Real-time monitoring system for photovoltaic glass coating thickness

Abstract

The application discloses a real-time monitoring system for photovoltaic glass coating thickness, and relates to the technical field of real-time monitoring, comprising: obtaining a first and a second spraying sub-task of a photovoltaic glass coating task and task information thereof, and determining first and second preset coating thickness real-time monitoring parameters; configuring a first spraying device to complete first layer coating spraying, real-time monitoring of coating thickness, and obtaining a first thickness monitoring result set; performing longitudinal compliance analysis and horizontal correlation analysis on the result set, calculating a first longitudinal influence coefficient and a first horizontal correlation influence coefficient; adjusting the first preset monitoring parameter, optimizing first spraying task monitoring; adjusting the second preset monitoring parameter, optimizing second spraying task monitoring configuration; performing second layer spraying, and real-time monitoring of the photovoltaic glass coating thickness. The application solves the technical problem that the prior art cannot accurately monitor the coating thickness distribution, and achieves the technical effect of accurate monitoring of the photovoltaic glass coating thickness.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of real-time monitoring, and particularly relates to a real-time monitoring system for the thickness of a photovoltaic glass coating. BACKGROUND

[0002] In the field of photovoltaic glass manufacturing, with the advancement of industrialization and the continuous improvement of product quality requirements, the demand for quality control of photovoltaic glass coatings is also rapidly increasing. These coatings usually include anti-reflection coatings, anti-reflection film coatings, etc., with multiple layers and complex characteristics, often requiring high uniformity and precise thickness control during the spraying process. If the coating thickness cannot be effectively monitored and adjusted, not only will it affect the optical performance of photovoltaic glass, but also may lead to a decline in product performance and increase production costs. Traditional coating thickness monitoring methods mainly rely on offline measurement and fixed parameter spraying, which lack real-time performance and are difficult to meet the demand for high-precision coating quality monitoring. SUMMARY

[0003] The present application provides a real-time monitoring system for the thickness of a photovoltaic glass coating, which aims to solve the technical problem that existing technologies cannot accurately monitor the coating thickness distribution.

[0004] In view of the above problems, the present application provides a real-time monitoring system for the thickness of a photovoltaic glass coating.

[0005] In a first aspect of the present application, a real-time monitoring system for the thickness of a photovoltaic glass coating is provided, which comprises:

[0006] The task acquisition module acquires a first spraying sub-task and a second spraying sub-task of a target photovoltaic glass coating task and corresponding first spraying sub-task information and second spraying sub-task information, wherein each spraying sub-task information in the first spraying sub-task information and the second spraying sub-task information comprises coating type, coating average thickness information, and coating thickness uniformity information; the retrieval module performs retrieval in a preset coating monitoring space with the first spraying sub-task information and the second spraying sub-task information as indexes to acquire first preset coating thickness real-time monitoring parameters and second preset coating thickness real-time monitoring parameters; the first thickness monitoring module performs parameter configuration on a first spraying device according to the first spraying sub-task information of the first spraying sub-task, performs coating spraying on a photovoltaic glass surface by using the first spraying device after the configuration is completed, performs coating thickness monitoring in a first preset monitoring window according to the first preset coating thickness real-time monitoring parameters, and obtains a first thickness monitoring result set; the horizontal and vertical analysis module performs vertical compliance analysis and horizontal correlation analysis on the first thickness monitoring result set to determine a first vertical influence coefficient and a first horizontal correlation influence coefficient; the second thickness monitoring module adjusts the first preset coating thickness real-time monitoring parameters according to the size of the first vertical influence coefficient to obtain first adjusted coating thickness real-time monitoring parameters, and performs real-time monitoring on the photovoltaic glass coating thickness of the first spraying sub-task in a next preset monitoring window according to the first adjusted coating thickness real-time monitoring parameters; the monitoring parameter adjustment module adjusts the second preset coating thickness real-time monitoring parameters according to the size of the first horizontal correlation influence coefficient to obtain second influence coating thickness real-time monitoring parameters; the coating spraying module performs parameter configuration on a second spraying device according to the second spraying sub-task information of the second spraying sub-task, performs coating spraying on the photovoltaic glass surface in the first preset monitoring window that has completed the first spraying sub-task by using the second spraying device after the configuration is completed, and performs real-time monitoring on the photovoltaic glass coating thickness according to the second influence coating thickness real-time monitoring parameters.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0008] The application obtains a first spraying sub-task and a second spraying sub-task of a target photovoltaic glass coating task, and corresponding first spraying sub-task information and second spraying sub-task information, wherein each spraying sub-task information in the first spraying sub-task information and the second spraying sub-task information comprises coating type, coating average thickness information, and coating thickness uniformity information; the first preset coating thickness real-time monitoring parameter and the second preset coating thickness real-time monitoring parameter are obtained by searching in a preset coating monitoring space with the first spraying sub-task information and the second spraying sub-task information as indexes respectively; the first spraying equipment is configured according to the first spraying sub-task information of the first spraying sub-task, the photovoltaic glass surface is coated by using the configured first spraying equipment, the coating thickness is monitored in the first preset monitoring window according to the first preset coating thickness real-time monitoring parameter, and a first thickness monitoring result set is obtained; the first thickness monitoring result set is subjected to longitudinal compliance analysis and horizontal correlation analysis to determine a first longitudinal influence coefficient and a first horizontal correlation influence coefficient; the first preset coating thickness real-time monitoring parameter is adjusted according to the size of the first longitudinal influence coefficient, a first adjusted coating thickness real-time monitoring parameter is obtained, and the photovoltaic glass coating thickness in the next preset monitoring window is monitored in real time according to the first adjusted coating thickness real-time monitoring parameter; the second preset coating thickness real-time monitoring parameter is adjusted according to the size of the first horizontal correlation influence coefficient, a second influence coating thickness real-time monitoring parameter is obtained; the second spraying equipment is configured according to the second spraying sub-task information of the second spraying sub-task, the photovoltaic glass surface on which the first spraying sub-task is completed in the first preset monitoring window is coated by using the configured second spraying equipment, and the photovoltaic glass coating thickness is monitored in real time according to the second influence coating thickness real-time monitoring parameter. The application solves the technical problem that the prior art cannot accurately monitor the coating thickness distribution, and achieves the technical effect of accurately monitoring the photovoltaic glass coating thickness by task information acquisition, coating monitoring parameter searching, spraying equipment configuration and real-time thickness monitoring, combined with longitudinal compliance analysis and horizontal correlation analysis, dynamic adjustment of monitoring parameters, optimization of spraying process, and the like. BRIEF DESCRIPTION OF DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort.

[0010] Figure 1 The real-time monitoring system structure schematic diagram of the photovoltaic glass coating thickness provided by the embodiment of the present application.

[0011] Figure 2This is a schematic diagram of the real-time monitoring method for the coating thickness of photovoltaic glass provided in an embodiment of this application.

[0012] Explanation of reference numerals in the attached figures: Task acquisition module 11, retrieval module 12, first thickness monitoring module 13, horizontal and vertical analysis module 14, second thickness monitoring module 15, monitoring parameter adjustment module 16, coating spraying module 17. Detailed Implementation

[0013] This application provides a real-time monitoring system for photovoltaic glass coating thickness, addressing the technical problem that existing technologies cannot accurately monitor coating thickness distribution. By acquiring task information, retrieving coating monitoring parameters, configuring spraying equipment, and monitoring thickness in real time, combined with vertical compliance analysis and horizontal correlation analysis, the system dynamically adjusts monitoring parameters and optimizes the spraying process, thereby achieving the technical effect of accurately monitoring the thickness of photovoltaic glass coating.

[0014] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0015] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0016] Example 1, as Figure 1 As shown in the embodiments of this application, a real-time monitoring system for the coating thickness of photovoltaic glass is provided for performing tasks such as... Figure 2 The method for real-time monitoring of photovoltaic glass coating thickness shown includes:

[0017] The task acquisition module 11 acquires the first and second spraying sub-tasks of the target photovoltaic glass coating task, as well as the corresponding first and second spraying sub-task information. Each spraying sub-task information in the first and second spraying sub-task information includes coating type, average coating thickness information, and coating thickness uniformity information.

[0018] In the embodiments of the present application, the task acquisition module receives a first spraying sub-task and a second spraying sub-task of a target photovoltaic glass coating task from a production management system, and corresponding first spraying sub-task information and second spraying sub-task information. The first spraying sub-task is used to coat a primer coating, such as an anti-reflection coating, on the photovoltaic glass. The second spraying sub-task is performed after the first spraying sub-task is completed, and is usually used to coat a top coating. The first spraying sub-task information and the second spraying sub-task information include operating parameters of the equipment when the corresponding task is performed.

[0019] Each spraying sub-task information in the received first spraying sub-task information and second spraying sub-task information includes coating type, coating average thickness information, and coating thickness uniformity information. The coating type specifies the characteristics and purpose of the spraying material; the coating average thickness information indicates the overall thickness of the target coating after spraying, which is usually in units of microns; and the coating thickness uniformity information reflects the uniformity of the thickness distribution of the coating on the glass surface.

[0020] The retrieval module 12 retrieves in a preset coating monitoring space with the first spraying sub-task information and the second spraying sub-task information as indexes, and obtains first preset coating thickness real-time monitoring parameters and second preset coating thickness real-time monitoring parameters.

[0021] In the embodiments of the present application, the retrieval module retrieves in a preset coating monitoring space with the first spraying sub-task information and the second spraying sub-task information as indexes, and locates a first target space point and a second target space point, respectively.

[0022] Then, the k-nearest neighbor algorithm is used to select k nearest neighbor sample points closest to the target space point from a plurality of sample space points, to form a first nearest neighbor sample set and a second nearest neighbor sample set. Finally, the sample coating thickness real-time monitoring parameters in each set are iteratively calculated, and the mean value is taken to generate the first preset coating thickness real-time monitoring parameters and the second preset coating thickness real-time monitoring parameters.

[0023] Further, in the system provided by the embodiments of the present application, the retrieval module 12 is further used for:

[0024] The preset coating monitoring space is constructed, wherein the coating monitoring space includes a plurality of sample space points, each sample space point corresponds to a sample spraying sub-task information, and each sample space point has an identification of a sample preset coating thickness real-time monitoring parameter; the first spraying sub-task information and the second spraying sub-task information are respectively input into the preset coating monitoring space, to obtain a first target space point and a second target space point; k sample space points closest to the first target space point and the second target space point are respectively extracted from the plurality of sample space points, to obtain a first neighbor sample space point set and a second neighbor sample space point set, wherein k is an integer greater than or equal to 3; a mean value of a sample coating thickness real-time monitoring parameter set corresponding to the first neighbor sample space point set is iteratively calculated, to obtain a first preset coating thickness real-time monitoring parameter; a mean value of a sample coating thickness real-time monitoring parameter set corresponding to the second neighbor sample space point set is iteratively calculated, to obtain a second preset coating thickness real-time monitoring parameter.

[0025] In the embodiments of the present application, a three-dimensional coating monitoring space is first constructed based on historical spraying task data, and each historical spraying sub-task information is stored in a structured manner by using database storage technology. Specifically, the coating type is mapped to the x-axis of the three-dimensional coordinate system, the coating average thickness is mapped to the y-axis, and the coating thickness uniformity is mapped to the z-axis, thereby generating a sample space point. Each sample space point corresponds to a historical spraying task and is labeled with a sample preset coating thickness real-time monitoring parameter, such as a monitoring frequency and a window time.

[0026] Next, the retrieval module receives the first spraying sub-task information, including the coating type, the coating average thickness, and the coating thickness uniformity, and the like. By using the coordinate mapping method, the first spraying sub-task information is projected into the coating monitoring space, and the first target space point corresponding thereto is located. Similarly, the second spraying sub-task information is input into the coating monitoring space in the same way, and the second target space point is determined. These target space points are matching positions of the current spraying task and the historical task in parameters, and serve as a basis for further retrieval.

[0027] In order to extract the samples most similar to the target points from the monitoring space, the k-nearest neighbor algorithm is used to calculate the Euclidean distance between the target space points and all sample space points. The sample points are sorted according to the calculated distances, and the k sample points with the smallest distances (k is an integer greater than or equal to 3) are selected. These sample points form the first neighbor sample space point set and the second neighbor sample space point set, respectively.

[0028] Next, the monitoring parameters in the near neighbor sample set are iteratively calculated. For the first near neighbor sample point set, the sample coating thickness real-time monitoring parameters corresponding to each sample point are extracted, and the mean value is calculated, which is taken as the first preset coating thickness real-time monitoring parameter. Similarly, the same calculation is performed on the second near neighbor sample point set to obtain the second preset coating thickness real-time monitoring parameter.

[0029] Further, the system provided by the application embodiment is further used for:

[0030] A three-dimensional space coordinate system is obtained, wherein the x-axis of the three-dimensional space coordinate system is a coating type, the y-axis is coating average thickness information, and the z-axis is coating thickness uniformity information; a plurality of sample spraying sub-task information and a plurality of corresponding sample coating thickness real-time monitoring parameters are obtained; the plurality of sample spraying sub-task information is input into the three-dimensional space coordinate system to obtain a plurality of sample space points, and the plurality of sample coating thickness real-time monitoring parameters are used to identify the plurality of sample space points; and the preset coating monitoring space is obtained according to the identified plurality of sample space points and the three-dimensional space coordinate system.

[0031] In the application embodiment, a three-dimensional space coordinate system is first obtained, and the x-axis of the coordinate system represents a coating type, the y-axis represents coating average thickness information, and the z-axis represents coating thickness uniformity information. The three dimensions jointly define the characteristics of a coating task, wherein the coating type distinguishes different material coatings (for example, anti-reflective coatings, anti-reflection film coatings, etc.); the coating average thickness information is the target value of the overall thickness after spraying, in units of microns; and the coating thickness uniformity information represents the uniformity of the coating thickness distribution.

[0032] Next, a plurality of sample spraying sub-task information and a plurality of corresponding sample coating thickness real-time monitoring parameters are obtained from a historical spraying task database. The sample spraying sub-task information includes a coating type, a coating average thickness, and a coating thickness uniformity, and describes the key characteristics of a spraying task. At the same time, monitoring parameters related to each task are extracted, such as a sampling frequency, a monitoring window time, and a target thickness range, etc., and these parameters are recorded to provide a reference for subsequent monitoring processes. Through data collection, a sample data set containing task characteristics and monitoring parameters is formed.

[0033] Subsequently, the plurality of sample spraying sub-task information is input into a three-dimensional space coordinate system, each sample spraying sub-task information is traversed, and the sample spraying sub-task information is mapped into a point in the three-dimensional space according to the coating type (x-axis), the average coating thickness (y-axis), and the coating thickness uniformity (z-axis) to obtain a plurality of sample space points. Then, each sample space point is identified, and the corresponding coating thickness real-time monitoring parameter is attached to each sample point. The identification process associates each sample point with its sampling frequency, monitoring window time, and thickness target range through an index mechanism.

[0034] After the identification is completed, the plurality of sample space points and the three-dimensional space coordinate system jointly constitute a preset coating monitoring space.

[0035] The first thickness monitoring module 13 performs parameter configuration on the first spraying device according to the first spraying sub-task information of the first spraying sub-task, uses the first spraying device after the parameter configuration to perform coating spraying on the surface of the photovoltaic glass, and performs coating thickness monitoring according to the first preset coating thickness real-time monitoring parameter within the first preset monitoring window to obtain a first thickness monitoring result set.

[0036] In the embodiment of the present application, the first thickness monitoring module performs parameter configuration on the first spraying device according to the device running parameters determined in the first spraying sub-task information of the first spraying sub-task. Parameter configuration refers to adjusting the running parameters of the spraying device, such as spraying pressure, coating flow rate, and spraying area range.

[0037] After the device configuration is completed, the first spraying device starts the spraying process to coat the surface of the photovoltaic glass. During the spraying process, the first thickness monitoring module starts thickness monitoring based on the first preset monitoring window. The first preset monitoring window is a time range that defines the start time and end time of the monitoring, for example, a time period from 5 seconds to 15 seconds after the spraying starts.

[0038] According to the first preset coating thickness real-time monitoring parameter, the monitoring module determines the sampling frequency (such as the number of measurements per second) and the measurement time interval (such as measuring once every 100 milliseconds) within the monitoring window. Within the time range of the monitoring window, the thickness of the coating on the surface of the photovoltaic glass is continuously measured by a dedicated thickness measurement device, such as a laser thickness measurement instrument or an ultrasonic thickness detector. These measurement data are recorded in chronological order to form a dynamic data sequence of the coating thickness. Finally, the data of the entire monitoring process are arranged into a first thickness monitoring result set, including the coating thickness value at each time point within the monitoring window and the corresponding time stamp information.

[0039] The horizontal and vertical analysis module 14 performs longitudinal compliance analysis and horizontal correlation analysis on the first thickness monitoring result set to determine a first longitudinal influence coefficient and a first horizontal correlation influence coefficient.

[0040] In the embodiment of the present application, the transverse and longitudinal analysis module first extracts the first coating average thickness information and the first coating thickness uniformity information in the first spraying sub-task information. In longitudinal compliance analysis, the transverse and longitudinal analysis module calculates the difference between the thickness measurement value in the first thickness monitoring result set and the first coating average thickness information point by point to generate a first coating thickness deviation value set. Then, the deviation value set is subjected to central value screening to determine the first coating thickness deviation central value representing the overall deviation trend. At the same time, the fluctuation variance of the thickness deviation value set is calculated to obtain the first deviation variance, which is used to quantify the overall stability of the thickness deviation. In combination with the thickness deviation central value and the deviation variance, the module completes the longitudinal compliance analysis and finally calculates the first longitudinal influence coefficient.

[0041] In transverse correlation analysis, the transverse and longitudinal analysis module directly calculates the fluctuation variance of the thickness measurement value of the first thickness monitoring result set to quantify the amplitude of the thickness change between each monitoring point. According to the size of the fluctuation variance, the first transverse correlation influence coefficient is determined.

[0042] Further, the system provided by the application embodiment is used, and the transverse and longitudinal analysis module 14 is further used for:

[0043] extracting the first coating average thickness information and the first coating thickness uniformity information of the first spraying sub-task information; performing longitudinal compliance analysis on the first thickness monitoring result set according to the first coating average thickness information and the first coating thickness uniformity information to determine a first longitudinal influence coefficient; and calculating the fluctuation variance of the first thickness monitoring result set to obtain a first transverse correlation influence coefficient.

[0044] In the embodiment of the present application, the first coating average thickness information and the first coating thickness uniformity information of the first spraying sub-task information are first extracted. The first coating average thickness information is the target thickness value of the spraying task; and the first coating thickness uniformity information is the requirement for the uniformity of the coating distribution.

[0045] In longitudinal compliance analysis, the actual thickness value of each monitoring point is first extracted from the first thickness monitoring result set, compared with the first coating average thickness information, and the thickness deviation value is calculated point by point to form a thickness deviation value set. Subsequently, the thickness deviation value set is subjected to statistical processing, and the overall deviation trend is extracted by calculating the central value. In addition, the dispersion degree of the thickness deviation value set is quantified, the variance is calculated, and the fluctuation amplitude of the data distribution is evaluated. The smaller the variance, the more concentrated the thickness data is distributed around the target value, and the lower the overall deviation degree. In combination with the analysis results of the central value and the variance, the first longitudinal influence coefficient is finally calculated.

[0046] In the transverse correlation analysis, the distribution of the thickness measurement data in space is further analyzed. By extracting the thickness values of adjacent monitoring points in the first thickness monitoring result set one by one, the thickness difference between adjacent points is calculated to form a point-to-point difference set for recording the spatial variation of the coating thickness. Then, the point-to-point difference set is statistically processed, and the stability of the thickness variation between adjacent monitoring points is quantified by calculating the variance. The smaller the variance, the more uniform the thickness variation, and the more consistent the distribution of the coating on the surface. Finally, the first transverse correlation influence coefficient is obtained according to the calculation result.

[0047] Further, the system provided by the application embodiment is further provided, wherein the transverse and longitudinal analysis module 14 is further used for:

[0048] According to the first coating average thickness information and the first thickness monitoring result set, a difference calculation is performed to obtain a first coating thickness deviation value set; a central value screening is performed on the first coating thickness deviation value set to determine a first coating thickness deviation central value; a fluctuation variance of the first coating thickness deviation value set is calculated to obtain a first deviation variance; and a longitudinal compliance analysis is performed in combination with the first coating thickness deviation central value, the first deviation variance and first coating thickness uniformity information to obtain the first longitudinal influence coefficient.

[0049] In the application embodiment, first, the actual thickness value of each monitoring point is extracted from the first thickness monitoring result set, and a difference calculation is performed with the first coating average thickness information to generate a first coating thickness deviation value set.

[0050] Next, the first coating thickness deviation value set is subjected to central value screening to determine the main trend of thickness deviation, and a first coating thickness deviation central value is obtained.

[0051] Then, the first coating thickness deviation value set is subjected to quantitative analysis of fluctuation amplitude to calculate a first deviation variance. Specifically, each deviation value is subtracted from the first coating thickness deviation central value to calculate its deviation degree, and then the deviation degrees are squared to eliminate the influence of positive and negative differences on fluctuation analysis. Then, the squared data is accumulated, and the average value of the accumulated result is taken to obtain the first deviation variance.

[0052] After the calculation of the central value and the deviation variance is completed, the longitudinal compliance analysis is performed in combination with the first coating thickness deviation central value, the first deviation variance, and the first coating thickness uniformity information. Specifically, the three features are input into a pre-constructed feedforward neural network model. The feedforward neural network performs comprehensive analysis on the input features through multiple layers of nonlinear mapping. The training data of the neural network model includes the coating thickness deviation central value, the thickness deviation variance, and the coating thickness uniformity information of the historical spraying task records as input features, and the corresponding longitudinal influence coefficient as the target output label. The longitudinal influence coefficient is provided by technical experts based on the actual quality performance of the historical tasks, and reflects the longitudinal compliance level of the task. During the training process, the neural network optimizes the weight and bias values through multiple forward propagation and back propagation, and learns the complex correlation between the input features and the longitudinal influence coefficient.

[0053] The fully trained feedforward neural network receives the feature input of the current task in the actual task, and calculates the first longitudinal influence coefficient through calculation.

[0054] Through the foregoing process, the first longitudinal influence coefficient is finally obtained.

[0055] Further, in the system provided by the application embodiment, the transverse and longitudinal analysis module 14 is further used for:

[0056] The mean value of the first coating thickness deviation value set is calculated to obtain the first coating thickness deviation mean value; the initial weighted center is updated in the first coating thickness deviation value set in combination with the weighted center updating formula with the first coating thickness deviation mean value as the initial weighted center, until the preset iteration update times are met, and the updated weighted center obtained last time is taken as the first coating thickness deviation central value.

[0057] In the application embodiment, first, the mean value is calculated from the first coating thickness deviation value set to obtain the first coating thickness deviation mean value.

[0058] Then, the first coating thickness deviation mean value is taken as the initial weighted center, and the initial weighted center is updated in combination with the weighted center updating formula. In each iteration, the center position is gradually moved to the true center of the data distribution by redistributing the weight. The iteration update process continues until the preset iteration update times are met. The preset iteration update times are the maximum iteration times pre-set, which are used to limit the calculation cost of the iteration process and prevent excessive iteration from causing efficiency decline.

[0059] The updated weighted center obtained last time is taken as the first coating thickness deviation central value.

[0060] Further, in the system provided by the application embodiment, the transverse and longitudinal analysis module 14 is further used for:

[0061] The weighted center update formula is obtained, wherein the weighted center update formula is:

[0062] ;

[0063] wherein, is the updated weighted center, is a plurality of first coating thickness deviation values in the first coating thickness deviation value set, is an i-th first coating thickness deviation value in the first coating thickness deviation value set, is a first coating thickness deviation mean value, is a weight kernel function constructed based on a Gaussian function.

[0064] In the embodiments of the present application, the weighted center update formula is obtained, and the weighted center update formula is wherein is the updated weighted center, indicating that the center value is recalculated according to all deviation values and their weights in this iteration. is a plurality of first coating thickness deviation values in the first coating thickness deviation value set, is an i-th first coating thickness deviation value in the first coating thickness deviation value set, is a first coating thickness deviation mean value, is a weight kernel function constructed based on a Gaussian function, which is used to calculate the weight according to the distance between the current weighted center and each data point. The closer the distance, the greater the weight of the point, and the farther the distance, the smaller the weight of the point.

[0065] In specific calculation, the first coating thickness deviation mean value is taken as the initial weighted center value, and the iteration calculation process is started. In each iteration, the current weighted center value is compared with each deviation value, and the distance between them is calculated. In order to weight these distances, a weight function constructed by a Gaussian kernel function is used. After completing the weight distribution, the new weighted center value is calculated by combining the weighted center update formula. The update formula multiplies all deviation values by their corresponding weights to calculate the weighted sum; then the weighted sum is divided by the sum of all weights to obtain the new weighted center value. The iteration process continues, and the steps of weight distribution and center update are repeated in each iteration. When the stopping condition is met, the iteration is ended, and the weighted center value obtained in the last calculation is determined as the first coating thickness deviation mean value.

[0066] The second thickness monitoring module 15 adjusts the first preset coating thickness real-time monitoring parameter according to the size of the first longitudinal influence coefficient, obtains a first adjusted coating thickness real-time monitoring parameter, and monitors the photovoltaic glass coating thickness of the first spraying sub-task in the next preset monitoring window in real time according to the first adjusted coating thickness real-time monitoring parameter.

[0067] In the embodiment of the present application, first, a plurality of historical data are used as training samples, including historical longitudinal influence coefficients, historical preset coating thickness real-time monitoring parameters and historical adjusted coating thickness real-time monitoring parameters. These data are input into the longitudinal influence network layer model constructed based on the feedforward neural network for supervised training. Through training, the complex relationship between the longitudinal influence coefficient and the real-time monitoring parameter adjustment is learned, and the network parameters are continuously optimized until the model converges, forming the trained longitudinal influence network layer.

[0068] In actual tasks, the second thickness monitoring module inputs the first longitudinal influence coefficient and the first preset coating thickness real-time monitoring parameter into the trained longitudinal influence network layer for analysis. The model calculates the first adjusted coating thickness real-time monitoring parameter based on the input data.

[0069] Finally, according to the first adjusted coating thickness real-time monitoring parameter, the coating thickness of the first spraying sub-task of the photovoltaic glass is monitored in real time in the next preset monitoring window.

[0070] Further, the system provided by the embodiment of the application further includes that the second thickness monitoring module 15 is further used for:

[0071] obtaining a plurality of historical longitudinal influence coefficients, a plurality of historical preset coating thickness real-time monitoring parameters and a plurality of historical adjusted coating thickness real-time monitoring parameters as training data; using the training data to supervise the training of the network layer constructed based on the feedforward neural network, and updating the network parameters according to the training output result until the training converges, obtaining the trained longitudinal influence network layer; using the trained longitudinal influence network layer to analyze the first longitudinal influence coefficient and the first preset coating thickness real-time monitoring parameter, and obtaining the first adjusted coating thickness real-time monitoring parameter.

[0072] In the embodiment of the present application, first, a plurality of historical longitudinal influence coefficients, a plurality of historical preset coating thickness real-time monitoring parameters and a plurality of historical adjusted coating thickness real-time monitoring parameters are obtained from historical spraying task data as training data of the neural network model.

[0073] Then, the training data are used to supervise the training of the longitudinal influence network layer constructed based on the feedforward neural network. In the process of supervised training, the training data are used for multiple forward propagation and backward propagation. The predicted value is calculated through forward propagation, compared with the actual historical adjusted coating thickness real-time monitoring parameter, and the loss value (such as mean square error) is calculated. In the process of backward propagation, the network parameters are updated according to the loss value, and the weight and bias values are gradually optimized. The training process is repeated until the loss value converges, indicating that the model has been able to accurately capture the mapping relationship between the input features and the target output. At this time, the trained longitudinal influence network layer is obtained.

[0074] Finally, the trained longitudinal influence network layer is used to analyze the first longitudinal influence coefficient of the current task and the first preset coating thickness real-time monitoring parameter, and through calculation, the first adjusted coating thickness real-time monitoring parameter is output.

[0075] The monitoring parameter adjustment module 16 adjusts the second preset coating thickness real-time monitoring parameter according to the size of the first transverse correlation influence coefficient, and obtains a second influence coating thickness real-time monitoring parameter.

[0076] In the embodiment of the application, the monitoring parameter adjustment module adjusts the second preset coating thickness real-time monitoring parameter according to the size of the first transverse correlation influence coefficient. Specifically, the second preset coating thickness real-time monitoring parameter is adjusted by subtracting the difference value of the first transverse correlation influence coefficient from 1 and then multiplying the second preset coating thickness real-time monitoring parameter, and a second influence coating thickness real-time monitoring parameter is obtained.

[0077] The coating spraying module 17 configures parameters of the second spraying device according to the second spraying sub-task information of the second spraying sub-task, sprays a coating on the surface of the photovoltaic glass in the first preset monitoring window which has completed the first spraying sub-task by using the configured second spraying device, and monitors the coating thickness of the photovoltaic glass in real time according to the second influence coating thickness real-time monitoring parameter.

[0078] In the embodiment of the application, the coating spraying module first receives the second spraying sub-task information of the second spraying sub-task. The second spraying sub-task information includes coating type, coating average thickness target value, and coating thickness uniformity requirement, etc.

[0079] According to the second spraying sub-task information, the coating spraying module configures parameters of the second spraying device. The second spraying device usually includes a spraying gun, a spraying pressure control system, a spraying flow rate adjustment module, etc. The parameter configuration process includes adjusting the spraying pressure, setting the coating flow rate, and defining the spraying path.

[0080] After the device configuration is completed, the second spraying device is started to spray a second layer of coating on the surface of the photovoltaic glass in the first preset monitoring window which has completed the first spraying sub-task. At the same time, the coating thickness is monitored in real time according to the second influence coating thickness real-time monitoring parameter during the second spraying task.

[0081] In the embodiment of the application, as described above, the embodiment of the application has at least the following technical effects:

[0082] The application obtains a first spraying sub-task and a second spraying sub-task of a target photovoltaic glass coating task, and corresponding first spraying sub-task information and second spraying sub-task information, wherein each spraying sub-task information in the first spraying sub-task information and the second spraying sub-task information comprises coating type, coating average thickness information, and coating thickness uniformity information; the first preset coating thickness real-time monitoring parameter and the second preset coating thickness real-time monitoring parameter are obtained by searching in a preset coating monitoring space with the first spraying sub-task information and the second spraying sub-task information as indexes respectively; the first spraying equipment is configured according to the first spraying sub-task information of the first spraying sub-task, the photovoltaic glass surface is coated by using the configured first spraying equipment, the coating thickness is monitored in the first preset monitoring window according to the first preset coating thickness real-time monitoring parameter, and a first thickness monitoring result set is obtained; the first thickness monitoring result set is subjected to longitudinal compliance analysis and horizontal correlation analysis to determine a first longitudinal influence coefficient and a first horizontal correlation influence coefficient; the first preset coating thickness real-time monitoring parameter is adjusted according to the size of the first longitudinal influence coefficient, a first adjusted coating thickness real-time monitoring parameter is obtained, and the photovoltaic glass coating thickness in the next preset monitoring window is monitored in real time according to the first adjusted coating thickness real-time monitoring parameter; the second preset coating thickness real-time monitoring parameter is adjusted according to the size of the first horizontal correlation influence coefficient, a second influence coating thickness real-time monitoring parameter is obtained; the second spraying equipment is configured according to the second spraying sub-task information of the second spraying sub-task, the photovoltaic glass surface on which the first spraying sub-task is completed in the first preset monitoring window is coated by using the configured second spraying equipment, and the photovoltaic glass coating thickness is monitored in real time according to the second influence coating thickness real-time monitoring parameter. The application solves the technical problem that the prior art cannot accurately monitor the coating thickness distribution, and achieves the technical effect of accurately monitoring the photovoltaic glass coating thickness by task information acquisition, coating monitoring parameter searching, spraying equipment configuration, real-time thickness monitoring, longitudinal compliance analysis, and horizontal correlation analysis, dynamic adjustment of monitoring parameters, and optimization of the spraying process.

[0083] It should be noted that the above sequence of the embodiments of the application is only for description, and does not represent the advantages and disadvantages of the embodiments. The above describes a specific embodiment of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0084] The above only describes the preferred embodiments of the application and does not limit the application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

[0085] The specification and drawings are, of course, to be regarded in an illustrative rather than a restrictive sense. It is to be understood that any such modifications, variations, combinations or equivalents that fall within the scope of the application are intended to be embraced herein.

Claims

1. A system for real-time monitoring of the thickness of a photovoltaic glass coating, characterized in that, The system comprises: a task acquisition module, which acquires a first spraying sub-task and a second spraying sub-task of a target photovoltaic glass coating task, and corresponding first spraying sub-task information and second spraying sub-task information, wherein each spraying sub-task information in the first spraying sub-task information and the second spraying sub-task information comprises coating type, coating average thickness information, and coating thickness uniformity information; a retrieval module, which performs retrieval in a preset coating monitoring space respectively with the first spraying sub-task information and the second spraying sub-task information as indexes, and acquires first preset coating thickness real-time monitoring parameters and second preset coating thickness real-time monitoring parameters; a first thickness monitoring module, which performs parameter configuration on a first spraying device according to the first spraying sub-task information of the first spraying sub-task, performs coating spraying on a photovoltaic glass surface by using the first spraying device after the parameter configuration, performs coating thickness monitoring in a first preset monitoring window according to the first preset coating thickness real-time monitoring parameters, and obtains a first thickness monitoring result set; a horizontal and vertical analysis module, which performs vertical compliance analysis and horizontal correlation analysis on the first thickness monitoring result set, and determines a first vertical influence coefficient and a first horizontal correlation influence coefficient; a second thickness monitoring module, which adjusts the first preset coating thickness real-time monitoring parameters according to the size of the first vertical influence coefficient, obtains first adjusted coating thickness real-time monitoring parameters, and performs real-time monitoring on the coating thickness of the photovoltaic glass in a next preset monitoring window according to the first adjusted coating thickness real-time monitoring parameters; a monitoring parameter adjustment module, which adjusts the second preset coating thickness real-time monitoring parameters according to the size of the first horizontal correlation influence coefficient, and obtains second influence coating thickness real-time monitoring parameters; a coating spraying module, which performs parameter configuration on a second spraying device according to the second spraying sub-task information of the second spraying sub-task, performs coating spraying on the photovoltaic glass surface in the first preset monitoring window, which has completed the first spraying sub-task, by using the second spraying device after the parameter configuration, and performs real-time monitoring on the coating thickness of the photovoltaic glass according to the second influence coating thickness real-time monitoring parameters; the preset coating monitoring space is constructed, wherein the coating monitoring space comprises a plurality of sample space points, each sample space point corresponds to a sample spraying sub-task information, and each sample space point has an identifier of sample preset coating thickness real-time monitoring parameters; the first spraying sub-task information and the second spraying sub-task information are respectively input into the preset coating monitoring space, and a first target space point and a second target space point are obtained; k sample space points closest to the first target space point and the second target space point are extracted from the plurality of sample space points, and a first neighbor sample space point set and a second neighbor sample space point set are obtained, wherein k is an integer greater than or equal to 3. Traverse the mean value of the sample coating thickness real-time monitoring parameter set corresponding to the first near neighbor sample space point set to obtain a first preset coating thickness real-time monitoring parameter; Traverse the mean value of the sample coating thickness real-time monitoring parameter set corresponding to the second near neighbor sample space point set to obtain a second preset coating thickness real-time monitoring parameter; Obtain a three-dimensional space coordinate system, wherein the x-axis of the three-dimensional space coordinate system is the coating type, the y-axis is the coating average thickness information, and the z-axis is the coating thickness uniformity information; Obtain a plurality of sample spraying sub-task information and a plurality of corresponding sample coating thickness real-time monitoring parameters; Input the plurality of sample spraying sub-task information into the three-dimensional space coordinate system to obtain a plurality of sample space points, and identify the plurality of sample space points using the plurality of sample coating thickness real-time monitoring parameters; According to the identified plurality of sample space points and the three-dimensional space coordinate system, the preset coating monitoring space is obtained.

2. The real time monitoring system of the thickness of the photovoltaic glass coating according to claim 1, characterized in that, Including: Extract the first coating average thickness information and the first coating thickness uniformity information of the first spraying sub-task information; According to the first coating average thickness information and the first coating thickness uniformity information, the first thickness monitoring result set is subjected to longitudinal compliance analysis to determine a first longitudinal influence coefficient; The fluctuation variance of the first thickness monitoring result set is calculated to obtain a first transverse correlation influence coefficient.

3. A real-time monitoring system of the thickness of a photovoltaic glass coating according to claim 2, characterized in that, Including: According to the first coating average thickness information and the first thickness monitoring result set, a difference value calculation is performed to obtain a first coating thickness deviation value set; The first coating thickness deviation value set is subjected to central value screening to determine a first coating thickness deviation central value; The fluctuation variance of the first coating thickness deviation set is calculated to obtain a first deviation variance; The first coating thickness deviation central value, the first deviation variance, and the first coating thickness uniformity information are combined for longitudinal compliance analysis to obtain the first longitudinal influence coefficient.

4. The real-time monitoring system of the thickness of the photovoltaic glass coating according to claim 3, characterized in that, Including: The mean value of the first coating thickness deviation value set is calculated to obtain a first coating thickness deviation mean value; The first coating thickness deviation mean value is taken as an initial weighted center, and the initial weighted center is iteratively updated in the first coating thickness deviation value set combined with a weighted center update formula until a preset iteration update number is satisfied, and the last obtained updated weighted center is taken as the first coating thickness deviation central value.

5. The real-time monitoring system of the thickness of the photovoltaic glass coating according to claim 4, characterized in that, Including: Obtain a weighted center update formula, wherein the weighted center update formula is: ； wherein is an updated weighted center, is a first coating thickness deviation value of the plurality of first coating thickness deviation values in the first coating thickness deviation value set, is an i-th first coating thickness deviation value of the plurality of first coating thickness deviation values in the first coating thickness deviation value set, is a first coating thickness deviation mean value, is a weight kernel function constructed based on a Gaussian function.

6. The real time monitoring system of the thickness of the photovoltaic glass coating according to claim 1, characterized in that, Including: Obtain a plurality of historical longitudinal influence coefficients, a plurality of historical preset coating thickness real-time monitoring parameters, and a plurality of historical adjusted coating thickness real-time monitoring parameters as training data; The network layer constructed based on the feedforward neural network is subjected to supervised training using the training data, and the network parameters are updated according to the training output result until the training converges, and a trained longitudinal influence network layer is obtained; The first adjusted coating thickness real-time monitoring parameter is obtained by analyzing the first longitudinal influence coefficient and the first preset coating thickness real-time monitoring parameter using the trained longitudinal influence network layer.

Images