A flexible etching production method and system for a refrigerator glass panel

Abstract

The application provides a flexible etching production method of a refrigerator glass panel, and the method comprises the following steps: S1, film feeding and initial state evaluation; S2, first decision and parameter presetting based on a process mapping model; S3, film pasting and visual positioning detection under the film; S4, precise film repairing decision based on defect positioning; S5, collaborative etching and polishing based on preset parameters and real-time feedback; and S6, final processing. The technical problems of large resource waste, poor quality consistency, slow response to incoming material fluctuation and process abnormality, lack of preventive control ability and the like in the traditional fixed AG glass etching production process are solved.

Description

Technical Field

[0001] This invention relates to the field of glass deep processing technology, specifically to an etching (AG) production method and system for glass panels of household appliances such as refrigerators, and more particularly to a flexible etching production method and system based on predictive process mapping and real-time feedback optimization. Background Technology

[0002] The glass panels of home appliances such as refrigerators and ovens are often etched to form an anti-glare (AG) surface to improve aesthetics and tactile feel. Traditional AG glass etching production lines are usually fixed linear processes, in which all glass substrates go through processes such as loading, cleaning, applying protective film, etching, polishing, cleaning, and unloading in sequence. Process parameters (such as conveyor speed and chemical pressure) are set to fixed values ​​or only allow for small-scale manual adjustment.

[0003] This fixed approach has significant drawbacks: First, it ignores the differences in the condition of incoming glass sheets (such as cleanliness and surface energy). Unnecessary cleaning of clean glass leads to a waste of water, electricity, and chemicals; while using the same etching parameters for slightly contaminated glass as for clean glass can result in uneven etching, affecting the final gloss and surface roughness consistency. Second, defects such as bubbles and wrinkles generated during the lamination process are often only discovered during final inspection, causing the entire sheet of glass, along with the consumed expensive etching solution (such as HF), to be scrapped, or requiring complex overall film peeling and rework, which is costly and inefficient. Furthermore, quality drift caused by factors such as chemical consumption and environmental fluctuations during production can only be corrected after a batch of defective products has been produced, lacking preventative control measures.

[0004] While existing technologies have attempted to introduce inspection stations into production lines for simple sorting or alarm functions, these are mostly limited to a single-stage "inspection-rejection" logic, failing to form a data-driven, personalized process decision-making and collaborative optimization system covering the entire process. Information is isolated between processes, making it impossible to achieve intelligent production that "predicts consequences based on causes and pre-adjusts parameters."

[0005] Therefore, developing a flexible etching production method and system that can adapt to fluctuations in incoming materials, achieve personalized processing, and possess online learning and optimization capabilities is of great significance for improving product quality, reducing production costs, and enhancing the intelligence level of production lines. Summary of the Invention

[0006] The present invention aims to overcome the shortcomings of the prior art and solve the technical problems existing in the traditional fixed AG glass etching production process, such as large resource waste, poor quality consistency, slow response to fluctuations in incoming materials and process abnormalities, and lack of preventive control capabilities.

[0007] This invention provides a flexible etching production method for refrigerator glass panels, the method comprising the following steps:

[0008] S1: Initial State Assessment of Glass Sheet Loading and Quantization: Obtain the quantization feature vector of the glass sheet surface, wherein the quantization feature vector includes at least the cleanliness index CI and the surface energy approximation SE; the cleanliness index CI is obtained by calculating the variance of gray values ​​from the diffuse reflection light image of the glass surface; the surface energy approximation SE is indirectly calculated by a contact angle measuring instrument.

[0009] S2: First decision and parameter preset based on the process mapping model: The quantized feature vector (CI, SE) is input into a pre-stored process mapping relationship database for matching. The database defines the mapping relationship between the feature vector and the process parameter package. The process parameter package includes at least a pretreatment necessity flag, a recommended etching transfer speed V_e, and a recommended etching solution spray pressure P_e. Based on the pretreatment necessity flag in the matching result, the glass is controlled to skip or enter the pretreatment cleaning station, and the matched (V_e, P_e) parameter package is sent to the etching section controller for preset. The process mapping relationship database is constructed in the following way: Multiple etching process experiments are conducted on glass samples with different (CI, SE) combinations. The optimization objective is to ensure that the finished product gloss and roughness meet the requirements and have the smallest fluctuation. The corresponding optimal process parameter package is determined for each feature vector interval.

[0010] S3: Film application and visual positioning detection under the film: After the film is applied, a near-infrared light source that can penetrate the protective film and an imaging unit are used to acquire the film-glass interface image, identify and locate the defect area, and generate a defect positioning map; the wavelength of the near-infrared light source is 900-1100nm.

[0011] S4: Precise film repair decision based on defect location: According to the defect location map, control the glass to enter the main etching line, or control the online local film repair station to perform local film repair, or divert the glass to the film peeling and rework line; the control of the online local film repair station to perform local film repair specifically includes: generating the optimal path plan according to the contour coordinates in the defect location map, controlling the multi-axis actuator to drive the micro negative pressure suction nozzle to remove the protective film of the defect area along the path, and using a hot air knife with closed-loop control of temperature and wind speed for local re-application.

[0012] S5: Collaborative Etching and Polishing Based on Preset Parameters and Real-Time Feedback: The etching section controller controls the etching process according to a preset (V_e, P_e) parameter package; after polishing, the gloss of the glass is measured, and the measured value is input into the adaptive process model. The model dynamically fine-tunes the preset parameter package of the glass with similar quantized feature vectors (CI, SE) that has not yet entered the etching section; the operating logic of the adaptive process model includes: performing feedforward compensation fine-tuning of the preset etching conveying speed V_e of the glass to be processed with similar (CI, SE) vectors based on the gloss value measured at the intermediate quality pre-inspection station.

[0013] S6: Final processing.

[0014] The present invention also provides a flexible etching production system for implementing the above-described method, comprising: a wafer loading machine, an initial state detection station, a pre-treatment cleaning station, a film attaching machine, a film layer detection station, an online local film repair station, a film peeling and rework line, a main etching and polishing section, an intermediate quality pre-inspection station, a final cleaning machine, and a wafer unloading machine arranged sequentially.

[0015] The initial state detection station is configured to output the quantized feature vector (CI, SE) of the glass;

[0016] The process mapping database stores the mapping relationship between feature vectors and process parameter packages;

[0017] The membrane inspection station includes a near-infrared penetrating imaging unit for outputting defect location maps;

[0018] The online local film repair station includes a multi-axis precision film repair device that can perform actions according to path planning;

[0019] The central control unit is connected to the initial state detection station, the film layer detection station, the intermediate quality pre-inspection station, the conveyor belt diversion device of each station, the driver of the main etching and polishing section, and the database signal, respectively.

[0020] The central control unit is configured to: perform the process mapping matching; control branches and parameter presets according to the matching results; control precise film repair according to the defect location map; and run the adaptive process model to achieve coordinated optimization of etching parameters.

[0021] The beneficial effects of this invention are as follows:

[0022] 1. By quantifying the initial state and mapping it to a personalized process parameter package, cleaning requirements and etching intensity are "tailor-made" for each piece of glass, ensuring quality consistency from the source and avoiding resource waste caused by over-cleaning.

[0023] 2. Clean glass skips pretreatment, directly saving water, electricity and chemicals; under-film visual inspection can accurately identify film defects at an early stage, and online local film repair can avoid the entire glass from being taken off the production line for rework or scrapped in the etching process, which greatly saves expensive consumables such as HF acid.

[0024] 3. By using the “predictive mapping” and “feedforward-feedback composite optimization” models, the production line is equipped with the ability to adapt and optimize in response to raw material fluctuations and process drift, transforming “post-inspection” into “process prevention” and improving overall equipment efficiency (OEE).

[0025] 4. The quantitative detection, near-infrared imaging, multi-axis motion control, database and control model used are all mature industrial technologies. Through the specific combination and logical design of this invention, a complete, efficient and easy-to-upgrade solution for existing production lines is constructed. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments. It should be understood that the following embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0027] This embodiment 1 provides a flexible etching production method for refrigerator glass panels.

[0028] This embodiment details the implementation process of the method. The method is executed on a continuous production line consisting of a series of conveyor belts and mainly includes steps such as quantitative initial state assessment, intelligent process decision-making, film defect detection and repair, and adaptive etching and polishing.

[0029] S1: Initial state assessment of wafer loading and quantization.

[0030] The raw glass sheet is loaded onto a conveyor belt by a loading machine and moves along the conveying direction. When the glass reaches the initial state inspection station, the conveyor belt pauses or maintains a constant speed. At this station, a linear light source at a specific angle (e.g., 30° to 60°) to the glass surface illuminates the glass surface, and a high-resolution linear or area array camera acquires an image of the diffuse reflected light from the other side. The system calculates the variance of pixel grayscale values ​​in several pre-defined regions of interest (such as the center and four corner areas of the panel) in this image to obtain the cleanliness index CI. The smaller the variance value, the cleaner and more uniform the glass surface.

[0031] Simultaneously, the contact angle measuring instrument integrated within this workstation (which can employ the inclined plate method or indirect optical analysis) projects a standard test droplet (such as deionized water) onto the glass surface and quickly captures an image of the droplet's contour. By analyzing the base width and height of the droplet in the image, or by directly using a fitting algorithm, the approximate surface energy SE of the glass surface (typically in mN / m) is indirectly calculated. At this point, the system generates a unique quantified feature vector (CI, SE) for the current glass, used to characterize its initial physicochemical state.

[0032] S2: First decision and parameter preset based on process mapping model.

[0033] The central control unit (such as an industrial computer or a programmable logic controller, PLC) receives the (CI, SE) vector from the initial state detection station. The central control unit then performs a query and matching within its pre-stored process mapping database.

[0034] The database was constructed based on preliminary process experiment data. Specifically, during the production process debugging phase, a large number of glass samples with different (CI, SE) combinations were collected, and etching and polishing experiments were conducted under various combinations of etching transfer speed (V_e) and etchant spray pressure (P_e). The optimization objective was to simultaneously meet preset target ranges (e.g., gloss 11-13 GU, Ra value 0.8-1.2 μm) and minimize batch-to-batch variation in both the final product's gloss (e.g., measured using a 60° angle gloss meter) and surface roughness (Ra value). Through data regression analysis or empirical lookup table methods, an optimal set of process parameters was determined for each (CI, SE) value range. This parameter set contains at least three key pieces of information: a pretreatment necessity indicator (Boolean value, indicating whether cleaning is required), a recommended etching transfer speed V_e, and a recommended etchant spray pressure P_e.

[0035] The matching process refers to the central control unit comparing the current glass (CI, SE) value with the boundary values ​​of each feature vector interval defined in the database to determine which specific interval it falls into, and then calling the process parameter package corresponding to that interval.

[0036] Based on the matching results, the central control unit executes the first-level decision: if the "pretreatment necessity flag" is "false", the diversion device is controlled to direct the glass directly to the subsequent lamination station; if it is "true", the diversion device is controlled to first direct it to the pretreatment cleaning station (e.g., through a cleaning unit containing an ultrasonic cleaning tank and deionized water spray). Simultaneously, the central control unit sends the matched (V_e, P_e) parameter packet to the etching section controller in the middle of the production line as the initial process parameters for that piece of glass.

[0037] S3: Film application and visual positioning detection under the film.

[0038] The cleaned glass enters the laminating machine (or is skipped). The laminating machine smoothly applies a whole sheet of transparent polymer protective film (usually polyethylene PE or polyethylene terephthalate PET) to the surface of the glass to be etched.

[0039] After the film is applied, the glass enters the film layer inspection station. This station uses a near-infrared LED array with a wavelength in the 900-1100nm range as the backlight or sidelight source. Light of this wavelength has good penetration through the aforementioned transparent protective film, while forming a clear interface reflection or transmission to the glass substrate. The camera is equipped with a narrowband filter of the corresponding wavelength to shield against ambient light interference and capture images of the interface between the film and the glass.

[0040] The image processing system (which can be integrated locally at the workstation or run by the central control unit) analyzes the acquired images. Using algorithms such as edge detection, threshold segmentation, and contour extraction, it identifies potential defects that may occur during the film application process, such as bubbles (typically appearing as approximately circular dark areas in the image) and wrinkles (appearing as bright lines or stripes). The system not only identifies the presence of defects but also precisely calculates the pixel position of each defect contour in the glass coordinate system and converts it to mechanical coordinates, ultimately generating a defect location map containing all defect types, sizes, and precise coordinates.

[0041] S4: Precise film repair decision based on defect location.

[0042] 4.1 The central control unit receives and analyzes the defect location map, and performs a second-level decision. There are three possible branch paths: Acceptance: If no defects are detected, or the total area of ​​all defects is less than a preset micro-threshold (e.g., less than 5 square millimeters), the glass film is deemed acceptable. The central control unit controls the conveyor system to directly enter the main etching and polishing section.

[0043] 4.2 Online Local Film Repair: If only localized, repairable defects exist (such as individual bubbles or short wrinkles), the online local film repair program is initiated. The central control unit plans an optimal trajectory based on the contour coordinates in the defect location map, minimizing the movement path of the repair tool and completely covering the defect area. The multi-axis actuator at the online local film repair station (e.g., an XY-axis gantry or a SCARA robot) drives its end-effector's micro-negative pressure suction nozzle along this planned trajectory. During movement, the suction nozzle applies negative pressure, precisely lifting and removing the protective film from the defect area. Subsequently, a hot air knife with closed-loop temperature and airflow control moves above the exposed area, blowing hot air at the appropriate temperature and volume according to the protective film material type (the optimal activation temperature curve is pre-stored in the database), ensuring the glass surface and the new film segment are in a state conducive to adhesion, and then local re-application is performed. After the film repair is completed, the glass can be returned to the film inspection station for rapid re-inspection.

[0044] 4.3 Diversion and Rework: If the defect area is too large, the number is too many, or the density is too dense, it is determined that online repair is not possible. The central control unit then controls the diversion device to guide the glass to the film peeling and rework line. On the rework line, the protective film is completely peeled off, and the glass is cleaned and returned to the loading machine queue, waiting to re-enter the production process.

[0045] S5: Collaborative etching and polishing based on preset parameters and real-time feedback.

[0046] The glass entering the main etching and polishing section is etched by the etching section controller according to the preset (V_e, P_e) parameter package in step S2. That is, it is conveyed at a matching speed and sprayed with etching solution (usually a hydrofluoric acid-based solution) at a matching pressure, followed by hydrofluoric acid polishing.

[0047] After the polishing process and before the final cleaning process, the glass passes through an intermediate quality pre-inspection station. This station uses a non-contact online gloss meter to quickly scan multiple preset measurement points (such as the center point and several edge points) on the glass panel to obtain the current gloss measurement value G_m of the glass.

[0048] The G_m value is fed into the adaptive process model running in the central control unit in real time. The core logic of this model is based on feedforward statistical process control, and its operation involves two key definitions:

[0049] Definition of "similar" quantized feature vector: In this embodiment, if the absolute value of the difference between the CI value of a piece of glass to be processed and the CI value of the currently measured glass is less than the threshold ΔCI, and the absolute value of the difference between its SE value and the SE value of the current glass is less than the threshold ΔSE, then the two pieces of glass are determined to "have similar quantized feature vectors". For example, ΔCI = 0.5, ΔSE = 2mN / m.

[0050] Dynamic Fine-Tuning Logic: The adaptive process model maintains a dynamic correction coefficient table in memory. Upon receiving the G_m value of the current glass, the model searches for all glasses in the "processing queue" that meet the aforementioned "similarity" condition before entering the etching section. If the G_m values ​​of N consecutive glasses with similar characteristics (e.g., N=3) all show deviations in the same direction (e.g., all exceeding the upper limit of the target range), the model determines that there is a systematic drift in the process. At this time, the model automatically generates a small compensation correction ΔV_e (e.g., -0.3 m / min) and updates the speed correction value for the corresponding (CI, SE) interval in the correction coefficient table. Subsequently, for newly entered glasses that match the (CI, SE) interval, the original V_e recommended value retrieved from the database is superimposed with this dynamic correction value before being sent to the etching section controller. This achieves feedforward compensation fine-tuning of process parameters for subsequent similar glasses based on real-time quality feedback, thereby actively suppressing batch-specific quality fluctuations.

[0051] S6: Final processing.

[0052] After adaptive etching and polishing, the glass enters a final cleaning machine to remove all chemical residues from the surface. Following this, a final quality inspection can be performed manually or via machine vision. Qualified refrigerator glass panels are automatically unloaded and packaged by a sheet-forming machine, completing the entire flexible manufacturing process.

[0053] Example 2: A flexible etching production system for implementing the method of Example 1

[0054] This embodiment describes the physical system for implementing the above method. The system is a linearly laid-out automated production line, with the main equipment arranged sequentially along a straight or circular conveyor belt according to the process order.

[0055] The system comprises, from the beginning to the end of the production process, the following:

[0056] Film loading machine: Used to automatically pick up and place glass sheets onto the conveyor belt.

[0057] Initial state detection station: Immediately following the setup of the loading machine. This station integrates a diffuse reflection imaging module (including a specific angle light source and a high-resolution camera) and a contact angle measurement module, used to perform step S1 and output the (CI, SE) vector.

[0058] Pre-treatment cleaning station: Located after the inspection station, it may include units such as ultrasonic cleaning tank, spray arm, and air knife drying, and is used to pre-treat the glass that needs to be cleaned.

[0059] Film applicator: Used to automatically apply protective film to glass surfaces.

[0060] Film layer inspection station: Located after the laminating machine. This station includes a near-infrared light source array with a wavelength of 900-1100nm, an industrial camera equipped with a narrowband filter, and an image processing unit for performing step S3 to generate a defect location map.

[0061] Online local film repair station: Located after the film layer inspection station. This station mainly includes a multi-axis precision robot (such as a gantry or articulated robot), with a micro negative pressure suction nozzle and an independently controlled hot air knife installed at the end of the robot, used to perform the local film repair operation in step S4.

[0062] Film peeling and rework line: A return line connected in parallel with or branching from the main conveyor line, used to receive and process glass with severely substandard film application and send it back to the starting point.

[0063] Main etching and polishing section: This is the core processing area of ​​the system, including the etching slurry spray chamber, the polishing slurry spray chamber, and a precise conveyor belt drive system. Its operation is managed by an independent etching section controller.

[0064] Intermediate quality pre-inspection station: Located after polishing, it integrates an online gloss meter for performing gloss measurement in step S5.

[0065] Final cleaning machine: Used to thoroughly clean the etched glass.

[0066] Unloading machine: Used to automatically unload, stack, or package finished glass products.

[0067] The core of the system is a central control unit (e.g., an industrial computer or a high-performance PLC). This central control unit is connected via an industrial network (e.g., Ethernet, fieldbus) to all the aforementioned inspection stations (2, 5, 9), all the execution devices (1, 4, 6, 8, 10, 11), and the signal distribution devices between the stations. The central control unit's internal memory stores a process mapping database and an adaptive process model program. It is programmed to receive signals ((CI,SE) vectors, defect location maps, gloss values) from each inspection station, make decisions based on the logic described in Example 1, and send control commands to each execution device, thereby coordinating the entire system to complete fully automated flexible production from S1 to S6.

[0068] The embodiments described above merely illustrate preferred embodiments of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A flexible etching production method for refrigerator glass panels, characterized in that, The method includes the following steps: S1: Initial state assessment of glass substrate loading and quantization: Obtain the quantization feature vector of the glass substrate surface, wherein the quantization feature vector includes at least the cleanliness index CI and the surface energy approximation value SE; S2: First decision and parameter preset based on the process mapping model: The quantized feature vector (CI, SE) is input into a pre-stored process mapping relationship database for matching. The database defines the mapping relationship between the feature vector and the process parameter package. The process parameter package includes at least a pretreatment necessity flag, a recommended etching transfer speed V_e, and a recommended etching solution spray pressure P_e. According to the pretreatment necessity flag in the matching result, the glass is controlled to skip or enter the pretreatment cleaning station, and the matched (V_e, P_e) parameter package is sent to the etching section controller for preset. S3: Film application and visual positioning detection under the film: After the film is applied, a near-infrared light source that can penetrate the protective film and an imaging unit are used to acquire the film-glass interface image, identify and locate the defect area, and generate a defect location map; S4: Precise film repair decision based on defect location: According to the defect location map, control the glass to enter the main etching line, or control the online local film repair station to perform local film repair, or divert the glass to the film stripping and rework line. S5: Collaborative etching and polishing based on preset parameters and real-time feedback: The etching section controller controls the etching process according to the preset (V_e, P_e) parameter package; after polishing, the gloss of the glass is measured and the measured value is input into the adaptive process model, which dynamically fine-tunes the preset parameter package of the glass with similar quantized feature vectors (CI, SE) that has not yet entered the etching section. S6: Final processing.

2. The flexible etching production method for a refrigerator glass panel according to claim 1, characterized in that, The cleanliness index CI is obtained by calculating the variance of gray values ​​from the diffuse reflection image of the glass surface; the surface energy approximation SE is obtained indirectly by a contact angle measuring instrument.

3. The flexible etching production method for a refrigerator glass panel according to claim 1, characterized in that, The process mapping database is constructed in the following way: multiple etching process experiments are conducted on glass samples with different (CI, SE) combinations, with the optimization goal of meeting the requirements for finished product gloss and roughness with minimal fluctuation, and the corresponding optimal process parameter package is determined for each feature vector interval.

4. The flexible etching production method for a refrigerator glass panel according to claim 1, characterized in that, In step S3, the wavelength of the near-infrared light source is 900-1100nm.

5. The flexible etching production method for a refrigerator glass panel according to claim 1, characterized in that, In step S4, the control of the online local film repair station for local film repair specifically includes: generating an optimal path plan based on the contour coordinates in the defect location map, controlling the multi-axis actuator to drive the micro negative pressure suction nozzle to remove the protective film of the defect area along the path, and using a hot air knife with closed-loop control of temperature and wind speed for local re-application.

6. The flexible etching production method for a refrigerator glass panel according to claim 1, characterized in that, In step S5, the operating logic of the adaptive process model includes: performing feedforward compensation fine-tuning on the preset etching transfer speed V_e of the glass to be processed with similar (CI, SE) vectors based on the gloss value measured at the intermediate quality pre-inspection station.

7. A flexible etching production system for implementing the method according to any one of claims 1-6, characterized in that, include: The following components are set up in sequence: a film loading machine, an initial state inspection station, a pre-treatment cleaning station, a film laminating machine, a film layer inspection station, an online local film repair station, a film peeling and rework line, a main etching and polishing section, an intermediate quality pre-inspection station, a final cleaning machine, and a film unloading machine. The initial state detection station is configured to output the quantized feature vector (CI, SE) of the glass; The process mapping database stores the mapping relationship between feature vectors and process parameter packages; The membrane inspection station includes a near-infrared penetrating imaging unit for outputting defect location maps; The online local film repair station includes a multi-axis precision film repair device that can perform actions according to path planning; The central control unit is connected to the initial state detection station, the film layer detection station, the intermediate quality pre-inspection station, the conveyor belt diversion device of each station, the driver of the main etching and polishing section, and the database signal, respectively. The central control unit is configured to: perform the process mapping matching; and control branches and parameter presets based on the matching results; Precise film repair is controlled based on the defect location map; And by running the adaptive process model, the etching parameters are optimized in a coordinated manner.