Automatic control method and system for slit extrusion coating glue supply process

Abstract

The application relates to the technical field of automatic control, in particular to a kind of automatic control method and system for slit extrusion coating glue supply process, comprising: according to monitoring data, identifying candidate defect area of substrate surface after coating, and according to the morphological characteristics and gray distribution characteristics of candidate defect area in image data with the change characteristics of time, determining temporary interference defect and real coating defect in candidate defect area.For the real coating defect determined, according to the pump speed difference of glue supply pump under historical coating state with similarity degree satisfying preset condition to current coating state, determine the pump speed adjustment value of current time, wherein, historical coating state is determined from historical operation data according to the comprehensive similarity degree of the characteristics of candidate defect area and monitoring data, and the coating state with similarity degree satisfying preset condition to current coating state.According to pump speed adjustment value, the pump speed of glue supply pump is controlled.The application provides a method for improving the accuracy of glue supply pump speed control.

Description

Technical Field

[0001] This application relates to the field of automatic control technology, specifically to an automatic control method and system for a slot extrusion coating adhesive supply process. Background Technology

[0002] Slit extrusion coating technology is widely used in precision coating fields such as lithium-ion battery electrode manufacturing, optical thin films, and flexible electronics. It utilizes a coating die with a precision cavity and slits to deliver functional coating slurry to the surface of a moving substrate at a stable and controllable flow rate and pressure, achieving high-precision continuous coating. The stability of the adhesive supply during the coating process directly determines the uniformity of the coating thickness and the surface quality; therefore, real-time monitoring and automatic control of the adhesive supply process are crucial for ensuring coating quality.

[0003] In existing technologies, automatic control of the glue supply process in slot extrusion coating is usually based on feedback adjustment using a single-dimensional detection method. For example, an online thickness gauge is installed on the coating production line to detect the coating thickness in real time. When the detected value deviates from the set range, the speed of the glue supply pump is adjusted through a proportional-integral-derivative control (PID) controller to adjust the glue supply amount so that the coating thickness returns to the target value. This control method can play a certain role in production scenarios with relatively stable operating conditions.

[0004] However, the actual industrial coating process is a complex process with multiple coupled factors and compound interference. Existing control methods based on single detection means are difficult to accurately identify the actual nature of coating abnormalities. There is a mismatch between the adjustment response and the actual situation, resulting in frequent adjustments to the glue supply pump speed but failing to effectively suppress coating fluctuations, and insufficient control accuracy. Summary of the Invention

[0005] To address the technical problems of frequent adjustments to the glue supply pump speed, difficulty in effectively suppressing coating fluctuations, and insufficient control precision, this application aims to provide an automatic control method and system for the slit extrusion coating glue supply process. The specific technical solution adopted is as follows: In a first aspect, an automatic control method for a slot extrusion coating adhesive supply process is provided, comprising: acquiring monitoring data of the surface of a coated substrate, the monitoring data including image data acquired sequentially over time; identifying candidate defect regions on the surface of the coated substrate based on the monitoring data, and determining temporary interference defects and actual coating defects within the candidate defect regions based on the changes in morphological features and grayscale distribution characteristics of the candidate defect regions in the image data over time; for the identified actual coating defects, determining the pump speed adjustment value at the current moment based on the pump speed difference of the adhesive supply pump under historical coating states that meet preset conditions of similarity with the current coating state, wherein the historical coating state is a coating state that meets preset conditions of similarity with the current coating state determined from historical operating data based on the characteristics of the candidate defect regions and the comprehensive similarity of the monitoring data; and controlling the pump speed of the adhesive supply pump according to the pump speed adjustment value.

[0006] As one possible approach, the monitoring data also includes thickness measurement data, and the process for identifying temporary interference defects and actual coating defects, including: Based on the deviation of pixel grayscale values ​​from the average grayscale value of the surrounding image, the deviation of pixel grayscale values ​​from the grayscale values ​​of pixels at the same coordinates in the previous frame, and the thickness deviation of the corresponding thickness measurement data from the preset substrate thickness reference value, candidate defect pixels are determined, and candidate defect regions composed of these pixels are extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel. Candidate defect regions are matched between adjacent frames of the image data. Based on the contour sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions, temporary interference defects and actual coating defects are determined.

[0007] One possible approach to determining candidate defective pixels includes: for any pixel within any frame of the image data, calculating a first deviation of the pixel's grayscale value from the average grayscale value of that frame, and a second deviation of the pixel's grayscale value from the grayscale value of a pixel at the same coordinates in the previous frame. Obtaining the thickness deviation between the thickness measurement data at the corresponding time in any frame and a preset substrate thickness reference value. Determining the defect probability value of any pixel based on the first deviation, the second deviation, and the thickness deviation. Marking pixels with defect probability values ​​greater than a preset threshold as candidate defective pixels.

[0008] One possible approach is to determine temporary interference defects and genuine coating defects based on the contour sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions. This includes: comparing the area of ​​the matched candidate defect region at its first appearance with its current area to determine the area change; using the standard deviation of the region's defect probability value across all image frames as the grayscale distribution fluctuation; using the mean of the grayscale gradient magnitude of the edge pixels of the candidate defect region as the contour sharpness; and determining the probability that the candidate defect region is a genuine coating defect based on the area change, grayscale distribution fluctuation, contour sharpness, and number of consecutively stable occurrences.

[0009] As one possible approach, the process of determining the pump speed adjustment value at the current moment includes: acquiring the deviation degree of each of multiple monitoring indicators, including glue supply pressure, glue supply flow rate, and slurry temperature. Using newly detected candidate defect areas as segmentation points, the coating process is divided into multiple time periods, and the defect severity index for each time period is determined based on the number and area of ​​candidate defect areas within each time period. Based on the temporal correlation between the deviation degree of multiple monitoring indicators and the defect severity index, the contribution index of each monitoring indicator to actual coating defects is determined. Based on each contribution index and the defect authenticity of each candidate defect area, the comprehensive defect risk index for actual coating defects at the current moment is determined. Based on the differences in the number and location of candidate defect areas and the differences in the deviation degree of multiple monitoring indicators, the comprehensive similarity between the current moment and each historical moment is determined. Several historical similar moments that meet preset conditions are selected, and based on the pump speed difference between each historical similar moment and the current pump speed, and the comprehensive defect risk index of each historical similar moment, reference weights are determined based on the comprehensive similarity, and the pump speed adjustment value is calculated using a weighted average.

[0010] One possible approach is to determine the contribution index of each monitoring indicator to actual coating defects, including: obtaining the mean deviation sequence of each monitoring indicator over each time period, and obtaining the time series sequence of the defect severity index. For each monitoring indicator, the correlation coefficient between the mean deviation sequence and the time series sequence of the defect severity index is calculated. For each monitoring indicator, the response lag time from the start of the time period to the moment when it is first marked as deviating is determined. Based on the correlation coefficient and the response lag time, the contribution index of the monitoring indicator at the current moment is determined.

[0011] One possible approach involves determining reference weights based on comprehensive similarity and then calculating a weighted average to obtain the pump speed adjustment value. This includes: acquiring the pump speed difference between all historical similar times and the current pump speed; using the product of the comprehensive similarity of the historical similar times and the corresponding comprehensive defect risk index as weights to perform a weighted average of the pump speed differences, resulting in a weighted result; and generating a pump speed adjustment value based on the weighted result and the current pump speed, wherein the single adjustment range of the pump speed adjustment value does not exceed a preset proportion of the rated pump speed.

[0012] One possible approach to acquiring monitoring data includes: receiving multi-source sensor data transmitted from an industrial fieldbus via an edge computing node, whereby the multi-source sensor data includes at least image data; the edge computing node also performs timestamp alignment and filtering / denoising processing on the received multi-source sensor data to obtain the monitoring data.

[0013] Secondly, an automatic control system for a slot extrusion coating adhesive supply process is provided, comprising: The data acquisition unit is used to acquire monitoring data of the substrate surface after coating. The monitoring data includes image data acquired in a time sequence.

[0014] The defect differentiation unit is used to identify candidate defect areas on the surface of the substrate after coating based on monitoring data, and to determine temporary interference defects and real coating defects in the candidate defect areas based on the morphological characteristics and grayscale distribution characteristics of the candidate defect areas in the image data over time.

[0015] The adjustment value determination unit is used to determine the pump speed adjustment value at the current moment for the identified actual coating defects, based on the difference in pump speed of the glue supply pump under the historical coating state that meets the preset conditions of similarity with the current coating state. The historical coating state is determined from historical operating data based on the characteristics of the candidate defect area and the comprehensive similarity of the monitoring data.

[0016] The control unit is used to control the pump speed of the glue supply pump according to the pump speed adjustment value.

[0017] As one possible approach, the monitoring data also includes thickness measurement data. The defect differentiation unit is specifically used for: Based on the deviation of pixel grayscale values ​​from the average grayscale value of the image, the deviation of pixel grayscale values ​​from the grayscale values ​​of pixels at the same coordinates in the previous frame, and the thickness deviation of the corresponding thickness measurement data from the preset substrate thickness reference value, candidate defect pixels are determined, and candidate defect regions composed of candidate defect pixels are extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel.

[0018] Candidate defect regions are matched between adjacent frames of image data. Based on the contour clarity, grayscale distribution fluctuation, area change, and number of consecutively stable frames of the matched candidate defect regions, temporary interference defects and real coating defects are determined.

[0019] This application offers the following advantages: By acquiring time-series image data as the basis for monitoring, candidate defect regions are identified. Then, based on the dynamic changes in the morphological and gray-scale distribution characteristics of these candidate defect regions within the time-series images, temporary interference defects and genuine coating defects are distinguished. Furthermore, for the identified genuine coating defects, the pump speed adjustment value for the current moment is obtained by combining the differences in the glue supply pump speed under historical coating conditions determined from historical operating data based on comprehensive similarity. This allows for the control of the glue supply pump speed. In this way, by combining the identification of coating anomalies with data-driven historical state matching, pump speed adjustment is targeted only at coating defects requiring intervention. Moreover, the adjustment range is referenced by pump speed parameters under similar conditions in historical operating data, effectively reducing ineffective adjustments caused by misjudged temporary interference and improving the accuracy of glue supply pump speed control and the stability of the coating process. Attached Figure Description

[0020] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic flowchart illustrating an automatic control method for a slot extrusion coating adhesive supply process according to an embodiment of this application. Figure 2 This is a schematic diagram of the system architecture of an automatic control system for a slot extrusion coating adhesive supply process provided in one embodiment of this application. Detailed Implementation

[0022] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an automatic control method for a slot extrusion coating adhesive supply process proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0024] The following describes in detail, with reference to the accompanying drawings, a specific scheme for an automatic control method for a slot extrusion coating adhesive supply process provided in this application.

[0025] Please see Figure 1 It illustrates a flowchart of an automatic control method for a slot extrusion coating adhesive supply process according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps: S101-S104.

[0026] S101. Obtain monitoring data of the substrate surface after coating.

[0027] The monitoring data includes image data collected in a time sequence.

[0028] In slot extrusion coating production lines, the surface condition of the coated substrate directly affects the coating quality of the final product. To achieve automated control of the coating adhesive supply process, monitoring data reflecting the coating surface condition and the operating parameters of the adhesive supply process must first be acquired.

[0029] In some implementations, monitoring points are located behind the coating die and in front of the drying section inlet. Industrial cameras are installed at these monitoring points, perpendicular to the coating film surface, to acquire images sequentially of the coated substrate surface. The image data includes multiple consecutively acquired images of the coating film surface, with each frame containing grayscale information for each pixel on the coating film surface.

[0030] A thickness gauge is installed at the monitoring point, with its measurement point aligned with the center of the industrial camera's field of view, to acquire thickness measurement data of the coated film at the corresponding location. For example, the thickness gauge can be... X-ray thickness gauges, laser thickness gauges, or other similar instruments are used to achieve non-contact online thickness measurement. If the sampling frequency of the thickness gauge is typically lower than that of an industrial camera, then the current output data of the thickness gauge represents the thickness data of multiple frames within a time window.

[0031] In addition, pressure sensors, flow meters, and temperature sensors can be installed on the main glue supply line to obtain glue supply pressure, glue supply flow rate, and slurry temperature in real time, respectively; a speed sensor can be installed at the substrate drive roller end of the coating machine to monitor the substrate running speed; and the current pump speed of the glue supply pump can be obtained based on the encoder.

[0032] In some embodiments, the monitoring data may also include thickness measurement data, adhesive supply pressure during coating, adhesive supply flow rate, slurry temperature, substrate running speed, and current pump speed.

[0033] As one possible approach, the process of acquiring monitoring data includes: S1011. Receive multi-source sensor data transmitted from the industrial fieldbus via the edge computing node. The multi-source sensor data includes image data.

[0034] In some embodiments, the edge computing node also performs timestamp alignment and filtering / denoising on the multi-source sensor data it receives to obtain monitoring data.

[0035] Specifically, an edge computing node includes a processor, memory, and a data interface. Data acquired by industrial cameras, thickness gauges, pressure sensors, flow meters, temperature sensors, speed sensors, and encoders is transmitted to the edge computing node via an industrial fieldbus. The edge computing node timestamps the received multi-source sensor data to ensure that data from different sensors at the same time can correspond to each other; simultaneously, it filters and denoises the data from each channel to eliminate electromagnetic interference and random noise in the industrial environment, thereby obtaining reliable monitoring data.

[0036] S102. Based on the monitoring data, identify candidate defect areas on the surface of the substrate after coating, and determine the temporary interference defects and real coating defects in the candidate defect areas based on the morphological characteristics and grayscale distribution characteristics of the candidate defect areas in the image data over time.

[0037] As one possible approach, the process for determining temporary interference defects and actual coating defects includes: S1021. Based on the deviation of the pixel gray value from the average gray value of the image, the deviation of the pixel gray value from the gray value of the same coordinate pixel in the previous frame image, and the thickness deviation of the thickness measurement data at the corresponding position from the preset substrate thickness reference value, candidate defect pixels are determined, and candidate defect regions composed of candidate defect pixels are extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel.

[0038] S1022. Extract the candidate defect region composed of candidate defect pixels.

[0039] S1023. Match candidate defect regions between adjacent frames of image data. Based on the outline clarity, grayscale distribution fluctuation, area change, and number of consecutively stable frames of the matched candidate defect regions, determine temporary interference defects and real coating defects.

[0040] As one possible approach, the process of determining candidate defective pixels includes: S201. For any pixel in any frame of the image data, calculate the first deviation of the gray value of any pixel from the average gray value of the frame, and the second deviation of the gray value of any pixel from the gray value of the pixel with the same coordinates as the pixel in the previous frame.

[0041] Specifically, with the first Taking a frame image as an example, let's record the pixels. In the The grayscale value in the frame image , No. The average gray level of the frame image is pixel In the The gray values ​​at the same coordinates in a frame image Then the pixel The first deviation is the grayscale value. With grayscale mean The absolute value of the difference This is used to characterize the degree to which the grayscale value of a pixel deviates from the overall grayscale level of the image. The second deviation is the grayscale value. grayscale value of the pixel with the same coordinates as the previous frame The absolute value of the difference This is used to characterize the degree of grayscale change of a pixel between adjacent frames.

[0042] S202. Obtain the thickness deviation between the thickness measurement data at any given moment in any frame of the image and the preset substrate thickness reference value.

[0043] Specifically, record the first The difference between the coating film thickness measured by the thickness gauge at the corresponding moment of the frame image and the preset substrate thickness reference value is... Take the absolute value of the difference. As a thickness deviation, when When the thickness deviation exceeds the preset threshold, it is considered that there may be a coating abnormality at that location, covering two defect types: coating thickening and coating thinning. The substrate thickness reference value is the standard thickness required by the coating process. For example, in the coating process of lithium-ion battery electrodes, the substrate thickness reference value can be set to 100μm. The specific value of the substrate thickness reference value can be determined according to the specific coating process requirements.

[0044] S203. Determine the defect probability value of any pixel based on the first deviation, the second deviation, and the thickness deviation.

[0045] Specifically, pixels In the Defect probability value in frame image The method for determining this is: the thickness deviation amount First deviation Second deviation Multiplying the three values ​​together yields the probability value of the defect. ,Right now: When the gray value of any pixel deviates significantly from the average gray value of the image, deviates significantly from the gray value of the pixel at the same coordinates in the previous frame, and has a significant thickness deviation at the corresponding position, the probability of the pixel being defective is high, indicating that the pixel is likely to have a coating defect.

[0046] S204. Mark pixels with a defect probability value greater than a preset threshold as candidate defect pixels.

[0047] In some embodiments, pixels with a defect probability value greater than a preset threshold are marked as candidate defect pixels. Specifically, the defect probability value is normalized based on a preset normalization algorithm. The maximum and minimum values ​​referenced for normalization are the statistical maximum and minimum values ​​of the defect probability values ​​of all pixels in the current frame image. Pixels whose normalization result is greater than a first preset threshold are marked as candidate defect pixels.

[0048] The first preset threshold value ranges from 0.3 to 0.7. The specific value can be determined according to the coating process accuracy requirements. For example, the first preset threshold value can be 0.5.

[0049] The specific method for extracting candidate defect regions composed of candidate defect pixels is as follows: Connectivity analysis is performed on all candidate defect pixels in a single frame image, and spatially adjacent candidate defect pixels are clustered into a single candidate defect region. Taking the first... Candidate defect regions in frame images For example, record the probability value of defects for all pixels in this region. The mean value is the area defect probability value. This is used to characterize the candidate defect region in the first... The overall probability of defects in a frame image.

[0050] When matching candidate defect regions between adjacent frames, the substrate movement direction is used as a reference. The marker in the frame image is the same as the first Candidate defect regions in frame image The region with the largest overlap area is selected as the candidate defect region. In the Matching regions in a frame image.

[0051] As one possible approach, temporary interference defects and real coating defects are determined based on the contour sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions. These include: S205. Compare the area of ​​the matched candidate defect region when it first appears with the area at the current moment to determine the amount of area change.

[0052] Specifically, record candidate defect regions The area at the time of its first appearance was In the current frame (the first frame) The area in the frame image is Through calculation and The absolute value of the difference determines the area change of the candidate defect region, which is used to characterize the area evolution trend of the region over time. When the candidate defect region is a temporary interference defect, its area may gradually shrink during the leveling process; when the candidate defect region is a real coating defect, its area usually remains stable or changes symmetrically as the substrate moves into / out of the camera's field of view.

[0053] S206. The standard deviation of the probability value of the candidate defect region in all image frames in which it appears is used as the degree of grayscale distribution fluctuation.

[0054] Specifically, record candidate defect regions Region defect probability value in all image frames in which it appears The standard deviation is This is used to characterize the degree of grayscale distribution fluctuation of the candidate defect region over time. When the candidate defect region is a temporary interference defect (orange peel pattern), it gradually disappears or undergoes significant morphological changes during the leveling process, resulting in its absence in some image frames or significant fluctuations in the region's defect probability value. The value is relatively large; when the candidate defect region is a real coating defect, its morphology is stable, its outline is clear, and it can be stably detected in consecutive frames, therefore The value is relatively small.

[0055] S207. The average grayscale gradient magnitude of the edge pixels in the candidate defect region is used as the contour sharpness.

[0056] Specifically, let the mean gray-level gradient magnitude of all edge pixels of candidate defect region j in the k-th frame be denoted as . This is used to characterize the clarity of the candidate defect region's outline. Edge pixels are the pixels on the boundary of the candidate defect region, and the gray-level gradient magnitude is the modulus of the gray-level change rate at each edge pixel. The clearer the outline of the candidate defect region, the larger the gray-level gradient magnitude of its edge pixels. The larger the value, the better.

[0057] S208. Based on the area change, grayscale distribution fluctuation, outline clarity, and number of consecutively stable frames, determine the probability value that the candidate defect area is a real coating defect.

[0058] Specifically, with the first Candidate defect regions in frame images For example, up to the [number]th The probability value of the candidate defect region being a real coating defect at the corresponding time in the frame image. The method for determining this is: [The text abruptly ends here, likely due to an incomplete sentence or a formatting error.] With the number of consecutively stable frames The product of and is used as the numerator. Add a preset constant as the denominator, and divide the numerator by the denominator to obtain the probability value. ,Right now in, Candidate defect region The number of consecutive and stable frames up to the current moment is used to characterize the stability and persistence of the region in the time dimension; when the region is not detected in a certain frame, the consecutive frame count is reset to zero and the count is restarted. Candidate defect region The standard deviation of the probability value of regional defects in all image frames in which it appears is used to characterize the degree of fluctuation in grayscale distribution; Candidate defect region Area at first appearance; Candidate defect region In the Area in a frame image; preset constant To avoid the denominator being zero, a value of 0.002 can be used. The resulting calculation error is within acceptable limits. The higher the clarity of the candidate defect region's outline, the more consecutive and stable frames it appears in, the smaller the fluctuation in grayscale distribution, and the more stable its area change, the higher its probability value. The larger the value, the greater the likelihood that the area represents a genuine coating defect.

[0059] Based on a pre-defined normalization algorithm, the probability values ​​are... Normalized to a preset range, candidate defect regions are obtained. In the Defect realism in frame images The probability value is used to characterize the credibility of the candidate defect region as a real coating defect. All parameters need to be pre-standardized before the probability value is calculated. The pre-standardization can be performed using the Min-Max standardization method, which maps each parameter to the interval [0,1]. The maximum and minimum values ​​referenced for normalization are determined based on the statistical results of the parameters corresponding to all candidate defect regions at the current time. When the number of candidate defect regions is less than the preset number, the maximum and minimum values ​​of the corresponding parameters in the historical statistical data are used as a supplement. When the statistical maximum and minimum values ​​are equal, all normalization results of the parameter are set to 1 to avoid calculation anomalies caused by the denominator being zero.

[0060] S103. For the identified actual coating defects, based on the pump speed difference of the glue supply pump under the historical coating state that meets the preset conditions of similarity with the current coating state, determine the pump speed adjustment value at the current moment.

[0061] After identifying the actual coating defect, it is necessary to further assess the severity of the defect and, in conjunction with historical experience in adjusting the pump speed under similar coating conditions, determine the pump speed adjustment value at the current moment.

[0062] As one possible approach, the process of determining the pump speed adjustment value at the current moment includes: S1031. Obtain the deviation of each of the multiple monitoring indicators, including glue supply pressure, glue supply flow rate and slurry temperature.

[0063] Specifically, for each monitoring indicator, the degree to which its monitored value deviates from its historical average at each time point is obtained. (Based on monitoring indicators...) For example, if it is at time The monitoring value is Deadline The average value of this indicator is Average of monitored values The calculation uses a sliding window method, with the window size being all monitored values ​​within a preset time period (e.g., the previous 5 minutes) from the current moment. This reflects recent operating conditions and avoids the influence of older historical data on current judgments. Therefore, this indicator at time... degree of deviation The method for determining it is: calculate the monitoring value. Compared with the average monitoring value The absolute value of the difference yields the degree of deviation. .

[0064] The deviation is normalized to a preset range based on a pre-defined normalization algorithm. The maximum and minimum values ​​referenced for normalization are the historical statistical maximum and minimum values ​​of the deviation of the monitoring indicator at each time point. When the statistical maximum and minimum values ​​are equal, the normalization result is set to 1. Time points where the normalization result is greater than a second preset threshold are marked as monitoring indicators. The deviation time is used to characterize the abnormal fluctuation of the indicator at that moment. When the deviation at the current moment exceeds the historical maximum value, the normalization result is set to 1; when the deviation at the current moment is lower than the historical minimum value, the normalization result is set to 0.

[0065] The second preset threshold ranges from 0.3 to 0.7. The specific value can be determined according to the sensitivity requirements for abnormal fluctuations in the monitoring indicators. For example, the second preset threshold can be 0.5.

[0066] S1032. Using the newly detected candidate defect areas as segmentation points, the coating process is divided into multiple time periods, and the defect severity index of each time period is determined based on the number and area of ​​candidate defect areas in each time period.

[0067] Specifically, the coating process of the current substrate is divided into multiple time periods, with the detection of newly added candidate defect areas as the time segmentation point. Taking a specific time period as an example, the severity index of defects during that period... The determination method can be: in, For the first The number of candidate defect regions in each time period; For this region in the 1st The maximum area within a time period For this region in the 1st Maximum defect accuracy within a given time period. and Each is the previous time period (the first) In each time period, the maximum area and the maximum defect accuracy of the same candidate defect region in the previous time period.

[0068] S1033. Based on the correlation between the deviation of multiple monitoring indicators and the defect severity index in time series, determine the contribution index of each monitoring indicator to the actual coating defects.

[0069] As one possible approach, the contribution index of each monitoring indicator to actual coating defects is determined, including: S301. Obtain the average deviation sequence of each monitoring indicator in each time period, and obtain the time series sequence of the defect severity index.

[0070] Specifically, record the first Deviation of monitoring indicator z in each time period The mean is The mean deviation of the index is obtained for each time period to form a mean deviation sequence. Simultaneously, the defect severity index for each time period is obtained. The time series that constitutes the defect severity index.

[0071] S302. For each monitoring indicator, calculate the correlation coefficient between the mean deviation sequence and the time series of the defect severity index.

[0072] Specifically, for each monitoring indicator, a preset correlation analysis algorithm is used to calculate the mean deviation sequence of that indicator and the time series sequence of the defect severity index. Pearson correlation coefficient between The correlation coefficient is used to characterize the linear correlation between the degree of deviation of the monitoring indicator and the severity of the defect. The larger the absolute value of the correlation coefficient, the stronger the correlation between the degree of deviation of the monitoring indicator and the severity of the defect.

[0073] S303. For each monitoring indicator, determine the response lag time from the start of the time period to the moment when it is first marked as a deviation.

[0074] Specifically, for each monitoring indicator, the time point at which the indicator is first marked as deviating within each time period is determined. The time interval between this time point and the start time of the time period is calculated, and the average of this time interval across all time periods is taken as denoted as . This is used to characterize the response speed of the monitoring indicator to the occurrence of defects. The shorter the average response lag time, the more timely the response of the monitoring indicator to the occurrence of defects, and the greater the contribution of the indicator to the actual coating defects.

[0075] S304. Based on the correlation coefficient and response lag time, determine the contribution index of the monitoring indicator at the current moment.

[0076] Specifically, for each monitoring indicator, based on the Pearson correlation coefficient of that indicator... and the average response lag time Determine the contribution index of the monitoring indicator at the current moment. The determination method can be adopted using the following formula: in, The absolute value of the Pearson correlation coefficient between the mean sequence of the deviation of the monitoring index z and the time series sequence of the defect severity index; The normalized mean response lag time is dimensionless. For monitoring indicators The average response lag time from the start of the time period is initially marked as the deviation time in each time period. This is the total duration of the sliding window used to calculate the mean of the monitored values. When the monitored indicators... The larger the absolute value of the Pearson correlation coefficient and the smaller the average response lag time, the greater the contribution index of the indicator, indicating that the indicator contributes more to the actual coating defects at the current moment.

[0077] S1034. Based on each contribution index and the defect authenticity of each candidate defect area, determine the comprehensive defect risk index for the actual coating defect at the current moment.

[0078] Specifically, at the current moment For example, based on the contribution index of each monitoring indicator... The defect realism of all candidate defect regions at the current moment. mean Determine the comprehensive defect risk index at the current moment. The determination method is as follows: multiply the contribution index of each monitoring indicator by the mean defect authenticity of all candidate defect areas, and then sum them to obtain the comprehensive defect risk index, i.e.: in, For monitoring indicators Contribution index at the current moment; The total number of monitoring indicators, i.e., the number of indicators involved in the contribution index calculation; For a moment The defect accuracy of all candidate defect regions detected The mean of the values ​​is used to characterize the overall credibility of all candidate defect regions at the current moment. For monitoring indicators At any moment The deviation is the value after normalization. The higher the comprehensive defect risk index, the higher the risk of real defects in the current coating process, and the greater the urgency to adjust the glue supply pump speed.

[0079] S1035. Based on the differences in the number of candidate defect regions, their locations, and the degree of deviation of multiple monitoring indicators, determine the comprehensive similarity between the current moment and each historical moment.

[0080] Specifically, at the current moment With a certain historical moment Taking this as an example, we determine the overall similarity between the two. The determination of overall similarity considers the differences in the following three dimensions: The first dimension, the difference in the number of candidate defect regions, is used to characterize the degree of difference in the number of defect regions between the current time and historical time.

[0081] The second dimension is the positional difference of the candidate defect regions. Specifically, this involves comparing a candidate defect region at the current moment with that at historical moments. The candidate defect region closest to the center point of the coordinate system is matched, the distance between the two regions is calculated, and the distances between all corresponding regions are summed to obtain the positional difference, which is used to characterize the degree of difference in the spatial distribution of defect regions between the current time and the historical time.

[0082] The third dimension is the difference in the degree of deviation of multiple monitoring indicators. Specifically, the difference in the degree of deviation of each monitoring indicator between the current time and the historical time is calculated, and the difference in the degree of deviation of all monitoring indicators is summed to obtain the comprehensive difference in the degree of deviation of the monitoring indicators, which is used to characterize the degree of difference between the current time and the historical time in the multi-dimensional monitoring indicators of the coating process.

[0083] Based on the differences in the three dimensions mentioned above, the method for determining the comprehensive similarity between the current moment and historical moments is as follows: using the natural constant as the base, and the absolute value of the quantitative difference... The absolute value of positional differences The absolute value of the overall difference between the deviation from the monitoring indicators The negative of the weighted sum of the three factors is used as the exponent to calculate the overall similarity. ,Right now: in, For the present moment and historical moments The sum of distances after matching each candidate defect region. , , The weight coefficients for each dimension of difference satisfy... + + =1, for example, all three can be set to 1 / 3, or the values ​​can be differentiated according to the importance of each dimension to the coating state.

[0084] When the differences in the above three dimensions decrease, the overall similarity approaches 1, indicating that the current coating state is highly similar to the historical coating state; when the difference in any dimension increases, the overall similarity decreases accordingly.

[0085] All parameters must undergo pre-standardization before comprehensive similarity calculation. The pre-standardization process can adopt the Z-score standardization method, in which the mean and standard deviation of each parameter are determined based on the statistical results of the parameter at each historical time, so that the parameters of each dimension have the same scale, thereby eliminating the influence of the difference in the dimensions of different parameters on the calculation results.

[0086] The comprehensive similarity is normalized to a preset interval based on a pre-defined normalization algorithm. The maximum and minimum values ​​referenced for normalization are the statistical maximum and minimum values ​​of the comprehensive similarity at each historical time. Time moments with normalization results greater than a third preset threshold are marked as historical similarity moments. When the comprehensive similarity at the current time exceeds the historical statistical maximum value, the normalization result is set to 1; when the comprehensive similarity at the current time is lower than the historical statistical minimum value, the normalization result is set to 0.

[0087] The third preset threshold ranges from 0.3 to 0.9. The specific value can be determined based on the strictness of matching historical similar states. For example, the third preset threshold can be 0.6.

[0088] S1036. Select several historical similar moments that meet the preset conditions for comprehensive similarity, and determine the reference weight based on the pump speed difference between the pump speed at each historical similar moment and the current pump speed, as well as the comprehensive defect risk index at each historical similar moment, and calculate the pump speed adjustment value by weighting.

[0089] Specifically, if the current pump speed of the glue supply pump is lower than the pump speed at similar historical moments, the current glue supply pump will be accelerated; if the current pump speed of the glue supply pump is higher than the pump speed at similar historical moments, the current glue supply pump will be decelerated.

[0090] As one possible approach, reference weights are determined based on comprehensive similarity, and the pump speed adjustment value is obtained through weighted calculation, including: S305. Obtain the pump speed difference between the pump speed at all historically similar times and the pump speed at the current time.

[0091] Specifically, taking the current moment as an example, for each historically similar moment, the difference between the pump speed at that historically similar moment and the pump speed at the current moment is calculated. When the difference A value greater than zero indicates that the pump speed at similar historical moments is higher than the current pump speed; when the difference is greater than zero... A value less than zero indicates that the pump speed at similar historical moments is lower than the pump speed at the current moment.

[0092] S306. Using the product of the comprehensive similarity at similar historical moments and the corresponding comprehensive defect risk index as weights, the pump speed difference is weighted and averaged to obtain the weighted result.

[0093] Specifically, for each historically similar moment, a reference weight is determined. The reference weight is the product of the overall similarity of that historically similar moment and the overall defect risk index of that historically similar moment. The greater the overall similarity and the greater the overall defect risk index, the greater the reference weight for that historically similar moment, indicating that the pump speed adjustment experience at that historically similar moment has higher reference value for the current moment. Since the overall defect risk index reflects the severity of coating defects at that historical moment, incorporating the overall defect risk index into the weight calculation ensures that, under the premise of overall similarity in coating conditions, historical moments with higher defect risks receive higher reference weights, thus guiding the current pump speed to reference effective adjustment experience under high-risk conditions.

[0094] The pump speed difference is weighted and averaged using reference weights from similar historical moments to obtain a weighted result. Specifically, the pump speed difference at all similar historical moments is multiplied by its corresponding reference weight, summed, and then divided by the sum of all reference weights to obtain the weighted result. If the sum of the reference weights at all similar historical moments is zero, no pump speed adjustment is performed at the current moment; that is, the pump speed adjustment value is set to zero.

[0095] S307. Based on the weighted result and the pump speed at the current moment, generate a pump speed adjustment value, wherein the single adjustment range of the pump speed adjustment value does not exceed a preset proportion of the rated pump speed.

[0096] The preset ratio ranges from 3% to 10%, for example, it can be 7%, to ensure a smooth adjustment process and avoid exceeding the physical load of the equipment, which could cause film breakage or unstable glue supply.

[0097] Specifically, the pump speed adjustment formula is to directly add the weighted result to the current pump speed in a limited manner, that is: in, This represents the maximum allowable adjustment step size in a single operation. To the rated pump speed of the glue supply pump, The preset ratio (the value ranges from 3% to 10%, for example, the value can be 7%). For the limiting function, Limited to the range Inside. When the weighted result When the pump speed is positive, the pump speed is increased; when the pump speed is negative, the pump speed is decreased, and the adjustment range shall not exceed [a certain value]. The adjustment range should never exceed the preset ratio of the rated pump speed to ensure a smooth adjustment process and avoid exceeding the physical load of the equipment, which could lead to membrane breakage or unstable glue supply.

[0098] To ensure the safe operation of the glue supply pump, the single adjustment range of the pump speed adjustment value shall not exceed the preset ratio of the rated pump speed of the glue supply pump, so as to avoid exceeding the physical load of the equipment, which may lead to film breakage or unstable glue supply.

[0099] S104. Control the pump speed of the glue supply pump according to the pump speed adjustment value.

[0100] Specifically, an adjustment command is generated at the edge computing node based on the pump speed adjustment value, and the adjustment command is transmitted to the industrial control computer. The industrial control computer then drives the glue supply pump to perform pump speed adjustment, thereby achieving precise control of the glue supply process in the slot extrusion coating process. Each adjustment process must meet the physical limitations of the equipment to ensure a smooth adjustment process and avoid exceeding the physical load of the equipment, which could lead to film breakage or unstable glue supply.

[0101] Please see Figure 2 This diagram illustrates the system architecture of an automatic control system for a slot extrusion coating adhesive supply process according to an embodiment of the present invention. The automatic control system includes: a data acquisition unit 21, a defect differentiation unit 22, an adjustment value determination unit 23, and a control unit 24. The units communicate bidirectionally via a communication link to ensure real-time interaction of collected data and analysis results. The communication link can employ wired or wireless transmission methods to meet the communication needs of different monitoring scenarios.

[0102] The data acquisition unit 21 is used to acquire monitoring data of the substrate surface after coating. The monitoring data includes image data acquired in a time sequence.

[0103] The defect differentiation unit 22 is used to identify candidate defect areas on the surface of the substrate after coating based on monitoring data, and to determine temporary interference defects and real coating defects in the candidate defect areas based on the morphological characteristics and grayscale distribution characteristics of the candidate defect areas in the image data over time.

[0104] The adjustment value determination unit 23 is used to determine the pump speed adjustment value at the current moment for the identified real coating defects, based on the difference in pump speed of the glue supply pump under the historical coating state that meets the preset conditions of similarity with the current coating state. The historical coating state is determined from historical operating data based on the characteristics of the candidate defect area and the comprehensive similarity of the monitoring data.

[0105] Control unit 24 is used to control the pump speed of the glue supply pump according to the pump speed adjustment value.

[0106] As one possible approach, the monitoring data also includes thickness measurement data. The defect differentiation unit 22 is specifically used for: Based on the deviation of the pixel gray value from the average gray value of the image, the deviation of the pixel gray value from the gray value of the same coordinate pixel in the previous frame, and the thickness deviation of the thickness measurement data at the corresponding position from the preset substrate thickness reference value, candidate defect pixels are determined, and candidate defect regions composed of candidate defect pixels are extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel.

[0107] Candidate defect regions are matched between adjacent frames of image data. Based on the contour clarity, grayscale distribution fluctuation, area change, and number of consecutively stable frames of the matched candidate defect regions, temporary interference defects and real coating defects are determined.

[0108] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0109] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An automatic control method for a slot extrusion coating adhesive supply process, characterized in that, include: Acquire monitoring data of the substrate surface after coating, the monitoring data including image data acquired in a time sequence; Based on the monitoring data, candidate defect areas on the surface of the coated substrate are identified, and based on the changes in the morphological features and grayscale distribution features of the candidate defect areas in the image data over time, temporary interference defects and real coating defects in the candidate defect areas are determined. For the identified actual coating defects, the pump speed adjustment value at the current moment is determined based on the pump speed difference of the glue supply pump under the historical coating state that meets the preset conditions of similarity with the current coating state. The historical coating state is a historical coating state that meets the preset conditions of similarity with the current coating state, determined from historical operating data based on the characteristics of the candidate defect area and the comprehensive similarity of the monitoring data. The pump speed of the glue supply pump is controlled according to the pump speed adjustment value.

2. The automatic control method for the slit extrusion coating adhesive supply process according to claim 1, characterized in that, The monitoring data also includes thickness measurement data. The process for determining the temporary interference defects and the actual coating defects includes: Based on the deviation of the pixel grayscale value from the average grayscale value of the image, the deviation of the pixel grayscale value from the grayscale value of the same coordinate pixel in the previous frame image, and the thickness deviation of the thickness measurement data at the corresponding position from the preset substrate thickness reference value, candidate defect pixels are determined, and the candidate defect region composed of the candidate defect pixels is extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel. Candidate defect regions are matched between adjacent frames of the image data. Based on the outline sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions, the temporary interference defects and the actual coating defects are determined.

3. The automatic control method for the slit extrusion coating adhesive supply process according to claim 2, characterized in that, The process of determining candidate defective pixels includes: For any pixel in any frame of the image data, calculate the first deviation of the gray value of the pixel from the average gray value of the image in any frame, and the second deviation of the gray value of the pixel from the gray value of the pixel with the same coordinates in the previous frame. Obtain the thickness deviation between the thickness measurement data and the preset substrate thickness reference value at the corresponding moment of any frame image; The defect probability value of any pixel is determined based on the first deviation, the second deviation, and the thickness deviation. Pixels with a defect probability value greater than a preset threshold are marked as candidate defect pixels.

4. The automatic control method for the slit extrusion coating adhesive supply process according to claim 2, characterized in that, The process of determining the temporary interference defect and the real coating defect based on the outline sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions includes: The area of ​​the matched candidate defect region at the first appearance is compared with the area at the current moment to determine the amount of area change; The standard deviation of the probability value of the candidate defect region in all image frames in which it appears is used as the degree of grayscale distribution fluctuation. The mean value of the grayscale gradient of the edge pixels in the candidate defect region is used as the contour sharpness. Based on the area change, the degree of grayscale distribution fluctuation, the outline clarity, and the number of consecutively stable occurrences, the probability value of the candidate defect region being a real coating defect is determined.

5. The automatic control method for the slit extrusion coating adhesive supply process according to claim 4, characterized in that, The process of determining the pump speed adjustment value at the current moment includes: The deviation of each of the multiple monitoring indicators is obtained, including glue supply pressure, glue supply flow rate and slurry temperature; The coating process is divided into multiple time periods based on the detected new candidate defect areas, and the defect severity index for each time period is determined according to the number and area of ​​the candidate defect areas in each time period. Based on the temporal correlation between the deviation of the multiple monitoring indicators and the defect severity index, the contribution index of each monitoring indicator to the actual coating defect is determined. Based on the contribution index and the defect authenticity of each candidate defect region, determine the comprehensive defect risk index for the actual coating defect at the current moment; Based on the differences in the number and location of the candidate defect regions and the differences in the degree of deviation of the multiple monitoring indicators, the comprehensive similarity between the current moment and each historical moment is determined. Several historical similar moments that meet the preset conditions for comprehensive similarity are selected. Based on the pump speed difference between the pump speed at each historical similar moment and the current pump speed, and the comprehensive defect risk index at each historical similar moment, a reference weight is determined based on the comprehensive similarity, and the pump speed adjustment value is obtained by weighted calculation.

6. The automatic control method for the slit extrusion coating adhesive supply process according to claim 5, characterized in that, The determination of the contribution index of each monitoring indicator to the actual coating defect includes: Obtain the mean deviation sequence of each monitoring indicator in each time period, and obtain the time series sequence of the defect severity index; For each of the monitoring indicators, the correlation coefficient between the mean deviation sequence and the time series sequence of the defect severity index is calculated. For each of the monitoring indicators, determine the response lag time from the start of the time period to the moment when it is first marked as a deviation; Based on the correlation coefficient and the response lag time, the contribution index of the monitoring indicator at the current moment is determined.

7. The automatic control method for the slit extrusion coating adhesive supply process according to claim 5, characterized in that, The step of determining reference weights based on the comprehensive similarity and calculating the pump speed adjustment value by weighting includes: Obtain the pump speed difference between the pump speed at all historically similar moments and the pump speed at the current moment; The pump speed difference is weighted and averaged using the product of the comprehensive similarity at the historical similar time and the corresponding comprehensive defect risk index as the weight, to obtain the weighted result. Based on the weighted result and the pump speed at the current moment, the pump speed adjustment value is generated, wherein the single adjustment range of the pump speed adjustment value does not exceed a preset proportion of the rated pump speed.

8. The automatic control method for the slit extrusion coating adhesive supply process according to claim 1, characterized in that, The process of acquiring the monitoring data includes: The edge computing node receives multi-source sensor data transmitted from the industrial fieldbus, and the multi-source sensor data includes at least the image data. The edge computing node also performs timestamp alignment and filtering / denoising on the multi-source sensor data it receives to obtain the monitoring data.

9. An automatic control system for a slot extrusion coating adhesive supply process, characterized in that, include: A data acquisition unit is used to acquire monitoring data of the substrate surface after coating, the monitoring data including image data acquired in a time sequence; The defect differentiation unit is used to identify candidate defect areas on the surface of the coated substrate based on the monitoring data, and to determine temporary interference defects and real coating defects in the candidate defect areas based on the changes in the morphological features and grayscale distribution features of the candidate defect areas in the image data over time. The adjustment value determination unit is used to determine the pump speed adjustment value at the current moment based on the difference in pump speed of the glue supply pump under the historical coating state that meets the preset conditions of similarity with the current coating state, for the determined real coating defects. The historical coating state is a coating state that meets the preset conditions of similarity with the current coating state, determined from historical operating data based on the characteristics of the candidate defect area and the comprehensive similarity of the monitoring data. The control unit is used to control the pump speed of the glue supply pump according to the pump speed adjustment value.

10. The automatic control system for the slit extrusion coating adhesive supply process according to claim 9, characterized in that, The monitoring data also includes thickness measurement data; The defect differentiation unit is specifically used for: Based on the deviation of the pixel grayscale value from the average grayscale value of the image, the deviation of the pixel grayscale value from the grayscale value of the same coordinate pixel in the previous frame image, and the thickness deviation of the thickness measurement data at the corresponding position from the preset substrate thickness reference value, candidate defect pixels are determined, and the candidate defect region composed of the candidate defect pixels is extracted. The region defect probability value of the candidate defect region is obtained from the defect probability value of each candidate defect pixel. Candidate defect regions are matched between adjacent frames of the image data. Based on the outline sharpness, grayscale distribution fluctuation, area change, and number of consecutively stable occurrences of the matched candidate defect regions, the temporary interference defects and the actual coating defects are determined.

Images