A method and system for on-line monitoring of zinc layer of galvanized sheet based on laser thickness measurement

Abstract

This invention relates to the field of optical measurement technology and proposes an online monitoring method and system for zinc coating on galvanized steel sheets based on laser thickness measurement. The method includes: collecting time-series distance data of multiple measurement points along the movement direction of the galvanized steel sheet to construct a composite thickness time-series dataset; establishing a multi-dimensional dynamic fluctuation feature matrix by combining the thickness change trajectory and lateral differences; simultaneously analyzing two different response thickness measurement signals at the same measurement point to obtain peak intensity feature values; coupling to generate dynamic zinc coating strength characterization coefficients; adaptively stripping the substrate drift reference value; and extracting the transient zinc coating thickness change sequence; comparing the instantaneous thickness with the allowable fluctuation range point by point; and outputting a trigger command with a time marker to the air knife controller when the limit is exceeded, thereby realizing real-time monitoring and control of zinc coating thickness anomalies. This invention can improve the efficiency of online monitoring of zinc coating on galvanized steel sheets.

Description

Technical Field

[0001] This invention relates to the field of optical measurement technology, and in particular to an online monitoring method and system for zinc layer of galvanized steel sheets based on laser thickness measurement. Background Technology

[0002] Currently, online detection of zinc layer thickness in galvanized steel sheets primarily employs laser triangulation or confocal methods to obtain the total sheet thickness. This involves arranging multiple fixed measuring points along the width of the production line to continuously collect distance variation data on the galvanized sheet surface, and then indirectly estimating the zinc layer thickness by combining this data with a substrate reference value. These methods rely on prior calibration of the substrate thickness or the assumption that the substrate thickness remains constant over a short period, using a single thickness measurement signal as the sole information source for thickness calculation. However, in actual continuous galvanizing production lines, dynamic factors such as tension fluctuations, sheet shape oscillations, and uneven transverse thickness distribution occur during high-temperature annealing and air-knife zinc scraping. Detection methods using a single measuring point or a single signal cannot distinguish the respective contributions of substrate thickness drift and surface zinc layer changes to the distance measurement data, resulting in inherent systematic biases in the measurement results that cannot be effectively identified.

[0003] This inherent limitation is particularly pronounced under complex operating conditions such as frequent changes in production line speed, large fluctuations in incoming plate thickness, or dynamic adjustments to air knife parameters. Because substrate thickness drift and zinc layer thickness changes manifest as frequency band aliasing and coupling in the measurement signal, calculation methods based on fixed reference values ​​or single signal sources can lead to continuous thickness deviations due to substrate fluctuations being incorrectly attributed to zinc layer changes, thus triggering frequent false alarms or missed alarms. Simultaneously, the lack of rapid response data to changes in zinc layer condition in the air knife control process makes it difficult to track actual coating deviations in a timely manner, ultimately affecting the lateral uniformity of galvanized products and the finished product yield. Summary of the Invention

[0004] This invention provides a method and system for online monitoring of zinc layer on galvanized steel sheets based on laser thickness measurement, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides an online monitoring method for zinc coating on galvanized steel sheets based on laser thickness measurement, comprising: Pt.1. Along the direction of movement of the galvanized sheet, multiple measuring points are arranged in the width direction of the galvanized sheet to simultaneously acquire distance measurement data of each measuring point position changing over time, and arrange them according to the sampling time to form a composite thickness time series dataset. Pt.2. Determine the thickness change trajectory and width direction difference value based on the composite thickness time series dataset, and perform spatiotemporal correlation between the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. Pt.3. Obtain two thickness measurement signals with different response characteristics synchronously output at the same detection point in the galvanized sheet, and synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. Pt.4. Couple and correlate the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients. Based on the dynamic zinc layer strength characterization coefficients, perform adaptive low-frequency stripping on the multidimensional dynamic fluctuation characteristic matrix to obtain the substrate drift reference value. Remove the reference value from the multidimensional dynamic fluctuation characteristic matrix to obtain the transient zinc layer thickness change sequence. Pt.5. Compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, generate a trigger command containing an identifier of the time of the abnormality and output it to the air knife controller.

[0006] In a preferred embodiment, the distance measurement data of each measuring point location changing over time is simultaneously acquired at multiple measuring points arranged in the width direction of the galvanized sheet, and arranged according to the sampling time to form a composite thickness time-series dataset, including: Distance sampling is triggered at each measuring point, and the sampled distance measurement data is bound to the timestamp of the sampling time to obtain the distance measurement data of the galvanized sheet; The distance measurement data of each measuring point are arranged in chronological order according to the timestamps to obtain the original distance sequence of the galvanized sheet; The continuity of adjacent data points in the original distance sequence is checked, and data points that do not meet the continuity condition are removed to obtain the distance sequence of the galvanized sheet; The distance measurement data in the distance sequence are arranged sequentially according to the timestamp to obtain the composite thickness time series dataset of the galvanized sheet.

[0007] In a preferred embodiment, the step of determining the thickness change trajectory and width direction difference value based on the composite thickness time-series dataset, and spatiotemporally correlating the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix includes: Based on the composite thickness time series dataset, differential analysis is performed on the distance measurement data at consecutive time points along the movement direction of the galvanized sheet to obtain the thickness gradient sequence of the galvanized sheet; Based on the composite thickness time series dataset, the distance measurement data at different measuring points at the same time point are subjected to lateral difference mapping to obtain the lateral difference distribution sequence of the galvanized sheet; The gradient values ​​in the thickness gradient sequence are spatiotemporally coupled with the width direction difference values ​​in the lateral difference distribution sequence to obtain the spatiotemporal correlation feature pairs of the galvanized sheet; The spatiotemporal correlation feature pairs are arranged sequentially according to the time dimension and then matrixed according to the measurement point dimension to obtain a multidimensional dynamic fluctuation feature matrix.

[0008] In a preferred embodiment, the step of simultaneously analyzing the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet includes: The two thickness measurement signals with different response characteristics are separated to obtain the first thickness measurement signal and the second thickness measurement signal. Peak location is performed on the first and second thickness measurement signals respectively, and the peak intensity characteristic values ​​of the first and second channels of the galvanized sheet are read at the highest point of the signal amplitude on the time axis. The first peak strength value and the second peak strength value are synchronously verified to obtain the first verified peak strength value and the second verified peak strength value of the galvanized sheet. The first verified peak strength value and the second verified peak strength value are associated and packaged together to form the two peak strength characteristic values ​​of the galvanized sheet.

[0009] In a preferred embodiment, the step of coupling and correlating the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients includes: Obtain a preset strength reference value and adjustment coefficient, wherein the adjustment coefficient includes a scaling coefficient and an offset coefficient; The dynamic zinc layer strength characterization coefficient is calculated based on the verified peak strength value of the first path and the verified peak strength value of the second path, the strength benchmark value, and the adjustment coefficient of the two peak strength characteristic values.

[0010] In a preferred embodiment, the formula for calculating the dynamic zinc layer strength characterization coefficient is as follows: ; In the formula, The coefficient representing the dynamic zinc layer strength is... The first verified peak intensity value among the two peak intensity characteristic values. The second verified peak intensity value among the two peak intensity characteristic values. The preset strength benchmark value, The preset scaling factor. This is the preset offset coefficient.

[0011] In a preferred embodiment, the step of adaptively stripping the multidimensional dynamic fluctuation feature matrix based on the dynamic zinc layer strength characterization coefficient to obtain a substrate drift reference value, and removing the reference value from the multidimensional dynamic fluctuation feature matrix to obtain a transient zinc layer thickness change sequence, includes: The preset frequency band threshold is dynamically corrected based on the dynamic zinc layer strength characterization coefficient to generate the corrected frequency band threshold. Based on the modified frequency band threshold, the low-frequency components of the multidimensional dynamic fluctuation feature matrix are screened, and the fluctuation components with frequencies lower than the frequency band threshold are extracted to obtain the low-frequency fluctuation components of the galvanized sheet. The low-frequency fluctuation component is assigned as the substrate drift reference value of the plate, and the substrate drift reference value of the plate is removed from the multidimensional dynamic fluctuation feature matrix to obtain the high-frequency residual fluctuation sequence of the galvanized plate. The amplitude matching calibration of the high-frequency residual fluctuation sequence is performed to obtain the transient zinc layer thickness change sequence of the galvanized sheet.

[0012] In a preferred embodiment, the step of comparing each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point includes: Obtain the upper limit and lower limit of the allowed fluctuation range; For each instantaneous thickness value in the transient zinc layer thickness change sequence, a boundary comparison is performed to determine whether the instantaneous thickness value is higher than the upper limit of the allowable fluctuation range or lower than the lower limit of the allowable fluctuation range, and a point-by-point comparison status mark is generated. When the point-by-point comparison status flag indicates that the instantaneous thickness value exceeds the allowable fluctuation range, the corresponding instantaneous thickness value and its time location information are marked as abnormal, and an abnormal point record is generated. The abnormal point records are associated with the point-by-point comparison status markers, and the comparison result sequence with abnormal markers is output.

[0013] In a preferred embodiment, the step of generating a trigger command containing an identifier of the time of the anomaly when an instantaneous thickness value exceeding the range is output to the air knife controller includes: Analyze the anomaly records corresponding to the instantaneous thickness values ​​marked as exceeding the limit in the comparison result sequence, and extract the time location information from the anomaly record; Based on the time and location information, a trigger command containing an identifier of the time when the abnormality occurred is generated, and the trigger command is transmitted to the command receiving port of the air knife controller.

[0014] To address the aforementioned problems, the present invention also provides an online monitoring system for zinc coating on galvanized steel sheets based on laser thickness measurement, the system comprising: The time-series acquisition module is used to simultaneously acquire distance measurement data of each measuring point position changing over time at multiple measuring points arranged along the width direction of the galvanized sheet, and arrange them according to the sampling time to form a composite thickness time-series dataset. The feature modeling module is used to determine the thickness change trajectory and width direction difference value based on the composite thickness time series dataset, and to spatiotemporally correlate the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. The signal analysis module is used to acquire two thickness measurement signals with different response characteristics that are synchronously output at the same detection point in the galvanized sheet, and to synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. The substrate decoupling module is used to couple and correlate the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients. Based on the dynamic zinc layer strength characterization coefficients, the multidimensional dynamic fluctuation feature matrix is ​​subjected to adaptive low-frequency stripping to obtain the substrate drift reference value of the plate, and the reference value is removed from the multidimensional dynamic fluctuation feature matrix to obtain the transient zinc layer thickness change sequence. The anomaly driving module is used to compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, a trigger command containing an identifier of the time of anomaly occurrence is generated and output to the air knife controller.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention acquires two thickness measurement signals with different response characteristics at the same detection point and performs synchronous analysis. The peak intensity characteristic values ​​of the two signals are coupled and correlated to generate a dynamic zinc layer strength characterization coefficient. Based on this coefficient, an adaptive low-frequency stripping is applied to the multidimensional dynamic fluctuation characteristic matrix, allowing the substrate drift reference value to be dynamically separated and removed from the original measurement data, thus obtaining a transient zinc layer thickness change sequence. Therefore, the fundamental limitation of being unable to distinguish between substrate thickness drift and zinc layer thickness change due to frequency band aliasing during measurement, and the inability of a single signal to differentiate their respective contributions, is effectively overcome. Substrate fluctuations are no longer incorrectly attributed to zinc layer changes, and the accuracy and reliability of thickness calculation are improved.

[0016] 2. After low-frequency stripping, this invention further performs amplitude matching calibration on the high-frequency residual fluctuation sequence, aligning the amplitude of the transient zinc layer thickness change sequence with the physical dimensions of the actual zinc layer thickness. This calibration operation, based on decoupling, further enhances the physical authenticity and numerical reliability of the transient sequence, enabling the input data for point-by-point comparison and judgment to have higher fidelity, thereby providing a more reliable basis for anomaly identification. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating an online monitoring method for zinc layer of galvanized steel sheet based on laser thickness measurement, according to an embodiment of the present invention. Figure 2 This is a functional module diagram of an online zinc layer monitoring system for galvanized steel sheets based on laser thickness measurement, provided in an embodiment of the present invention. Figure 3 This is a schematic diagram of the two peak intensities and dynamic zinc layer strength characterization coefficients provided in an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0018] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0019] This application provides an online monitoring method for the zinc coating of galvanized steel sheets based on laser thickness measurement. The executing entity of this online monitoring method for the zinc coating of galvanized steel sheets based on laser thickness measurement includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application embodiment: a server, a terminal, etc. In other words, the online monitoring method for the zinc coating of galvanized steel sheets based on laser thickness measurement can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0020] Reference Figure 1 The diagram shown is a flowchart illustrating an online monitoring method for the zinc layer of galvanized steel sheets based on laser thickness measurement, according to an embodiment of the present invention. In this embodiment, the online monitoring method for the zinc layer of galvanized steel sheets based on laser thickness measurement includes: Pt.1. Along the direction of movement of the galvanized sheet, multiple measuring points are arranged in the width direction of the galvanized sheet to simultaneously acquire distance measurement data of each measuring point position changing over time, and arrange them according to the sampling time to form a composite thickness time series dataset. In this embodiment of the invention, in a continuous hot-dip galvanizing production line, the galvanized sheet travels along the direction of motion at a speed of several meters to tens of meters per second. Simultaneously, due to factors such as tension fluctuations within the annealing furnace, air knife impacts, and sheet shape oscillations, the sheet surface exhibits dynamic fluctuations in both the width and direction of motion. Existing single-point or sparse-point measurement methods struggle to capture the thickness distribution differences in the width direction. Furthermore, if there is a sampling delay between measurement points, the fluctuation information in the direction of motion and the width direction will overlap, failing to provide a synchronized and aligned original data foundation for subsequent spatiotemporal correlation analysis. Therefore, this step involves arranging multiple measurement points along the width direction of the galvanized sheet. Each measurement point uses a laser rangefinder sensor to simultaneously trigger distance sampling, and the sampling time is synchronously locked by the same clock source to ensure a consistent time reference for each measurement point.

[0021] In one embodiment, the number of measuring points in the width direction is 20 to 50. This number can be set by those skilled in the art based on the width of the galvanized sheet and the desired lateral resolution. The wider the width or the higher the requirement for lateral uniformity, the more measuring points should be added.

[0022] The distance measurement data obtained at each measuring point represents the optical path length between the laser sensor and the surface of the galvanized sheet at that measuring point.

[0023] After each measuring point completes a distance sampling, the sampled distance measurement data is bound to the timestamp of the sampling time. The timestamp is uniformly generated by the system clock with an accuracy of microseconds or milliseconds, thereby obtaining the distance measurement data of the galvanized plate. This data includes three elements: measuring point location identifier, distance value, and sampling time.

[0024] Next, the distance measurement data of each measuring point is sorted according to the order of the timestamps. That is, the data of each measuring point belonging to the same sampling time is sorted into the data frame corresponding to that time, thus obtaining the original distance sequence of the galvanized plate. This sequence is arranged according to the sampling time, and each frame contains the distance measurement data of all measuring points at that time.

[0025] The continuity of adjacent data frames in the original distance sequence is checked. During the check, the distance change between the current data frame and the previous data frame at each measurement point is calculated. If the distance change at a certain measurement point exceeds the preset continuity threshold, the data point is determined to be a data point that does not meet the continuity condition and is removed.

[0026] Preferably, the continuity threshold is a distance change rate between two adjacent frames of no more than 5% to 15%. This threshold can be obtained by the operator based on the maximum plate shape fluctuation amplitude under normal working conditions of the galvanized sheet, or it can be determined by statistical analysis based on the distance change distribution in historical production data.

[0027] After removing data points that do not meet the continuity condition, the distance sequence of the galvanized sheet is obtained. This sequence removes outlier data caused by splashes, instantaneous vibration and impact, or signal loss, thus ensuring data quality.

[0028] Finally, the distance measurement data in the distance sequence are arranged in the order of timestamps, and the data of each measuring point at the same time are arranged in the order of width direction to form a two-dimensional data array. The row direction of this array is the time dimension, and the column direction is the measuring point dimension, which is the composite thickness time series dataset of galvanized sheet.

[0029] In one feasible implementation, the distance measurement data can also be arranged in a transpose form with the measurement point as the primary key and the time series of each measurement point as the row vector. That is, each row corresponds to the distance measurement data sequence of a measurement point on a continuous time axis, and each column corresponds to the distance data of all measurement points at the same sampling time. The two arrangements are transposes of each other and can completely preserve the spatiotemporal structure of the data. Those skilled in the art can choose either arrangement according to their habits of subsequent matrix operations.

[0030] In this embodiment of the application, by arranging multiple measuring points along the width direction for simultaneous sampling and aligning them according to the timestamp, the distance data of each measuring point in the width direction at each sampling time is completely synchronized in time. This avoids the false lateral differences introduced by the time difference in traditional inspection-type or time-division scanning methods, and provides a data foundation with both temporal continuity and spatial integrity for the subsequent steps of constructing a multidimensional dynamic fluctuation feature matrix with spatiotemporal correlation.

[0031] Pt.2. Determine the thickness change trajectory and width direction difference value based on the composite thickness time series dataset, and perform spatiotemporal correlation between the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. In this embodiment of the invention, the composite thickness time series dataset records the distance measurement data of each measuring point on a continuous time axis. However, the original distance data itself only characterizes the optical path length from the sensor to the plate surface, and does not explicitly express the degree of thickness fluctuation of the plate surface in the direction of motion and the degree of thickness distribution difference in the width direction.

[0032] During the continuous production process of galvanized steel sheets, the thickness fluctuations manifest as dynamic fluctuations along the direction of motion over time, as well as uneven lateral distribution along the width direction at the same moment. These two fluctuation modes are physically coupled – the thickness change trend at a certain width position in the direction of motion may have a phase difference or amplitude difference with other width positions. If only analyzed along a single dimension, the spatiotemporal evolution law of the sheet thickness cannot be fully described.

[0033] Therefore, this step first performs differential analysis on the distance measurement data at consecutive time points along the movement direction of the galvanized sheet based on the composite thickness time series dataset to obtain the thickness gradient sequence of the galvanized sheet.

[0034] Specifically, for each measurement point, the distance measurement data of the measurement point at the current sampling time is calculated by subtracting the distance measurement data at the previous sampling time to obtain the gradient value of the motion direction of the measurement point at that time.

[0035] In one feasible implementation, the difference operation uses first-order forward difference, where the gradient value is equal to the current value minus the previous value.

[0036] In other implementations, the difference operation can also use central difference, where the gradient value is equal to the value at the next time step minus the value at the previous time step, divided by twice the sampling interval. Central difference has higher numerical accuracy and is suitable for scenarios with high sampling frequencies.

[0037] In some other implementations, the difference operation can also use second-order difference to reflect the acceleration information of thickness change, which is suitable for working conditions where the plate oscillation amplitude is large and the curvature change needs to be monitored.

[0038] Regardless of which differential method is used, the time gradient sequence of each measuring point is calculated one by one on the continuous time axis. The time gradient sequences of all measuring points together constitute the thickness gradient sequence of the galvanized sheet, which expresses the rate of thickness change of each position on the sheet in the direction of movement.

[0039] Subsequently, based on the composite thickness time series dataset, the distance measurement data at different measuring points at the same time point were mapped laterally to obtain the lateral difference distribution sequence of the galvanized sheet.

[0040] Specifically, for the same sampling time, the distance measurement data of adjacent measuring points in the width direction are selected and the difference is calculated to obtain the lateral difference value between each pair of adjacent measuring points.

[0041] In one implementation, the lateral difference mapping uses first-order difference, meaning the lateral difference value is equal to the difference in distance between adjacent measuring points.

[0042] In another implementation, the lateral difference mapping can also use the deviation relative to the center value in the width direction. That is, the distance value of the measuring point at the center position in the width direction is used as a reference to calculate the degree of deviation of other measuring points relative to the reference value. This method is suitable for working conditions where the lateral distribution of the plate surface has symmetrical characteristics.

[0043] In another implementation, the lateral difference mapping can also use the relative rate of change between adjacent measuring points, that is, the difference is divided by the distance between adjacent measuring points and then normalized. This method can eliminate the dimensional influence caused by the different distances between measuring points.

[0044] The lateral difference data at all sampling times are arranged in chronological order to form a lateral difference distribution sequence of the galvanized sheet. This sequence expresses the degree of unevenness in the thickness distribution in the width direction of the sheet at each time point.

[0045] Next, the gradient values ​​in the thickness gradient sequence are spatiotemporally coupled with the width direction difference values ​​in the lateral difference distribution sequence to obtain the spatiotemporal correlation feature pairs of galvanized sheets.

[0046] In one embodiment, the spatiotemporal coupling adopts a feature splicing method, which combines the gradient value and the difference value in the width direction at the same measurement point at the same time into a two-dimensional feature vector. This feature vector contains both the change information in the motion direction and the difference information in the width direction at that point.

[0047] Optionally, the spatiotemporal coupling can also adopt a feature weighted summation method, that is, after assigning a preset first weight coefficient and a preset second weight coefficient to the gradient value and the difference value in the width direction, they are weighted and accumulated to obtain a one-dimensional scalar value of the spatiotemporal associated feature pair.

[0048] Optionally, the spatiotemporal coupling can also adopt a cross-product method, that is, multiplying the gradient value with the difference value in the width direction to form an interactive feature value. This method can characterize the cooperative change relationship between the gradient in the motion direction and the difference in the width direction.

[0049] In another feasible implementation, the spatiotemporal coupling can also adopt a nonlinear mapping method, that is, inputting the gradient value and the difference value in the width direction into a preset kernel function for high-dimensional mapping to obtain nonlinear spatiotemporal correlation feature pairs. This method is suitable for working conditions where the plate surface fluctuation pattern is complex and linear combination is insufficient to characterize its change law.

[0050] Finally, the spatiotemporal correlation feature pairs are arranged sequentially according to the time dimension and matrixed according to the measurement point dimension to obtain a multidimensional dynamic fluctuation feature matrix.

[0051] In one embodiment, the matrix arrangement uses the sampling time as the row index and the measurement point location as the column index, and fills the spatiotemporal correlation feature pairs of each measurement point at each sampling time into the corresponding column positions of the row, forming a two-dimensional matrix with the number of rows equal to the total number of sampling frames and the number of columns equal to the total number of measurement points.

[0052] In another embodiment, the matrix arrangement can also be formed as a transposed two-dimensional matrix with the measurement point location as the row index and the sampling time as the column index.

[0053] In another feasible implementation, the multidimensional dynamic fluctuation feature matrix can be extended into a three-dimensional tensor. That is, with the thickness gradient value and the width direction difference value as two independent channels, a three-dimensional data volume is formed by multiplying the sampling time by the measurement point position by the feature channel, so as to implement differentiated filtering strategies for the two-dimensional fluctuation features in subsequent processing.

[0054] In this embodiment, regardless of which matrix arrangement method is used, the motion direction fluctuation information in the thickness gradient sequence and the width direction difference information in the lateral difference distribution sequence are uniformly represented under the same mathematical framework. This allows the variation law of plate thickness in both time and space to be expressed in a coordinated manner, providing a more structured data carrier for subsequent identification and removal of the substrate drift component from the matrix.

[0055] Pt.3. Obtain two thickness measurement signals with different response characteristics synchronously output at the same detection point in the galvanized sheet, and synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. In this embodiment of the invention, during laser thickness measurement, the information carried by a single thickness measurement signal is modulated by multiple factors, including power fluctuations of the laser source, attenuation of reflectivity due to the oxide layer on the plate surface, differences in absorption of different wavelengths of laser light by the substrate material, and the nonlinear effect of zinc layer thickness variation on signal intensity. When these multiple factors act simultaneously on the same signal, the change in signal amplitude cannot be solely attributed to the zinc layer thickness variation. Fluctuations in substrate reflectivity may be misinterpreted as zinc layer thickening or thinning, leading to measurement deviation. Therefore, this step simultaneously acquires two thickness measurement signals with different response characteristics at the same detection point on the galvanized plate.

[0056] In one embodiment, the first of the two thickness measurement signals with different response characteristics uses a confocal thickness measurement signal, which is sensitive to changes in the surface position of the plate, and its peak intensity reflects the concentration of reflected light energy. The second uses a triangulation thickness measurement signal, which is sensitive to changes in the distribution of scattered light, and its peak intensity reflects surface roughness and coating structure information. Both signals are triggered synchronously by the same clock source at the same detection point to ensure that the starting positions of the two signals are perfectly aligned on the time axis.

[0057] In another feasible implementation, the first signal is a vertically incident reflected signal and the second signal is an obliquely incident scattered signal, with the two having different response functions to the substrate reflectivity and the surface structure, respectively.

[0058] In another feasible implementation, the first signal is a short-wavelength laser thickness measurement signal, and the second signal is a long-wavelength laser thickness measurement signal. The different wavelengths of lasers have different penetration depths and reflection characteristics to the zinc layer and the substrate material, thus providing complementary thickness information.

[0059] Those skilled in the art can choose any of the above combinations to acquire two thickness measurement signals with different response characteristics, based on actual detection accuracy requirements and cost considerations.

[0060] After acquiring the two thickness measurement signals, the two thickness measurement signals with different response characteristics are first separated to obtain the first thickness measurement signal and the second thickness measurement signal.

[0061] The signal separation operation separates the mixed components in the original acquired signal caused by circuit coupling or optical crosstalk, so that the two signals can independently represent the response information of their corresponding detection channels.

[0062] In one feasible implementation, the signal separation operation employs a time-domain filtering method, which involves designing a digital filter based on the time-domain characteristics differences between the two signals for separation.

[0063] In another feasible implementation, the signal separation operation adopts the frequency domain separation method, that is, after performing Fourier transform on the original signal, the energy distribution regions of the two signals are cut in the frequency domain and then inversely transformed to restore them.

[0064] In another feasible implementation, the signal separation operation employs a blind source separation algorithm, which separates the signal sources one by one through independent component analysis when the prior mixing model of the two signals is unknown.

[0065] Regardless of which signal separation method is used, the first and second thickness measurement signals obtained after separation are independent time-series signals, each retaining the complete waveform information of the corresponding detection channel.

[0066] Next, peak values ​​of the first and second thickness measurement signals are located, and the peak intensity characteristic values ​​of the first and second channels of the galvanized sheet are read at the highest point of the signal amplitude on the time axis.

[0067] In one embodiment, peak positioning adopts a fixed window search method, that is, a time window length is set to a preset value, and the maximum value of the signal amplitude is searched within the window as the peak point. The window length can be two to five times the laser pulse width.

[0068] In another feasible implementation, peak positioning adopts a sliding window adaptive search method, that is, the window length is dynamically adjusted according to the full width and height of the signal to adapt to the fluctuations in the signal pulse width caused by board vibration or speed changes during the production process.

[0069] In another feasible implementation, peak positioning uses interpolation fitting. After locating the interval where the highest signal amplitude is located, polynomial or Gaussian fitting is performed on the discrete data points within the interval. The extreme points of the fitted curve are then used as the precise peak positions. This method can obtain peak intensity values ​​with sub-sampling accuracy even with a limited sampling rate.

[0070] After peak location is determined, the signal amplitude corresponding to the peak location is read, which are the first peak intensity value and the second peak intensity value, respectively.

[0071] Subsequently, the first and second peak strength values ​​were simultaneously verified to obtain the first and second verified peak strength values ​​of the galvanized sheet.

[0072] The purpose of synchronization verification is to confirm that the peak intensity values ​​of the two signals at the same detection time are within the effective range, and that the peak position deviation of the two signals on the time axis does not exceed the preset synchronization tolerance threshold, thereby ensuring that the peak intensity values ​​of the two signals correspond to the same state at the same detection point.

[0073] In one feasible implementation, the synchronization verification includes time deviation verification, which involves calculating the time deviation between the peak position of the first signal and the peak position of the second signal. If the absolute value of the deviation is less than a preset time synchronization threshold, the synchronization verification is deemed to have passed. The time synchronization threshold can be set as the duration of a single sampling period or determined according to the laser pulse repetition frequency, which is generally tens to hundreds of microseconds.

[0074] In another feasible implementation, the synchronization verification also includes amplitude validity verification, that is, determining whether the peak intensity values ​​of the first channel and the second channel are both within the preset amplitude validity range. The lower limit of the amplitude validity range is greater than the system noise floor, and the upper limit is less than the signal saturation threshold. If any peak intensity value exceeds the range, the frame data is determined to be invalid and marked as an abnormal frame. Subsequent steps exclude the abnormal frame from the feature value calculation.

[0075] In another feasible implementation, the synchronous verification also includes signal quality verification, that is, calculating the signal-to-noise ratio of the first thickness measurement signal and the second thickness measurement signal near the peak position respectively. If the signal-to-noise ratio of either signal is lower than the preset signal-to-noise ratio threshold, the credibility of the peak intensity value at that moment is determined to be insufficient, and an indicator indicating that the verification failed is output.

[0076] After the synchronization verification is passed, the first verified peak intensity value and the second verified peak intensity value are obtained. Both of these values ​​have been confirmed to meet the input conditions for subsequent coupling operations in terms of time synchronization, amplitude validity and signal quality.

[0077] Finally, the first and second verified peak strength values ​​are associated and packaged to obtain the two peak strength characteristic values ​​of the galvanized sheet.

[0078] In one embodiment, the associated packaging is implemented in the form of a feature vector, that is, the first verified peak intensity value and the second verified peak intensity value are used as two components of the two-dimensional feature vector, which is output as a whole at the same time point.

[0079] In another embodiment, the associated packaging is implemented in the form of structured data, that is, a structured data unit containing a timestamp identifier and two strength values.

[0080] In this embodiment, by acquiring two thickness measurement signals with different response characteristics at the same detection point and performing a systematic analysis operation including signal separation, peak location, and synchronous verification, the two signals, each carrying different physical information, are calibrated into comparable and correlated peak intensity feature values ​​under the same time reference. This provides temporally aligned, amplitude-effective, and quality-reliable input data for subsequently coupling and correlating the two feature values ​​to extract zinc layer feature information.

[0081] Pt.4. Couple and correlate the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients. Based on the dynamic zinc layer strength characterization coefficients, perform adaptive low-frequency stripping on the multidimensional dynamic fluctuation characteristic matrix to obtain the substrate drift reference value. Remove the reference value from the multidimensional dynamic fluctuation characteristic matrix to obtain the transient zinc layer thickness change sequence. In this embodiment of the invention, after the synchronous analysis of the two peak intensity feature values ​​is completed, the first verified peak intensity value and the second verified peak intensity value are obtained. These two values ​​are extracted from thickness measurement signals with different response characteristics, and each carries differentiated information about the surface condition of the galvanized sheet.

[0082] However, the individual peak intensity values ​​of the two signals only reflect the signal strength of their respective detection channels and have not yet formed a unified quantitative characterization of the zinc layer thickness.

[0083] To integrate the differentiated information of the two signals into a unified zinc layer state metric, this step first couples and correlates the peak intensity characteristic values ​​of the two signals to generate dynamic zinc layer intensity characterization coefficients.

[0084] Specifically, a preset strength benchmark value and adjustment coefficient are obtained, wherein the adjustment coefficient includes a scaling coefficient and an offset coefficient.

[0085] The preset strength benchmark value characterizes the expected response level of the two signals under the benchmark zinc layer thickness. Its value is obtained by a person skilled in the art through calibration experiments using standard samples with known zinc layer thickness under the same testing conditions.

[0086] In one embodiment, the preset strength benchmark value is the arithmetic mean of the two peak strength measurements of the standard sample.

[0087] In another embodiment, the preset strength benchmark value is the weighted average of the two peak strength measurements of the standard sample. The weighting coefficient is determined according to the signal-to-noise ratio of the two signals, and the one with the higher signal-to-noise ratio is given a larger weight.

[0088] In another embodiment, the preset intensity benchmark value can also be the theoretical calculation value of the standard sample, that is, calculated by an optical simulation model based on the laser power, surface reflectivity and geometric optical path parameters.

[0089] The scaling factor and offset factor are used to perform a linear transformation on the coupling result, so that the numerical range of the dynamic zinc layer strength characterization factor falls within the preset range, which facilitates subsequent numerical calculations with the frequency band threshold.

[0090] In one embodiment, the scaling factor is between 0.8 and 1.2.

[0091] In another embodiment, the scaling factor is between 1.5 and 2.5.

[0092] In yet another embodiment, the scaling factor is between 0.3 and 0.7.

[0093] The specific value of the scaling factor is determined comprehensively based on the dynamic range of the amplitude difference between the two signals, the sensitivity requirements for subsequent frequency band threshold correction, and the system noise level.

[0094] The offset coefficient ranges from -0.5 to 0.5, and is used to adjust the dynamic zinc layer strength characterization coefficient to a range symmetrical about 0, so as to simplify the logical complexity of subsequent correction calculations.

[0095] Those skilled in the art can adaptively adjust the scaling factor and offset factor according to the actual detection accuracy requirements, signal sampling frequency, and the allowable range of zinc layer thickness deviation.

[0096] The dynamic zinc layer strength characterization coefficient is calculated based on the first and second verified peak strength values, the preset strength benchmark value, and the adjustment coefficient.

[0097] In one embodiment, the formula for calculating the dynamic zinc layer strength characterization coefficient is as follows: ; In the formula, This is a coefficient characterizing the dynamic zinc layer strength. This refers to the first verified peak intensity value among the two peak intensity characteristic values. This refers to the verified peak intensity value of the second of the two peak intensity characteristic values. The preset strength benchmark value, The preset scaling factor. This is the preset offset coefficient.

[0098] In this formula, and The difference is in After normalization, an exponential function mapping is used to nonlinearly amplify the difference between the two signals. When the difference between the two signals increases... The value increases exponentially when the difference between the two signals decreases. The value approaches zero, and the scaling factor... and offset coefficient Then the index mapping result is further adjusted to the target value range.

[0099] In another feasible implementation, the formula for calculating the dynamic zinc layer strength characterization coefficient can also be replaced by the following alternative form: ; This ratio form has lower computational complexity and is suitable for scenarios where the two signals differ significantly and there is no need to exponentially amplify the minute differences.

[0100] In another feasible implementation, the formula for calculating the dynamic zinc layer strength characterization coefficient can also be expressed as a logarithmic ratio: ; This logarithmic form is in and When the ratio spans multiple orders of magnitude, it can compress the variation range of the dynamic zinc layer strength characterization coefficient to a linear range, making it suitable for detection systems with significant differences in the response sensitivity of the two signals.

[0101] In another feasible implementation, the formula for calculating the dynamic zinc layer strength characterization coefficient can also be in the form of the following S-shaped function: ; This S-shaped function maps the dynamic zinc layer strength characterization coefficient to a monotonically increasing bounded interval, making it suitable for scenarios where the difference between two signals may produce extreme values ​​due to sensor saturation or cutoff, and it is necessary to control them within a bounded range to avoid overflow in subsequent numerical calculations.

[0102] After generating the dynamic zinc layer strength characterization coefficient, this coefficient is used to perform adaptive low-frequency stripping on the multidimensional dynamic fluctuation feature matrix.

[0103] The multidimensional dynamic fluctuation feature matrix contains low-frequency fluctuation components caused by the thickness drift of the substrate and high-frequency transient fluctuation components corresponding to the change in zinc layer thickness, but the frequency band boundaries of the two are not fixed.

[0104] When the operating speed of the galvanized sheet changes, the frequency of the same spatial periodic fluctuation in the time domain changes accordingly. The slow change of the substrate thickness along the direction of movement may exhibit a higher time frequency at certain speeds, while the transient change of the zinc layer caused by the air knife scraping may exhibit a lower time frequency at certain speeds. Therefore, the degree of frequency band overlap between substrate drift and zinc layer change dynamically with the change of operating conditions.

[0105] Therefore, this step first dynamically corrects the preset frequency band threshold based on the dynamic zinc layer strength characterization coefficient to generate the corrected frequency band threshold.

[0106] In one embodiment, the dynamic correction adopts a linear correction method, and the correction band threshold is equal to the product of the preset band threshold and the dynamic zinc layer strength characterization coefficient.

[0107] When the dynamic zinc layer strength characterization coefficient increases, it indicates that the difference between the two signals is increasing and the weight of the zinc layer signal in the total signal is enhanced. The frequency band threshold should be appropriately reduced to retain more high-frequency components and avoid misjudging the zinc layer change as substrate drift.

[0108] When the dynamic zinc layer strength characterization coefficient decreases, it indicates that the difference between the two signals is decreasing, and the weight of substrate reflectivity fluctuations or optical path interference in the total signal is increasing. The frequency band threshold should be appropriately increased to filter out more low-frequency substrate components and avoid misjudging substrate drift as zinc layer change.

[0109] In another feasible implementation, the dynamic correction adopts a non-linear correction method. The correction band threshold is equal to the preset band threshold multiplied by the positive power of the dynamic zinc layer strength characterization coefficient. The power value is greater than 0 and less than 1. The specific value of the power determines the correction sensitivity of the dynamic zinc layer strength characterization coefficient to the band threshold.

[0110] In another feasible implementation, the dynamic correction adopts a piecewise linear correction method. When the dynamic zinc layer strength characterization coefficient is less than the first preset value, the first correction slope is used; when the dynamic zinc layer strength characterization coefficient is greater than the second preset value, the second correction slope is used; and when it is in between, the third correction slope is used. This allows the correction frequency band threshold to have a differentiated response rate in different K value ranges.

[0111] The specific calculation method for the corrected frequency band threshold is set by those skilled in the art based on the relative relationship between the production line speed variation range, the substrate thickness fluctuation frequency distribution, and the zinc layer variation frequency distribution.

[0112] The initial value of the preset frequency band threshold is determined by a person skilled in the art after performing frequency domain analysis based on the maximum rate of change of substrate thickness drift and the minimum rate of change of zinc layer thickness under normal production conditions of galvanized sheet. Alternatively, it can be determined by performing power spectral density analysis on the thickness sequence in historical data and setting the frequency corresponding to the valley point on the power spectral density curve as the initial frequency band threshold.

[0113] Based on the corrected frequency band threshold, the low-frequency components of the multidimensional dynamic fluctuation feature matrix are screened, and the fluctuation components with frequencies lower than the corrected frequency band threshold are extracted to obtain the low-frequency fluctuation components of the galvanized sheet.

[0114] In one embodiment, the low-frequency component screening adopts the Fourier transform method, that is, the multidimensional dynamic fluctuation feature matrix is ​​subjected to discrete Fourier transform along the time dimension to obtain frequency domain data, and frequency domain components with frequency values ​​lower than the corrected frequency band threshold are selected as low-frequency fluctuation components, and then restored to time domain signal by inverse Fourier transform.

[0115] In another embodiment, the low-frequency component screening adopts a digital low-pass filtering method, that is, a low-pass filter with a cutoff frequency equal to the modified frequency band threshold is designed, and the multidimensional dynamic fluctuation feature matrix is ​​passed through the low-pass filter along the time dimension, and the output signal is the low-frequency fluctuation component.

[0116] In another embodiment, the low-frequency component screening adopts wavelet decomposition, that is, multi-scale wavelet decomposition is performed on the multi-dimensional dynamic fluctuation feature matrix along the time dimension, and the approximate components with a scale larger than a preset scale after decomposition are taken as low-frequency fluctuation components.

[0117] The low-frequency fluctuation component reflects the slow drift of the substrate itself caused by factors such as changes in the thickness of the incoming plate and tension fluctuations in the annealing furnace. This drift is superimposed on the zinc layer thickness change in the thickness measurement signal and must be separated.

[0118] The low-frequency fluctuation component is assigned as the substrate drift reference value, which represents the degree of deviation of the substrate thickness from the reference value at the current time and location.

[0119] Next, the substrate drift reference value is removed from the multidimensional dynamic fluctuation feature matrix.

[0120] In one embodiment, the removal operation is a subtraction operation, which is to subtract the substrate drift reference value corresponding to the same measuring point position at the same time from each element in the multidimensional dynamic fluctuation feature matrix.

[0121] In another embodiment, the removal operation is a division operation, which is to divide each element in the multidimensional dynamic fluctuation feature matrix by the substrate drift reference value of the plate at the same measurement point at the same time and then normalize it.

[0122] In another embodiment, the removal operation is a logarithmic difference operation, which involves taking the logarithm of each element in the multidimensional dynamic fluctuation feature matrix and the corresponding substrate drift benchmark value and then performing a difference operation. This method is suitable for physical scenarios where substrate drift and zinc layer change are multiplicatively coupled rather than additively coupled.

[0123] After removing the substrate drift reference value, the residual information in the multidimensional dynamic fluctuation feature matrix is ​​the fluctuation component after removing the influence of the low-frequency drift of the substrate, thus obtaining the high-frequency residual fluctuation sequence of the galvanized sheet.

[0124] Ideally, the high-frequency residual wave sequence should only contain transient fluctuation information caused by changes in zinc layer thickness. However, the amplitude of the high-frequency residual wave sequence may be subject to dimensional shift or amplitude scaling due to mathematical operations after removing the reference value.

[0125] Therefore, amplitude matching calibration was performed on the high-frequency residual fluctuation sequence to restore its correspondence with the actual zinc layer thickness variation.

[0126] In one embodiment, amplitude matching calibration adopts a calibration curve mapping method, that is, using the amplitude of the high-frequency residual wave sequence obtained by a standard sample with known zinc layer thickness increment under the same detection conditions, a calibration curve between amplitude and thickness increment is established, and then the amplitude of the actual high-frequency residual wave sequence is mapped to the corresponding thickness change through the calibration curve.

[0127] In another embodiment, amplitude matching calibration adopts a scaling factor correction method, that is, calculating the ratio between the root mean square value of the high-frequency residual wave sequence and the root mean square value of the high-frequency residual wave sequence of the standard sample, and then multiplying the amplitude of the actual sequence by the reciprocal of the scaling factor to complete the amplitude calibration.

[0128] In another embodiment, the amplitude matching calibration adopts a linear regression method, that is, the amplitude of the high-frequency residual wave sequence is linearly fitted with the known thickness increment of the standard sample, and the fitting coefficient is then applied to the actual sequence.

[0129] After amplitude matching calibration is completed, the transient zinc layer thickness change sequence of the galvanized sheet is obtained. Each value in the sequence corresponds to the transient deviation of the zinc layer thickness at a certain sampling point relative to the reference value.

[0130] In this embodiment, the peak intensity characteristic values ​​of two different response characteristics are coupled and correlated into a dynamic zinc layer intensity characterization coefficient. This coefficient is updated in real time as the difference between the two signals changes. The frequency band threshold is then dynamically corrected using this coefficient, so that the frequency band threshold is adaptively adjusted as the zinc layer state changes. This achieves the adaptive separation of the boundary point between the substrate drift reference value and the zinc layer thickness change according to the current operating conditions. The substrate component that drifts with the operating conditions is removed from the multidimensional dynamic fluctuation characteristic matrix, and a transient sequence that is aligned with the actual zinc layer thickness change in terms of physical dimensions and amplitude is obtained.

[0131] Pt.5. Compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, generate a trigger command containing an identifier of the time of the abnormality and output it to the air knife controller.

[0132] In this embodiment of the invention, after obtaining the transient zinc layer thickness change sequence, each instantaneous thickness value in the sequence represents the transient deviation of the zinc layer thickness at the current measuring point position relative to the reference value. However, whether this deviation is within the acceptable production tolerance range has not yet been determined. Quality control of the zinc layer thickness of galvanized sheets requires controlling the actual thickness deviation within a preset tolerance zone. Once this tolerance zone is exceeded, it is considered a thickness anomaly, requiring the air knife actuator to adjust the air knife pressure or distance to achieve thickness correction. However, there is an inherent transmission delay and mechanical action delay between the time of the anomaly occurrence and the completion of the physical response by the air knife actuator. If the trigger command does not contain the position information at the time of the anomaly occurrence, the air knife will not be able to accurately align the compensation action with the actual location of the defect, resulting in compensation offset or compensation lag.

[0133] Therefore, this step first obtains the upper limit and lower limit of the allowed fluctuation range.

[0134] The permissible fluctuation range characterizes the maximum permissible positive deviation and the maximum permissible negative deviation of the zinc coating thickness relative to the target thickness. Its values ​​are determined by the quality grade standards of the galvanized products and customer contract requirements.

[0135] In one embodiment, the upper limit of the allowable fluctuation range is positive five percent of the target thickness value, and the lower limit of the allowable fluctuation range is negative five percent of the target thickness value. This symmetrical range is suitable for conventional galvanized products that have high requirements for zinc layer uniformity but moderate tolerance for absolute thickness deviation.

[0136] In another embodiment, the upper limit of the allowable fluctuation range is positive 10% of the target thickness value, and the lower limit of the allowable fluctuation range is negative 3% of the target thickness value. This asymmetric range is suitable for application scenarios that are sensitive to excessively thin zinc layers but have a greater tolerance for excessively thick layers, such as galvanized sheets that require subsequent deep drawing processing.

[0137] In another embodiment, the upper limit and lower limit of the allowable fluctuation range are set using fixed absolute thickness values. The upper limit is the target thickness value plus a first thickness threshold, and the lower limit is the target thickness value minus a second thickness threshold. The first thickness threshold and the second thickness threshold are set independently, and their specific values ​​are specified by the quality control specifications of the galvanizing production line. Generally, they are the equivalent thickness values ​​corresponding to the weight of zinc layer of tens to hundreds of grams per square meter.

[0138] The specific values ​​of the allowable fluctuation range are determined by those skilled in the art based on the product grade, customer agreement and production line capacity. Different allowable fluctuation range parameters can be configured for different specifications of galvanized sheets.

[0139] After obtaining the upper and lower limits of the allowable fluctuation range, each instantaneous thickness value in the transient zinc layer thickness change sequence is compared against the boundary. Each instantaneous thickness value is determined to be higher than the upper limit of the allowable fluctuation range or lower than the lower limit of the allowable fluctuation range, and a point-by-point comparison status mark is generated.

[0140] In one embodiment, the boundary comparison adopts a direct comparison operation. For the instantaneous thickness value at each measurement point position at each sampling time, it is compared with the upper limit and lower limit of the allowable fluctuation range. When the instantaneous thickness value is greater than the upper limit, a first state flag is output. When the instantaneous thickness value is less than the lower limit, a second state flag is output. When the instantaneous thickness value is between the upper limit and the lower limit, a third state flag is output.

[0141] In another embodiment, the boundary comparison adopts the difference operation method to calculate the upper deviation between the instantaneous thickness value and the upper limit value and the lower deviation between the lower limit value and the instantaneous thickness value respectively. The state interval of the point is determined according to the positive and negative combination of the upper and lower deviations and the corresponding state mark is generated.

[0142] In another embodiment, the boundary comparison adopts a normalized deviation calculation method, which subtracts the target thickness value from the instantaneous thickness value and then divides it by half the width of the allowable fluctuation range to obtain the normalized deviation value. If the absolute value of the normalized deviation value is greater than 1, it is determined to be out of range, and the direction of exceeding the range is distinguished according to its positive or negative sign.

[0143] When the point-by-point comparison status flag indicates that the instantaneous thickness value exceeds the allowable fluctuation range, the corresponding instantaneous thickness value and its time location information are marked as abnormal, and an abnormal point record is generated.

[0144] In one embodiment, the abnormal point record includes an abnormal timestamp, an abnormal measurement point location identifier, an instantaneous thickness value, and an out-of-direction identifier. The out-of-direction identifier is used to distinguish whether the abnormality is due to excessive thickness or insufficient thickness.

[0145] In another embodiment, the anomaly record also includes the instantaneous thickness values ​​of several frames before and after the anomaly, so as to provide information on the thickness change trend before and after exceeding the limit for subsequent air knife control.

[0146] The abnormal point records are associated with the point-by-point comparison status markers, and the comparison result sequence with abnormal markers is output. The comparison result sequence is arranged in chronological order. Each frame of data contains the comparison status markers of all measurement points at that time and the abnormal point records when an anomaly exists.

[0147] Next, the abnormal point records corresponding to the instantaneous thickness values ​​marked as exceeding the limit are analyzed in the comparison result sequence, and the time location information in the abnormal point records is extracted.

[0148] The time and location information includes the system timestamp of the time when the anomaly occurred, and the position coordinates of the galvanized plate along the direction of movement corresponding to the time when the anomaly occurred.

[0149] In one embodiment, the location coordinates are obtained by multiplying the system timestamp by the current running speed of the galvanized sheet.

[0150] In another embodiment, the position coordinates are obtained by accumulating the pulse count values ​​of an encoder installed upstream of the production line. The encoder is installed on the shaft of the tension roller or pinch roller, and its pulse frequency is proportional to the production line running speed. The position coordinates are equal to the pulse count value divided by the number of pulses corresponding to each unit length of the encoder.

[0151] Generate a trigger command that includes an identifier indicating when the anomaly occurred, based on the time and location information.

[0152] The time of anomaly is marked so that the air knife controller can identify the exact position of the galvanized sheet segment that needs thickness correction on the production line after receiving the trigger command, so that the defective segment can still be compensated even after the air knife action is delayed.

[0153] In one embodiment, the time of the exception is identified in the form of a timestamp, that is, the trigger command includes the system clock reading at the time of the exception.

[0154] In another embodiment, the time of an anomaly is identified by a plate segment number, that is, the galvanized sheet is divided into continuous plate segments along the direction of movement at a fixed length and each plate segment is assigned a unique number, and the time of an anomaly is identified by the number of the plate segment where the anomaly occurs.

[0155] In another embodiment, the time of the anomaly is identified in the form of location coordinates, that is, the trigger command includes the distance coordinate value of the anomaly location along the direction of movement.

[0156] In another embodiment, the time of an anomaly is identified by an encoder count value, that is, the trigger command includes the cumulative encoder pulse count value corresponding to the abnormal position. This method is suitable for scenarios where the air knife controller uses the same encoder signal for position synchronization, without the need for clock synchronization or coordinate conversion between the trigger command and the actuator.

[0157] After generating the trigger command, the trigger command is transmitted to the command receiving port of the air knife controller.

[0158] In one embodiment, the trigger command is transmitted to the air knife controller via an industrial fieldbus, which is a Profinet bus, an EtherNet / IP bus, or a Modbus TCP bus.

[0159] In another embodiment, the trigger command is transmitted to the air knife controller via a hardwired digital output port, carrying the trigger signal in the form of a digital pulse signal.

[0160] In another embodiment, the trigger command is transmitted to the air knife controller via an industrial Ethernet in the form of a data packet, which includes an identifier indicating when the anomaly occurred and an identifier indicating the out-of-direction.

[0161] In this embodiment, by comparing each instantaneous thickness value in the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point, and generating a trigger command carrying an anomaly occurrence time identifier when the limit is exceeded, the trigger command is transmitted to the air knife controller. This enables the thickness correction action of the air knife controller to determine the accurate position of the defect segment on the production line based on the anomaly occurrence time identifier, eliminating the compensation position offset caused by transmission delay and mechanical action delay, realizing the precise spatiotemporal correspondence between the over-limit anomaly and the actuator, and providing a position benchmark for the fixed-point and quantitative compensation of the air knife.

[0162] like Figure 2 The diagram shown is a functional block diagram of an online zinc layer monitoring system for galvanized steel sheets based on laser thickness measurement, provided in an embodiment of the present invention.

[0163] The online zinc coating monitoring system for galvanized steel sheets based on laser thickness measurement described in this invention can be installed in electronic devices. Depending on the functions implemented, the online zinc coating monitoring system for galvanized steel sheets based on laser thickness measurement may include a timing acquisition module, a feature construction module, a signal analysis module, a substrate decoupling module, and an anomaly driving module. The modules described in this invention can also be referred to as units, which are a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, stored in the memory of the electronic device.

[0164] In this embodiment, the functions of each module / unit are as follows: The time-series acquisition module is used to simultaneously acquire distance measurement data of each measuring point position changing over time at multiple measuring points arranged along the width direction of the galvanized sheet, along the direction of movement of the galvanized sheet, and arrange them according to the sampling time to form a composite thickness time-series dataset. The feature modeling module is used to determine the thickness change trajectory and the width direction difference value based on the composite thickness time series dataset, and to spatiotemporally correlate the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. The signal analysis module is used to acquire two thickness measurement signals with different response characteristics that are synchronously output at the same detection point in the galvanized sheet, and to synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. The substrate decoupling module is used to couple and correlate the two peak intensity characteristic values ​​to generate a dynamic zinc layer strength characterization coefficient. Based on the dynamic zinc layer strength characterization coefficient, the module performs adaptive low-frequency stripping on the multidimensional dynamic fluctuation characteristic matrix to obtain the substrate drift reference value of the plate, and removes the reference value from the multidimensional dynamic fluctuation characteristic matrix to obtain a transient zinc layer thickness change sequence. The anomaly driving module is used to compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, a trigger command containing an identifier of the time of anomaly occurrence is generated and output to the air knife controller.

[0165] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0166] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0167] like Figure 3 The figure shown is a schematic diagram of the two peak intensities and dynamic zinc layer strength characterization coefficients provided in an embodiment of the present invention.

[0168] Figure 3This diagram illustrates the peak intensity characteristics of two thickness measurement signals with different response characteristics obtained through synchronous analysis at the same detection point on a galvanized sheet, as well as the changing relationship of the dynamic zinc layer strength characterization coefficient generated through coupling correlation calculation. The horizontal axis typically represents time or detection location, while the vertical axis represents signal intensity or characterization coefficient values, intuitively reflecting the differences between different response signals and their impact on the dynamic zinc layer strength characterization coefficient.

[0169] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0170] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0171] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for online monitoring of zinc layer on galvanized steel sheets based on laser thickness measurement, characterized in that, The method includes: Pt.

1. Along the direction of movement of the galvanized sheet, multiple measuring points are arranged in the width direction of the galvanized sheet to simultaneously acquire distance measurement data of each measuring point position changing over time, and arrange them according to the sampling time to form a composite thickness time series dataset. Pt.

2. Determine the thickness change trajectory and width direction difference value based on the composite thickness time series dataset, and perform spatiotemporal correlation between the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. Pt.

3. Obtain two thickness measurement signals with different response characteristics synchronously output at the same detection point in the galvanized sheet, and synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. Pt.

4. Couple and correlate the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients. Based on the dynamic zinc layer strength characterization coefficients, perform adaptive low-frequency stripping on the multidimensional dynamic fluctuation characteristic matrix to obtain the substrate drift reference value. Remove the reference value from the multidimensional dynamic fluctuation characteristic matrix to obtain the transient zinc layer thickness change sequence. Pt.

5. Compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, generate a trigger command containing an identifier of the time of the abnormality and output it to the air knife controller.

2. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 1, characterized in that, The method involves simultaneously acquiring distance measurement data of each measuring point location over time at multiple measuring points arranged along the width direction of the galvanized sheet, and arranging the data according to the sampling time to form a composite thickness time-series dataset, including: Distance sampling is triggered at each measuring point, and the sampled distance measurement data is bound to the timestamp of the sampling time to obtain the distance measurement data of the galvanized sheet; The distance measurement data of each measuring point are arranged in chronological order according to the timestamps to obtain the original distance sequence of the galvanized sheet; The continuity of adjacent data points in the original distance sequence is checked, and data points that do not meet the continuity condition are removed to obtain the distance sequence of the galvanized sheet. The distance measurement data in the distance sequence are arranged sequentially according to the timestamp to obtain the composite thickness time series dataset of the galvanized sheet.

3. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 1, characterized in that, The step of determining the thickness change trajectory and width direction difference value based on the composite thickness time-series dataset, and spatiotemporally correlating the thickness change trajectory and width direction difference value to construct a multidimensional dynamic fluctuation feature matrix includes: Based on the composite thickness time series dataset, differential analysis is performed on the distance measurement data at consecutive time points along the movement direction of the galvanized sheet to obtain the thickness gradient sequence of the galvanized sheet; Based on the composite thickness time series dataset, the distance measurement data at different measuring points at the same time point are subjected to lateral difference mapping to obtain the lateral difference distribution sequence of the galvanized sheet; The gradient values ​​in the thickness gradient sequence are spatiotemporally coupled with the width direction difference values ​​in the transverse difference distribution sequence to obtain the spatiotemporal correlation feature pairs of the galvanized sheet; The spatiotemporal correlation feature pairs are arranged sequentially according to the time dimension and then matrixed according to the measurement point dimension to obtain a multidimensional dynamic fluctuation feature matrix.

4. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 1, characterized in that, The synchronous analysis of the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet includes: The two thickness measurement signals with different response characteristics are separated to obtain the first thickness measurement signal and the second thickness measurement signal. Peak location is performed on the first and second thickness measurement signals respectively, and the peak intensity characteristic values ​​of the first and second channels of the galvanized sheet are read at the highest point of the signal amplitude on the time axis. The first peak strength value and the second peak strength value are synchronously verified to obtain the first verified peak strength value and the second verified peak strength value of the galvanized sheet. The first verified peak strength value and the second verified peak strength value are associated and packaged together to form the two peak strength characteristic values ​​of the galvanized sheet.

5. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 1, characterized in that, The step of coupling and correlating the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients includes: Obtain a preset strength reference value and adjustment coefficient, wherein the adjustment coefficient includes a scaling coefficient and an offset coefficient; The dynamic zinc layer strength characterization coefficient is calculated based on the verified peak strength value of the first path and the verified peak strength value of the second path, the strength benchmark value, and the adjustment coefficient of the two peak strength characteristic values.

6. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 5, characterized in that, The formula for calculating the dynamic zinc layer strength characterization coefficient is as follows: ; In the formula, This is the characteristic coefficient for the dynamic zinc layer strength. The first verified peak intensity value among the two peak intensity characteristic values. The second verified peak intensity value among the two peak intensity characteristic values. The preset strength benchmark value, The preset scaling factor. This is the preset offset coefficient.

7. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 6, characterized in that, The process involves adaptive low-frequency stripping of the multidimensional dynamic fluctuation feature matrix based on the dynamic zinc layer strength characterization coefficient to obtain a substrate drift reference value, and removing the reference value from the multidimensional dynamic fluctuation feature matrix to obtain a transient zinc layer thickness change sequence, including: The preset frequency band threshold is dynamically corrected based on the dynamic zinc layer strength characterization coefficient to generate the corrected frequency band threshold. Based on the modified frequency band threshold, the low-frequency components of the multidimensional dynamic fluctuation feature matrix are screened, and the fluctuation components with frequencies lower than the frequency band threshold are extracted to obtain the low-frequency fluctuation components of the galvanized sheet. The low-frequency fluctuation component is assigned as the substrate drift reference value of the plate, and the substrate drift reference value of the plate is removed from the multidimensional dynamic fluctuation feature matrix to obtain the high-frequency residual fluctuation sequence of the galvanized plate. The amplitude matching calibration of the high-frequency residual fluctuation sequence is performed to obtain the transient zinc layer thickness change sequence of the galvanized sheet.

8. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 1, characterized in that, The step of comparing each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point includes: Obtain the upper limit and lower limit of the allowed fluctuation range; For each instantaneous thickness value in the transient zinc layer thickness change sequence, a boundary comparison is performed to determine whether the instantaneous thickness value is higher than the upper limit of the allowable fluctuation range or lower than the lower limit of the allowable fluctuation range, and a point-by-point comparison status mark is generated. When the point-by-point comparison status flag indicates that the instantaneous thickness value exceeds the allowable fluctuation range, the corresponding instantaneous thickness value and its time location information are marked as abnormal, and an abnormal point record is generated. The abnormal point records are associated with the point-by-point comparison status markers, and the comparison result sequence with abnormal markers is output.

9. The method for online monitoring of zinc layer in galvanized steel sheets based on laser thickness measurement as described in claim 8, characterized in that, When an instantaneous thickness value exceeding this range exists, a trigger command containing an identifier of the time of the anomaly is generated and output to the air knife controller, including: Analyze the anomaly records corresponding to the instantaneous thickness values ​​marked as exceeding the limit in the comparison result sequence, and extract the time location information from the anomaly record; Based on the time and location information, a trigger command containing an identifier of the time when the abnormality occurred is generated, and the trigger command is transmitted to the command receiving port of the air knife controller.

10. An online monitoring system for zinc coating on galvanized steel sheets based on laser thickness measurement, characterized in that, The system for implementing the online monitoring method for zinc layer of galvanized steel sheet based on laser thickness measurement as described in claim 1 includes: The time-series acquisition module is used to simultaneously acquire distance measurement data of each measuring point position changing over time at multiple measuring points arranged along the width direction of the galvanized sheet, and arrange them according to the sampling time to form a composite thickness time-series dataset. The feature modeling module is used to determine the thickness change trajectory and width direction difference value based on the composite thickness time series dataset, and to spatiotemporally correlate the thickness change trajectory and the width direction difference value to construct a multidimensional dynamic fluctuation feature matrix. The signal analysis module is used to acquire two thickness measurement signals with different response characteristics that are synchronously output at the same detection point in the galvanized sheet, and to synchronously analyze the two thickness measurement signals with different response characteristics to obtain the two peak intensity characteristic values ​​of the galvanized sheet. The substrate decoupling module is used to couple and correlate the two peak intensity characteristic values ​​to generate dynamic zinc layer strength characterization coefficients. Based on the dynamic zinc layer strength characterization coefficients, the multidimensional dynamic fluctuation feature matrix is ​​subjected to adaptive low-frequency stripping to obtain the substrate drift reference value of the plate, and the reference value is removed from the multidimensional dynamic fluctuation feature matrix to obtain the transient zinc layer thickness change sequence. The anomaly driving module is used to compare each instantaneous thickness value of the transient zinc layer thickness change sequence with a preset allowable fluctuation range point by point. When there is an instantaneous thickness value that exceeds the range, a trigger command containing an identifier of the time of anomaly occurrence is generated and output to the air knife controller.

Images