Online model measuring instrument dynamic monitoring system

Abstract

The invention discloses a dynamic monitoring system for an online version measuring instrument, and relates to the technical field of foil measurement, the dynamic monitoring system comprises a data acquisition module, a central processing unit, an online interaction module and a display alarm module, and the data acquisition module acquires historical data of offline foil film parameters and online foil film parameters; mathematical modeling and regression analysis are carried out based on the central processing unit, an off-line and on-line mixed model is built to obtain a regression equation, the effective degree of the regression equation is evaluated, real-time foil parameters are fed back to the central processing unit through the on-line interaction module, and whether the foil is qualified or not is judged by substituting the real-time foil parameters into the regression equation. According to the foil detection system, the on-line alarm prompt signal is generated and transmitted to the display alarm module, so that corresponding parameter display and alarm prompt operation are performed, the work efficiency of foil detection is improved, the comprehensiveness and integrity of foil detection are ensured, and the accuracy of alarm prompt in the foil processing process is improved.

Description

Technical Field

[0001] The invention relates to the technical field of foil material measurement, and in particular to a dynamic monitoring system of an online version measuring instrument. Background Art

[0002] Lithium battery production has strict requirements on the thickness, width, surface quality, etc. of foil. With the development of sensors, data processing and automation technology, a technical basis has been provided for the realization of online pattern measuring instruments, enabling real-time and high-precision monitoring of foil quality. Since it is difficult to monitor foil during the production of lithium batteries, the current production line usually adopts an offline on-machine sampling mode, which is cumbersome, time-consuming and labor-intensive to operate. The sampling area accounts for a low proportion of the entire foil area, and it is impossible to achieve large-area real-time dynamic monitoring. There are defects such as low foil monitoring efficiency and insufficient accuracy. Therefore, the traditional offline sampling mode cannot meet the needs of real-time monitoring, affecting production efficiency and product quality. In view of the above technical defects, a solution is now proposed. Summary of the invention

[0003] The purpose of the present invention is to solve the defects of low foil monitoring efficiency and insufficient accuracy in the prior art, improve the working efficiency of foil detection, ensure the comprehensiveness and integrity of foil detection, and further improve the accuracy of alarm prompts for the foil processing process.

[0004] In order to achieve the above object, the present invention adopts the following technical solutions: An online pattern measuring instrument dynamic monitoring system comprises a data acquisition module, a central processing unit, an online interaction module and a display alarm module, wherein the data acquisition module, the central processing unit, the online interaction module and the display alarm module are communicatively connected; The data acquisition module is used to collect foil data: the foil is tested offline and online respectively by an offline pattern measuring instrument and an online pattern measuring instrument, the historical data of offline foil parameters and online foil parameters are obtained, and the historical data of offline foil parameters and online foil parameters are integrated and marked as foil data; The central processing unit is used to construct an offline-online hybrid model: by performing preliminary analysis on the foil data, evaluating the offline foil flatness and the online foil flatness, and integrating and acquiring the pattern data through n groups of foil flatness data values, and by performing mathematical modeling and regression analysis on the pattern data, building an offline-online hybrid model to obtain the regression equation, and evaluating the effectiveness of the regression equation; The online interactive module is used to determine the online qualification of foil materials: through the online pattern measuring instrument, real-time detection is carried out to obtain the real-time data of online foil parameters and evaluate the online foil flatness; by setting the qualified standard of the foil material of the offline pattern measuring instrument and feeding it back to the central processor for substitution into the regression equation, the online alarm threshold is obtained; and then the online foil flatness is compared with the online alarm threshold to generate an online alarm prompt signal; The display alarm module is used to receive signals and display alarms: by receiving online alarm prompt signals, the prompt text is edited and corresponding parameter display and alarm prompt operations are performed.

[0005] Furthermore, the specific process of obtaining the pattern data is as follows: A1, offline foil parameters include offline foil thickness and offline foil range; The offline pattern measuring instrument collects single laser data by reciprocating and stopping, thereby collecting parameters of the produced foil material by reciprocating and stopping for multiple times; extracts Ny offline detection feature points through the motion trajectory of the reciprocating and stopping single laser rangefinder, marks any offline detection feature point as Dy, and marks the offline foil film thickness of the offline detection feature point Dy as Hy, thereby obtaining the offline foil film thickness Hy of the Ny offline detection feature points; Sort the offline foil thickness Hy of Ny offline detection feature points to obtain the highest value of the offline foil thickness Hy and the minimum value , thereby calculating the offline foil range Ry, evaluating the offline foil flatness Y through the offline foil range Ry, and constructing the offline pattern data set Qy through n offline foil flatness Y; A2, online foil parameters include online foil thickness and online foil range; The online pattern measuring instrument performs real-time multi-laser distance measurement during the foil production and transmission process, thereby collecting parameters of the foil in production; extracts Nx online detection feature points through the relative motion trajectory of the foil film dynamically transmitted on the multi-laser distance meter, marks any online detection feature point as Dx, and marks the online foil film thickness of the online detection feature point Dx as Hx, thereby obtaining the online foil film thickness Hx of the Nx online detection feature points; Sort the online foil thickness Hx of Nx online detection feature points to obtain the highest value of the online foil thickness Hx and the minimum value , thereby calculating the online foil range Rx, evaluating the online foil flatness X through the online foil range Rx, and constructing the online pattern number set Qx through n online foil flatness X; A3, integrate and mark the offline pattern data set Qy and the online pattern data set Qx as pattern data.

[0006] Furthermore, the specific process of building an offline-online hybrid model is as follows: B1, identify and prioritize pattern data through the Pareto chart of standardized effects, and determine the correlation between the offline pattern data set Qy and the online pattern data set Qx; B2, through Pearson modeling, obtain the overall Pearson correlation coefficient γxy to evaluate the correlation between the online pattern set Qx and the offline pattern set Qy; B3, perform regression analysis on the pattern data to evaluate the estimation accuracy of the model coefficients, the fitting effect and significance of the regression model, and thus comprehensively evaluate the effectiveness of the regression equation.

[0007] Furthermore, the specific process of regression analysis of pattern data is as follows: B3-1, mathematical function modeling is performed through the offline pattern data set Qy and the online pattern data set Qx to build a regression equation: offline pattern data Y= + *Online pattern data X; B3-2, build a scatter plot of the pattern data through the pattern data, and build a matrix through the scatter plot for regression analysis, so as to obtain the regression analysis data of the pattern data and generate a regression model: ; Then obtain the estimation accuracy evaluation index Xjd of the model coefficients to evaluate the estimation accuracy of the model coefficients; B3-3, evaluating the fitting effect of the regression model, obtaining the model fitting effect evaluation parameters, and then comprehensively obtaining the fitting effect evaluation index Xnh through the model fitting effect evaluation parameters to evaluate the fitting effect of the regression model; B3-4, perform variance analysis on the regression equation to obtain the model's significant factor evaluation parameters, obtain the significance evaluation index Xxz through the model's significant factor evaluation parameters, and evaluate the significance of the regression model; B3-5, and then by combining the estimation accuracy evaluation index Xjd of the model coefficient, the fitting effect evaluation index Xnh of the regression model, and the significance evaluation index Xxz, the regression equation effectiveness index Eff is obtained to comprehensively evaluate the effectiveness of the regression equation.

[0008] Furthermore, the specific process of obtaining the overall Pearson correlation coefficient γxy through Pearson modeling is as follows: B2-1, the online pattern data set Qx contains n data values ​​of online foil flatness X, and the overall mean is obtained through the online pattern data set Qx and standard deviation Sx; B2-2, the offline pattern data set Qy contains n offline foil flatness Y data values, and the overall mean is obtained through the offline pattern data set Qy and standard deviation Sy; B2-3, and then obtain the overall covariance and the overall Pearson correlation coefficient γxy; B2-4, set the evaluation interval of the overall Pearson correlation coefficient γxy, and evaluate the correlation between the online pattern set Qx and the offline pattern set Qy through interval comparison.

[0009] Furthermore, the specific parameters for regression analysis of pattern data are: Regression analysis data include the coefficients of the regression equation, standard errors of the coefficients, T values, P values, and variance inflation factors. The coefficients of the marked independent variables are , the coefficient of the labeled constant term is ; Mark the coefficient of the independent variable The standard error is , marking constant term coefficient The standard error is ; T value = coefficient / standard error of coefficient; Generate a regression model through regression analysis data: ; The model fitting effect evaluation parameters include the determination coefficient Rsq, the adjusted determination coefficient Radj, and the predicted determination coefficient Rpred; The parameters for evaluating significant factors in the model include degrees of freedom (DF), adjusted sum of squares (AdjSS), adjusted mean square (AdjMS), F-statistic and P-value.

[0010] Furthermore, the generation process of the online alarm prompt signal is as follows: The qualified standard of the foil material of the offline pattern meter is preset, and the offline alarm threshold value of the offline foil flatness Y of the offline pattern data set Qy is marked as Gy; Substitute the qualified condition of offline foil flatness Y ≤ offline alarm threshold Gy into the regression equation , thereby setting the online alarm threshold Gx. When the online foil flatness X of the online pattern set Qx is lower than the online alarm threshold Gx, an online alarm prompt signal is generated.

[0011] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: The present invention collects historical data of offline foil parameters and online foil parameters through a data acquisition module, performs mathematical modeling and regression analysis based on a central processing unit, builds an offline and online hybrid model to obtain a regression equation, evaluates the effectiveness of the regression equation, and then feeds back real-time foil parameters to the central processing unit through an online interactive module, substitutes them into the regression equation to determine whether the foil is qualified, generates an online alarm prompt signal and transmits it to a display alarm module, and performs corresponding parameter display and alarm prompt operations; Among them, the online pattern measuring instrument based on this system is separately set up in the foil processing production lines such as the slitting machine and the carbon coating machine. Real-time multi-laser ranging is carried out during the foil production and transmission process. Intelligent monitoring improves the work efficiency of foil detection; through the dynamic transmission of the foil production process, comprehensive real-time monitoring of the foil processing standards is achieved, ensuring the comprehensiveness and integrity of foil detection; by evaluating the effectiveness of the regression equation of the offline and online hybrid model, the accuracy of the alarm prompts for the foil processing process is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 A connection diagram of the system modules of the present invention is shown; Figure 2 A schematic diagram showing the steps of the solution process of the present invention is shown; Figure 3 A scatter plot of the pattern data of the present invention is shown. DETAILED DESCRIPTION

[0013] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0014] Embodiment 1: like Figure 1-Figure 3 As shown, an online pattern measuring instrument dynamic monitoring system includes a data acquisition module, a central processing unit, an online interaction module and a display alarm module, wherein the data acquisition module, the central processing unit, the online interaction module and the display alarm module are communicatively connected; At present, the industry generally uses offline pattern testers to test foils by sampling and monitoring the produced foils on the machine, and collects data on a small area of ​​the entire roll of foil; and under the premise that the dynamic transmission process of lithium batteries on the production line cannot be monitored in real time, this case designs and launches an online pattern tester, and realizes real-time detection of foils based on an online pattern measuring instrument dynamic monitoring system; The specific working steps of the online pattern measuring instrument dynamic monitoring system are as follows: S1, the data acquisition module collects foil data: by setting an offline pattern measuring instrument and an online pattern measuring instrument, the foil is tested offline and online respectively, the historical data of offline foil parameters and online foil parameters are obtained, and the historical data of offline foil parameters and online foil parameters are integrated and marked as foil data; S1-1, the offline pattern measuring instrument collects data by laser ranging through reciprocating stops, collects data once every certain distance, and reciprocates multiple times to collect the parameters of the produced foil; Offline foil film parameters include offline foil film thickness and offline foil film range; Extract Ny offline detection feature points through the motion trajectory of the reciprocating stop of the single laser rangefinder, mark any offline detection feature point as Dy, mark the offline foil film thickness of the offline detection feature point Dy as Hy, so as to obtain the offline foil film thickness Hy of the Ny offline detection feature points; Sort the offline foil thickness Hy of Ny offline detection feature points to obtain the highest value of the offline foil thickness Hy and the minimum value , and then calculate the offline foil range Ry: ; The offline foil flatness Y is evaluated by the offline foil range Ry: ;in, is the conversion coefficient and Greater than 0, conversion coefficient It refers to the preset constant for converting the offline foil film range Ry into the offline foil film flatness Y. When the offline foil film range Ry is higher, the offline foil film flatness Y is evaluated to be lower. Construct an offline pattern data set Qy through n offline foil flatness Y; S1-2, the online pattern measuring instrument is separately set up in the foil processing production line such as the slitting machine and the carbon coating machine, and performs real-time multi-laser ranging during the foil production and transmission process, so as to collect the parameters of the foil in production; Online foil parameters include online foil thickness and online foil range; Through the relative motion trajectory of the foil film dynamically transmitted on the multi-laser rangefinder, Nx online detection feature points are extracted, any online detection feature point is marked as Dx, and the online foil film thickness of the online detection feature point Dx is marked as Hx, so as to obtain the online foil film thickness Hx of the Nx online detection feature points; Sort the online foil thickness Hx of Nx online detection feature points to obtain the highest value of the online foil thickness Hx and the minimum value , and then calculate the online foil range Rx: ; Evaluation of online foil flatness X by online foil extreme difference Rx: ;in, is the conversion coefficient and Greater than 0, conversion coefficient It refers to the preset constant that converts the online foil film range Rx into the online foil film flatness X. When the online foil film range Rx is higher, the online foil film flatness is evaluated to be lower; Construct the online pattern set Qx through n online foil flatness X; S1-3, integrate and mark the offline pattern data set Qy and the online pattern data set Qx as pattern data.

[0015] S2, the central processing unit builds an offline-online hybrid model: by performing preliminary analysis on the foil data, evaluating the offline foil flatness and the online foil flatness, and integrating n groups of foil flatness data values ​​to obtain the pattern data, and by performing Pearson modeling and regression analysis on the pattern data, building an offline-online hybrid model to obtain the regression equation, and evaluating the effectiveness of the regression equation.

[0016] The specific process of building an offline-online hybrid model is as follows: S2-1, the Pareto chart of standardized effects is used to identify and prioritize pattern data. The response is offline pattern data, and the standardized effect value α=0.05, indicating that the factor with the greatest impact on the result is offline pattern data, which indicates that there is a correlation between the offline pattern data set Qy and the online pattern data set Qx; Then, through Pearson modeling, the overall Pearson correlation coefficient γxy is obtained. The specific process is as follows: S2-101, the online pattern data set Qx includes n data values ​​of online foil flatness X; Then the online version number set Qx: {X1, X2, X3...Xn}; The overall mean of the online version set Qx for: ; The standard deviation Sx of the online version number set Qx is: ; i is the element index of the index set; S2-102, the offline pattern data set Qy includes n data values ​​of offline foil flatness Y; Then the offline version number set Qy: {Y1, Y2, Y3...Yn}; The overall mean of the offline version set Qy for: ; The standard deviation Sy of the offline version data set Qy is: ; S2-103, the population covariance is: ; The overall Pearson correlation coefficient γxy is obtained from this: ; Set the evaluation interval of the overall Pearson correlation coefficient γxy, and evaluate the correlation between the online pattern set Qx and the offline pattern set Qy through interval comparison, so as to determine and perform regression analysis; The overall Pearson correlation coefficient γxy=0.906 was obtained by calculating the experimental data; It is assumed that when γxy∈(0.5,1.0), it means that there is a strong and positive correlation between the two, and thus further regression analysis is performed on the offline pattern set Qy and the online pattern set Qx.

[0017] S2-2, the specific process of regression analysis of pattern data is as follows: Some experimental values ​​of the pattern data are shown in Table 1: Table 1 Experimental data record table of pattern data

[0018] Mathematical function modeling is performed through the offline pattern data set Qy and the online pattern data set Qx to build a regression equation: offline pattern data Y=1.455+1.7664*online pattern data X.

[0019] S2-3, build a scatter plot of the pattern data through the experimental values ​​of the pattern data, such as Figure 3 As shown in Table 2, a matrix is ​​constructed through a scatter plot for regression analysis to obtain the regression analysis data of the pattern data. The regression analysis data of the pattern data includes coefficients, standard errors of coefficients, T-statistics, P-values, and variance inflation factors (VIF), as shown in Table 2: Table 2 Regression analysis data record table of pattern data

[0020] S2-301, the independent variable refers to the online pattern data, and the independent variable coefficient refers to the influence of the online pattern data on the offline pattern data. The coefficient of the independent variable is marked as : ; S2-302, the constant term refers to the intercept term of the regression model, which indicates the predicted value of the dependent variable when the independent variable takes the value of 0. The coefficient of the constant term is : ; S2-303, the standard error of the coefficient is used to measure the accuracy of the coefficient estimate. The smaller the standard error of the coefficient, the more accurate the estimate; Label the coefficients of the independent variables Standard error The formula is: ;in, is the estimate of the error variance, let ;RSS is the residual sum of squares, let ; Then the constant term coefficient Standard error The formula is: .

[0021] S2-304, T value is used to test whether the coefficient is significant, T value = coefficient / standard error of coefficient, the larger the T value, the more significant the coefficient; S2-305, generate a regression model through regression analysis of pattern data: ; S2-306, by independent variable coefficient Standard error and constant coefficient Standard error Combined, the estimated accuracy evaluation index Xjd of the model coefficients is obtained: ;in, , The coefficients of the independent variables are Standard error and constant coefficient Standard error The weight factor coefficient of , The preset values ​​of are all greater than 0; the evaluation interval of the estimation accuracy evaluation index Xjd is set. When the estimation accuracy evaluation index Xjd is higher, the estimation accuracy of the evaluation model coefficient is better.

[0022] S2-4, evaluating the fitting effect of the regression model to obtain model fitting effect evaluation parameters, the model fitting effect evaluation parameters include the determination coefficient Rsq, the adjusted determination coefficient Radj, and the predicted determination coefficient Rpred; S2-401, coefficient of determination Rsq: ,in, is the ith observation value, is the i-th predicted value, is the mean of the observed values; The coefficient of determination Rsq is between (0,1). When the coefficient of determination Rsq is closer to 1, the model fits the data better; when the coefficient of determination Rsq is closer to 0, the model fits the data worse. S2-402, adjusted coefficient of determination Radj: ; The adjusted coefficient of determination Radj is used to reflect the true degree of fit of the model. When the number of independent variables is adjusted, the differences in model fitting effects with different numbers of independent variables are compared. When the adjusted coefficient of determination Radj becomes higher, it means that the newly added independent variables have a significant impact and explanatory power on the model.

[0023] S2-403, predicted coefficient of determination Rpred: ,in, It is the predicted value of the ith sample after building a model using all the data except the ith sample. The higher the predicted coefficient of determination Rpred, the better the predictive ability of the regression model and the more accurate the predicted value.

[0024] The values ​​of the model fitting effect evaluation parameters are shown in Table 3: Table 3 Model fitting effect evaluation parameter record table

[0025] S2-404, comprehensively obtain the fitting effect evaluation index Xnh through the model fitting effect evaluation parameters: ;in, , , are the weight factor coefficients of the determination coefficient Rsq, the adjusted determination coefficient Radj, and the predicted determination coefficient Rpred, respectively, and , , The preset values ​​of are all greater than 0; the evaluation interval of the fitting effect evaluation index Xnh is set. When the fitting effect evaluation index Xnh is higher, the fitting effect of the evaluated regression model is better.

[0026] S2-5, then perform variance analysis on the regression equation to obtain the model significant factor evaluation parameters, which include degrees of freedom (DF), adjusted sum of squares (AdjSS), adjusted mean square (AdjMS), F-statistic and P-value, as shown in Table 4: Table 4 Model significant factor evaluation parameter record table

[0027] S2-501, degrees of freedom (DF) refers to the number of samples of independent information; the preset data samples are 85, among which the lack-of-fit error is used to evaluate the systematic deviation between the model and the data; pure error refers to the random variation of the data itself; S2-502, AdjSS refers to the contribution of each factor to the total variation; Mark the AdjSS of the regression model as ModelSS: ; Mark the error AdjSS as ErrorSS: ; S2-503, AdjMS is the adjusted mean square, which is calculated by dividing AdjSS by its degrees of freedom; S2-504, F value is used for model significance test, ; The larger the F value, the greater the proportion of variance that the regression model can explain to the total variance, that is, the better the model fits the data and the higher the significance; The F value of the regression model is labeled ModelF, and the F value of the lack-of-fit error is labeled ErrorF; S2-505, P value is used to judge the significance of the model. It refers to the probability of the current F value or a larger F value when the model is not significant. Therefore, when the model is significant, the P value of the regression model is 0; the P value of the lack-of-fit error is ; The smaller the P value, the more sufficient evidence there is that the model is significant; Label the P value of the regression model as ModelP and the P value of the lack-of-fit error as ErrorP.

[0028] S2-506, obtain the significance evaluation index Xxz through the model significance factor evaluation parameters: ;in, , , , are the weight factor coefficients of the F value ModelF of the regression model, the F value ErrorF of the lack-of-fit error, the P value ModelP of the regression model, and the P value ErrorP of the lack-of-fit error, respectively, and , , , The preset values ​​of are all greater than 0; the evaluation interval of the significance evaluation index Xxz is set. When the significance evaluation index Xxz is higher, the significance of the regression model is better.

[0029] S2-6, and then obtain the regression equation effective index Eff by combining the estimation accuracy evaluation index Xjd of the model coefficient, the fitting effect evaluation index Xnh of the regression model, and the significance evaluation index Xxz: ; in, , , They are the weight indexes of the estimation accuracy evaluation index Xjd of the model coefficient, the fitting effect evaluation index Xnh of the regression model, and the significance evaluation index Xxz, respectively. , , The preset values ​​of are all greater than 0; when the estimation accuracy evaluation index Xjd of the model coefficient, the fitting effect evaluation index Xnh of the regression model, and the significance evaluation index Xxz are higher, the regression equation effective index Eff is higher, and the effectiveness of the regression equation is higher; set the effective interval of the regression equation effective index Eff. When the regression equation effective index Eff is within the effective interval, it means that the regression equation =1.455+1.7664* Effective, thus real-time online foil production qualification monitoring can be carried out based on the regression equation.

[0030] S3, online interactive module online interactive judgment of foil quality: through the online pattern measuring instrument to perform real-time detection, obtain the real-time data of online foil parameters, and evaluate the online foil flatness; by setting the qualified standard of the foil of the offline pattern measuring instrument, and feeding it back to the central processor for substitution into the regression equation, the online alarm threshold is obtained; then the online foil flatness is compared with the online alarm threshold, so as to generate an online alarm prompt signal; The pre-set qualified standard of the foil material of the offline pattern meter is: let the offline alarm threshold value Gy of the offline foil flatness Y of the offline pattern data set Qy = 7mm; Substituting Y≤7mm into the regression equation Y=1.455+1.7664*X, we can deduce that X≤3.13mm; Then the online alarm threshold Gx is set to 3, and when the online foil flatness X of the online pattern set Qx is lower than the online alarm threshold Gx, an online alarm prompt signal is generated; Substitute the online foil film parameter feedback detected in real time in the foil production line into the regression equation: =1.455+1.7664* , calculate and obtain real-time offline foil film parameters.

[0031] S4, the display alarm module receives the signal and displays the alarm: by receiving the online alarm prompt signal, the prompt text is edited and the corresponding parameters are displayed, and the alarm prompt operation of the sound signal is performed to prompt the management personnel of the foil production line to perform the corresponding foil processing; The prompt text includes online foil parameters and inferred offline foil parameters, so as to visually monitor the deviation degree of foil flatness, and subsequently increase the data sample size of the regression equation to continuously improve the fitting degree of the regression equation.

[0032] In summary, the present invention collects historical data of offline foil parameters and online foil parameters through a data acquisition module, performs mathematical modeling and regression analysis based on a central processing unit, builds an offline online hybrid model to obtain a regression equation, evaluates the effectiveness of the regression equation, and then feeds back real-time foil parameters to the central processing unit through an online interactive module, substitutes them into the regression equation to determine whether the foil is qualified, generates an online alarm prompt signal and transmits it to a display alarm module, and performs corresponding parameter display and alarm prompt operations; Among them, the online pattern measuring instrument based on this system is separately set up in the foil processing production lines such as slitting machines and carbon coating machines. Real-time multi-laser ranging is carried out during the foil production and transmission process. Intelligent monitoring improves the work efficiency of foil detection; through the dynamic transmission of the foil production process, comprehensive real-time monitoring of the foil processing standards is achieved, ensuring the comprehensiveness and integrity of foil detection; by evaluating the effectiveness of the regression equation, the accuracy of the alarm prompts in the foil processing process is improved.

[0033] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technical personnel in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0034] The above formulas are all dimensionless and numerical calculations. The formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formula are set by technicians in this field according to actual conditions. The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An online pattern measuring instrument dynamic monitoring system, characterized in that: It includes a data acquisition module, a central processing unit, an online interaction module and a display alarm module, wherein the data acquisition module, the central processing unit, the online interaction module and the display alarm module are communicatively connected; The data acquisition module is used to collect foil data: the foil is tested offline and online respectively by an offline pattern measuring instrument and an online pattern measuring instrument, the historical data of offline foil parameters and online foil parameters are obtained, and the historical data of offline foil parameters and online foil parameters are integrated and marked as foil data; The central processing unit is used to construct an offline-online hybrid model: by performing preliminary analysis on the foil data, evaluating the offline foil flatness and the online foil flatness, and integrating and acquiring the pattern data through n groups of foil flatness data values, and by performing mathematical modeling and regression analysis on the pattern data, building an offline-online hybrid model to obtain the regression equation, and evaluating the effectiveness of the regression equation; The online interactive module is used to determine the online qualification of foil materials: through the online pattern measuring instrument, real-time detection is carried out to obtain the real-time data of online foil parameters and evaluate the online foil flatness; by setting the qualified standard of the foil material of the offline pattern measuring instrument and feeding it back to the central processor for substitution into the regression equation, the online alarm threshold is obtained; and then the online foil flatness is compared with the online alarm threshold to generate an online alarm prompt signal; The display alarm module is used to receive signals and display alarms: by receiving online alarm prompt signals, the prompt text is edited and corresponding parameter display and alarm prompt operations are performed.

2. The online pattern measuring instrument dynamic monitoring system according to claim 1, characterized in that: The specific process of obtaining pattern data is as follows: A1, offline foil parameters include offline foil thickness and offline foil range; The offline pattern measuring instrument collects single laser data by reciprocating and stopping, thereby collecting parameters of the produced foil material by reciprocating and stopping for multiple times; extracts Ny offline detection feature points from the motion trajectory of reciprocating and stopping in single laser ranging mode, marks any offline detection feature point as Dy, and marks the offline foil film thickness of the offline detection feature point Dy as Hy, thereby obtaining the offline foil film thickness Hy of Ny offline detection feature points; Sort the offline foil thickness Hy of Ny offline detection feature points to obtain the highest value of the offline foil thickness Hy and the minimum value , thereby calculating the offline foil range Ry, evaluating the offline foil flatness Y through the offline foil range Ry, and constructing the offline pattern data set Qy through n offline foil flatness Y; A2, online foil parameters include online foil thickness and online foil range; The online pattern measuring instrument performs real-time multi-laser distance measurement during the foil production and transmission process, thereby collecting parameters of the foil in production; extracts Nx online detection feature points through the relative motion trajectory of the foil film dynamically transmitted on the multi-laser distance meter, marks any online detection feature point as Dx, and marks the online foil film thickness of the online detection feature point Dx as Hx, thereby obtaining the online foil film thickness Hx of the Nx online detection feature points; Sort the online foil thickness Hx of Nx online detection feature points to obtain the highest value of the online foil thickness Hx and the minimum value , thereby calculating the online foil range Rx, evaluating the online foil flatness X through the online foil range Rx, and constructing the online pattern number set Qx through n online foil flatness X; A3, integrate and mark the offline pattern data set Qy and the online pattern data set Qx as pattern data.

3. The online pattern measuring instrument dynamic monitoring system according to claim 2 is characterized in that: The specific process of building an offline-online hybrid model is as follows: B1, identify and prioritize pattern data through the Pareto chart of standardized effects, and determine the correlation between the offline pattern data set Qy and the online pattern data set Qx; B2, through Pearson modeling, obtain the overall Pearson correlation coefficient γxy to evaluate the correlation between the online pattern set Qx and the offline pattern set Qy; B3, perform regression analysis on the pattern data to evaluate the estimation accuracy of the model coefficients, the fitting effect and significance of the regression model, and thus comprehensively evaluate the effectiveness of the regression equation.

4. The online pattern measuring instrument dynamic monitoring system according to claim 3 is characterized in that: The specific process of regression analysis of pattern data is as follows: B3-1, mathematical function modeling is performed through the offline pattern data set Qy and the online pattern data set Qx to build a regression equation: offline pattern data Y= + *Online pattern data X; B3-2, build a scatter plot of the pattern data through the pattern data, and build a matrix through the scatter plot for regression analysis, so as to obtain the regression analysis data of the pattern data and generate a regression model: ; Then obtain the estimation accuracy evaluation index Xjd of the model coefficients to evaluate the estimation accuracy of the model coefficients; B3-3, evaluating the fitting effect of the regression model, obtaining the model fitting effect evaluation parameters, and then comprehensively obtaining the fitting effect evaluation index Xnh through the model fitting effect evaluation parameters to evaluate the fitting effect of the regression model; B3-4, perform variance analysis on the regression equation to obtain the model's significant factor evaluation parameters, obtain the significance evaluation index Xxz through the model's significant factor evaluation parameters, and evaluate the significance of the regression model; B3-5, and then by combining the estimation accuracy evaluation index Xjd of the model coefficient, the fitting effect evaluation index Xnh of the regression model, and the significance evaluation index Xxz, the regression equation effectiveness index Eff is obtained to comprehensively evaluate the effectiveness of the regression equation.

5. The online pattern measuring instrument dynamic monitoring system according to claim 4 is characterized in that: The specific process of obtaining the overall Pearson correlation coefficient γxy through Pearson modeling is as follows: B2-1, the online pattern data set Qx contains n data values ​​of online foil flatness X, and the overall mean is obtained through the online pattern data set Qx and standard deviation Sx; B2-2, the offline pattern data set Qy contains n offline foil flatness Y data values, and the overall mean is obtained through the offline pattern data set Qy and standard deviation Sy; B2-3, and then obtain the overall covariance and the overall Pearson correlation coefficient γxy; B2-4, set the evaluation interval of the overall Pearson correlation coefficient γxy, and evaluate the correlation between the online pattern set Qx and the offline pattern set Qy through interval comparison.

6. The online pattern measuring instrument dynamic monitoring system according to claim 5, characterized in that: The specific parameters for regression analysis of pattern data are: Regression analysis data include the coefficients of the regression equation, standard errors of the coefficients, T values, P values, and variance inflation factors. The coefficients of the marked independent variables are , the coefficient of the labeled constant term is ; Mark the coefficient of the independent variable The standard error of , marking constant term coefficient The standard error of ; T value = coefficient / standard error of coefficient; Generate a regression model through regression analysis data: ; The model fitting effect evaluation parameters include the determination coefficient Rsq, the adjusted determination coefficient Radj, and the predicted determination coefficient Rpred; The parameters for evaluating significant factors in the model include degrees of freedom DF, adjusted sum of squares AdjSS, adjusted mean square AdjMS, F value and P value.

7. The online pattern measuring instrument dynamic monitoring system according to claim 6 is characterized by: The generation process of the online alarm prompt signal is: The qualified standard of the foil material of the offline pattern meter is preset, and the offline alarm threshold value of the offline foil flatness Y of the offline pattern data set Qy is marked as Gy; Substitute the qualified condition of offline foil flatness Y ≤ offline alarm threshold Gy into the regression equation , thereby setting the online alarm threshold Gx. When the online foil flatness X of the online pattern set Qx is lower than the online alarm threshold Gx, an online alarm prompt signal is generated.

Images