Vacuum coating equipment fault detection method and system based on coating quality

Abstract

The invention provides a vacuum coating equipment fault detection method and system based on coating quality, and relates to the technical field of vacuum coating equipment detection.The method comprises the steps that dependent variables are obtained by adjusting various fault types and independent variables in a normal state; according to the fault contrast data set, constructing a multiple linear regression model of the fault contrast data set; determining the optimal quantile of the fault contrast data set in combination with an information criterion; carrying out regression on each section of fault contrast data set by utilizing a multiple linear regression model to obtain multiple linear regression model coefficients under each section of fault contrast data set; calculating the influence weight of each piece of coating quality data on each influence factor; combining each influence weight to construct a naive Bayes classifier which is used for outputting a fault category and has a weighted value; acquiring real-time coating quality data; and taking the real-time coating quality data as input data of the improved naive Bayes classifier, and outputting a prediction fault category. And the fault detection efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of vacuum coating equipment detection technology, and in particular to a vacuum coating equipment fault detection method and system based on coating quality. Background Art

[0002] Vacuum coating is an industrial process used to apply thin films or coatings on the surface of substrates, usually in a controlled vacuum environment. These coatings can increase the wear resistance, corrosion resistance, anti-reflection or enhance the aesthetics of the material. The equipment usually involves complex mechanical, electrical and electronic systems, including vacuum pumps, deposition sources, control systems, etc. Vacuum coating equipment fault detection refers to the use of various technologies and methods to identify and predict possible faults or problems in vacuum coating equipment.

[0003] Fault detection of vacuum coating equipment is crucial to ensure the continuity and efficiency of the production process. Timely detection and resolution of faults can significantly reduce production interruptions, ensure product quality, reduce maintenance costs, and improve overall equipment reliability. In addition, fault detection can also help prevent possible safety accidents and protect the safety of operators.

[0004] However, the existing fault detection of vacuum coating equipment is still in the stage of manual inspection. This frequent inspection method is harmful to the health of inspectors, prone to subjective experience bias, insufficient fault detection accuracy, and this discrete form of inspection is discontinuous and prone to inspection loopholes, which leads to the inability to detect equipment failures in real time, which not only damages the coating equipment, but also greatly increases the production cost, seriously restricting the promotion and use of vacuum coating equipment. Summary of the invention

[0005] In order to solve the problem that the vacuum coating equipment fault detection in the prior art is still in the manual inspection stage, this frequent inspection method is harmful to the health of the inspectors, is prone to subjective experience bias, and has insufficient fault detection accuracy. In addition, this discrete form of inspection is discontinuous and prone to inspection loopholes, which in turn leads to the failure to discover equipment faults in real time, which not only damages the coating equipment, but also causes a significant increase in production costs, seriously restricting the promotion and use of vacuum coating equipment. The present invention provides a vacuum coating equipment fault detection method and system based on coating quality.

[0006] The technical solution provided by the embodiment of the present invention is as follows: First aspect An embodiment of the present invention provides a vacuum coating equipment fault detection method based on coating quality, comprising: S1: Taking the influencing factors of coating quality data as independent variables and the coating quality data of the substrate as dependent variables, the dependent variables are obtained by adjusting the independent variables under various fault categories and normal states, where the independent variables and dependent variables constitute the fault control data set; S2: Based on the fault control data set, a multivariate linear regression model of the fault control data set is constructed with minimizing the loss value of each influencing factor as the objective function; S3: Determine the optimal quantile of the fault comparison data set in combination with the information criterion, where the optimal quantile is the quantile value when the information criterion is the lowest; S4: segmenting the fault control data set according to the optimal quantile, and regressing each segment of the fault control data set using a multivariate linear regression model to obtain the multivariate linear regression model coefficients under each segment of the fault control data set; S5: Calculate the influence weight of each coating quality data on each influencing factor by combining the optimal quantile and the coefficient of the multivariate linear regression model under each fault control data set, where the influence weight is the coefficient of the independent variable that minimizes the objective function; S6: constructing a naive Bayes classifier with weighted values ​​for outputting a fault category by combining each influence weight, wherein the fault category is a correlation category of the influence factor; S7: Obtain real-time coating quality data; S8: Use the real-time coating quality data as input data of the improved naive Bayes classifier and output the predicted fault category.

[0007] Second aspect An embodiment of the present invention provides a vacuum coating equipment fault detection system based on coating quality, comprising: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the vacuum coating equipment fault detection method based on coating quality as described in the first aspect is implemented.

[0008] The third aspect An embodiment of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the program is executed by a processor, the method for detecting faults of vacuum coating equipment based on coating quality as described in the first aspect is implemented.

[0009] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: In the present invention, the coating quality data of the easily accessible and intuitive product terminal is taken as the detection object, and the equipment fault is obtained through data analysis and mining. In the process of data analysis and mining, a multivariate linear regression model of the fault control data set is constructed with the objective function of minimizing the loss value of each influencing factor. The optimal quantile of the fault control data set is determined in combination with the information criterion to segment the fault control data set that is difficult to classify. Then, the optimal quantile is used to intervene in the regression coefficient, and the influence weight of each coating quality data on each influencing factor is accurately refined and corrected. The influence weight is introduced into the traditional naive Bayes classifier as a weighted value. This can not only effectively avoid the problem that the regression classification process cannot be completed, but also effectively improve the accuracy of the law between the captured coating quality data and the fault category. Finally, the close correlation between the coating quality data and various fault categories is obtained, and the automatic fault detection is completed at all times, which can timely discover equipment faults and effectively improve the fault detection accuracy, reduce the situation of misdiagnosis, reduce the workload of staff, reduce the health hazards of the production environment to staff, and enhance the market competitiveness of coating equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0011] Figure 1 A schematic flow chart of a method for detecting a vacuum coating equipment fault based on coating quality provided by an embodiment of the present invention; Figure 2 A schematic structural diagram of a vacuum coating equipment fault detection system based on coating quality provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0012] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0013] In the embodiments of the present invention, words such as "exemplarily" and "for example" are used to indicate examples, illustrations or explanations. Any embodiment or design described as "example" in the present invention should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of the word "example" is intended to present the concept in a specific way. In addition, in the embodiments of the present invention, the meaning expressed by "and / or" can be both, or it can be either of the two.

[0014] In the embodiments of the present invention, "image" and "picture" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same. "of", "corresponding, relevant" and "corresponding" can sometimes be used interchangeably. It should be noted that when the difference between them is not emphasized, the meanings they intend to express are the same.

[0015] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript such as W1. When the difference is not emphasized, the meanings to be expressed are consistent.

[0016] In order to make the technical problems, technical solutions and advantages to be solved by the present invention more clear, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0017] Reference Manual Attached Figure 1 , showing a schematic flow chart of a vacuum coating equipment fault detection method based on coating quality provided by an embodiment of the present invention.

[0018] The embodiment of the present invention provides a method for detecting a vacuum coating equipment fault based on coating quality, which can be implemented by a vacuum coating equipment fault detection device based on coating quality, which can be a terminal or a server. The processing flow of the method for detecting a vacuum coating equipment fault based on coating quality can include the following steps: S1: The influencing factors of coating quality data are taken as independent variables, and the coating quality data of the substrate is taken as the dependent variable. The dependent variable is obtained by adjusting the independent variables under various fault categories and normal conditions.

[0019] Among them, the independent variables and dependent variables constitute the fault control data set.

[0020] It should be noted that the selection of the coating quality data of the substrate as the dependent variable directly reflects the function and performance of the vacuum coating equipment, and can accurately reveal the direct connection between the equipment operation status and the coating results, making fault diagnosis more intuitive and accurate. By analyzing the relationship between coating quality and various operating parameters, it is possible to effectively identify which factors will affect the final coating effect to what extent, thereby providing a basis for operation optimization and fault prevention. In addition, this method makes data analysis highly target-relevant, because improving coating quality is the ultimate and most important goal in the vacuum coating process. By directly linking fault diagnosis with product quality, it is easier to gain the understanding and support of production line staff, so that necessary adjustments can be quickly implemented in actual operations to reduce misdiagnosis.

[0021] In one possible implementation, the coating quality data includes coating thickness, coating uniformity, coating roughness, coating adhesion and coating wear resistance, and the influencing factors include the vacuum degree of the vacuum coating equipment, the substrate heating temperature, the deposition rate, the target material purity, the target material heating temperature and the target substrate distance.

[0022] It should be noted that when the vacuum degree is insufficient, the residual gas and water vapor in the vacuum chamber will react with the coating material or cause defects. The higher the vacuum degree, the less the impact of residual gas in the air on the quality of the film layer, and the higher the purity of the coating, that is, the better the coating quality. The temperature of the substrate will affect the crystal structure and adhesion of the film layer. Appropriate heating can improve the quality of the film layer. When the heating temperature of the substrate is insufficient, the quality of the coating will decrease. The speed of the deposition rate will affect the structure and density of the film layer. Too fast may lead to uneven film layer, and too slow may increase production costs. The appropriate deposition rate helps to improve the quality of the coating. The purity of the target material. The coating material with higher purity can reduce impurities in the film layer and improve the quality of the coating. Different coating materials meet different functional requirements. The physical and chemical properties of the material directly affect the quality of the film layer. Therefore, the purity of the target material is also an important factor affecting the quality of the coating, and it is also classified as a vacuum coating equipment failure here. During the physical vapor deposition process, the coating material, i.e., the target material, will be heated to its evaporation point, and the evaporated atoms or molecules will migrate to the surface of the substrate in a vacuum and condense. In sputtering, high-energy ions are used to bombard the target material of the coating material, which is atomically projected onto the substrate. In the chemical vapor deposition process, chemical gases decompose at high temperatures and the resulting coating material is deposited on the surface of the substrate. Coating under insufficient heating and temperature will affect the coating quality. Changing the distance between the target and the substrate can control the deposition rate. The closer the distance, the higher the deposition rate because the particles have less loss when reaching the substrate.

[0023] Specifically, each influencing factor is directly related to each basic device of the vacuum coating equipment. When problems occur in each device, it will directly or indirectly affect the coating quality, and the impact on the coating quality varies. The selected coating thickness, coating uniformity, coating roughness, coating adhesion and coating wear resistance are the basic evaluation indicators of coating quality, and they can all be quantified through equipment detection. There is a hidden relationship between the coating quality data and the fault type of the vacuum coating equipment. The selection of quantifiable coating quality data can improve the effectiveness and accuracy of fault detection of the vacuum coating equipment. Moreover, the influencing factors are also quantifiable, and the change in the influencing factors is directly related to the fault type, which is also the data basis for subsequent vacuum coating equipment fault detection.

[0024] In one possible implementation, the fault categories include vacuum pump failure, substrate heating system failure, target bombardment power regulator failure, gas flow controller failure, insufficient target purity, target heating system failure, and target substrate distance controller failure.

[0025] It should be noted that the change in the influencing factors is directly associated with the type of fault. The specific association is that insufficient vacuum is mapped to vacuum pump failure, unsatisfactory substrate heating temperature is mapped to substrate heating system failure, insufficient deposition rate can be mapped to target bombardment power regulator failure of the target bombardment power in physical deposition, or the gas flow rate in the chemical deposition process, i.e., the corresponding gas flow controller failure, insufficient target purity is directly determined as a target purity problem, unsatisfactory target heating temperature is mapped to target heating system failure, and target-substrate spacing controller failure is mapped to target-substrate spacing controller failure when the preset threshold is not met.

[0026] It should be noted that in step S1, quantifiable influencing factors are selected as dependent variables instead of fault types. The reason for this is that the influencing factors corresponding to the fault types are a range of variations, rather than just corresponding to one influencing factor value. When a fault type occurs, the influencing factors may have various parameter values. Through the multi-mapping relationship between the influencing factors and the coating quality data, the hidden relationship between the two can be fully explored, and the fault type can be determined more accurately.

[0027] S2: Based on the fault control data set, a multivariate linear regression model of the fault control data set is constructed with the objective function of minimizing the loss value of each influencing factor.

[0028] Among them, the multiple linear regression model is a widely used statistical method for analyzing the linear relationship between two or more predictor variables (independent variables) and a response variable (dependent variable). Constructing a multiple linear regression model on a fault control data set to minimize the loss value of each influencing factor as the objective function has obvious advantages. First, this method allows accurate quantification of the specific contribution of each influencing factor to the coating quality, thereby providing a scientific basis for in-depth understanding of how each parameter affects the final result. By minimizing the loss value, the model can optimize parameter estimation and ensure the accuracy and reliability of the prediction. In addition, the application of multiple linear regression helps to reveal the interactions and dependencies between different variables, making fault detection and prevention strategies more targeted and efficient. This method can also help identify and correct hidden problems that may lead to a decline in coating quality, thereby optimizing the production process and improving product quality and production efficiency.

[0029] In a possible implementation, S2 specifically includes: S201: Determine the basic multiple linear regression equation: .

[0030] in, Indicates i Coating quality data The inference influence factor value under w represents the coefficient of the independent variable, b Represents a constant.

[0031] S202: Determine the objective function: .

[0032] in, Respectively represent the independent variable coefficient and constant that minimizes the objective function, Indicates i The actual influencing factor value under the coating quality data, that is, the i Coating quality data The corresponding influencing factor value is N Indicates the quantity of coating quality data.

[0033] S203: Perform least squares estimation on the objective function to determine the optimal closed-form solution of the independent variable coefficient and the constant, that is, the independent variable coefficient and the constant with the smallest objective function: ; ; .

[0034] in, E represents the intermediate variable, It means to find partial derivatives.

[0035] S204: Substitute the independent variable coefficients and constants that minimize the objective function into the basic multiple linear regression equation to obtain a multiple linear regression model: .

[0036] It should be noted that by clearly defining the objective function and applying the least squares method, we can ensure that the model parameters that best fit the data are found in a statistical sense, providing a good bias-variance balance and maximizing the explanatory power and predictive ability of the model. Secondly, by accurately calculating the coefficients of each variable, we can directly understand the degree of influence of different factors on the results.

[0037] S3: Determine the optimal quantile of the fault comparison data set in combination with the information criterion.

[0038] Among them, the optimal quantile is the quantile value when the information criterion is the lowest.

[0039] Among them, the information criterion is a tool used to evaluate the goodness of fit of a statistical model. It not only considers the degree of fit of the model to the data, but also includes the penalty for model complexity. The purpose of the information criterion is to find a model that achieves the best balance between complexity and goodness of fit. Quantile is a measure that describes the location of the data distribution and is used to determine whether a certain proportion of data in the data set is lower than (or equal to) this value. Quantile is the point that divides the data into continuous intervals with equal probability. The lowest value of the information criterion represents that the model maintains appropriate simplicity while satisfying the fit of the data, thus achieving the best balance between model complexity and goodness of fit. Such a model is usually considered to be the best choice from the optional models because it can effectively explain the data and avoid the problem of overfitting. Using the information criterion to select the quantile that best fits the data, especially in quantile regression, can help determine the best fit of the model at a given quantile, thereby improving the accuracy and reliability of model predictions.

[0040] It should be noted that the use of information criteria to determine the optimal quantile of the fault control data set ensures that the model will neither overfit nor underfit when fitting the data set, thereby achieving the best balance between model simplicity and prediction accuracy. This strategy allows the model to more effectively capture and explain key changes in the data, thereby improving the accuracy and reliability of fault prediction. By optimizing the quantile selection, this method can also improve the adaptability of the model to different data segments, and enable the multivariate linear regression process to effectively discover certain nonlinear relationships in the data set, further enhancing the model's ability to identify complex fault patterns.

[0041] In a possible implementation, the information criterion includes Akaike information criterion and Bayesian information criterion.

[0042] S3 specifically includes: S301: Determine candidate quantiles, where the candidate quantiles include quartiles, quintiles, deciles, and percentiles.

[0043] S302: Determine the following under the candidate quantile: ; .

[0044] Among them, AIC and BIC represent the Akaike information criterion value and the Bayesian information criterion value, respectively. k represents the number of model parameters, that is, the number of model parameters of the multivariate linear regression model. represents the maximum likelihood estimate of the multivariate linear regression model under alternative quantiles, m represents the sample size at the alternative quantile.

[0045] Among them, the number of model parameters of the multivariate linear regression model here is the coefficient, also known as the slope and intercept. The maximum likelihood estimate of the multivariate linear regression model under the alternative quantile is the probability of occurrence of data in the corresponding segment of the fault control data set when the number of model parameters and sample size are determined under the segmentation of the alternative quantile. The greater the probability of data occurrence, the more accurate the quantile selection is.

[0046] S303: Select the candidate quantile corresponding to the minimum Akaike information criterion value or the Bayesian information criterion value as the optimal target quantile output.

[0047] It should be noted that in the establishment of the multivariate linear regression model, the use of information criteria (AIC and BIC) to select the optimal quantile can determine the quantile based on statistical principles, ensuring that the selected quantile is neither too complex (avoiding overfitting) nor too simple (avoiding underfitting) when fitting the data set, thereby achieving the best balance between the complexity of the model and the accuracy of prediction. Secondly, by comparing the AIC and BIC values ​​at different quantiles, the performance of each model can be objectively evaluated, and the model that best represents the key features of the data can be selected. This strategy not only improves the overall efficiency and effectiveness of the model, but also further enhances the model's ability to diagnose potential faults by focusing on important intervals of the data.

[0048] S4: segment the fault control data set according to the optimal quantile, and regress each segment of the fault control data set using a multivariate linear regression model to obtain the multivariate linear regression model coefficients under each segment of the fault control data set.

[0049] S5: Calculate the influence weight of each coating quality data on each influencing factor by combining the optimal quantile and the coefficient of the multivariate linear regression model under each fault control data set.

[0050] Among them, the influence weight is the independent variable coefficient that minimizes the objective function.

[0051] It should be noted that the optimal quantile is used to segment the fault control data set, and the multivariate linear regression model is applied to each segment. By performing specialized regression analysis on different parts of the data set, the specific behaviors and trends of the data in different intervals can be captured more finely. In addition, using the optimal quantile for data segmentation helps to ensure that each model part is focused on the key areas of the data, thereby improving the explanatory power and prediction accuracy of the overall model. This segmented regression can not only reveal potential patterns in complex data structures, but also provide a more flexible and adaptable modeling method when dealing with non-uniform or heterogeneous data, ultimately making the model more adaptable and responsive to abnormal or deviating from the norm.

[0052] In a possible implementation manner, the influence weight is calculated as follows: .

[0053] in, Indicates j The optimal quantile corresponding to the segment fault comparison data set, Indicates j Segment fault comparison data set i The weight of the coating quality data on the selected influencing factors, Indicates i The influence weight of each coating quality data on the selected influencing factors.

[0054] It should be noted that this calculation method combines the weights of specific influencing factors in each segment with the optimal quantile corresponding to the segment, allowing the weights to more accurately reflect the actual degree of influence of each factor in different data intervals. This method not only increases the sensitivity of the model to subtle changes in the data, but also improves the accuracy and pertinence of weight allocation, so that the final model can better adapt to and reveal the complex relationships in the data. In addition, by calculating specific weights for each data segment, the local characteristics of the data can be effectively captured and utilized, thereby improving the performance of the entire model in prediction and classification. This analysis method is particularly suitable for processing vacuum coating data with high nonlinearity and heterogeneity, ultimately promoting the improvement of fault detection accuracy.

[0055] S6: Combining various impact weights to construct an improved naive Bayes classifier with weighted values ​​for outputting fault categories.

[0056] The fault category is an associated category of the influencing factors.

[0057] Among them, the working mechanism of the naive Bayes classifier is based on Bayes' theorem, which is a basic theorem in statistics that describes conditional probability. Bayes' theorem provides a method to calculate the probability of an event occurring based on the probabilities of certain related events.

[0058] It should be noted that by combining the various influence weights to construct a weighted improved naive Bayes classifier, it can more accurately reflect the actual importance of each influencing factor in predicting the fault category, thereby improving the accuracy of classification. Weighted processing allows the classifier to pay more attention to more critical factors when facing variables with different importance, reducing the interference of less important factors. In addition, this method can effectively overcome the limitations of the feature independence assumption in the traditional naive Bayes classifier, making the model closer to the complex relationships in the real world, thereby providing higher reliability and utility in practical applications. Such a classifier is particularly suitable for processing complex data sets with obvious interdependence between features, improving the overall performance of the fault detection process.

[0059] In a possible implementation, the improved naive Bayes classifier is specifically as follows: .

[0060] in, Indicates that given coating quality data The posterior probability of the occurrence of fault category C is, Indicates the fault type C The prior probability of Indicates i The weight of the coating quality data on the selected influencing factors, Indicates the fault type C Coating quality data x i The probability of occurrence, Indicates a proportional relationship.

[0061] It should be noted that by introducing weight adjustment into the naive Bayes classifier, the classifier can not only consider the existence or non-existence of each influencing factor, but also consider the importance of each factor to the fault classification. Specifically, the contribution of each influencing factor is adjusted by its weight, so that factors with higher diagnostic value for the fault category have greater influence in the model. This weighted method increases the flexibility and accuracy of the model, making the classifier more sensitive to important features and subtle differences in the data. In addition, this method optimizes the strict assumption of feature independence in traditional naive Bayes, and allows certain dependencies between features through weight adjustment, so as to better adapt to the complexity of real-world data and improve the prediction performance of the model in practical applications.

[0062] S7: Obtain real-time coating quality data.

[0063] S8: Use the real-time coating quality data as input data of the improved naive Bayes classifier and output the predicted fault category.

[0064] In a possible implementation manner, the predicted fault category is the fault category corresponding to the maximum posterior probability.

[0065] Among them, the posterior probability is a core concept in Bayesian statistics. It describes the re-evaluation of the probability of a hypothesis or event occurring after certain evidence or information is known. Specifically, the posterior probability is the conditional probability of an event occurring after given observational data. Predicting the fault category to be the fault category corresponding to the largest posterior probability means that when using the naive Bayes classifier for classification, the fault category that maximizes the posterior probability will be selected as the prediction result. This is based on the maximum a posteriori probability criterion in Bayesian decision theory, that is, selecting the category that is most likely to produce the observed data. This method ensures that among all possible fault categories, the one selected is the fault category that is most likely to actually occur under given coating quality data.

[0066] In the actual application process, the specific process of this scheme is to first obtain the fault control data set, and then construct a multivariate linear regression model of the fault control data set with the objective function of minimizing the loss value of each influencing factor. In the process of direct regression analysis of the fault control data set by the constructed model, it may be difficult to find a linear relationship due to the multilateral nature of the data set. In order to compensate for this shortcoming, the optimal quantile is first determined to segment the data set, that is, the quantile value is determined with the lowest information criterion as the goal, and the segmented quantile that best suits the data is selected to automatically balance the complexity of the model and its ability to fit the data, thereby selecting a quantile that is neither overfitting nor underfitting, and then using The quantile divides the data set into segments. At this time, the probability of a linear relationship between each segment of the data set is greatly improved, and the regression coefficient obtained by the multivariate linear regression equation is more accurate and reliable. Then the obtained regression coefficient is multiplied by the optimal quantile as the influence weight of the corresponding coating quality data on the selected influencing factors, and the correlation between the coating quality data and the selected influencing factors is further accurately mined. Based on the influence weight, an improved naive Bayes classifier with a weighted value is established to improve the limitations and ambiguity of the conditional independence assumption of the traditional naive Bayes classifier, accurately correct the relationship between each coating quality data and the influencing factors, and improve the classification accuracy of the improved naive Bayes classifier. Finally, the improved naive Bayes classifier is used to classify the real-time coating quality data, obtain the fault category with the highest posterior probability and output it. Compared with the traditional vacuum coating equipment fault detection method, this method uses intuitive coating quality data as the independent variable. It will trigger the detection of fault categories only when the coating quality data does not meet the production target. It monitors the coating equipment at all times with the goal of ensuring that production efficiency and end products meet expectations, improves production efficiency, avoids frequent shutdowns caused by misdiagnosis, and effectively improves the production benefits and market competitiveness of the equipment.

[0067] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: In the present invention, the coating quality data of the easily accessible and intuitive product terminal is taken as the detection object, and the equipment fault is obtained through data analysis and mining. In the process of data analysis and mining, a multivariate linear regression model of the fault control data set is constructed with the objective function of minimizing the loss value of each influencing factor. The optimal quantile of the fault control data set is determined in combination with the information criterion to segment the fault control data set that is difficult to classify. Then, the optimal quantile is used to intervene in the regression coefficient, and the influence weight of each coating quality data on each influencing factor is accurately refined and corrected. The influence weight is introduced into the traditional naive Bayes classifier as a weighted value. This can not only effectively avoid the problem that the regression classification process cannot be completed, but also effectively improve the accuracy of the law between the captured coating quality data and the fault category. Finally, the close correlation between the coating quality data and various fault categories is obtained, and the automatic fault detection is completed at all times, which can timely discover equipment faults and effectively improve the fault detection accuracy, reduce the situation of misdiagnosis, reduce the workload of staff, reduce the health hazards of the production environment to staff, and enhance the market competitiveness of coating equipment.

[0068] Reference Manual Attached Figure 2 , showing a schematic structural diagram of a vacuum coating equipment fault detection system based on coating quality provided by the present invention.

[0069] The present invention further provides a vacuum coating equipment fault detection system 20 based on coating quality, which is applied to the above-mentioned vacuum coating equipment fault detection method based on coating quality, and comprises: Processor 201.

[0070] The memory 202 stores computer-readable instructions. When the computer-readable instructions are executed by the processor 201 , a vacuum coating equipment fault detection method based on coating quality as in the method embodiment is implemented.

[0071] The vacuum coating equipment fault detection system 20 based on coating quality provided by the present invention can execute the above-mentioned vacuum coating equipment fault detection method based on coating quality and achieve the same or similar technical effects. To avoid repetition, the present invention will not go into details.

[0072] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: In the present invention, the coating quality data of the easily accessible and intuitive product terminal is taken as the detection object, and the equipment fault is obtained through data analysis and mining. In the process of data analysis and mining, a multivariate linear regression model of the fault control data set is constructed with the objective function of minimizing the loss value of each influencing factor. The optimal quantile of the fault control data set is determined in combination with the information criterion to segment the fault control data set that is difficult to classify. Then, the optimal quantile is used to intervene in the regression coefficient, and the influence weight of each coating quality data on each influencing factor is accurately refined and corrected. The influence weight is introduced into the traditional naive Bayes classifier as a weighted value. This can not only effectively avoid the problem that the regression classification process cannot be completed, but also effectively improve the accuracy of the law between the captured coating quality data and the fault category. Finally, the close correlation between the coating quality data and various fault categories is obtained, and the automatic fault detection is completed at all times, which can timely discover equipment faults and effectively improve the fault detection accuracy, reduce the situation of misdiagnosis, reduce the workload of staff, reduce the health hazards of the production environment to staff, and enhance the market competitiveness of coating equipment.

[0073] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0074] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0075] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0076] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0077] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0078] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0079] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0080] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described equipment, devices and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0081] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0082] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0083] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0084] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0085] An embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method for detecting a vacuum coating equipment fault based on coating quality as described in the method embodiment is implemented.

[0086] A computer-readable storage medium provided by the present invention can implement the steps and effects of the vacuum coating equipment fault detection method based on coating quality of the above method embodiment. To avoid repetition, the present invention will not go into details.

[0087] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: In the present invention, the coating quality data of the easily accessible and intuitive product terminal is taken as the detection object, and the equipment fault is obtained through data analysis and mining. In the process of data analysis and mining, a multivariate linear regression model of the fault control data set is constructed with the objective function of minimizing the loss value of each influencing factor. The optimal quantile of the fault control data set is determined in combination with the information criterion to segment the fault control data set that is difficult to classify. Then, the optimal quantile is used to intervene in the regression coefficient, and the influence weight of each coating quality data on each influencing factor is accurately refined and corrected. The influence weight is introduced into the traditional naive Bayes classifier as a weighted value. This can not only effectively avoid the problem that the regression classification process cannot be completed, but also effectively improve the accuracy of the law between the captured coating quality data and the fault category. Finally, the close correlation between the coating quality data and various fault categories is obtained, and the automatic fault detection is completed at all times, which can timely discover equipment faults and effectively improve the fault detection accuracy, reduce the situation of misdiagnosis, reduce the workload of staff, reduce the health hazards of the production environment to staff, and enhance the market competitiveness of coating equipment.

[0088] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0089] There are a few points to note: (1) The drawings of the embodiments of the present invention only involve structures related to the embodiments of the present invention. Other structures may refer to conventional designs.

[0090] (2) For the sake of clarity, in the drawings used to describe the embodiments of the present invention, the thickness of layers or regions is exaggerated or reduced, that is, these drawings are not drawn according to the actual scale. It is understood that when an element such as a layer, film, region or substrate is referred to as being "on" or "under" another element, the element may be "directly" "on" or "under" the other element or there may be intermediate elements.

[0091] (3) In the absence of conflict, the embodiments of the present invention and the features therein may be combined with each other to obtain new embodiments.

[0092] The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. The protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for detecting faults of vacuum coating equipment based on coating quality, characterized in that: include: S1: taking the influencing factors of the coating quality data as independent variables and the coating quality data of the substrate as dependent variables, and obtaining the dependent variables by adjusting the independent variables under various fault categories and normal states, wherein the independent variables and the dependent variables constitute a fault control data set; S2: According to the fault control data set, a multivariate linear regression model of the fault control data set is constructed with minimizing the loss value of each influencing factor as the objective function; S3: Determine the optimal quantile of the fault comparison data set in combination with the information criterion, wherein the optimal quantile is the quantile value when the information criterion is the lowest; S4: segmenting the fault control data set according to the optimal quantile, and regressing each segment of the fault control data set using the multivariate linear regression model to obtain the multivariate linear regression model coefficients for each segment of the fault control data set; S5: Calculate the influence weight of each coating quality data on each influencing factor by combining the optimal quantile and the coefficient of the multivariate linear regression model under each fault control data set, wherein the influence weight is the independent variable coefficient that minimizes the objective function; S6: constructing an improved naive Bayes classifier with a weighted value for outputting a fault category in combination with each influence weight, wherein the fault category is a correlation category of the influence factor; S7: Obtain real-time coating quality data; S8: Using the real-time coating quality data as input data of the improved naive Bayes classifier, and outputting a predicted fault category.

2. The vacuum coating equipment fault detection method based on coating quality according to claim 1, characterized in that: The coating quality data includes coating thickness, coating uniformity, coating roughness, coating adhesion and coating wear resistance, and the influencing factors include the vacuum degree of vacuum coating equipment, substrate heating temperature, deposition rate, target purity, target heating temperature and target substrate distance.

3. The method for detecting vacuum coating equipment faults based on coating quality according to claim 1, characterized in that: The fault categories include vacuum pump failure, substrate heating system failure, target bombardment power regulator failure, gas flow controller failure, insufficient target purity, target heating system failure and target substrate distance controller failure.

4. The method for detecting vacuum coating equipment faults based on coating quality according to claim 1, characterized in that: The S2 specifically includes: S201: Determine the basic multiple linear regression equation: ； in, Indicates i Coating quality data The inference influence factor value under w represents the coefficient of the independent variable, b represents a constant; S202: Determine the objective function: ； in, represent the independent variable coefficient and constant that minimize the objective function, respectively. Indicates i The actual influencing factor value under the coating quality data, that is, the value of the first i Coating quality data The corresponding influencing factor value is N Indicates the quantity of coating quality data; S203: Perform least squares estimation on the objective function to determine the optimal closed-form solution of the independent variable coefficient and the constant, that is, the independent variable coefficient and the constant with the minimum objective function: ； ； ； in, E represents the intermediate variable, It means to find partial derivatives; S204: Substitute the independent variable coefficients and constants that minimize the objective function into the basic multiple linear regression equation to obtain the multiple linear regression model: 。 5. The method for detecting vacuum coating equipment faults based on coating quality according to claim 1, characterized in that: The information criteria include Akaike information criterion and Bayesian information criterion; The S3 specifically includes: S301: Determine candidate quantiles, wherein the candidate quantiles include quartiles, quintiles, deciles and percentiles; S302: Determine under the candidate quantile: ； ； Wherein, AIC and BIC represent the Akaike information criterion value and the Bayesian information criterion value, respectively. k represents the number of model parameters, i.e., the number of model parameters of the multivariate linear regression model, represents the maximum likelihood estimate of the multivariate linear regression model under the alternative quantile, m represents the sample size at the stated alternative quantile; S303: Select the candidate quantile corresponding to the minimum Akaike information criterion value or the Bayesian information criterion value as the optimal target quantile output.

6. The method for detecting vacuum coating equipment faults based on coating quality according to claim 1, characterized in that: The calculation method of the influence weight is specifically as follows: ； in, Indicates j The optimal quantile corresponding to the segment fault comparison data set, Indicates j Segment fault comparison data set i The weight of the coating quality data on the selected influencing factors, Indicates i The influence weight of each coating quality data on the selected influencing factors.

7. The method for detecting vacuum coating equipment faults based on coating quality according to claim 1, characterized in that: The improved naive Bayes classifier is specifically: ； in, Indicates that given coating quality data The posterior probability of the occurrence of fault category C is, Indicates the fault type C The prior probability of Indicates i The weight of the coating quality data on the selected influencing factors, Indicates the fault type C Coating quality data x i The probability of occurrence, Indicates a proportional relationship.

8. The method for detecting vacuum coating equipment faults based on coating quality according to claim 7, characterized in that: The predicted fault category is the fault category corresponding to the maximum posterior probability.

9. A vacuum coating equipment fault detection system based on coating quality, characterized in that: include: processor; A memory having computer-readable instructions stored thereon, wherein when the computer-readable instructions are executed by the processor, the vacuum coating equipment fault detection method based on coating quality as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the vacuum coating equipment fault detection method based on coating quality as described in any one of claims 1 to 8 is implemented.

Images