Magnesium oxide coating manufacturing control system and method based on image analysis

Abstract

The application discloses a magnesium oxide coating manufacturing control system and method based on image analysis, relates to the technical field of coating manufacturing process control, and collects a surface image of silicon steel to be coated, extracts silicon steel data based on the surface image of the silicon steel, analyzes a surface state of the silicon steel based on the silicon steel data, collects a magnesium oxide coating image, analyzes uniformity of the magnesium oxide coating based on the magnesium oxide coating image, simultaneously analyzes an influence of coating uniformity on current viscosity, collects historical coating process data, trains a bonding strength prediction model based on the historical coating process data, inputs the surface state of the silicon steel and the influence of the coating uniformity on the current viscosity, outputs predicted bonding strength, and analyzes an optimal adhesive increase amount based on target bonding strength and the predicted bonding strength. Before the magnesium oxide coating is coated, the surface state of the silicon steel and coating characteristics are analyzed, bonding strength is predicted, the addition amount of the adhesive is accurately adjusted, and the bonding quality is ensured to be stably within a target range.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coating manufacturing process control, in particular to a magnesium oxide coating manufacturing control system and method based on image analysis. BACKGROUND

[0002] In the manufacturing of electrical equipment such as power transformers and motors, the surface of oriented silicon steel sheets needs to be coated with a layer of inorganic coating mainly composed of magnesium oxide. This coating not only provides excellent insulation performance to reduce iron loss, but also ensures the key to long-term operation reliability of the equipment through the bonding strength between the coating and the silicon steel substrate. Insufficient bonding strength can lead to coating peeling, which in turn causes local short circuit of the core and energy efficiency decline, resulting in serious quality problems.

[0003] Currently, in the manufacturing and coating process of magnesium oxide coating, the control of the final bonding strength mostly relies on independent monitoring of a single process parameter and offline, post-quality inspection. The micro-topography of the silicon steel surface, cleanliness, and the uniformity of magnesium oxide particles in the coating all contribute to the underlying factors affecting the bonding strength. However, these factors are often ignored or only qualitatively judged in the existing control system. For example, it is impossible to quantify the specific impact of surface roughness or coating uniformity on interfacial bonding force. Moreover, the evaluation of bonding strength mostly relies on samples taken from the production line after coating and curing, which are then destructively tested in the laboratory. Therefore, it is difficult to establish a precise prediction model that can comprehensively reflect the substrate state, coating characteristics, and process history.

[0004] Therefore, there is an urgent need in the field for an intelligent control method that can predict the bonding strength in real time and adjust the amount of adhesive accordingly, in order to fundamentally improve the coating quality of magnesium oxide coating. SUMMARY

[0005] To overcome the shortcomings of the prior art, the present application provides a magnesium oxide coating manufacturing control system and method based on image analysis. The present application predicts the bonding strength by analyzing the surface state of the silicon steel and the characteristics of the coating before the magnesium oxide coating is applied. By accurately adjusting the amount of adhesive, the bonding quality is ensured to be stable within the target range, while avoiding waste of adhesive.

[0006] To achieve the above-mentioned purpose, the present application provides a magnesium oxide coating manufacturing control method based on image analysis, which comprises the following specific steps:

[0007] S1, acquiring an image of the surface of the silicon steel to be coated, extracting silicon steel data based on the silicon steel surface image, and analyzing the surface state of the silicon steel based on the silicon steel data;

[0008] S2, acquiring an image of the magnesium oxide coating, analyzing the uniformity of the magnesium oxide coating based on the image, and analyzing the influence of the coating uniformity on the current viscosity;

[0009] S3, collecting historical coating process data, training the bonding strength prediction model based on the historical coating process data, inputting the influence of the silicon steel surface state and the coating uniformity on the current viscosity, and outputting the predicted bonding strength;

[0010] S4, analyzing the optimal adhesive increase amount based on the target bonding strength and the predicted bonding strength.

[0011] Preferably, the S1 comprises the following specific steps:

[0012] S11, collecting a silicon steel surface image to be coated;

[0013] S12, extracting silicon steel data based on the silicon steel surface image, the silicon steel data including silicon steel plate temperature, surface wettability, and surface roughness, the silicon steel plate temperature being an average temperature, the surface wettability being a contact angle, and the surface roughness being a root mean square roughness;

[0014] S13, obtaining a temperature abnormal value based on the ratio of the average temperature to a safety temperature threshold, obtaining a contact angle abnormal value based on the ratio of the contact angle to a standard contact angle, and obtaining a roughness abnormal value based on the ratio of the root mean square roughness to a standard roughness, and obtaining a silicon steel surface state abnormal value by weighted sum of the temperature abnormal value, the contact angle abnormal value, and the roughness abnormal value.

[0015] Preferably, the S2 comprises the following specific steps:

[0016] S21, collecting a magnesium oxide coating image;

[0017] S22, analyzing the uniformity of the magnesium oxide coating based on the magnesium oxide coating image, the uniformity of the magnesium oxide coating including particle uniformity and color uniformity, the particle uniformity being a standard deviation of particle area ratio, and the color uniformity being a color standard deviation;

[0018] S23, obtaining a particle uniformity abnormal value based on the ratio of the standard deviation of particle area ratio to a standard particle area ratio, obtaining a color uniformity abnormal value based on the ratio of the color standard deviation to a standard color difference value, and obtaining a coating uniformity abnormal value by weighted sum of the particle uniformity abnormal value and the color uniformity abnormal value;

[0019] S24, collecting a viscosity measurement value of the current magnesium oxide coating, and obtaining an actual viscosity based on the product of the viscosity measurement value, the coating uniformity abnormal value, and a coating uniformity influence coefficient.

[0020] Preferably, the S3 comprises the following specific steps:

[0021] S31, collecting historical coating process data, the historical coating process data including historical silicon steel surface state abnormal values, historical magnesium oxide coating actual viscosities, and historical bonding strength values;

[0022] S32, divide the historical coating process data into a training set, a validation set, and a test set, wherein the training set is 70%, the validation set is 15%, and the test set is 15%, build and train a bond strength prediction model, wherein the number of input layer nodes is equal to the number of features, input standardization is performed using a Batck Normalization layer, the first hidden layer of the hidden layer contains 128 nodes, the second hidden layer contains 64 nodes, and the third hidden layer contains 32 nodes, each layer is followed by a Dropout layer with a dropout rate of 0.2, the output layer outputs a bond strength prediction value, the mean squared error is used as the loss function, the Adam optimizer is used, the initial learning rate is 0.001, the learning rate is reduced to half of the original every 20 rounds, the main evaluation indicator is the mean absolute error, the training batch size is set to 32, the number of training rounds is set to 200, the validation set loss is stopped when it does not improve for 10 consecutive rounds, the forward propagation calculates the prediction value, the loss function value is calculated, the gradient is calculated by back propagation, and the optimizer updates the weight parameters. Record the training set and validation set loss every round, save a model checkpoint every 10 rounds, display the training progress and indicator changes in real time, evaluate the model performance on the validation set, analyze the residual distribution of the prediction value and the true value, check for systematic bias, calculate the mean absolute error on the test set, draw a scatter plot of the prediction value and the true value, analyze the error distribution characteristics, and verify the generalization ability under different working conditions. The output strategy formula of the kth neuron of the h+1 layer is: , wherein, represents the output of the kth neuron of the h+1 layer of the bond strength prediction model, represents the number of neurons of the h layer, represents the connection weight of the bth neuron of the h layer and the kth neuron of the h+1 layer of the bond strength prediction model, represents the output of the bth neuron of the h layer of the bond strength prediction model, represents the bias of the linear relationship between the bth neuron of the h layer and the kth neuron of the h+1 layer of the bond strength prediction model, represents a Sigmoid activation function;

[0023] S33, obtain the trained bond strength prediction model, input the silicon steel surface state abnormal value and the actual viscosity of the coating, and output the predicted bond strength.

[0024] Preferably, the S4 comprises the following specific steps:

[0025] Obtain the target bond strength and the predicted bond strength, and obtain the adhesive increase amount based on the difference between the target bond strength and the predicted bond strength divided by the increase amount of the adhesive per unit bond strength.

[0026] The application also provides a magnesia coating manufacturing control system based on image analysis, comprising:

[0027] A silicon steel surface analysis module is configured to collect a silicon steel surface image of a silicon steel to be coated, extract silicon steel data based on the silicon steel surface image, and analyze a silicon steel surface state based on the silicon steel data.

[0028] A coating influence analysis module is configured to collect a magnesia coating image, analyze uniformity of the magnesia coating based on the magnesia coating image, and analyze an influence of the coating uniformity on a current viscosity.

[0029] A prediction model training module is configured to collect historical coating process data, and train a bonding strength prediction model based on the historical coating process data.

[0030] A bonding strength prediction module is configured to obtain the bonding strength prediction model, input the silicon steel surface state and the influence of the coating uniformity on the current viscosity, and output a predicted bonding strength.

[0031] An optimal dosage analysis module is configured to analyze an optimal adhesive dosage based on a target bonding strength and the predicted bonding strength.

[0032] The application also provides an electronic device comprising a processor and a memory, wherein the memory stores a computer program that can be invoked by the processor, and the processor executes the above-mentioned magnesia coating manufacturing control method based on image analysis by invoking the computer program stored in the memory.

[0033] The application also provides a computer-readable storage medium storing instructions, which, when executed on a computer, cause the computer to execute the above-mentioned magnesia coating manufacturing control method based on image analysis.

[0034] Compared with the prior art, the application has the following beneficial effects: the application collects a silicon steel surface image of a silicon steel to be coated, extracts silicon steel data based on the silicon steel surface image, analyzes a silicon steel surface state based on the silicon steel data, collects a magnesia coating image, analyzes uniformity of the magnesia coating based on the magnesia coating image, analyzes an influence of the coating uniformity on a current viscosity, collects historical coating process data, trains a bonding strength prediction model based on the historical coating process data, inputs the silicon steel surface state and the influence of the coating uniformity on the current viscosity, outputs a predicted bonding strength, and analyzes an optimal adhesive dosage based on a target bonding strength and the predicted bonding strength. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below only constitute some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0036] Figure 1 Flowchart of the image analysis-based magnesium oxide coating manufacturing control method of the present application;

[0037] Figure 2 Flowchart of S1 of the image analysis-based magnesium oxide coating manufacturing control method of the present application;

[0038] Figure 3 Flowchart of S2 of the image analysis-based magnesium oxide coating manufacturing control method of the present application;

[0039] Figure 4 Overall framework of the image analysis-based magnesium oxide coating manufacturing control system of the present application. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art belong to the scope of protection of the present application.

[0041] Please refer to Figure 1 The image analysis-based magnesium oxide coating manufacturing control method provided in the embodiments of the present application includes the following specific steps:

[0042] S1, collecting the surface image of the silicon steel to be coated, extracting silicon steel data based on the surface image of the silicon steel, and analyzing the surface state of the silicon steel based on the silicon steel data;

[0043] Please refer to Figure 2 In the present embodiment, S1 includes the following specific steps:

[0044] S11, building a multi-modal image collection system, collecting the surface image of the silicon steel to be coated, capturing the temperature distribution map by an infrared thermal imager, scanning the silicon steel to be collected by a 3D contour instrument to generate a height map, taking the detailed image of the silicon steel surface under the best lighting by a 2D visible light camera, storing the silicon steel surface image data and attaching the position and time stamp, and transmitting to a central processing server;

[0045] S12, extracting silicon steel data based on the silicon steel surface image, the silicon steel data including silicon steel plate temperature, surface wettability and surface roughness, the silicon steel plate temperature being an average temperature, loading a preset silicon steel emissivity parameter, generating a temperature map using a thermal imager, and calculating the average temperature in the silicon steel area to be collected, the surface wettability being a contact angle, before the camera takes a picture, a liquid droplet is deposited on a specific position on the surface of the silicon steel by a liquid droplet spraying device, a high-speed 2D camera captures the profile image of the liquid droplet from the side, the image is converted to grayscale, Gaussian blur is used for denoising, a Canny operator is used to accurately extract the profile edge of the liquid droplet, the extracted edge points are fitted into a circle, the contact line (baseline) between the liquid droplet and the solid surface is identified, the tangent lines at the two intersection points of the liquid droplet profile and the baseline are calculated respectively, and the included angle between the two tangent lines is the contact angle, the smaller the contact angle, the better the wettability, indicating that the magnesium oxide coating has strong spreading ability on the surface of the silicon steel, has a larger effective contact area, has stronger intermolecular force, and improves the bonding strength between the coating and the silicon steel substrate, the surface roughness is the root mean square roughness, the height information of each point on the surface is directly output by using a 3D profiler to form a point cloud or a height map, Gaussian filtering is used to remove outliers in the point cloud, a reference plane is fitted, the entire point cloud data is subtracted from the plane to eliminate the overall tilt caused by the uneven placement of the silicon steel plate, the average height of the silicon steel area to be collected is calculated first, then the average value of the square of the height deviation of each point is calculated, and finally the root mean square roughness is obtained by taking the square root, the root mean square roughness is more sensitive to the peaks and valleys of the surface, and therefore can better reflect the true characteristics of the surface in evaluating the adhesion, tribological performance and other properties of the coating.

[0046] S13, obtaining temperature abnormal values based on the ratio of the average temperature to the safety temperature threshold, obtaining contact angle abnormal values based on the ratio of the contact angle to the standard contact angle, and obtaining roughness abnormal values based on the ratio of the root mean square roughness to the standard roughness, and obtaining silicon steel surface state abnormal values by weighted sum of the temperature abnormal values, the contact angle abnormal values and the roughness abnormal values.

[0047] Preferably, the temperature abnormal values, the contact angle abnormal values and the roughness abnormal values are calculated by using the coefficient of variation method, wherein the coefficient of variation calculation formula is: , wherein, represents the coefficient of variation of the i-th abnormal value, reflecting the dispersion degree of the abnormal value, represents the standard deviation of the i-th abnormal value, represents the average value of the i-th abnormal value, the weight of each abnormal value is obtained by calculating the proportion of each coefficient of variation in the total coefficient of variation.

[0048] The obtaining step of the safety temperature threshold is: preparing silicon steel test samples at different temperatures, measuring the bonding strength at each temperature point, determining the critical temperature at which the strength starts to decrease, and obtaining ±10°C of the bonding strength peak temperature as the process window. The safety temperature threshold in the embodiment can be set to 450°C.

[0049] The obtaining step of the standard contact angle is: preparing silicon steel test samples under different contact angle conditions, measuring the corresponding bonding strength values, establishing a contact angle-bonding strength relationship curve, and selecting a contact angle range with a bonding strength greater than or equal to 95%. The standard contact angle in the embodiment can be set to 72°.

[0050] The obtaining step of the standard roughness is: measuring the surface roughness value of each silicon steel test sample using a roughness meter, using the same coating and process for coating and curing, testing the bonding strength of each silicon steel test sample, and setting the roughness corresponding to the peak region of the bonding strength as the standard roughness. The standard roughness in the embodiment can be set to 0.45 μm.

[0051] S2, collect the magnesium oxide coating image, analyze the uniformity of the magnesium oxide coating based on the magnesium oxide coating image, and analyze the influence of the coating uniformity on the current viscosity;

[0052] Please refer to Figure 3 In the embodiment, S2 includes the following specific steps:

[0053] S21, build a high-precision microscopic image acquisition system to collect the magnesium oxide coating image. When the magnesium oxide coating is relatively uniform, a static sampling method can be used to extract a small amount of coating with a sampler, drop a drop of coating on a clean glass slide, cover it with a cover glass to form a uniform thickness thin layer. When the magnesium oxide coating is not uniform, a flow observation method can be used. A flow cell with a transparent observation window is connected to the coating circulation pipeline. The coating is pumped into the flow cell at a constant and slow flow rate by a micro pump, and the microscope directly aims at the center of the observation window for shooting.

[0054] S22, analyze the uniformity of the magnesium oxide coating based on the magnesium oxide coating image, the uniformity of the magnesium oxide coating including particle uniformity and color uniformity, the uniformity mainly refers to the dispersion degree of particles in the liquid base, whether there is agglomeration, sedimentation or local concentration, also refers to the uniformity of the color, when adding pigment, uneven color means that the pigment is not completely dispersed, which will change the rheological properties of the system, oil stains, moisture or different batches of magnesium oxide, color difference caused by different calcination temperatures, and part of the additives may react slightly with magnesium oxide, resulting in color change, thereby affecting the viscosity, the particle uniformity is the standard deviation of the particle area ratio, the particles are separated from the background by threshold segmentation, a binary image is generated, the whole image is evenly divided into a grid, the area ratio of particles in each grid is calculated, and then the standard deviation of the area ratio of all grids is calculated, the smaller the standard deviation of the particle area ratio, the more uniform the particle distribution, the color uniformity is the color standard deviation, the RGB image of the magnesium oxide coating collected by the camera is converted to the Lab color space, including brightness (L*), red-green hue (a*), and yellow-blue hue (b*), the image is divided into an MxN grid, the average (L*, a*, b*) value of all pixels in each grid is calculated, and then the standard deviation of the (L*, a*, b*) value of all grids is calculated, the smaller the standard deviation, the more uniform the color;

[0055] S23, obtain particle uniformity outliers based on the ratio of the standard deviation of the particle area ratio to the standard particle area ratio, obtain color uniformity outliers based on the ratio of the color standard deviation to the standard color difference value, and obtain coating uniformity outliers by weighted sum of the particle uniformity outliers and the color uniformity outliers;

[0056] The standard particle area ratio and the standard color difference value are obtained by: collecting microscopic images of each magnesium oxide coating sample, using image analysis software to calculate the particle area ratio and color difference value of each magnesium oxide coating sample, using a rotational rheometer to measure the real and comprehensive viscosity curve of each magnesium oxide coating sample, preparing each magnesium oxide coating sample into a coating layer, curing under standard conditions, testing the bonding strength, and obtaining the particle area ratio and color difference value corresponding to the best bonding strength as the standard value, the standard particle area ratio of the embodiment can be set to 21%, and the standard color difference value can be set to 1.5;

[0057] S24, collect the viscosity measurement value of the current magnesium oxide coating using a viscometer, and obtain the actual viscosity based on the product of the viscosity measurement value, the coating uniformity outlier value and the coating uniformity influence coefficient.

[0058] Preferably, the experimental proportion and environmental conditions are fixed, the preparation is carried out with particles and color as two main variables, the microscopic images of the paint are taken by using a standardized microscopic imaging system, the particle area proportion and color difference value of each sample are calculated by using image analysis software, the paint uniformity is obtained, the viscosity data is collected, a fixed shear rate related to the actual coating process is set, the apparent viscosity under the shear rate is measured and recorded, the scatter plot is drawn with the paint uniformity index as the abscissa (X) and the viscosity as the ordinate (Y), the one-dimensional linear regression analysis is carried out, and the slope of the equation is the paint uniformity influence coefficient;

[0059] Preferably, the embodiment uses the coefficient of variation method to calculate the weight corresponding to the particle uniformity abnormal value and the color uniformity abnormal value.

[0060] S3, collecting historical coating process data, training a bonding strength prediction model based on the historical coating process data, inputting the surface state of silicon steel and the influence of paint uniformity on current viscosity, and outputting the predicted bonding strength;

[0061] In the embodiment, S3 includes the following specific steps:

[0062] S31, collecting historical coating process data, the historical coating process data including historical silicon steel surface state abnormal value, historical magnesium oxide paint actual viscosity and historical bonding strength value measured in the laboratory, data cleaning and preprocessing of the historical coating process data, and using Z-score standardization;

[0063] S32, divide the historical coating process data into a training set, a validation set, and a test set, wherein the training set is 70%, the validation set is 15%, and the test set is 15%, build and train a bond strength prediction model, wherein the number of input layer nodes is equal to the number of features, input standardization is performed using a Batck Normalization layer, the first hidden layer of the hidden layer contains 128 nodes, the second hidden layer contains 64 nodes, and the third hidden layer contains 32 nodes, each layer is followed by a Dropout layer with a dropout rate of 0.2, the output layer outputs a bond strength prediction value, the mean squared error is used as the loss function, the Adam optimizer is used, the initial learning rate is 0.001, the learning rate is reduced to half of the original every 20 rounds, the main evaluation indicator is the mean absolute error, the training batch size is set to 32, the number of training rounds is set to 200, the validation set loss is stopped when it does not improve for 10 consecutive rounds, the forward propagation calculates the prediction value, the loss function value is calculated, the gradient is calculated by back propagation, and the optimizer updates the weight parameters. Record the training set and validation set loss every round, save a model checkpoint every 10 rounds, display the training progress and indicator changes in real time, evaluate the model performance on the validation set, analyze the residual distribution of the prediction value and the true value, check if there is a systematic bias, calculate the mean absolute error on the test set, draw a scatter plot of the prediction value and the true value, analyze the error distribution characteristics, and verify the generalization ability under different working conditions. The output strategy formula of the kth neuron of the h+1th layer is: , wherein, represents the output of the kth neuron of the h+1th layer of the bond strength prediction model, represents the number of neurons in the hth layer, represents the connection weight of the bth neuron of the hth layer and the kth neuron of the h+1th layer of the bond strength prediction model, represents the output of the bth neuron of the hth layer of the bond strength prediction model, represents the bias of the linear relationship between the bth neuron of the hth layer and the kth neuron of the h+1th layer of the bond strength prediction model, represents a Sigmoid activation function;

[0064] S33, obtain the trained bond strength prediction model, input the silicon steel surface state abnormal value and the actual viscosity of the coating, and output the predicted bond strength.

[0065] S4, analyze the optimal amount of adhesive based on the target bond strength and the predicted bond strength.

[0066] In this embodiment, S4 includes the following specific steps:

[0067] Obtain the target bond strength and the predicted bond strength, and obtain the adhesive increase amount based on the difference between the target bond strength and the predicted bond strength divided by the increase amount of the unit adhesive to the bond strength.

[0068] Preferably, all other components (magnesium oxide, solvent) and all process parameters (coating, curing) are fixed, only the content of the binder is changed, a concentration gradient is set, for example, five experimental groups with binder mass fraction of 1.0%, 1.5%, 2.0%, 2.5% and 3.0% are designed, 3-5 repeated coating samples are prepared for each experimental group, curing is carried out under the same conditions, the average bonding strength is taken as the ordinate (Y) and the binder content as the abscissa (X), a scatter plot is drawn, the scatter plot is approximated to a straight line, one-dimensional linear regression analysis is carried out, and the regression equation Y=kX+b is obtained, wherein k represents the increase of the bonding strength per unit of binder, and b represents the basic bonding strength without binder.

[0069] Please refer to Figure 4 The embodiment of the present application also provides a magnesium oxide coating manufacturing control system based on image analysis, comprising:

[0070] A silicon steel surface analysis module is configured to acquire a silicon steel surface image of a silicon steel to be coated, extract silicon steel data based on the silicon steel surface image, and analyze a silicon steel surface state based on the silicon steel data.

[0071] A coating influence analysis module is configured to acquire a magnesium oxide coating image, analyze uniformity of the magnesium oxide coating based on the magnesium oxide coating image, and analyze an influence of the coating uniformity on a current viscosity.

[0072] A prediction model training module is configured to acquire historical coating process data, and train a bonding strength prediction model based on the historical coating process data.

[0073] A bonding strength prediction module is configured to acquire the bonding strength prediction model, input the silicon steel surface state and the influence of the coating uniformity on the current viscosity, and output a predicted bonding strength.

[0074] An optimal dosage analysis module is configured to analyze an optimal binder increase amount based on a target bonding strength and the predicted bonding strength.

[0075] The embodiment of the present application also provides an electronic device, comprising a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes the above-mentioned magnesium oxide coating manufacturing control method based on image analysis by calling the computer program stored in the memory.

[0076] The electronic device can have a large difference due to configuration or performance, and can include one or more processors (Central Processing Units, CPUs) and one or more memories, wherein the memory stores at least one computer program loaded and executed by the processor to implement the image analysis-based magnesium oxide coating manufacturing control method provided by the above method embodiments. The electronic device can also include other components for implementing device functions, for example, the electronic device can also have a wired or wireless network interface and an input and output interface, and the like, so as to perform data input and output. This embodiment will not be described here.

[0077] The embodiment of the present application also provides a computer readable storage medium, which stores instructions, and when a computer program runs on a computer device, the computer device executes the above-mentioned image analysis-based magnesium oxide coating manufacturing control method.

[0078] For example, the computer readable storage medium can be a Read-Only Memory (ROM), a Random Access Memory (RAM), a Compact Disc Read-Only Memory (CD-ROM), a magnetic tape, a floppy disk and an optical data storage device, etc.

[0079] It should be understood that in various embodiments of the present application, the size of the sequence number of each process described above does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0080] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof, and when implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product, which includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the above-described embodiments can be implemented in whole or in part to generate a flow or function according to the embodiments of the present application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center through a wired network or a wireless network. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server, data center, or the like, which includes one or more available medium collections. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0081] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in the present application can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0082] In several embodiments provided by the present application, it should be understood that the disclosed system, device, and method can be implemented in other ways. For example, the above-described device embodiments are merely illustrative. For example, the division of units is merely one, and actual implementation can have another division manner. For example, multiple units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, devices, or units, which can be electrical, mechanical, or other forms.

[0083] In the description of the present specification, the description of the terms "one embodiment", "an example", "a specific example" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application, and the illustrative description of the above terms does not necessarily refer to the same embodiment or example, and the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0084] The above disclosed preferred embodiments of the present application are only used to help explain the present application, and the preferred embodiments do not describe all the details and limit the present application to the specific embodiments. Obviously, many modifications and changes can be made according to the content of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can well understand and utilize the present application. The present application is limited only by the claims and their full scope and equivalents.

Claims

1. A method for controlling the manufacturing of magnesium oxide coatings based on image analysis, characterized in that, The specific steps include the following: S1. Acquire images of the silicon steel surface to be coated, extract silicon steel data based on the silicon steel surface images, and analyze the surface condition of the silicon steel based on the silicon steel data; S2. Acquire images of magnesium oxide coating, analyze the uniformity of magnesium oxide coating based on the images, and analyze the influence of coating uniformity on the current viscosity. S3. Collect historical coating process data, train the bonding strength prediction model based on the historical coating process data, input the influence of silicon steel surface condition and coating uniformity on the current viscosity, and output the predicted bonding strength. S4. Analyze the optimal amount of adhesive to add based on the target bond strength and the predicted bond strength; S1 includes the following specific steps: S11. Acquire images of the silicon steel surface to be coated; S12. Extract silicon steel data based on silicon steel surface image. The silicon steel data includes silicon steel plate temperature, surface wettability and surface roughness. The silicon steel plate temperature is the average temperature, the surface wettability is the contact angle, and the surface roughness is the root mean square roughness. S13. Obtain temperature anomaly values ​​based on the ratio of average temperature to safe temperature threshold, obtain contact angle anomaly values ​​based on the ratio of contact angle to standard contact angle, obtain roughness anomaly values ​​based on the ratio of root mean square roughness to standard roughness, and obtain silicon steel surface condition anomaly values ​​by weighted summation of temperature anomaly values, contact angle anomaly values, and roughness anomaly values. S2 includes the following specific steps: S21. Acquire images of magnesium oxide coating; S22. Analyze the uniformity of magnesium oxide coating based on the image of magnesium oxide coating. The uniformity of magnesium oxide coating includes particle uniformity and color uniformity. The particle uniformity is the standard deviation of particle area ratio, and the color uniformity is the standard deviation of color. S23. Obtain particle uniformity anomalies based on the ratio of the standard deviation of particle area ratio to the standard particle area ratio; obtain color uniformity anomalies based on the ratio of the standard deviation of color to the standard color difference; and obtain coating uniformity anomalies by weighted summation of particle uniformity anomalies and color uniformity anomalies. S24. Collect the viscosity measurement value of the current magnesium oxide coating, and obtain the actual viscosity based on the product of the viscosity measurement value, the coating uniformity anomaly value and the coating uniformity influence coefficient.

2. The method for controlling the manufacturing of magnesium oxide coatings based on image analysis according to claim 1, characterized in that, S3 includes the following specific steps: S31. Collect historical coating process data, including historical abnormal values ​​of silicon steel surface condition, historical actual viscosity of magnesium oxide coating, and historical bonding strength value; S32. Divide the historical coating process data into training, validation, and test sets, with the training set comprising 70%, the validation set 15%, and the test set 15%. Construct and train a bonding strength prediction model, where the number of input layer nodes equals the number of features, and the output layer outputs the predicted bonding strength value. The output strategy formula for the k-th neuron in the (h+1)th layer is: In the formula, This represents the output of the k-th neuron in the (h+1)th layer of the bond strength prediction model. This represents the number of neurons in the h-th layer. This represents the connection weight between neuron b in layer h and neuron k in layer h+1 of the adhesion strength prediction model. This represents the output of neuron b in the h-th layer of the bond strength prediction model. This represents the bias in the linear relationship between neuron b in layer h and neuron k in layer h+1 of the bond strength prediction model. This represents the Sigmoid activation function; S33. Obtain the trained bonding strength prediction model, input the abnormal values ​​of the silicon steel surface condition and the actual viscosity of the coating, and output the predicted bonding strength.

3. The method for controlling the manufacturing of magnesium oxide coatings based on image analysis according to claim 2, characterized in that, S4 includes the following specific steps: Obtain the target bond strength and the predicted bond strength. The amount of adhesive added is obtained by dividing the difference between the target bond strength and the predicted bond strength by the amount of adhesive added per unit of bond strength.

4. A magnesium oxide coating manufacturing control system based on image analysis, used to implement the magnesium oxide coating manufacturing control method based on image analysis as described in any one of claims 1-3, characterized in that, include: The silicon steel surface analysis module is used to acquire images of the silicon steel surface to be coated, extract silicon steel data based on the silicon steel surface images, and analyze the surface condition of the silicon steel based on the silicon steel data. The coating impact analysis module is used to acquire images of magnesium oxide coatings, analyze the uniformity of magnesium oxide coatings based on the images, and analyze the impact of coating uniformity on the current viscosity. The prediction model training module is used to collect historical coating process data and train a bond strength prediction model based on the historical coating process data. The bond strength prediction module is used to obtain a bond strength prediction model. It takes the influence of silicon steel surface condition and coating uniformity on the current viscosity as input and outputs the predicted bond strength. The optimal dosage analysis module is used to analyze the optimal amount of adhesive to add based on the target bond strength and the predicted bond strength.

5. An electronic device, characterized in that, include: A memory and a processor, wherein the memory stores a computer program that can be called by the processor, and the processor executes the image analysis-based magnesium oxide coating manufacturing control method according to any one of claims 1-3 by calling the computer program stored in the memory.

6. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the image analysis-based magnesium oxide coating manufacturing control method as described in any one of claims 1-3.

Images