Production parameter tracing and control system and method for gypsum board face defects

Abstract

The application relates to the technical field of production process control, in particular to a production parameter tracing and control system and method for board surface defects of a gypsum board. The method comprises the following steps: acquiring a surface texture reference point of a to-be-tested gypsum board in a forming stage, and acquiring texture features and corresponding displacement coordinates of a surface defect area of a finished gypsum board obtained by cutting the to-be-tested gypsum board; determining a target timestamp of the surface defect area by using the displacement coordinates, the texture features and the surface texture reference point; acquiring corresponding edge thickness gradient curves, surface temperature rise curves and internal glue curing degree values by using the target timestamp, so as to determine a working condition classification result of the to-be-tested gypsum board, and controlling a forming flow guide device or a drying hot air device. Through the technical scheme, the production parameters of the board surface of the gypsum board can be traced and controlled.

Description

Technical Field

[0001] This application relates to the field of production process control technology, and in particular to a production parameter traceability and control system and method for gypsum board surface defects. Background Technology

[0002] Gypsum board is a lightweight building material made from building gypsum as the main raw material. Water is added and stirred to form a gypsum slurry. Special protective paperboard is then applied to both sides of the slurry. The slurry is then formed, solidified, cut, dried, and trimmed through continuous industrial processes. In addition to the main hemihydrate gypsum, its internal components usually include functional additives such as retarders, foaming agents, starch, fibers, and water-reducing agents. These additives are used to regulate the slurry setting speed, reduce the density of the board, and improve the toughness of the blank and the adhesion performance of the board surface. Gypsum board is one of the most widely used wall and ceiling materials in the modern industrialized building system.

[0003] During the continuous production process, gypsum board is prone to various surface defects such as bubbles, dents, bulges, scratches, grout leakage, peeling, uneven thickness, paper wrinkles, and stains. Gypsum board with surface defects not only has an unsatisfactory appearance, but localized surface defects also weaken the bonding strength between the facing paper and the gypsum core, making it prone to paper warping, cracking, and detachment during use, thus reducing the overall bending and impact resistance of the board.

[0004] Abnormal process parameters on gypsum board surfaces often correspond to specific types of surface defects. The coupling effect of multiple factors further increases the difficulty of defect cause investigation. Traditional production models mostly involve post-defect rectification, relying on operators' experience to blindly adjust processes. This rectification is delayed and has high trial-and-error costs, failing to prevent the mass production of defects from the source. Therefore, it is necessary to conduct source analysis on the production parameters of surface defects, locate the key abnormal process parameters that induce various surface defects, and carry out targeted closed-loop parameter control to stabilize the production line operation and reduce the incidence of surface defects. Summary of the Invention

[0005] To trace and control the production parameters of gypsum board surface defects, this application provides a system and method for tracing and controlling the production parameters of gypsum board surface defects.

[0006] According to a first aspect of the embodiments of this application, a method for tracing and controlling production parameters of gypsum board surface defects is provided, comprising: acquiring surface texture reference points of the gypsum board to be tested during the forming stage, and acquiring texture features and corresponding displacement coordinates of the surface defect area of ​​the finished gypsum board obtained by the cutting process of the gypsum board to be tested; determining the theoretical drift time based on the displacement coordinates and the production line conveying speed, and defining a historical time window centered on the theoretical drift time, and performing sliding pixel matching between the texture features and the surface texture reference points within the historical time window; when the matching texture overlap is greater than a preset overlap threshold, determining the target timestamp of the surface defect area. From the pre-stored production data of the gypsum board under test during the drying stage, the edge thickness gradient curve, surface temperature rise curve, and internal adhesive curing degree value corresponding to the target timestamp are determined. The edge slope value is determined using the edge thickness gradient curve, and the heat penetration delay parameter is determined using the surface temperature rise curve and the internal adhesive curing degree value. The edge slope value and the heat penetration delay parameter are input into a pre-established working condition mapping database for spatial coordinate mapping to determine the working condition classification result. The working condition classification result is used to match the target equipment control parameters in a pre-established equipment control mapping table, and control commands for the molding guide equipment or drying hot air equipment are output.

[0007] This allows for traceability and control of production parameters related to surface defects in gypsum boards.

[0008] Optionally, the surface temperature rise curve and the internal adhesive curing degree value are determined by the following method: acquiring infrared thermal images and microwave dielectric signals of the edge of the gypsum board under test during the drying stage; using the infrared thermal images to determine the surface temperature rise curve; extracting the actual dielectric loss factor from the microwave dielectric signal; performing a difference operation between the actual dielectric loss factor and the pre-acquired liquid starch adhesive dielectric loss factor and solid starch adhesive dielectric loss factor; and outputting the relative dielectric difference characteristic value; performing a mapping operation on the relative dielectric difference characteristic value to determine the internal adhesive curing degree value used to characterize the cross-linking state of the adhesive line at the edge of the gypsum core layer.

[0009] Optionally, the method further includes: determining that the corresponding face paper is in an abnormal deformation state when the texture overlap is less than or equal to a preset overlap threshold, thereby triggering a degradation compensation mode; inputting the displacement coordinates into a dynamic velocity compensation matrix containing historical transmission speed data to perform spatial displacement extrapolation calculations, and determining the target timestamps of the updated surface defect areas in the forming and drying stages.

[0010] Optionally, the edge slope value can be determined using the edge thickness gradient curve, including: extracting the maximum vertical drop value and the horizontal width value from the edge thickness gradient curve, and using the ratio of the maximum vertical drop value to the horizontal width value as the edge slope value; the edge slope value is used to characterize the steepness of the molding shape of the gypsum core layer.

[0011] Optionally, the heat penetration delay parameter is determined using the surface temperature rise curve and the internal adhesive curing degree value, including: extracting the starting time point when the surface temperature rise curve reaches the preset temperature reference on the time axis, and simultaneously extracting the response time point when the internal adhesive curing degree value reaches the preset curing reference; subtracting the starting time point from the response time point to obtain the lag time difference value, and using the ratio of the lag time difference value to the preset characteristic thermal response time constant of the face paper as the heat penetration delay parameter.

[0012] In this way, the heat transfer time is transformed into a standard evaluation indicator that can be directly called by the controller, thus improving the scientific nature of the evaluation of thermal parameters in multiple temperature zones.

[0013] Optionally, the working condition mapping database is pre-divided into multiple distribution intervals corresponding to different working conditions; the edge slope value and heat penetration delay parameter are synchronously input into the pre-established working condition mapping database for spatial coordinate mapping to determine the working condition classification result, including: determining the spatial distribution position of the monitoring mapping coordinate points in the working condition mapping database after spatial coordinate mapping, determining the target distribution interval in which the spatial distribution position falls among multiple distribution intervals, and taking the working condition corresponding to the target distribution interval as the working condition classification result; wherein, the monitoring mapping coordinate points are coordinate points determined according to the edge slope value and heat penetration delay parameter of the gypsum board to be tested; different working conditions include airflow separation damage at the edge of the gypsum board and insufficient hot air penetration at the edge of the gypsum board.

[0014] This allows for the determination of the working conditions in the gypsum board production process, enabling targeted control of the production process.

[0015] Optionally, if the working condition classification result indicates that the airflow separation at the edge of the gypsum board is disrupted, the target equipment control parameters are matched in a pre-established equipment control mapping table using the working condition classification result, and control instructions are output for the molding guide equipment or the drying hot air equipment. These instructions include: using the difference between the edge slope value and the preset slope threshold as the slope excess, and using the mechanical deflection angle that matches the slope excess in the preset calibration database as the target equipment control parameter; the calibration database contains the mapping relationship between the deflection angle of the molding guide equipment and the edge slope value, as well as the dimensional constraint condition of the total edge thickness that prevents interference from the folding of the gypsum board facing paper; outputting a first control instruction for driving the molding guide equipment to deflect outward by a mechanical deflection angle, and outputting a pressure lock instruction to the glue pump equipment to prevent an increase in the amount of edge glue injected.

[0016] In this way, by mapping the slope overshoot to the calibration database, energy waste caused by simply increasing the hot air volume is avoided.

[0017] Optionally, if the working condition classification result indicates insufficient hot air penetration at the edge of the gypsum board, the target equipment control parameters are matched in a pre-established equipment control mapping table using the working condition classification result, and control instructions are output for the molding guide equipment or drying hot air equipment. These instructions include: using the difference between the heat penetration delay parameter and the preset thermal resistance threshold as the thermal resistance excess; using the mechanical opening adjustment value corresponding to the thermal resistance excess in the pre-established equipment control mapping table as the target equipment control parameter; outputting a second control instruction to increase the mechanical opening of the lateral drying hot air equipment in the corresponding drying temperature zone, so that the increased mechanical opening value is equal to the mechanical opening adjustment value; and outputting a pressure lock instruction to the glue pump equipment to prohibit increasing the amount of edge glue injection.

[0018] Optionally, after outputting a pressure lock command to the glue pump equipment to prohibit increasing the amount of edge glue injection, the method further includes: initiating a timeout self-recovery monitoring process for the pressure lock command, and obtaining the number of degummed boards in the surface defect area during the production of a preset monitored quantity of subsequent finished gypsum boards; when the number of degummed boards is greater than or equal to a preset degumming warning threshold, releasing the pressure lock state of the glue pump equipment and outputting a verification alarm command.

[0019] In this way, by setting a moving monitoring window with continuous output quantity and early warning threshold, different degumming situations caused by thermal fluid factors and mechanical wear can be effectively distinguished, preventing line stoppage accidents caused by long-term lockout commands.

[0020] Optionally, the method further includes: obtaining the instantaneous tension of the facing paper during the unwinding path of the paper feeding roll upstream of the forming stage, and obtaining the first micro-wave spatial spacing value of the facing paper surface and the second undulating wave spatial spacing value of the finished gypsum board surface; using cross-correlation calculation to dynamically compare the first micro-wave spatial spacing value and the second undulating wave spatial spacing value to determine the cross-correlation peak value; when the cross-correlation peak value falls within the preset tolerance range, and the instantaneous tension has a transient fluctuation corresponding to the matching frequency, determining that the corresponding undulating wave is a type of equipment failure caused by mechanical damping jamming of the facing paper.

[0021] Optionally, the method further includes: acquiring an internal X-ray transmission image of the gypsum board under test during the drying stage; using a region extraction algorithm based on grayscale distribution to locate the high-density agglomeration area of ​​gypsum in the internal X-ray transmission image; and extracting the spatial absolute coordinates corresponding to the high-density agglomeration area of ​​gypsum; acquiring the brittle fracture acoustic signal of the finished gypsum board during the cutting stage; separating the gypsum board cutting position coordinates corresponding to the abnormal acoustic spectrum; performing a spatial overlap comparison calculation between the gypsum board cutting position coordinates and the spatial absolute coordinates; and outputting a third control command to increase the excitation force of the vibrating fluidized bed of the gypsum powder homogenization equipment upstream of the molding stage when the gypsum board cutting position coordinates match the spatial absolute coordinates.

[0022] According to a second aspect of the present application, a production parameter traceability and control system for gypsum board surface defects is provided, comprising: a processor and a memory, wherein the memory stores computer program instructions, and the computer program instructions, when executed by the processor, implement the steps of the production parameter traceability and control method for gypsum board surface defects provided in the first aspect of the present application.

[0023] The technical solutions provided by the embodiments of this application may include the following beneficial effects: obtaining the surface texture reference points of the gypsum board under test during the molding stage, and obtaining the texture features and corresponding displacement coordinates of the surface defect area of ​​the finished gypsum board obtained by the cutting process of the gypsum board under test, so as to determine the working condition classification result of the gypsum board production process according to the performance of the gypsum board at different production stages; using the working condition classification result to match the target equipment control parameters in the pre-established equipment control mapping table, and output control commands for the molding guide equipment or drying hot air equipment, so as to realize the traceability control of the production parameters of the gypsum board surface defects.

[0024] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating a method for tracing and controlling production parameters of gypsum board surface defects according to an exemplary embodiment;

[0026] Figure 2 This is a schematic diagram illustrating the spatial distribution mapping of a working condition mapping database; Figure 3 This is a schematic diagram of a production parameter traceability and control system for gypsum board surface defects, according to an exemplary embodiment. Detailed Implementation

[0027] To trace and control the production parameters of gypsum board surface defects, this application provides a system and method for tracing and controlling the production parameters of gypsum board surface defects. Figure 1 This is a flowchart illustrating a method for tracing and controlling production parameters for defects on gypsum board surfaces, according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps.

[0028] In step S101, sliding pixel matching is performed on the gypsum board to be tested.

[0029] The surface texture reference points of the gypsum board under test during the forming stage are obtained, and the texture features and corresponding displacement coordinates of the surface defect area of ​​the finished gypsum board obtained by the cutting process are also obtained. The theoretical drift time is determined based on the displacement coordinates and the production line conveying speed, and a historical time window is defined with the theoretical drift time as the center. Within the historical time window, the texture features and the surface texture reference points are matched by sliding pixels.

[0030] The total length of a continuous gypsum board manufacturing production line in an industrial setting is typically greater than 150 meters. The conveyor belt at the bottom carries the semi-solid gypsum board to be tested from the forming station to the multi-temperature zone drying kiln at the far end. An industrial digital camera array is arranged above the mechanical pressing plate in the forming stage. The industrial digital camera array is used to acquire images of the continuously moving surface of the facing paper. The acquired raw optical image data is transmitted to the programmable logic controller and industrial control computer.

[0031] The industrial control computer reads the lens calibration parameters pre-stored in the internal register, performs radial distortion correction and spatial perspective transformation on the original optical image, and restores the true texture of the face paper surface in a two-dimensional horizontal plane. Because the face paper undergoes interlacing and pressing of wood pulp fibers in the papermaking process, the face paper surface has a randomly distributed and structurally unique fine fiber texture.

[0032] Industrial control computers extract pixel regions with significant grayscale distribution gradients as surface texture reference points for continuous gypsum boards. These surface texture reference points provide coordinate references for continuous movement, ensuring that subsequent traceability of quality issues at the physical cut-off point at the end of the production line is supported by coordinates.

[0033] Finished gypsum boards undergo surface optical scanning inspection at the quality inspection station at the end of the production line. When the quality inspection system detects abnormal gray-level bulging or color peeling in a local area, it marks the abnormal area as a surface defect area. The control system can extract the gray-level co-occurrence matrix features or local binary pattern features inside and around the surface defect area as texture features for comparative analysis.

[0034] The control system reads the current absolute mechanical position of the finished gypsum board at the quality inspection station and the corresponding current timestamp. Combined with the running speed parameters of the dryer conveyor belt and the forming conveyor belt recorded in the production execution system, or the cumulative pulse value recorded by the incremental photoelectric encoder installed on the transmission roller shaft, the system calculates the drift time required for the surface defect area to spatially revert back to the forming station. The control system extends a preset time span in the forward and backward time axis directions with the theoretical drift time as the center node, defining a historical time window with redundancy and fault tolerance capabilities.

[0035] Within the historical time window, the control system uses the texture features extracted from the surface defect area as a detection template with a fixed pixel size, and performs pixel-by-pixel step-by-pixel translation comparison on the panoramic reference image composed of consecutive image frames corresponding to the historical time window.

[0036] The control system calculates the spatially normalized cross-correlation values ​​of the overlapping areas of the detection template and the panoramic reference image. These values ​​are used to characterize the statistical similarity of the pixel grayscale distribution between two independent image regions.

[0037] The spatially normalized cross-correlation coefficient is used to characterize the statistical similarity of the overlapping regions of the detection template and the panoramic reference image in terms of pixel gray level distribution. This coefficient is used as a specific indicator of texture overlap in subsequent steps. Texture overlap is the numerical representation of the spatially normalized cross-correlation coefficient.

[0038] By retrieving the maximum peak position of the spatially normalized cross-correlation values, the precise coordinates of the surface defect area on the current finished board at the initial formation stage can be located, thus achieving precise time axis alignment of spatial physical defects on long-distance production lines.

[0039] The calculation process of the spatially normalized cross-correlation coefficient is as follows: The pixel gray values ​​of the overlapping areas of the detection template and the panoramic reference image are respectively processed to zero-mean value to eliminate the influence of the overall change in illumination intensity on the matching result, and the sum of the dot product between the detection template and the overlapping area after zeroing is calculated.

[0040] Calculate the sum of squares of the gray values ​​of each pixel in the detection template after zeroing and the sum of squares of the gray values ​​of each pixel in the overlapping area. Divide the sum of the dot products by the square root of the product of the two sums of squares to obtain the spatially normalized cross-correlation coefficient.

[0041] During the sliding pixel matching process, the image is translated point by point on the panoramic reference image with a step size of a single pixel. The spatial normalized cross-correlation coefficient of each translation position is calculated. The maximum coefficient value among all translation positions is used as the texture overlap, and the position corresponding to the maximum coefficient value is used as the matching and localization result.

[0042] During the molding and extrusion process, the gypsum slurry undergoes a slight overflow and expansion deformation from the side area between the two layers of facing paper. A linear laser emitter installed on the side of the molding equipment projects structured light stripes onto the undried edge. An auxiliary industrial camera installed at a set angle captures the reflected visible light image. The changing lighting conditions in the industrial environment cause a large amount of environmental background interference in the visible light image. Image processing algorithms can be used to separate the edge shadow area of ​​the gypsum slurry.

[0043] The morphological changes of the edge shadow area reflect the rheological properties of the gypsum slurry under lateral extrusion. By extracting the pixel displacement of the edge shadow area relative to the calibrated zero point, the original two-dimensional image data reflecting the dynamic changes of the gypsum board edge thickness is obtained. The preset triangulation space transformation matrix integrates camera intrinsic parameters and laser plane equation parameters. Through matrix multiplication, the two-dimensional pixel displacement is transformed into depth information in three-dimensional space.

[0044] The vertical height difference output reflects the height bulge of the gypsum board side caused by grout accumulation at the current moment. The height difference values ​​obtained in the continuous time period are smoothly spliced ​​and interpolated according to the preset system time series to generate an edge thickness gradient curve that can continuously reflect the dynamic changes in the edge thickness of the gypsum board.

[0045] The edge thickness gradient curve provides reliable data support for subsequent analysis of the uniformity of gypsum slurry distribution in the edge region. Frequency domain filtering transforms the original spatial domain visible light image matrix to the spatial frequency domain through two-dimensional discrete Fourier transform. In industrial manufacturing plants, the ambient light during the molding stage is often severely interfered with by a large amount of free dust and high-temperature water vapor, which may result in randomly distributed spatial high-frequency noise in the acquired visible light image.

[0046] The corresponding frequency domain attenuation function can be set for high-frequency noise components at different spatial scales to suppress interference from sudden changes in local image grayscale caused by water vapor condensation or dust diffuse reflection. The visible light image after inverse Fourier transform restoration retains the low-frequency information of the original fiber structure on the surface of the face paper. The industrial control system identifies and extracts the fiber texture intersections as surface texture reference points with high illumination invariance, which improves the positioning reliability of reference point extraction under complex working conditions.

[0047] In step S102, the target timestamp of the surface defect area of ​​the gypsum board is determined.

[0048] When the matching texture overlap is greater than the preset overlap threshold, the target timestamp of the surface defect area is determined. From the pre-stored production data of the gypsum board under test during the drying stage, the edge thickness gradient curve, surface temperature rise curve and internal glue curing degree value corresponding to the target timestamp are determined.

[0049] During stable production line operation, the texture of surface defect areas and the texture of reference points in the forming stage of more than 500 normal finished boards can be continuously collected. The normalized cross-correlation coefficient of each board can be calculated to establish a statistical distribution, and the preset overlap threshold can be determined by using the statistical analysis results.

[0050] Based on the determined target timestamp, a structured query command is executed in the distributed industrial database. The distributed industrial database is then written with the production process data collected by the sensor nodes of the production line in time sequence. The drying stage is the core thermal process that determines the final physical strength of the gypsum board. Multiple independent temperature control zones are arranged inside the drying kiln.

[0051] It can locate the data storage block corresponding to the target timestamp, extract the edge thickness gradient curve data directly bound to the target timestamp into the cache memory for preprocessing, and simultaneously extract the surface temperature rise curve data containing the continuous recorded values ​​of the temperature sensor to obtain the internal glue curing degree value used to monitor the edge sealing quality. The combination of parameters in each independent dimension constitutes a complete digital twin mapping describing the production history of a specific defect area from three perspectives: morphological structure, thermodynamic transfer state, and chemical crosslinking progress.

[0052] In one embodiment, the surface temperature rise curve and the internal adhesive curing degree value are determined by the following method: acquiring an infrared thermal image and microwave dielectric signal of the edge of the gypsum board under test during the drying stage; using the infrared thermal image to determine the surface temperature rise curve; extracting the actual dielectric loss factor from the microwave dielectric signal; performing a difference operation between the actual dielectric loss factor and the pre-acquired liquid starch adhesive dielectric loss factor and solid starch adhesive dielectric loss factor; and outputting a relative dielectric difference characteristic value; performing a mapping operation on the relative dielectric difference characteristic value to determine the internal adhesive curing degree value used to characterize the crosslinking state of the adhesive line at the edge of the gypsum core layer.

[0053] An infrared thermal imager installed in the observation window on the side of the drying kiln can capture the distribution of thermal radiation energy in the edge area of ​​the gypsum board under test during the drying stage. Each pixel grayscale value of the infrared thermal image corresponds to a specific surface Celsius temperature value. By analyzing a continuous sequence of infrared thermal images, the surface temperature rise curve of the gypsum board edge as the conveying time increases can be reconstructed.

[0054] The surface temperature rise curve reflects the combined efficiency of convective and radiative heat transfer from the external heat source to the surface of the facing paper. Since the upper and lower layers of facing paper at the edge of the gypsum board are bonded and sealed with special modified starch adhesive, the moisture inside the starch adhesive gradually evaporates after absorbing heat, while the polymer undergoes a cross-linking and curing reaction. Under the action of microwave electromagnetic field, liquid water molecules undergo intense polarization relaxation and generate significant microwave energy absorption loss.

[0055] A microwave sensor placed at a specific location in the drying kiln emits a microwave detection signal toward the edge of the gypsum board and receives the microwave dielectric signal after it is transmitted or reflected. The actual dielectric loss factor in the microwave dielectric signal is extracted as a key physical indicator for evaluating the proportion of free water contained in the edge glue. The value of the actual dielectric loss factor is positively correlated with the water content in the glue.

[0056] The extracted actual dielectric loss factor is linearly differentially calculated with the dielectric loss factor of liquid starch glue and the dielectric loss factor of solid starch glue in a completely dry state, which were previously measured in a laboratory under standard constant temperature and humidity conditions. By calculating the relative deviation of the current actual value between the liquid and solid references, the relative dielectric difference characteristic value is output to eliminate the interference of the overall environmental humidity background.

[0057] The relative permittivity difference characteristic value is a dimensionless proportionality coefficient between 0 and 1. The relative permittivity difference characteristic value can be nonlinearly mapped by referring to the curing kinetics test calibration table pre-entered in the register. The internal glue curing degree value used to characterize the cross-linking state of the glue line at the edge of the gypsum core layer was determined. The magnitude of the internal glue curing degree value can characterize whether the edge of the finished gypsum board will have serious mechanical performance defects such as cracking or delamination.

[0058] In one implementation, if the texture overlap is less than or equal to a preset overlap threshold, the corresponding face paper can be determined to be in an abnormal deformation state to trigger a degradation compensation mode; the displacement coordinates are input into a dynamic velocity compensation matrix containing historical transmission speed data to perform spatial displacement extrapolation calculations, and the target timestamps of the updated surface defect areas in the forming and drying stages are determined.

[0059] As the paper passes through multiple mechanical transmission rollers, it continuously bears longitudinal tensile stress. Under specific high temperature and high humidity conditions, the paper fibers in some areas may undergo plastic elongation deformation or nonlinear creep. Tensile deformation causes the texture features carried by the surface defect area to be locally distorted in terms of physical size and spatial arrangement, which in turn causes a significant reduction in texture overlap during the sliding pixel matching process.

[0060] When the extracted texture overlap is less than or equal to the set overlap threshold, such as 0.8, the accuracy of timestamp backtracking cannot be guaranteed by relying on texture similarity. It can be determined that the corresponding face paper is in an abnormal stretching and twisting deformation state, triggering the degradation compensation mode to ensure the continuous operation of the monitoring system. Those skilled in the art can adjust the specific value of the exemplary overlap threshold according to the actual working conditions.

[0061] The dynamic speed compensation matrix, which contains historical transmission speed data of each upstream transmission node, can be called. The dynamic speed compensation matrix is ​​a piecewise integral operation data structure. Each row of the matrix corresponds to one transmission node in the forming conveyor belt, transition conveyor belt, and drying conveyor belt, and each column corresponds to a sampling time interval.

[0062] The matrix elements record the actual linear velocity values ​​of the corresponding transmission nodes at the corresponding sampling time. The sampling time interval is set to one second. When performing spatial displacement extrapolation calculations, the dynamic velocity compensation matrix is ​​traversed in reverse from the current time. The passage time of the surface defect area on each transmission node is integrated piecewise and accumulated. The integration step size is consistent with the sampling time interval.

[0063] During the accumulation process, when the deviation between the actual linear velocity value and the nominal linear velocity value of a certain transmission node exceeds 10%, the actual linear velocity value of the transmission node is used to replace the nominal value in the integration calculation, thereby correcting the displacement calculation error caused by velocity fluctuation. The result of the integration accumulation is the actual spatial displacement of the surface defect area from the quality inspection station to the forming station. Based on the actual spatial displacement and the spatial layout parameters of each station, the target timestamp of the updated surface defect area in the forming stage and the drying stage is determined.

[0064] Inverse discrete integration was performed on the velocity data in the time dimension to perform spatial displacement extrapolation based on mechanical dynamics characteristics. The influence of visual recognition deviation caused by texture distortion was eliminated. The target timestamps of the updated surface defect region in the forming and drying stages were determined by solving the dynamic equations. This ensured that the start time of defect generation could be accurately located under complex material deformation conditions, and improved the robustness of the system under harsh production conditions.

[0065] In step S103, the working condition classification result of the defects of the gypsum board is determined.

[0066] The edge slope value is determined by using the edge thickness gradient curve, and the heat penetration delay parameter is determined by using the surface temperature rise curve and the internal adhesive curing degree value. The edge slope value and the heat penetration delay parameter are input into the pre-established working condition mapping database for spatial coordinate mapping to determine the working condition classification result.

[0067] For example, determining the edge slope value using the edge thickness gradient curve includes: extracting the maximum vertical drop value and the horizontal width value from the edge thickness gradient curve, and using the ratio of the maximum vertical drop value to the horizontal width value as the edge slope value; the edge slope value is used to characterize the steepness of the molding shape of the gypsum core layer.

[0068] The edge thickness gradient curve reflects the spatial geometric distribution contour of the gypsum slurry as it overflows to both sides and gradually thins after being squeezed by the upper and lower facing papers; the microprocessor of the control system automatically retrieves the highest point and the lowest reference point of the edge thickness gradient curve in the vertical thickness direction through the peak detection algorithm, and calculates the distance difference between the two to obtain the maximum vertical drop value.

[0069] The horizontal width value is obtained by calculating the distance span between the start and end points of the curve in the width direction of the cross section. The specific edge slope value is obtained by dividing the maximum vertical drop value reflecting the degree of thickness change by the horizontal width value reflecting the lateral extension range. The edge slope value is equivalent to the tangent function value of the slurry edge distribution section. The larger the edge slope value, the more severe the thickness attenuation of the slurry has occurred within a very short horizontal distance.

[0070] The edge slope value is used to visually characterize the steepness of the molding shape of the gypsum core layer. The steep molding shape leads to complex local aerodynamic turbulence resistance when the core is purged by strong lateral convection hot air inside the high-temperature drying kiln.

[0071] Furthermore, the heat penetration delay parameter is determined using the surface temperature rise curve and the internal adhesive curing degree value. This includes: extracting the starting time point when the surface temperature rise curve reaches the preset temperature reference on the time axis, and simultaneously extracting the response time point when the internal adhesive curing degree value reaches the preset curing reference; subtracting the starting time point from the response time point to obtain the lag time difference; and using the ratio of the lag time difference to the preset characteristic thermal response time constant of the facing paper as the heat penetration delay parameter. The heat penetration delay parameter is a dimensionless characteristic index that combines time delay and material thermal resistance effect. A value greater than the preset parameter threshold indicates a stronger resistance to heat penetration through the facing paper to the gypsum core layer. The heat penetration delay parameter is used to characterize the degree of influence of the time delay of hot air penetrating the facing paper layer to the interior of the gypsum core layer on the heat conduction process.

[0072] The characteristic thermal response time constant of the facing paper can be predetermined based on the characteristic thermal response time of the facing paper during the manufacturing process of qualified gypsum board. The preset temperature reference can be, for example, equal to 85℃. An infrared thermal imager can be placed at the inlet of the drying kiln to record the temperature rise curves of 100 normal board edges. The first derivative of the curve is calculated, and the inflection point temperature at which the slope changes from flat to steep is taken to determine the preset temperature reference.

[0073] In the initial stage of gypsum board entering the drying kiln, the surface temperature rises as it absorbs heat from the surrounding environment. The control system retrieves the specific moment on the time axis when the surface temperature rise curve changes from flat to rapid and reaches the exemplary temperature benchmark of 85 degrees Celsius, and marks it as the starting point when heat begins to accumulate in large quantities. Because the facing paper contains a large number of porous fiber structures that hinder the rapid conduction of heat energy, the curing reaction of the internal adhesive has a certain time lag relative to the surface temperature rise.

[0074] The response time point corresponding to the internal adhesive curing degree value reaching the exemplary curing benchmark is extracted synchronously. The curing benchmark can be equal to 60% for example. The lag time difference is calculated by subtracting the relatively advanced start time point from the relatively lagging response time point. The lag time difference reflects the total time consumed by heat energy to pass through the cover paper of a specific thickness.

[0075] The ratio of the lag time difference to the preset characteristic thermal response time constant of the facing paper is used as the heat penetration delay parameter. The time dimension variable and the material property variable are integrated by using dimensionless transformation. The obtained heat penetration delay parameter takes into account the cumulative effect of material property resistance and actual transmission time. The heat penetration delay parameter is used to characterize the influence of the time delay of hot air penetrating the facing paper layer to reach the interior of the gypsum core layer on the heat conduction process.

[0076] When the heat penetration delay parameter is greater than the preset threshold, it means that the internal adhesive is in a low-temperature, non-crosslinked state for a long time. It is easy for the adhesive to fail and cause delamination defects due to the rapid boiling and escape of moisture during the subsequent drying process. The dimensionless characteristics of the heat penetration delay parameter provide a solid physical basis for the comparative analysis of the control system.

[0077] In one implementation, the working condition mapping database is pre-divided into multiple distribution intervals corresponding to different working conditions; the edge slope value and the heat penetration delay parameter are synchronously input into the pre-established working condition mapping database for spatial coordinate mapping to determine the working condition classification result, including: determining the spatial distribution position of the monitoring mapping coordinate points in the working condition mapping database after spatial coordinate mapping, determining the target distribution interval in which the spatial distribution position falls among multiple distribution intervals, and taking the working condition corresponding to the target distribution interval as the working condition classification result; wherein, the monitoring mapping coordinate points are coordinate points determined according to the edge slope value and the heat penetration delay parameter of the gypsum board to be tested; different working conditions include airflow separation damage at the edge of the gypsum board and insufficient hot air penetration at the edge of the gypsum board.

[0078] The working condition mapping database is a two-dimensional Cartesian coordinate system data structure stored in non-volatile memory. The horizontal axis represents the edge slope value and the vertical axis represents the heat penetration delay parameter. The database is based on a large amount of prior industrial fluid dynamics simulation data and on-site historical fault statistics. It uses an unsupervised classification algorithm to divide the two-dimensional space into regions and pre-defines multiple distribution interval boundary lines corresponding to different potential quality risk working conditions.

[0079] The control system combines the edge slope value obtained from real-time acquisition and calculation with the heat penetration delay parameter into a two-dimensional array, and synchronously inputs it into the working condition mapping database for spatial coordinate mapping processing. This materializes the abstract process parameter into a monitoring mapping coordinate point on a two-dimensional plane. By comparing the relative spatial topological relationship between the monitoring mapping coordinate point and the boundary line of each distribution interval, the control system accurately determines the target distribution interval into which the spatial distribution location actually falls.

[0080] The control system uses the working conditions corresponding to the physical state labels pre-bound to the target distribution range containing the monitoring and mapping coordinate points as the current working condition classification result. Different working conditions include two typical abnormal production states that lead to defects: airflow separation failure at the edge of the gypsum board and insufficient hot air penetration at the edge of the gypsum board. Through spatial two-dimensional mapping, the serious problems of missed reports and misjudgments that are easy to be generated by linear threshold comparison based on a single edge slope parameter or a single temperature parameter are avoided.

[0081] Figure 2 This is a spatial distribution mapping diagram of a working condition mapping database, such as... Figure 2 As shown, the working condition mapping database is constructed based on a two-dimensional coordinate system. Its horizontal axis represents the quantified value of the steepness of the cross-sectional shape, i.e., the edge slope value, and the vertical axis represents the thermal energy penetration resistance hysteresis ratio, i.e., the thermal penetration delay parameter. A preset slope threshold is pre-calibrated and set in the coordinate system. Figure 2 The morphological warning threshold and the penetration hysteresis warning threshold are included.

[0082] By using the intersection of preset slope thresholds and penetration hysteresis warning thresholds, the two-dimensional space is divided into four distribution intervals corresponding to different working conditions. Specifically: when the edge slope value is less than the preset slope threshold and the thermal penetration delay parameter is less than the penetration hysteresis warning threshold, the monitored mapping coordinate point falls into the normal interval, and the working condition is determined to be normal; when the edge slope value is greater than the preset slope threshold and the thermal penetration delay parameter is less than the penetration hysteresis warning threshold, the working condition is determined to be airflow separation failure at the edge of the gypsum board. In this way, the real-time parameter combination can be intuitively and accurately mapped to the specific working condition classification results.

[0083] In step S104, the production parameters of the gypsum board surface are adjusted.

[0084] Using the operating condition classification results, the target equipment control parameters are matched in a pre-established equipment control mapping table, and control commands for the forming guide equipment or drying hot air equipment are output.

[0085] The pre-stored equipment control mapping table inside the control system is a key-value pair data structure dictionary. The dictionary data structure uses a lookup algorithm to map different operating condition classification results to specific mechanical control variables of the underlying actuators. Based on the currently determined abnormal operating condition classification results, the target equipment control parameters that can offset or compensate for the abnormal state can be quickly retrieved.

[0086] Control commands conforming to the general communication protocol are output to the underlying programmable logic controller via the Ethernet bus in the industrial field. After parsing the control parameters contained in the commands, the underlying programmable logic controller generates corresponding analog voltage signals or high-power pulse width modulation signals, which drive the stepper motor of the molding guide device installed in the molding stage or the servo motor that controls the damper of the drying kiln to perform corresponding mechanical displacement adjustment actions.

[0087] For example, when the working condition classification result is that the airflow separation at the edge of the gypsum board is damaged, the target equipment control parameters are matched in a pre-established equipment control mapping table using the working condition classification result, and control instructions for the molding guide equipment or the drying hot air equipment are output. These instructions include: taking the difference between the edge slope value and the preset slope threshold as the slope excess, and taking the mechanical deflection angle that matches the slope excess in the preset calibration database as the target equipment control parameter; the calibration database contains the mapping relationship between the deflection angle of the molding guide equipment and the edge slope value, as well as the dimensional constraint condition of the total edge thickness that prevents the folding interference of the gypsum board facing paper; outputting a first control instruction for driving the molding guide equipment to deflect the mechanical deflection angle outward, and outputting a pressure lock instruction to the glue pump equipment to prevent the increase of edge glue injection.

[0088] When the working condition classification result is determined to be airflow separation failure at the edge of the gypsum board, it means that the current edge slope value, which is greater than the safe range, causes the high-speed hot airflow blown from the side of the drying kiln to cause severe hydrodynamic airflow separation when it comes into contact with the steep edge of the gypsum board. The airflow separation forms a local vacuum vortex, which hinders the effective exchange of heat. This makes it impossible for the surface temperature of the edge area to reach the minimum reaction temperature required for glue cross-linking, which may lead to edge delamination defects.

[0089] The difference between the current abnormally large edge slope value and the upper limit of the safe range preset slope threshold is used as the slope overscaling amount, which reflects the degree of morphological deterioration. By querying the calibration database that stores a large amount of wind tunnel purging test data in advance, a linear interpolation algorithm is used to find the outward expansion mechanical deflection angle required to offset the slope overscaling amount. The mechanical deflection angle is determined as the target equipment control parameter for smoothing the edge morphology.

[0090] The preset slope threshold can be between 0.18 and 0.22; the calibration database not only contains the positive mapping relationship between the deflection angle of the forming guide device and the edge slope value, but also integrates the limit size constraint condition of the total edge thickness that prevents excessive folding interference between the upper and lower facing paper of the gypsum board, thus preventing safety accidents caused by blindly increasing the mechanical deflection angle, which would lead to the gypsum slurry losing lateral support and causing the facing paper to tear.

[0091] The first control command containing a specific angle value is output to the stepper motor driver located at the molding station to drive the molding guide device to deflect outward by a set mechanical deflection angle to forcibly reduce the steepness of the edge thickness gradient curve, so that the strong lateral wind can flow smoothly over the edge area, thereby eliminating the airflow separation vortex that hinders heat exchange.

[0092] The pressure lock command outputs a pressure lock signal containing a specific register address to the glue pump equipment responsible for supplying edge adhesive. Due to insufficient heat caused by airflow separation, if the operator blindly increases the amount of edge adhesive injected due to seeing delamination defects, a large amount of liquid water that cannot solidify will accumulate at the edge, causing severe steam bubbling and expansion. The pressure lock command can prevent manual adjustment operations that increase the amount of edge adhesive injected, preventing the defects from worsening.

[0093] Furthermore, when the working condition classification result indicates insufficient hot air penetration at the edge of the gypsum board, the target equipment control parameters are matched in the pre-established equipment control mapping table using the working condition classification result, and control instructions are output for the molding guide equipment or drying hot air equipment. These instructions include: using the difference between the heat penetration delay parameter and the preset thermal resistance threshold as the thermal resistance excess, and using the mechanical opening adjustment value corresponding to the thermal resistance excess in the pre-established equipment control mapping table as the target equipment control parameter; outputting a second control instruction to increase the mechanical opening of the lateral drying hot air equipment in the corresponding drying temperature zone, so that the increased mechanical opening value is equal to the mechanical opening adjustment value, and outputting a pressure lock instruction to the glue pump equipment to prohibit the increase of edge glue injection.

[0094] When the working condition classification result is determined to be insufficient hot air penetration at the edge of the gypsum board, it indicates that although the edge shape is relatively flat and there is no airflow vortex obstruction, the current batch of facing paper has a moisture content that exceeds the standard or the paper's air permeability parameter has decreased. This causes the energy carried by the hot air to be excessively consumed in the process of penetrating the paper to reach the internal adhesive layer, resulting in a serious lag in heat transfer.

[0095] The difference obtained by subtracting the current excessively large thermal penetration delay parameter from the preset thermal resistance threshold used to characterize the normal penetration time can characterize the thermal resistance excess of the additional thermal resistance level; the control system uses a lookup table algorithm to retrieve the mechanical opening adjustment value of the air supply duct louvers corresponding to the thermal resistance excess value from a pre-established equipment control mapping table.

[0096] The process of establishing the equipment control mapping table is as follows: For edge airflow separation failure conditions, the lateral hot air flow state under different edge slope values ​​is simulated in a laboratory wind tunnel environment. The wind speed distribution in the edge area is measured using a hot-wire anemometer, and the critical slope value at which significant airflow separation occurs is recorded. Based on the critical slope value, the slope is increased incrementally according to the slope excess, and the edge slope improvement effect of the forming guide device is tested at different outward deflection angles. The minimum mechanical deflection angle that can reduce the edge slope value below the critical slope value is recorded, and a correspondence table between slope excess and mechanical deflection angle is established.

[0097] To address the issue of insufficient edge hot air penetration, the mechanical opening of the lateral drying hot air equipment is adjusted on-site in the drying kiln, gradually increasing to 100%. Simultaneously, the changes in the heat penetration delay parameter under different opening conditions are measured, and the minimum mechanical opening adjustment value at which the heat penetration delay parameter decreases below the preset thermal resistance threshold is recorded. A correspondence table between the thermal resistance exceeding the standard and the mechanical opening adjustment value is established. The two sets of correspondence tables are stored in the equipment control mapping table in the form of key-value pairs. The dimensional constraint condition of the total edge thickness is set as a safety verification item. When the calculated mechanical deflection angle causes the total edge thickness to exceed the dimensional constraint condition, an alarm is triggered and the deflection angle is limited to a safe range.

[0098] The preset thermal resistance threshold can be equal to the average value of the historical heat penetration delay parameters in the normal gypsum board production process; the mechanical opening adjustment value can be used as the target equipment control parameter for increasing the total heat radiation air volume; a second control command is output to the servo cylinder controller of the lateral drying hot air equipment corresponding to the specific drying temperature zone that produces insufficient heat penetration defects. The control command drives the cylinder to extend so that the mechanical opening value of the air supply louver is equal to the calculated mechanical opening adjustment value.

[0099] Increasing the cross-sectional ventilation area enhances the total dynamic and static energy of the hot air blowing towards the edge of the gypsum board, allowing more heat to penetrate the gaps in the facing paper fibers, which have high thermal resistance. Simultaneously, a pressure lock command is output to the glue pump equipment to prevent the increase of edge glue injection, thus avoiding the complete absorption of limited penetrating heat due to excessive injection of uncrosslinked cold glue before the heat penetration capacity returns to normal, which could cause damage to the edge due to local moisture saturation.

[0100] By acquiring the surface texture reference points during the forming stage and performing sliding pixel matching within the historical time window, precise timestamp positioning of defect areas on long-distance production lines was achieved, providing a reliable time reference for subsequent parameter tracing. By extracting the edge thickness gradient curve and surface temperature rise curve, the physical state of the defect area was characterized from two dimensions: morphological structure and thermodynamic transfer, overcoming the one-sidedness of single-parameter evaluation.

[0101] By establishing a two-dimensional mapping relationship between edge slope value and heat penetration delay parameter, process anomalies are transformed into intuitive coordinate space distributions, facilitating the control system to quickly identify the type of working condition and match the corresponding control strategy. By distinguishing between two typical working conditions—edge airflow separation failure and insufficient edge hot air penetration—and outputting differentiated equipment control commands, energy waste caused by simply increasing the hot air volume is avoided, as is the risk of tearing of the face paper caused by blindly adjusting the molding equipment. By setting pressure lockout commands and timeout self-recovery monitoring processes, the baseline drift fault of the glue pump flow can be identified in a timely manner, balancing the relationship between defect suppression and equipment protection.

[0102] In one embodiment, after outputting a pressure lock command to the glue pump equipment to prohibit increasing the amount of glue injected at the edges, a timeout self-recovery monitoring process for the pressure lock command can also be initiated, and the number of degummed boards with surface defects in the process of producing 200 consecutive finished gypsum boards can be obtained; if the number of degummed boards is greater than or equal to a preset degumming warning threshold of 10 boards, it is determined that there is a flow baseline drift fault in the glue pump equipment, the pressure lock state of the glue pump equipment is released, and a verification alarm command is output.

[0103] The pressure lock command, which prohibits increasing the amount of adhesive injected at the edges, effectively prevents edge bubbling caused by moisture overload. However, if the actual adhesive flow rate of the adhesive pump drifts and decreases due to mechanical wear of the equipment over a long period of time, maintaining the pressure lock state for a long time will cause large-area physical delamination and degumming at the edges of the gypsum board due to insufficient absolute amount of adhesive.

[0104] The background timer module of the control system can initiate a timeout self-recovery monitoring process for pressure lockout commands. The monitoring process is equipped with a moving monitoring sliding window. The control system continuously acquires the number of delaminated boards with severe delamination areas during the production of 200 exemplary finished gypsum boards through the end vision inspection camera system.

[0105] When the number of degummed boards obtained from the statistics is greater than or equal to the set example preset degumming warning threshold of 10 boards, it can be determined that the continuously increasing number of degummed defective products is no longer due to insufficient heat or airflow separation causing the glue to not cure, but rather due to mechanical wear failure of insufficient flow of the glue pump body. An unlock reset signal can be sent to release the pressure lock state of the glue pump equipment.

[0106] The control system can grant manual and automatic control permissions to increase glue flow to restore material supply, and simultaneously drive the audible and visual alarms in the industrial site and the host computer software in the control room to output verification alarm commands, reminding maintenance engineers to go to the site to check whether there is hardening blockage in the glue application pipeline or abnormal wear of the gear pump body.

[0107] In one embodiment, the instantaneous tension of the facing paper during its travel in the paper unwinding path of the paper feeding roll upstream of the forming stage can also be obtained, along with the first micro-wave spatial spacing value of the facing paper surface and the second undulating wave spatial spacing value of the finished gypsum board surface. The cross-correlation calculation is used to dynamically compare the first micro-wave spatial spacing value and the second undulating wave spatial spacing value to determine the cross-correlation peak value. When the cross-correlation peak value falls within a preset tolerance range, and the instantaneous tension exhibits transient fluctuations at a corresponding matching frequency, it is determined that the corresponding undulating wave is a type of equipment failure caused by mechanical damping jamming of the facing paper.

[0108] At the very front of the forming stage is a huge paper feeding roll for the face sheet. A force sensor installed at the rear end of the floating tension roller on the paper ejection path can collect the instantaneous tension signal of the face sheet during high-speed travel in real time. Before the face sheet is completely flattened, its surface exhibits a microscopic fine wrinkle structure. An optical scanning device obtains the first micro-wave crease space spacing value on the surface of the face sheet.

[0109] After the finished gypsum board is dried and cooled at high temperature, the shrinkage of the inner gypsum core layer amplifies the surface morphology difference of the outer facing paper. The end three-dimensional laser scanner obtains the second undulating ripple spatial spacing value of the finished gypsum board surface. The first micro-ripple spatial spacing value and the second undulating ripple spatial spacing value constitute a dynamic transmission mapping relationship in the spatial geometric distribution.

[0110] The specific process of cross-correlation operation is as follows: The spatial spacing values ​​of the first and second undulating ripples are respectively zero-mean processed to eliminate the overall amplitude difference between the two sequences. Fast Fourier Transform (FFT) is then performed on the two zero-meaned sequences to convert the time-domain sequences to the frequency domain. The conjugate product of the two frequency-domain sequences is calculated, and an inverse FFT is performed on the conjugate product result to obtain the cross-correlation function sequence. The peak point in the cross-correlation function sequence is retrieved; the value corresponding to the peak point is the cross-correlation peak value, and the x-coordinate corresponding to the peak point is the optimal time delay or spatial offset between the two sequences.

[0111] The two spatial spacing value sequences collected are converted into frequency domain waveform sequences by cross-correlation operation and then dynamically compared. The cross-correlation peak value, which reflects the correlation between the frequencies of the two ripple distributions, is obtained. When the cross-correlation peak value falls within a preset tolerance range, for example, with 0.95 as the reference value, it is determined that the spatial distribution frequency of the huge undulating ripples on the surface of the finished product is consistent with the spatial distribution frequency of the microwave ripples on the upstream original paper.

[0112] Instantaneous tension data can be extracted to verify whether there are transient tension fluctuations corresponding to the frequency of the ripple distribution. When the tension fluctuation frequency matches, it can be determined that the corresponding undulating ripples are not caused by uneven pulp mixing or abnormal temperature, but by mechanical damping jamming caused by lack of oil in the bearing of the paper feed roll or uneven wear of the friction plate of the mechanical brake. This pinpoints the surface ripple defects caused by periodic mechanical jamming and pulling of the paper.

[0113] In one embodiment, an internal X-ray transmission image of the gypsum board under test during the drying stage can be obtained. A region extraction algorithm based on grayscale distribution is used to locate the high-density agglomeration area of ​​gypsum in the internal X-ray transmission image, and the spatial absolute coordinates corresponding to the high-density agglomeration area of ​​gypsum are extracted. The brittle fracture acoustic signal of the finished gypsum board during the cutting stage is obtained, and the gypsum board cutting position coordinates corresponding to the abnormal acoustic spectrum are separated. The spatial overlap of the gypsum board cutting position coordinates and the spatial absolute coordinates is compared. When the gypsum board cutting position coordinates match the spatial absolute coordinates, a third control command is output to increase the excitation force of the vibrating fluidized bed of the gypsum powder homogenization equipment upstream of the molding stage.

[0114] An X-ray transmission scanner arranged inside a multi-layer hot air drying kiln continuously emits low-dose rays onto the moving gypsum board to be tested. The amount of ray attenuation after penetrating the core layer of the gypsum board is received by the bottom detector array, thus converting it into an internal X-ray transmission map with grayscale changes.

[0115] An abnormal dark spot region with a gray value less than a preset transmission threshold is screened out using a region extraction algorithm based on gray distribution. This locates the high-density agglomeration area of ​​gypsum formed when the gypsum slurry fails to disperse sufficiently during the mixing process, and extracts the spatial absolute coordinates of the high-density agglomeration area.

[0116] When the finished gypsum board passes through the high-speed rotating flying saw cutting blade assembly at the end of the production line, the blades cut the hard gypsum core layer, exciting a high-frequency brittle fracture acoustic signal with specific frequency characteristics. The microphone array arranged at the cutting station can acquire the acoustic signal and transmit it to the signal processor for fast Fourier transform, separating the abnormally sharp sound wave spectrum generated by the blade hitting hard impurities. Simultaneously, the encoder records the gypsum board cutting position coordinates corresponding to the abnormal sound wave spectrum.

[0117] The processing procedure for the acoustic signal of brittle fracture can be as follows: Three microphones are installed in an equilateral triangle layout around the cutting station. The distance between the microphones and the flying saw cutting blade assembly is one meter. The output signals of the three microphones are collected synchronously.

[0118] The three acoustic signals were pre-emphasized and windowed, with a frame length of 20 milliseconds and a frame shift of 10 milliseconds. A Hamming window was used as the window function. A Fast Fourier Transform was performed on each frame to calculate the power spectral density. The power spectral density was compared with the reference power spectrum under normal cutting conditions to calculate the energy anomaly ratio of each frequency band.

[0119] When the energy anomaly ratio of a certain frequency band exceeds three times, the frame is marked as an anomalous frame. Spectral features are extracted from consecutive anomalous frames, including the center frequency, bandwidth, and spectral centroid. The extracted spectral features are then matched with a pre-established anomalous acoustic wave spectrum template library. This library contains spectral features of three typical anomalous types: tool impact with hard impurities, tool wear and dulling, and cavities within the gypsum core layer. When the matching degree exceeds 80%, the anomalous type is determined to be tool impact with hard impurities, and the encoder position information corresponding to the anomalous frame is recorded as the gypsum board cutting position coordinates.

[0120] The system can read the spatial absolute coordinates obtained from the previous scan stored in the network database, and compare the spatial overlap between the gypsum board cutting position coordinates identified at the end and the spatial absolute coordinates in the established three-dimensional virtual digital dashboard. When the Euclidean distance error between the gypsum board cutting position coordinates and the spatial absolute coordinates is less than the matching threshold and thus a successful match is achieved, it can be determined that the hard damage area encountered by the end cutting is affected by the high-density gypsum blocks left in the upstream molding.

[0121] The control system outputs control messages to drive the action of the upstream dry powder material system in the molding stage, and outputs a third control command to the frequency converter driver of the gypsum powder homogenization equipment responsible for feeding, to increase the working voltage of the vibrating motor at the bottom of the vibrating fluidized bed to enhance the excitation force; the enhanced vibration fluidization disperses the free water and lumps inside the gypsum powder before entering the mixer, eliminating the uneven molding thickness and tool chipping quality hazards caused by raw material lumps.

[0122] Figure 3 This is a schematic diagram illustrating the structure of a production parameter traceability and control system 1000 for gypsum board surface defects according to an exemplary embodiment. (Refer to...) Figure 3 The production parameter traceability and control system 1000 for gypsum board surface defects includes a processor 1100 and a memory 1200. The memory 1200 stores computer program instructions. When the computer program instructions are executed by the processor 1100, they implement all or part of the steps of the production parameter traceability and control method for gypsum board surface defects in this application.

[0123] The system's hardware connections are as follows: The industrial digital camera array is connected to the industrial control computer via a gigabit Ethernet interface, transmitting the acquired raw optical image data to the industrial control computer for image processing. The line structure laser emitter and auxiliary industrial camera are connected to the programmable logic controller (PLC) via a universal serial bus interface. The PLC interacts with the industrial control computer via an industrial Ethernet bus.

[0124] Infrared thermal imagers and microwave sensors are connected to a distributed data acquisition module via analog output interfaces or digital communication interfaces. The data acquisition module is connected to an industrial control computer via a fieldbus. Incremental photoelectric encoders are installed on the drive rollers of the forming conveyor belt and the drying conveyor belt. The encoder pulse signals are directly connected to the high-speed counting port of the programmable logic controller.

[0125] The analog output port of the programmable logic controller is connected to the stepper motor driver of the molding and guiding equipment, the servo cylinder controller of the drying hot air equipment, and the frequency converter driver of the glue application pump equipment through shielded cables. It outputs analog voltage signals or pulse width modulation signals to drive the actuators. The industrial control computer and the programmable logic controller are both connected to the industrial Ethernet switch of the production management system to realize the uploading of production data and the storage of quality traceability information.

[0126] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only.

[0127] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method for tracing and controlling production parameters for defects on gypsum board surfaces, characterized in that, include: The surface texture reference points of the gypsum board under test during the forming stage are obtained, and the texture features and corresponding displacement coordinates of the surface defect area of ​​the finished gypsum board obtained by the cutting process of the gypsum board under test are obtained. The historical time window is defined by the displacement coordinates, and the texture features and surface texture reference points are matched by sliding pixels within the historical time window. When the matching texture overlap is greater than the preset overlap threshold, the target timestamp of the surface defect area is determined. From the production data pre-stored for the gypsum board under test during the drying stage, the edge thickness gradient curve, surface temperature rise curve and internal glue curing degree value corresponding to the target timestamp are determined. The edge slope value is determined by using the edge thickness gradient curve, and the heat penetration delay parameter is established by using the surface temperature rise curve and the internal adhesive curing degree value. The edge slope value and the heat penetration delay parameter are input into the pre-established working condition mapping database for spatial coordinate mapping to establish the working condition classification result. Using the operating condition classification results, the target equipment control parameters are matched in a pre-established equipment control mapping table, and control commands for the forming guide equipment or drying hot air equipment are output.

2. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, The edge thickness gradient curve and surface texture reference points are determined in the following way: The visible light image of the undried edge of the gypsum board under test during the forming stage is obtained. The pixel displacement of the shadow area of ​​the edge containing the gypsum slurry in the visible light image is extracted. The pixel displacement is substituted into the preset triangulation space transformation matrix for solution. The vertical height difference is output. The continuously obtained height difference is spliced ​​according to the time series to generate the edge thickness gradient curve. A multi-scale spatial frequency domain filtering algorithm is used to filter and denoise visible light images, and the intersection points of the paper fiber texture in the processed visible light images are used as the surface texture reference points.

3. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, The surface temperature rise curve and the internal adhesive curing degree value were determined in the following way: Infrared thermal images and microwave dielectric signals of the edge of the gypsum board under test during the drying stage are acquired. The surface temperature rise curve is determined using the infrared thermal images. The actual dielectric loss factor in the microwave dielectric signal is extracted. The actual dielectric loss factor is then compared with the pre-acquired dielectric loss factors of liquid starch glue and solid starch glue to perform a differential operation and output the relative dielectric difference characteristic value. By mapping and calculating the relative permittivity difference characteristic values, an internal adhesive curing degree value is established to characterize the cross-linking state of the adhesive lines at the edge of the gypsum core layer.

4. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, The method further includes: When the matching texture overlap is less than or equal to the preset overlap threshold, the corresponding face paper is determined to be in an abnormal deformation state to trigger the degradation compensation mode. The displacement coordinates are input into a dynamic velocity compensation matrix containing historical transmission speed data to perform spatial displacement extrapolation calculations, and the target timestamps for the updated surface defect regions in the forming and drying stages are established.

5. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, Determining edge slope values ​​using an edge thickness gradient curve includes: The maximum vertical drop and horizontal width values ​​in the edge thickness gradient curve are extracted, and the ratio of the maximum vertical drop value to the horizontal width value is used as the edge slope value; the edge slope value is used to characterize the steepness of the molding shape of the gypsum core layer.

6. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, The thermal penetration delay parameters were established using surface temperature rise curves and internal adhesive curing values, including: Extract the starting time point when the surface temperature rise curve reaches the preset temperature benchmark on the time axis, and simultaneously extract the response time point when the internal adhesive curing degree value reaches the preset curing benchmark. The lag time difference is obtained by subtracting the start time from the response time point. The ratio of the lag time difference to the preset characteristic thermal response time constant of the face paper is used as the heat penetration delay parameter. The heat penetration delay parameter is used to characterize the influence of the time delay of hot air penetrating the face paper layer to reach the interior of the gypsum core layer on the heat conduction process.

7. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 1, characterized in that, The working condition mapping database is pre-divided into multiple distribution intervals corresponding to different working conditions; The edge slope value and heat penetration delay parameter are synchronously input into a pre-established working condition mapping database for spatial coordinate mapping to establish the working condition classification results, including: After determining the spatial coordinate mapping, monitor the spatial distribution of the mapped coordinate points in the working condition mapping database, determine the target distribution interval in which the spatial distribution position falls among multiple distribution intervals, and take the working condition corresponding to the target distribution interval as the working condition classification result. Among them, the monitoring mapping coordinate points are determined based on the edge slope value and heat penetration delay parameter of the gypsum board to be tested; different working conditions include airflow separation damage at the edge of the gypsum board and insufficient hot air penetration at the edge of the gypsum board.

8. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 7, characterized in that, When the working condition classification result indicates that the airflow separation at the edge of the gypsum board is disrupted, the target equipment control parameters are matched using the working condition classification result in a pre-established equipment control mapping table, and control commands for the molding guide equipment or drying hot air equipment are output, including: The difference between the edge slope value and the preset slope threshold is taken as the slope overshoot. The mechanical deflection angle that matches the preset calibration database with the slope overshoot is taken as the target equipment control parameter. The calibration database contains the mapping relationship between the deflection angle of the forming guide device and the edge slope value, as well as the size constraint condition of the total edge thickness that has the function of preventing the folding interference of the gypsum board facing paper. Output a first control command to drive the molding guide device to deflect the mechanical deflection angle outward, and output a pressure lock command to the glue pump device to prevent the increase of edge glue injection.

9. The method for tracing and controlling production parameters for gypsum board surface defects according to claim 7, characterized in that, If the working condition classification result indicates insufficient hot air penetration at the edge of the gypsum board, the target equipment control parameters are matched using the working condition classification result in a pre-established equipment control mapping table, and control commands are output for the forming guide equipment or drying hot air equipment, including: The difference between the heat penetration delay parameter and the preset thermal resistance threshold is taken as the thermal resistance excess, and the mechanical opening adjustment value corresponding to the pre-established equipment control mapping table and the thermal resistance excess is taken as the target equipment control parameter. The system outputs a second control command to increase the mechanical opening of the lateral drying hot air equipment in the corresponding drying temperature zone, so that the increased mechanical opening value is equal to the mechanical opening adjustment value, and outputs a pressure lock command to the glue pump equipment to prevent the increase of edge glue injection.

10. A production parameter traceability and control system for gypsum board surface defects, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions, which, when executed by the processor, implement the method for tracing and controlling production parameters of gypsum board surface defects according to any one of claims 1-9.

Images