct experimental design for the key quality characteristics of key processes and establish transfer functions, which includes the following steps: Step a. Determine the key quality characteristic Y; Step b. Preliminarily determine the influencing factor X of Y and the level of the influencing factor X according to the key quality characteristic Y, wherein the number of influencing factors X is multiple; Step c. Select an orthogonal test table according to the number of levels of each influencing factor X; Step d. Collect data based on the test table test to verify whether the degree of influence of the influencing factor X and its impact size meet the preset standards. If so, execute step e; if not, repeat steps bc; Step e. Establish a transfer function model through regression analysis to determine the relationship between the key quality characteristic Y and its influencing factor X.

2. The method for monitoring product quality of a digital production line according to claim 1, characterized in that: Collect equipment data and production process data, specifically by selecting the corresponding preset data collection method based on the type of production equipment to collect equipment data and production process data; If the production equipment has a data interface, the first preset data collection method is selected to transmit the data to the gateway via the bus or industrial Ethernet network for data collection; If the production equipment does not have a data interface, the second preset data collection method is selected to collect data through an external sensor.

3. The method for monitoring product quality of a digital production line according to claim 2, wherein: Equipment data at least includes equipment temperature, vibration, current, stress and strain data; The production process data at least includes the characteristics of the production parts and the production data of each production unit or production equipment in the production process.

4. The method for monitoring product quality of a digital production line according to claim 1, wherein: The prediction results of product quality are specifically displayed in a visual form.

5. The method for monitoring product quality of a digital production line according to claim 1, characterized in that: when with y i If the area overlap is greater than the preset qualified threshold, it is predicted that the product quality meets the quality management requirements and the prediction result is displayed; like with y i If the area overlap is between the preset risk threshold and the preset qualified threshold, it is predicted that the product quality has a risk of poor quality, the prediction result is displayed, and a risk warning is issued; like with y i If the area overlap is lower than the preset risk threshold, it is predicted that the product quality does not meet the quality management requirements, the prediction result is displayed, and the product is isolated.

6. A digital production line product quality monitoring system, characterized in that: include: The data acquisition module is used to collect equipment data and production process data and build a database; the database is used to store at least the collected data, and the database includes at least a real-time database and a historical database. The real-time database is also used to build a real-time quality monitoring system, and the historical database is also used to store collected data, query historical sample values ​​of real-time production data, and conduct subsequent analysis and modeling; The experimental design module is used to obtain historical data of key quality characteristics of key processes based on data collection results, conduct experimental design for the key quality characteristics of key processes, and establish a transfer function. The module specifically includes the following steps: step a. determining key quality characteristic Y; step b. preliminarily determining the influencing factor X of Y and the level of influencing factor X based on the key quality characteristic Y, wherein the number of influencing factors X is multiple; step c. selecting an orthogonal experimental table based on the number of levels of each influencing factor X; step d. experimentally collecting data based on the experimental table to verify whether the degree of influence of influencing factor X and its influence size meet the preset standards. If so, step e is executed; if not, steps b and c are repeated; step e. establishing a transfer function model through regression analysis to determine the relationship between key quality characteristic Y and its influencing factor X. A key process is a process that has a direct impact and plays a decisive role on at least one or more of product quality, performance, function, life, reliability, and cost, and is also a process that forms key quality characteristics; based on the data collection results, obtaining historical data of key quality characteristics of key processes, specifically using the SOV source of variation analysis statistical model combined with the embedded method to select key quality characteristics, and then obtaining corresponding historical data; The real-time monitoring module is used to monitor the feedback results of the transfer function in real time and determine whether the transfer function analysis results meet the quality management requirements. If so, it will switch to reverse control based on the transfer function. If not, the product will be isolated. If there is a risk, a risk warning will be issued. Reverse control includes displaying the predicted results of product quality and controlling the subsequent processes. Monitor the feedback results of the transfer function in real time to determine whether the transfer function analysis results meet the quality management requirements, specifically: The following transfer function prediction model is established: Among them, the response variable y i represents the i-th key quality characteristic, explanatory variable x ip represents the pth influencing factor of the i-th response variable, β0, β1…, β p is the regression model coefficient, ε i is a random error term, which is used to represent the influence of random factors on the model, ε1,ε2,…,ε n iidN(0,σ 2 ) is the model assumption, that is, the error terms are independent of each other and normally distributed, with an expected value of 0 and a constant variance; The input X of the transfer function prediction model is divided into two parts. One part is the influencing factors involved in the completed process, which is recorded as The value of is certain; the other part is the influencing factors involved in the unfinished process It is unknown and is obtained by using historical data distribution; Then use Monte Carlo simulation to predict the distribution of predicted values ​​when the product is offline according to with y i The area overlap of the specification distribution can be used to predict whether the quality of the product meets the quality management requirements; The display module is used to display the prediction results of product quality based on the judgment results.

Images