A method for initial inspection and parameter verification of glass surface grinding quality

Abstract

The application provides a glass surface grinding quality preliminary detection and parameter verification method, adopts high-purity deionized water + superfine fiber dust-free cloth to wipe the surface, cooperates temperature and humidity and anti-static measures, eliminates the interference of pollutants, observes the residual area through S-shaped coating of a carbon black ink pen and reflection of light, combines laser scanning to obtain the Ra value and image recognition residual rate, adjusts the grinding pressure, avoidance distance and local grinding frequency based on the detection result, calibrates the Z-axis precision and pressure sensor of the equipment, continuously produces 100 samples, verifies the stability through a control chart, solidifies mature parameters to a PLC system, constructs an MES process database, mines the correlation between the grinding parameters and quality indexes based on regression analysis, realizes parameter self-adaptive iterative optimization, improves the grinding quality consistency, reduces the trial and error cost, and meets the batch production demand of high-precision glass products.

Description

Technical Field

[0001] This invention relates to the field of intelligent manufacturing technology, and in particular to a method for initial inspection of glass surface grinding quality and parameter verification. Background Technology

[0002] In the field of precision glass grinding, the initial inspection of surface quality and the verification of process parameters directly affect the optical performance and yield of products. The traditional ink stain method relies on manual visual judgment of residues. The human eye has a high discrimination error in low-contrast ink stains and cannot quantify the evaluation. Parameter adjustment lacks real-time data support. For example, the correlation between grinding pressure and surface roughness can only be adjusted after offline testing, which leads to an increase in batch defect rate. Vibration monitoring is mostly used for fault early warning, but it is not dynamically correlated with the wear state of the grinding disc, resulting in 20% over-maintenance or delayed replacement. Process optimization relies on the experience of engineers, and the historical parameter database has not established a multivariate regression model, which makes the parameter adaptation cycle long.

[0003] To address the aforementioned issues, existing technologies lack a closed-loop data system for the entire process, and parameter optimization relies on small-sample trial and error, making it difficult to meet the large-scale production demands of high-precision optical glass. In contrast, this method introduces an ink coating method combined with 3D laser scanning, precise control of environmental parameters, dynamic vibration monitoring, and regression analysis using an MES database. This enables quantitative control of the entire grinding quality process, from pretreatment to automatic parameter calibration, solving the technical bottlenecks of detection lag, blind parameter adjustment, and uncontrollable environmental interference in traditional processes, thereby improving process stability and production efficiency. Summary of the Invention

[0004] The purpose of this invention is to provide a method for initial inspection of glass surface grinding quality and parameter verification.

[0005] The problem to be solved by this invention is to address the issues of low quality inspection efficiency, delayed parameter adjustment, and insufficient process stability during the glass surface grinding process.

[0006] A method for initial inspection and parameter verification of glass surface grinding quality, the technical solution of which is as follows: S1: Perform surface pretreatment and environmental parameter control. Wipe the glass surface with deionized water and a lint-free cloth to remove contaminants including oil and fingerprints. Use infrared heating to raise the surface temperature to ≤30℃. Maintain the workshop temperature at 20~25℃ and humidity at 45%~55%. Lay anti-static mats at the grinding station and wear anti-static gloves. S2: Apply the carbon black pigment-containing ink pen evenly to the glass surface in an S-shaped trajectory, covering an area of ​​≥80%. Allow it to air dry for 30 seconds. Observe the ink stain under reflected light and judge its quality. If the ink stain is completely invisible, it is considered qualified. If the residual ink stain area is >5%, the parameters need to be adjusted. S3: Use a portable laser roughness meter to scan the 3D morphology of the glass surface and obtain the Ra value. Set the upper limit of the Ra value to 0.1μm. If it exceeds the limit, the parameter adjustment will be triggered. Take pictures of the ink residue area with a camera and identify the ink pixel ratio. Residual rate = (number of ink pixels / total number of pixels) × 100%. If it is >3%, it is determined that optimization is needed. S4: If either Ra value > 0.1μm or residual rate > 3% is met, adjust the parameters, increase the grinding pressure by 5%-10%, and reduce the clearance distance by 0.1mm. If the ink residue is concentrated in a local area, increase the number of grinding cycles in that area by 20%. Before adjusting the parameters, calibrate the equipment by using a standard glass plate to calibrate the Z-axis position and pressure sensor of the grinding equipment. During the grinding process, monitor the wear status of the grinding disc in real time through a vibration sensor. If the vibration frequency is abnormal, trigger a pause and prompt to replace the grinding disc. S5: If the residual ink area is <1% and the Ra value is ≤0.1μm, the parameter verification is deemed successful. Produce 100 glass samples continuously, record the Ra value and residual rate each time, and draw a control chart. When the control chart shows that the data for 20 consecutive times are within the Ra value fluctuation <0.02μm control limit, the current parameter is solidified as the standard process parameter. S6: Upload the Ra value, residual rate, parameter settings, and wear status data of each test to the MES system to form a process database. If the Ra value of a certain batch suddenly increases by 0.05μm, the system will automatically push an alert to the process engineer. Based on historical data, regression analysis will be used to determine the optimal parameter combination and continuously optimize the process parameters.

[0007] Furthermore, in step S1, deionized water and a lint-free cloth are used to wipe the glass surface, an anti-static mat is laid at the grinding station, and anti-static gloves are worn, including: Deionized water with a purity ≥18MΩ·cm and a resistivity that has passed testing is used as the cleaning solution. The glass surface is wiped with a lint-free cloth made of microfiber with a weight ≥600g / m² and a pile length ≤0.1mm. The wiping direction is a zigzag pattern to ensure that more than 95% of the glass surface is covered. Avoid cross-wiping and repeated wiping to prevent particle residue. After wiping, the surface is immediately blown with clean compressed air to remove residual water droplets. If there are oily contaminants on the surface, use a cleaner containing fluorocarbon surfactants to wipe the area, then rinse with deionized water and dry. For fingerprint residue, wipe away the residue with a lint-free cloth. The surface resistivity of the grinding station is 10. 6 ~10 8 Antistatic floor mats with a resistance of Ω and a thickness of ≥2mm are used. The edges of the mats are connected to the workshop ground wire via grounding copper busbars. Workers wear antistatic gloves. The inner layer is made of conductive fiber woven fabric, and the outer layer is made of polyester fiber. The surface resistance is ≤10 Ω. 7Ω, the glove wrist needs to be in contact with a resistance of 10. 6 The Ω anti-static wrist strap is connected, and the wrist strap is grounded through a wire. The gloves must be tested for static electricity before use and should be replaced once per shift.

[0008] Furthermore, in step S2, the ink pen containing carbon black pigment is uniformly coated onto the glass surface in an S-shaped trajectory, and the ink stain is observed and determined under reflected light, including: Apply coating to the glass surface in an S-shaped trajectory, with a single trajectory width of 10-15mm and a spacing of ≤5mm between adjacent trajectories, ensuring a coverage area of ​​≥80%. During coating, maintain a constant contact pressure between the pen tip and the glass surface. The coating area should avoid the edge area after infrared heating as described in S1, and be at least 20mm away from the edge to prevent thermal stress from causing the ink to crack. After coating, place the glass under the controlled environmental parameters of 20~25℃ and 45%~55% as described in S1 and allow it to air dry naturally for 30 seconds. A reflective light detection platform is set up in the area of ​​the anti-static floor mat laid as described in S1. An LED cold light source with a color temperature of 5500K and an illuminance of ≥1000 lux is used, and the light source is irradiated at a 45° angle to the glass surface. The grid method was used to estimate the residual area. The glass surface was divided into 100 10mm×10mm grids. The number of grids covered by the residual ink was counted and the percentage was calculated. If the ink was completely invisible under reflected light, that is, there was no color difference between the ink and the glass background, it was determined that the surface roughness met the requirements of subsequent processes. If the residual ink area was observed to be >5%, it was determined that the parameters needed to be adjusted. For areas with more than 5% residual ink, a temporary mark was made on the glass surface using a red marker pen.

[0009] Furthermore, in step S3, scanning the 3D morphology of the glass surface to obtain the Ra value, and capturing images of the ink residue area with a camera to identify the ink pixel ratio, includes: The laser scanning head moves along the glass surface in two dimensions along the XY axis with a step accuracy of 0.1 mm, generating 3D point cloud data of the surface with a resolution of ≤0.01 μm. At least 3 areas are scanned for each piece of glass, avoiding the edge within 20 mm. The average Ra value is taken as the final result. If the Ra value is ≤0.1 μm, it is considered qualified. If the Ra value is >0.1 μm, parameter adjustment is triggered. The camera lens is positioned at a 45° angle to the glass surface to ensure that the contrast between the ink and the background is greater than 2:1. Grayscale processing is performed based on OpenCV, and the grayscale value of the ink pixel is set to be less than 128. The residual ink area is segmented, and the residual rate is calculated as (number of ink pixels / total number of pixels) × 100%. If it is greater than 3%, it is determined that optimization is needed.

[0010] Furthermore, in step S4, parameter adjustment is performed, and equipment calibration is conducted before parameter adjustment, including: Increase the pressure of the grinding equipment by 5%-10%, preferably by 5% to avoid over-grinding and surface damage. Reduce the clearance between the grinding disc and the glass surface by 0.1mm to ensure that the grinding disc does not contact the glass surface to prevent scratches. For local areas with concentrated residual ink, increase the number of grinding cycles by 20% and locate the specific location using an area scanner. Using a standard glass plate with a known surface roughness and Ra value ≤0.05μm after cleaning in S1 as the calibration reference, fix the standard glass plate on the worktable of the grinding equipment, start the calibration program, adjust the Z-axis height to make the grinding disk parallel to the glass surface with an error ≤0.01mm, calibrate the pressure sensor, and ensure that the applied pressure value deviates from the set value by <1%; A vibration sensor with a sampling frequency of 1kHz is installed at the bottom of the grinding disc to monitor the vibration frequency of the grinding disc in real time. The normal vibration frequency range is set to 100-200Hz. If an abnormal frequency >300Hz or <80Hz is detected, the system will pause and prompt the user to replace the grinding disc. If the average vibration frequency of a batch of grinding discs suddenly increases by 10% compared with the historical average, the system will pause and prompt the user to replace the grinding disc.

[0011] Furthermore, in step S5, 100 glass samples are continuously produced, the Ra value and residual rate are recorded each time, a control chart is plotted, and the current parameters are solidified into standard process parameters, including: Start the automated production line and process 100 glass samples continuously. Ensure that each sample goes through the entire process from S1 to S4. After processing 10 samples, retest the Ra value and residual rate. If abnormal data is found, immediately trigger the parameter adjustment process. All data is uploaded to the MES system in real time and compared with historical data for analysis. Plot a control chart with Ra value on the x-axis and sample number on the y-axis. The upper control limit UCL = average Ra value + 0.02 μm and the lower control limit LCL = average Ra value - 0.02 μm. If the data fluctuates between UCL and LCL for 20 consecutive times, the process stability is considered to be up to standard. If a single point exceeds the control limit or 7 consecutive points deviate to one side, a parameter adjustment warning is triggered. When the control chart shows that the Ra value fluctuation is within the range of <0.02μm for 20 consecutive data points and the residual rate is stable at <1%, it is determined that the process parameters are mature and can be solidified as standard parameters. The current grinding pressure, avoidance distance, and number of local grinding times are written into the equipment PLC as the default process settings. A standardized work instruction is generated in the MES system, which includes parameter settings, equipment calibration steps, and contingency plans for handling abnormalities.

[0012] Furthermore, in step S6, a process database is formed. Based on historical data, regression analysis is used to determine the optimal parameter combination, and continuous optimization of process parameters is performed, including: The database structure is designed to include a main table and sub-tables. The main table stores the core data of each batch of glass samples, including Ra value, residual rate, and grinding parameters. The sub-tables record environmental parameters such as temperature and humidity, vibration frequency, equipment status, and historical parameter adjustment records. Indexes are created by batch number, production date, and equipment number to support fast queries. Ra value, residual rate, grinding pressure, and avoidance distance variables were extracted from the process database. Irrelevant fields such as batch number were excluded, outliers were removed, and the numerical range was standardized. A linear relationship between Ra value and grinding pressure and avoidance distance was established. A polynomial fitting analysis was used to analyze the nonlinear relationship between residual rate and ink coating coverage. The recommended parameters are output based on the regression model, and an optimization report is generated, which includes the prediction effect and risk assessment. The recommended parameters are then imported into the adjustment process described in S4 and verified in continuous production as described in S5. The Ra value fluctuation of 100 samples is <0.02μm, and finally solidified into standard process parameters.

[0013] The beneficial effects of this invention are: it enables rapid qualitative and quantitative determination of grinding quality, triggers parameter adjustment through real-time data feedback to form closed-loop control, significantly shortens the process debugging cycle, ensures the accuracy of grinding parameters and the stability of equipment status, and reduces batch quality fluctuations caused by equipment drift or wear. The stability of the process is verified by statistical process control. When the data meets the standard for 20 consecutive times, the parameters are solidified to avoid human experience bias and ensure the consistency of mass production. The sensitivity of parameters is predicted by combining historical data analysis to avoid potential quality problems in advance. Recommended parameters are generated and imported into the production process to achieve iterative optimization of the process. Attached Figure Description

[0014] Figure 1 This is a flowchart of a method for initial inspection and parameter verification of glass surface grinding quality. Detailed Implementation

[0015] The present invention will be further described clearly and completely below, but the scope of protection of the present invention is not limited thereto.

[0016] Example 1

[0017] A method for initial inspection and parameter verification of glass surface grinding quality, the technical solution of which is as follows: S1: Perform surface pretreatment and environmental parameter control. Wipe the glass surface with deionized water and a lint-free cloth to remove contaminants including oil and fingerprints. Use infrared heating to raise the surface temperature to ≤30℃. Maintain the workshop temperature at 20-25℃ and humidity at 45%-55%. Lay anti-static mats at the grinding station and wear anti-static gloves. S2: Apply the carbon black pigment-containing ink pen evenly to the glass surface in an S-shaped trajectory, covering an area of ​​≥80%. Allow it to air dry for 30 seconds. Observe the ink stain under reflected light and judge its quality. If the ink stain is completely invisible, it is considered qualified. If the residual ink stain area is >5%, the parameters need to be adjusted. S3: Use a portable laser roughness meter to scan the 3D morphology of the glass surface and obtain the Ra value. Set the upper limit of the Ra value to 0.1μm. If it exceeds the limit, the parameter adjustment will be triggered. Take pictures of the ink residue area with a camera and identify the ink pixel ratio. Residual rate = (number of ink pixels / total number of pixels) × 100%. If it is >3%, it is determined that optimization is needed. S4: If either Ra value > 0.1μm or residual rate > 3% is met, adjust the parameters, increase the grinding pressure by 5%-10%, and reduce the clearance distance by 0.1mm. If the ink residue is concentrated in a local area, increase the number of grinding cycles in that area by 20%. Before adjusting the parameters, calibrate the equipment by using a standard glass plate to calibrate the Z-axis position and pressure sensor of the grinding equipment. During the grinding process, monitor the wear status of the grinding disc in real time through a vibration sensor. If the vibration frequency is abnormal, trigger a pause and prompt to replace the grinding disc. S5: If the residual ink area is <1% and the Ra value is ≤0.1μm, the parameter verification is deemed successful. Produce 100 glass samples continuously, record the Ra value and residual rate each time, and draw a control chart. When the control chart shows that the data for 20 consecutive times are within the Ra value fluctuation <0.02μm control limit, the current parameter is solidified as the standard process parameter. S6: Upload the Ra value, residual rate, parameter settings, and wear status data of each test to the MES system to form a process database. If the Ra value of a certain batch suddenly increases by 0.05μm, the system will automatically push an alert to the process engineer. Based on historical data, regression analysis will be used to determine the optimal parameter combination and continuously optimize the process parameters.

[0018] refer to Figure 1 The diagram shown is a flowchart of a method for initial inspection and parameter verification of glass surface grinding quality.

[0019] Furthermore, in step S1, deionized water and a lint-free cloth are used to wipe the glass surface, an anti-static mat is laid at the grinding station, and anti-static gloves are worn, including: Deionized water with a purity ≥18MΩ·cm and a resistivity that has passed testing is used as the cleaning solution. The glass surface is wiped with a lint-free cloth made of microfiber with a weight ≥600g / m² and a pile length ≤0.1mm. The wiping direction is a zigzag pattern to ensure that more than 95% of the glass surface is covered. Avoid cross-wiping and repeated wiping to prevent particle residue. After wiping, the surface is immediately blown with clean compressed air to remove residual water droplets. If there are oily contaminants on the surface, use a cleaner containing fluorocarbon surfactants to wipe the area, then rinse with deionized water and dry. For fingerprint residue, wipe away the residue with a lint-free cloth. The surface resistivity of the grinding station is 10. 6 ~10 8 Antistatic floor mats with a resistance of Ω and a thickness of ≥2mm are used. The edges of the mats are connected to the workshop ground wire via grounding copper busbars. Workers wear antistatic gloves. The inner layer is made of conductive fiber woven fabric, and the outer layer is made of polyester fiber. The surface resistance is ≤10 Ω. 7 Ω, the glove wrist needs to be in contact with a resistance of 10. 6 The Ω anti-static wrist strap is connected, and the wrist strap is grounded through a wire. The gloves must be tested for static electricity before use and should be replaced once per shift.

[0020] Furthermore, in step S2, the ink pen containing carbon black pigment is uniformly coated onto the glass surface in an S-shaped trajectory, and the ink stain is observed and determined under reflected light, including: Apply coating to the glass surface in an S-shaped trajectory, with a single trajectory width of 10-15mm and a spacing of ≤5mm between adjacent trajectories, ensuring a coverage area of ​​≥80%. During coating, maintain a constant contact pressure between the pen tip and the glass surface. The coating area should avoid the edge area after infrared heating as described in S1, and be at least 20mm away from the edge to prevent thermal stress from causing the ink to crack. After coating, place the glass under the controlled environmental parameters of 20~25℃ and 45%~55% as described in S1 and allow it to air dry naturally for 30 seconds. A reflective light detection platform is set up in the area of ​​the anti-static floor mat laid as described in S1. An LED cold light source with a color temperature of 5500K and an illuminance of ≥1000 lux is used, and the light source is irradiated at a 45° angle to the glass surface. The grid method was used to estimate the residual area. The glass surface was divided into 100 10mm×10mm grids. The number of grids covered by the residual ink was counted and the percentage was calculated. If the ink was completely invisible under reflected light, that is, there was no color difference between the ink and the glass background, it was determined that the surface roughness met the requirements of subsequent processes. If the residual ink area was observed to be >5%, it was determined that the parameters needed to be adjusted. For areas with more than 5% residual ink, a temporary mark was made on the glass surface using a red marker pen.

[0021] Furthermore, in step S3, scanning the 3D morphology of the glass surface to obtain the Ra value, and capturing images of the ink residue area with a camera to identify the ink pixel ratio, includes: The laser scanning head moves along the glass surface in two dimensions along the XY axis with a step accuracy of 0.1 mm, generating 3D point cloud data of the surface with a resolution of ≤0.01 μm. At least 3 areas are scanned for each piece of glass, avoiding the edge within 20 mm. The average Ra value is taken as the final result. If the Ra value is ≤0.1 μm, it is considered qualified. If the Ra value is >0.1 μm, parameter adjustment is triggered. The camera lens is positioned at a 45° angle to the glass surface to ensure that the contrast between the ink and the background is greater than 2:1. Grayscale processing is performed based on OpenCV, and the grayscale value of the ink pixel is set to be less than 128. The residual ink area is segmented, and the residual rate is calculated as (number of ink pixels / total number of pixels) × 100%. If it is greater than 3%, it is determined that optimization is needed.

[0022] Furthermore, in step S4, parameter adjustment is performed, and equipment calibration is conducted before parameter adjustment, including: Increase the pressure of the grinding equipment by 5%-10%, preferably by 5% to avoid over-grinding and surface damage. Reduce the clearance between the grinding disc and the glass surface by 0.1mm to ensure that the grinding disc does not contact the glass surface to prevent scratches. For local areas with concentrated residual ink, increase the number of grinding cycles by 20%. For example, if the original plan was to grind 10 times, adjust it to 12 times. Use an area scanner to locate the specific location. Using a standard glass plate with a known surface roughness and Ra value ≤ 0.05 μm after cleaning in S1 as the calibration reference, fix the standard glass plate on the worktable of the grinding equipment, start the calibration program, adjust the Z-axis height to make the grinding disk parallel to the glass surface with an error ≤ 0.01 mm, calibrate the pressure sensor, and ensure that the applied pressure value deviates from the set value by < 1%. For example, if the set pressure is 0.5 N, the actual output should be 0.495~0.505 N. A vibration sensor with a sampling frequency of 1kHz is installed at the bottom of the grinding disc to monitor the vibration frequency of the grinding disc in real time. The normal vibration frequency range is set to 100-200Hz. If an abnormal frequency >300Hz or <80Hz is detected, the system will pause and prompt the user to replace the grinding disc. If the average vibration frequency of a batch of grinding discs suddenly increases by 10% compared with the historical average, the system will pause and prompt the user to replace the grinding disc.

[0023] Furthermore, in step S5, 100 glass samples are continuously produced, the Ra value and residual rate are recorded each time, a control chart is plotted, and the current parameters are solidified into standard process parameters, including: Start the automated production line and process 100 glass samples continuously. Ensure that each sample goes through the entire process from S1 to S4. After processing 10 samples, retest the Ra value and residual rate. If abnormal data is found, immediately trigger the parameter adjustment process. All data is uploaded to the MES system in real time and compared with historical data for analysis. Plot a control chart with Ra value on the x-axis and sample number on the y-axis. The upper control limit UCL = average Ra value + 0.02 μm and the lower control limit LCL = average Ra value - 0.02 μm. If the data fluctuates between UCL and LCL for 20 consecutive times, the process stability is considered to be up to standard. If a single point exceeds the control limit or 7 consecutive points deviate to one side, a parameter adjustment warning is triggered. When the control chart shows that the Ra value fluctuation is within the range of <0.02μm for 20 consecutive data points and the residual rate is stable at <1%, it is determined that the process parameters are mature and can be solidified as standard parameters. The current grinding pressure, avoidance distance, and number of local grinding times are written into the equipment PLC as the default process settings. A standardized work instruction is generated in the MES system, which includes parameter settings, equipment calibration steps, and contingency plans for handling abnormalities.

[0024] Furthermore, in step S6, a process database is formed. Based on historical data, regression analysis is used to determine the optimal parameter combination, and continuous optimization of process parameters is performed, including: The database structure is designed to include a main table and sub-tables. The main table stores the core data of each batch of glass samples, including Ra value, residual rate, and grinding parameters. The sub-tables record environmental parameters such as temperature and humidity, vibration frequency, equipment status, and historical parameter adjustment records. Indexes are created by batch number, production date, and equipment number to support fast queries, such as Ra value fluctuation analysis for batch 20250730-001. Ra value, residual rate, grinding pressure, and avoidance distance variables were extracted from the process database. Irrelevant fields such as batch number were excluded, outliers were removed, and the numerical range was standardized. A linear relationship between Ra value and grinding pressure and avoidance distance was established: when the grinding pressure increased by 5%, the Ra value decreased by 0.015 μm; when the avoidance distance decreased by 0.1 mm, the Ra value decreased by 0.01 μm. Polynomial fitting analysis was used to analyze the nonlinear relationship between residual rate and ink coating coverage: when the coverage increased from 80% to 90%, the residual rate decreased from 3% to 1.5%. Based on the regression model, recommended parameters are output: grinding pressure = 0.6N, avoidance distance = 0.2mm. An optimization report is generated, which includes the predicted effect and risk assessment. The recommended parameters are then imported into the adjustment process described in S4 and verified in continuous production as described in S5. The Ra value fluctuation of 100 samples is <0.02μm, and finally solidified into standard process parameters.

[0025] Example 2 The product is aluminosilicate glass with dimensions of 300mm × 200mm × 0.7mm. A lint-free cloth soaked in deionized water with a resistivity of 18.2 MΩ·cm is used to wipe the surface in a zigzag pattern, achieving 98% coverage. Residual water stains are then removed by compressed air. The surface is then heated to 28℃ using infrared heating, with an ambient temperature of 22℃ and humidity of 50%. The operator wears anti-static gloves. The surface resistance is 8 × 10⁻⁶. 6 Ω, standing on an anti-static mat, grounding resistance 10 Ω. 7 Ω.

[0026] Apply ink in an S-shaped pattern using a charcoal black ink pen. The pattern should be 12mm wide and 4mm apart, covering 85% of the area. Avoid the 25mm edge area. After drying for 30 seconds, observe under a 5500K light source. You will find residual ink in the upper right corner, covering 6.2% of the area. Mark this area in red.

[0027] Laser roughness meter scanned 3 areas: Area 1 (top left): Ra=0.13μm; Area 2 (center): Ra=0.09μm; Area 3 (top right): Ra=0.15μm; Average Ra=0.12μm (>0.1μm threshold); Camera captured the residual area in the top right corner: Image resolution: 2480×1748 pixels; Ink smudge pixels: 42,150 (total pixels 4,335,040); Residual rate = 0.97% (<3% threshold).

[0028] Grinding pressure: +8% (320kPa→345kPa), clearance distance: -0.1mm (1.0mm→0.9mm), number of grinding cycles in the upper right corner: +20% (3 times→4 times), Z-axis calibration using a standard glass plate (Ra=0.04μm) (parallelism error 0.008mm), pressure sensor deviation: 0.5% (<1% threshold), vibration frequency during grinding: 175Hz (normal range 100-200Hz), no abnormal alarms.

[0029] Data from 100 consecutively produced samples were recorded, including sample intervals, mean Ra (μm), residual rate (%), and control chart status. For the first 18 consecutive points (1-20, 0.092, 0.85), the values ​​were within control. For the next 40 points (21-40, 0.095, 0.91), a single point exceeded the UCL (triggering a retest). For the next 60 points (41-100, 0.088±0.01, 0.72±0.15), the values ​​were within control. Control chart parameters: UCL=0.098μm, LCL=0.078μm. The final data fluctuation for the 60 consecutive points was <0.02μm.

[0030] Database regression analysis (historical data from 200 batches), variables, regression coefficients, and p-values ​​are as follows: grinding pressure (kPa), -0.12, 0.003; avoidance distance (mm), 0.25, 0.001; ambient humidity (%), 0.08, 0.042; optimization recommendation: when humidity > 60%, the avoidance distance needs to be additionally reduced by -0.05mm (after verification, Ra fluctuation decreased by 40%).

[0031] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0032] This invention provides a method for initial inspection and parameter verification of glass surface grinding quality. The method involves wiping the surface with high-purity deionized water and a microfiber lint-free cloth, along with temperature and humidity control and anti-static measures to eliminate contaminant interference. The residual area is observed by applying an S-shaped coating with a carbon black ink pen and reflecting light. Laser scanning is used to obtain the Ra value and image recognition of the residual rate. Based on the detection results, the grinding pressure, avoidance distance, and number of local grinding cycles are adjusted. The Z-axis accuracy and pressure sensor of the equipment are calibrated. One hundred samples are continuously produced, and stability is verified using control charts. Mature parameters are then integrated into the PLC system, and a MES process database is constructed. Regression analysis is used to explore the correlation between grinding parameters and quality indicators, enabling adaptive iterative optimization of parameters, improving grinding quality consistency, reducing trial-and-error costs, and meeting the mass production needs of high-precision glass products.

Claims

1. A method for initial inspection and parameter verification of glass surface grinding quality, characterized in that, include: S1: Perform surface pretreatment and environmental parameter control. Wipe the glass surface with deionized water and a lint-free cloth to remove contaminants including oil and fingerprints. Use infrared heating to raise the surface temperature to ≤30℃. Maintain the workshop temperature at 20-25℃ and humidity at 45%-55%. Lay anti-static mats at the grinding station and wear anti-static gloves. S2: Apply the carbon black pigment-containing ink pen evenly to the glass surface in an S-shaped trajectory, covering an area of ​​≥80%. Allow it to air dry for 30 seconds. Observe the ink stain under reflected light and judge its quality. If the ink stain is completely invisible, it is considered qualified. If the residual ink stain area is >5%, the parameters need to be adjusted. S3: Use a portable laser roughness meter to scan the 3D morphology of the glass surface and obtain the Ra value. Set the upper limit of the Ra value to 0.1μm. If it exceeds the limit, the parameter adjustment will be triggered. Take pictures of the ink residue area with a camera and identify the ink pixel ratio. Residual rate = (number of ink pixels / total number of pixels) × 100%. If it is >3%, it is determined that optimization is needed. S4: If either Ra value > 0.1μm or residual rate > 3% is met, adjust the parameters, increase the grinding pressure by 5%-10%, and reduce the clearance distance by 0.1mm. If the ink residue is concentrated in a local area, increase the number of grinding cycles in that area by 20%. Before adjusting the parameters, calibrate the equipment by using a standard glass plate to calibrate the Z-axis position and pressure sensor of the grinding equipment. During the grinding process, monitor the wear status of the grinding disc in real time through a vibration sensor. If the vibration frequency is abnormal, trigger a pause and prompt to replace the grinding disc. S5: If the residual ink area is <1% and the Ra value is ≤0.1μm, the parameter verification is deemed successful. Produce 100 glass samples continuously, record the Ra value and residual rate each time, and draw a control chart. When the control chart shows that the data for 20 consecutive times are within the Ra value fluctuation <0.02μm control limit, the current parameter is solidified as the standard process parameter. S6: Upload the Ra value, residual rate, parameter settings, and wear status data of each test to the MES system to form a process database. If the Ra value of a certain batch suddenly increases by 0.05μm, the system will automatically push an alert to the process engineer. Based on historical data, regression analysis will be used to determine the optimal parameter combination and continuously optimize the process parameters.

2. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S1, the glass surface is wiped with deionized water and a lint-free cloth. An anti-static mat is laid at the grinding station, and anti-static gloves are worn. Deionized water with a purity of ≥18MΩ·cm and a resistivity that has passed the test is used as the cleaning solution. The glass surface is wiped with a lint-free cloth made of microfiber with a weight of ≥600g / m² and a pile length of ≤0.1mm. The wiping direction is a zigzag pattern. After wiping, the surface is immediately blown with clean compressed air to remove residual water droplets. If there are oily contaminants on the surface, use a cleaner containing fluorocarbon surfactants to wipe the area, then rinse with deionized water and dry. For fingerprint residue, wipe away the residue with a lint-free cloth. The surface resistivity of the grinding station is 10. 6 ~10 8 Antistatic floor mats with a resistance of Ω and a thickness of ≥2mm are used. The edges of the mats are connected to the workshop ground wire via grounding copper busbars. Workers wear antistatic gloves. The inner layer is made of conductive fiber woven fabric, and the outer layer is made of polyester fiber. The surface resistance is ≤10 Ω. 7 Ω, the glove wrist needs to be in contact with a resistance of 10. 6 The Ω anti-static wrist strap is connected, and the wrist strap is grounded through a wire. The gloves must be tested for static electricity before use and should be replaced once per shift.

3. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S2, an ink pen containing carbon black pigment is uniformly coated onto the glass surface in an S-shaped trajectory. The ink stain is then observed and judged under reflected light, including: Apply coating to the glass surface in an S-shaped trajectory, with a single trajectory width of 10-15mm and a spacing of ≤5mm between adjacent trajectories. Avoid the edge area after infrared heating as described in S1, and keep a distance of ≥20mm from the edge. After coating, place the glass under the controlled environmental parameters of 20-25℃ and 45%-55% as described in S1 and allow it to air dry naturally for 30 seconds. A reflective light detection platform is set up in the area of ​​the anti-static floor mat laid as described in S1. An LED cold light source with a color temperature of 5500K and an illuminance of ≥1000 lux is used, and the light source is irradiated at a 45° angle to the glass surface. The grid method was used to estimate the residual area. The glass surface was divided into 100 10mm×10mm grids. The number of grids covered by the residual ink was counted and the percentage was calculated. If the ink was completely invisible under reflected light, the surface roughness was determined to meet the requirements of subsequent processes. If the residual ink area was observed to be >5%, the parameters were determined to need to be adjusted. For areas with more than 5% residual ink, a temporary mark was made on the glass surface using a red marker.

4. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S3, the 3D morphology of the glass surface is scanned to obtain the Ra value, and the area of ​​ink residue is photographed by a camera to identify the percentage of ink pixels, including: The laser scanning head moves along the glass surface in two dimensions along the XY axis with a step accuracy of 0.1 mm, generating 3D point cloud data of the surface with a resolution of ≤0.01 μm. At least 3 areas are scanned for each piece of glass, avoiding the edge within 20 mm. The average Ra value is taken as the final result. If the Ra value is ≤0.1 μm, it is considered qualified. If the Ra value is >0.1 μm, parameter adjustment is triggered. The camera lens illuminates the glass surface at a 45° angle, and the contrast between the ink and the background is greater than 2:

1. Grayscale processing is performed based on OpenCV, and the grayscale value of the ink pixel is set to be less than 128. The residual ink area is segmented, and the residual rate is calculated as (number of ink pixels / total number of pixels) × 100%. If it is greater than 3%, it is determined that optimization is needed.

5. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S4, parameter adjustment is performed. Before adjusting the parameters, equipment calibration is performed, including: Increase the pressure of the grinding equipment by 5%-10% and reduce the clearance between the grinding disc and the glass surface by 0.1mm. For local areas with concentrated residual ink, increase the number of grinding cycles by 20% and locate the specific location using an area scanner. Using a standard glass plate with a known surface roughness and Ra value ≤ 0.05 μm after cleaning in S1 as the calibration reference, fix the standard glass plate on the worktable of the grinding equipment, start the calibration program, adjust the Z-axis height to make the grinding disk parallel to the glass surface with an error ≤ 0.01 mm, calibrate the pressure sensor, and ensure that the applied pressure value deviates from the set value by < 1%; A vibration sensor with a sampling frequency of 1kHz is installed at the bottom of the grinding disc to monitor the vibration frequency of the grinding disc in real time. The normal vibration frequency range is set to 100-200Hz. If an abnormal frequency is detected, i.e. >300Hz and <80Hz, the process is paused and a prompt to replace the grinding disc is triggered. If the average vibration frequency of a batch of grinding discs suddenly increases by 10% compared with the historical average, the process is paused and a prompt to replace the grinding disc is triggered.

6. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S5, 100 glass samples are continuously produced, and the Ra value and residual rate are recorded each time. A control chart is plotted, and the current parameters are solidified as standard process parameters, including: Start the automated production line and continuously process 100 glass samples. Each sample goes through the full process described in S1-S4. After processing 10 samples, retest the Ra value and residual rate. If abnormal data is found, trigger the parameter adjustment process. All data is uploaded to the MES system in real time and compared with historical data for analysis. Plot a control chart with Ra value on the x-axis and sample number on the y-axis. The upper control limit UCL = average Ra value + 0.02 μm and the lower control limit LCL = average Ra value - 0.02 μm. If the data fluctuates between UCL and LCL for 20 consecutive times, the process stability is considered to be up to standard. If either a single point exceeds the control limit or 7 consecutive points deviate to one side, a parameter adjustment warning is triggered. When the control chart shows that the Ra value fluctuation is within the range of <0.02μm for 20 consecutive data points and the residual rate is stable at <1%, it is determined that the process parameters are mature and solidified as standard parameters. The current grinding pressure, avoidance distance, and number of local grinding times are written into the equipment PLC as the default process settings. A standardized work instruction is generated in the MES system, which includes parameter settings, equipment calibration steps, and contingency plans for handling abnormalities.

7. The method for initial inspection and parameter verification of glass surface grinding quality as described in claim 1, characterized in that, In step S6, a process database is formed. Based on historical data, regression analysis is used to determine the optimal parameter combination, and continuous optimization of process parameters is performed, including: The database structure is designed to include a main table and sub-tables. The main table stores the core data of each batch of glass samples, including Ra value, residual rate, and grinding parameters. The sub-tables record environmental parameters such as temperature and humidity, vibration frequency, equipment status, and historical parameter adjustment records. Indexes are created by batch number, production date, and equipment number to support fast queries. Ra value, residual rate, grinding pressure, and avoidance distance variables were extracted from the process database. Irrelevant fields such as batch number were excluded, outliers were removed, and the numerical range was standardized. A linear relationship between Ra value and grinding pressure and avoidance distance was established. A polynomial fitting analysis was used to analyze the nonlinear relationship between residual rate and ink coating coverage. The recommended parameters are output based on the regression model, and an optimization report is generated, which includes the prediction effect and risk assessment. The recommended parameters are then imported into the adjustment process described in S4 and verified in continuous production as described in S5. The Ra value fluctuation of 100 samples is <0.02μm, and finally solidified into standard process parameters.

Images