LCD screen bonding manufacturing method

By aligning and analyzing the decompression time constant and cover plate flexural response of the vacuum bonding fixture, the problem of uneven optical adhesive layer flow in large-size LCD display modules was solved, achieving higher consistency and stability and improving display quality.

CN123113183APending Publication Date: 2026-11-17GUANGDONG RUIHUA OPTOELECTRONICS TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611157191.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-31
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In the mass production bonding process of large-size, ultra-thin LCD display modules, the existing process has the problem that the micron-level instantaneous deformation of the cover plate causes uneven flow rate of the optical adhesive layer, resulting in the misalignment of fine lines and uneven display of local Mura, which affects the display consistency.

Method used

By collecting the graded depressurization time constants of each evacuation zone of the vacuum bonding fixture and the displacement sequence of the local flexural response of the cover plate, migration impedance parameters are established, the depressurization rate and flexural response time sequence are aligned, bonding response units are divided, and interface migration impedance evolution data analysis is performed to achieve coordinated control of the differential distribution of the restricted flow path of the optical adhesive.

Benefits of technology

It improves the consistency and interface stability of large-size LCD bonding, reduces the risk of local glue buildup and bubble retention, and improves bonding quality and yield.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113183A_ABST
    Figure CN123113183A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of display module manufacturing technology. It provides a method for manufacturing LCD screen bonding, comprising: dividing bonding response units according to the variation trend of migration impedance parameters between adjacent evacuation zones; collecting the gap closing rate, local deflection change, and optical adhesive front advancement displacement trajectory of each bonding response unit to generate interface migration impedance evolution data; performing second-order deflection change analysis on the interface migration impedance evolution data to obtain the differential distribution of the unidirectional confined flow path of the adhesive; performing zoned dynamic correction control on the depressurization duration and local advancement displacement amplitude of each evacuation zone based on the differential distribution of the unidirectional confined flow path; and, based on the bonding interface state after zoned dynamic correction control, collecting the optical adhesive layer thickness distribution, display area edge alignment deviation, and pixel mapping error distribution, and correcting the process control parameters based on the depressurization cycle and migration impedance parameters.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of display module manufacturing technology, specifically an LCD liquid crystal display screen bonding manufacturing method. Background Technology

[0002] Currently, LCD display modules and protective covers generally adopt vacuum dispensing optical full lamination technology. Compared with traditional frame bonding, it can eliminate the interface air layer, effectively improve the screen display contrast, and enhance the module's dustproof and moisture-proof capabilities. This process involves vacuum pressing after dispensing, so that the liquid optical adhesive fills the gap between the cover and the LCD display surface, and after curing, the two are bonded and fixed. It has been widely used in various display products.

[0003] However, in the mass production bonding process of large-size, ultra-thin LCD display modules, the existing process still has obvious defects: due to the influence of the cover glass support structure and the pressure difference of the vacuum adsorption fixture, the cover glass is prone to micron-level instantaneous deformation during the bonding stage, resulting in uneven flow rate of the optical adhesive layer; in particular, the adhesive spreading in the edge area of ​​the module is restricted, which can easily cause sub-pixel-level relative offset between the upper and lower substrates. The final product will have defects such as fine line offset and uneven local Mura display, which seriously reduces the display consistency. Conventional improvement methods such as thickening the cover glass, fine adjusting the fixture pressure, and optimizing the dispensing parameters cannot completely solve the bonding deviation problem caused by the instantaneous deformation of the cover glass.

[0004] Therefore, it is necessary to provide an LCD screen bonding manufacturing method to address the mismatch in the unidirectional restricted flow path of optical adhesive caused by the asynchronous decompression difference between vacuum zones and the local flexural response of the cover plate, so as to improve the consistency of the bonding interface. Summary of the Invention

[0005] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0006] The technical solution adopted by this invention to solve its technical problem is: The graded decompression time constants and local flexural response displacement sequences of each evacuation zone of the vacuum bonding fixture were collected. The decompression rate and flexural response time sequence were aligned and mapped to establish the migration impedance parameters characterizing the air path damping characteristics of each evacuation zone. The bonding response units were divided according to the variation trend of the migration impedance parameters between adjacent evacuation zones. During the controlled pressing stage of the cover plate relative to the display module, the gap closing rate, local deflection change and optical adhesive front advancement displacement trajectory of each bonding response unit are collected. The continuous advancement trajectory of the optical adhesive front is used as a unified timing reference to perform spatial consistency correction on the gap closing rate and local deflection change, and generate interface migration impedance evolution data of each bonding response unit. Second-order flexural change analysis was performed on the interface migration impedance evolution data. When the local second-order flexural change rate changed from negative to positive and continued to decrease, the corresponding bonding response unit was identified as the local elastic release trigger node of the cover plate. Based on the spatial distribution of the trigger node, the propulsion hysteresis of the optical adhesive front edge between adjacent bonding response units was spatially weighted and corrected. The corrected propulsion hysteresis was used as the unidirectional flow boundary constraint condition, and the difference distribution of the unidirectional restricted flow path of the adhesive between each bonding response unit was analyzed. Based on the differential distribution of unidirectional confined flow paths, the execution sequence of depressurization cycle and local propulsion displacement of cover plate in each extraction zone is configured in a coordinated manner. The impedance convergence reference value obtained by calibration of each bonding response unit is used as the process constraint condition to perform dynamic correction control of depressurization duration and local propulsion displacement amplitude in each extraction zone. Based on the bonding interface state after partitioned dynamic correction control, the optical adhesive layer thickness distribution, display area edge alignment deviation and pixel mapping error distribution are collected. When any parameter exceeds the limit, the corresponding bonding response unit is located, and the process control parameters are corrected based on the pressure relief cycle and migration impedance parameters, and written into the vacuum bonding fixture controller.

[0007] As a further aspect of the present invention: The process of establishing the migration impedance parameters characterizing the gas path damping properties of each extraction zone is as follows: The pressure decay curves of each evacuation zone of the vacuum bonding fixture during the staged depressurization process were obtained, and the corresponding depressurization time constants were extracted using an exponential fitting method. The local flexural displacement sequence of the cover plate under the action of each air extraction zone was collected synchronously, and the pressure decay sequence and displacement response sequence were time-stamped and aligned based on a unified sampling frequency. The aligned pressure change gradient and flexural response amplitude are normalized, and the dynamic response ratio coefficient between the two is calculated. Using the phase lag mapping relationship between the local flexural response of the cover plate and the pressure attenuation process as a constraint, an inter-regional pressure and deformation coupling consistency correction term is introduced to correct the spatiotemporal consistency of the dynamic response ratio coefficient. Based on the dynamic response ratio coefficient, the gas path damping characterization function of each extraction zone is constructed, and the coupling relationship between pressure decay rate and structural flexural response is parameterized to form migration impedance parameters.

[0008] The process of dividing the fitted response unit based on the variation trend of the migration impedance parameter between adjacent pumping zones is as follows: The migration impedance parameters corresponding to each extraction zone are serialized and rearranged according to the spatial topological adjacency relationship to construct a continuous impedance distribution sequence. Based on the difference in migration impedance parameters between adjacent extraction zones, the local impedance gradient and the acceleration of change are calculated. The impedance gradient change process is equivalent to the vacuum path equivalent damped propagation process, and an impedance propagation delay factor is introduced to correct the difference between adjacent partitions. When the local impedance gradient changes from a steady change to a discontinuous transition state, and the acceleration of the change exceeds the preset structural sensitivity threshold, the corresponding location is determined as the interface impedance fracture boundary point. Using boundary points as constraints, the bonding area of ​​the display module is reconstructed in a structured segmentation to form multiple bonding response units.

[0009] The process of generating the interface migration impedance evolution data for each mating response unit is as follows: During the controlled pressing process, the real-time gap change sequence of each bonding response unit and the local deflection response sequence of the cover plate are collected. Using the spatial position of the optical adhesive leading edge trajectory as the main reference coordinate, the sampling data of different bonding response units are resampled and reconstructed in a spatiotemporal manner. Phase alignment correction is performed on response data from different regions; Among them, the phase alignment correction uses the activation moment of local interface contact caused by the advancement of the optical adhesive front edge as the trigger reference; Based on the corrected gap closing rate and deflection change, the interface compression response evolution coefficient per unit time is calculated to form interface migration impedance evolution data.

[0010] The process of identifying the corresponding bonding response unit as the cover plate local elastic release trigger node is as follows: Second-order difference calculations were performed on the interface migration impedance evolution data to obtain the local flexural acceleration variation sequence. Analyze the sign transition points of the second-order rate of change of flexure and identify the inflection point region where the value changes from negative to positive. A duration window analysis was performed on the inflection point region. When the positive rate of change showed a decreasing trend within the continuous sampling period, it was determined that the inflection point region changed from a compression-dominated state to an elastic release state. The bonding response unit that meets the above conditions is marked as a local elastic release trigger node.

[0011] The process of analyzing the difference distribution of the colloidal unidirectional confined flow path between each bonding response unit is as follows: Extract the propulsion time difference and spatial displacement difference of the optical adhesive front edge between adjacent bonding response units, and calculate the propulsion hysteresis. Based on the spatial distribution of the triggering nodes, spatial weighting coefficients are assigned to the propulsion lag between each adjacent mating response unit; The spatial weight coefficient is negatively correlated with the distance from the trigger node to the boundary of the corresponding fitting response unit, so as to perform spatial weight correction on the initial propulsion lag and obtain the corrected propulsion lag. Using the corrected propulsion hysteresis as a flow retardation constraint factor, a flow priority matrix of colloids is constructed among different bonding response units. After completing the construction of the flow priority matrix, a correction based on real physical field measurement data is applied to the matrix. Constraint propagation analysis of the colloidal propagation path is performed based on the modified flow priority matrix to determine the dominant and restricted unidirectional flow paths and output the distribution of differences in colloidal flow paths among each bonding response unit.

[0012] The process of coordinating the execution sequence of depressurization cycle and local advance displacement of cover plate in each extraction zone is as follows: Based on the differential distribution of colloidal flow paths, the flow retardation coefficient corresponding to each bonding response unit is calculated, and the topological connection matrix of flow dependence between each bonding response unit is constructed. Based on the weight distribution of the hysteresis coefficients of each fitted response unit in the topology connection matrix, a pressure relief priority sequence is generated. Based on the pressure relief priority sequence, the pressure relief duration and pressure relief rate of each extraction zone are configured in a graded manner to form the zone pressure relief cycle parameters; Based on the flow direction vector distribution of each bonding response unit, a sequence of local propulsion displacement directions of the cover plate is generated; Based on the temporal correspondence between the depressurization cycle parameters and the propulsion displacement direction sequence, the execution sequence of the local propulsion displacement of the cover plate is determined, and the depressurization execution cycle of each extraction zone is synchronously matched and configured.

[0013] The process of dynamically correcting and controlling the depressurization duration and local propulsion displacement of each extraction zone is as follows: The migration impedance sequence data of each mating response unit is collected in real time, and the impedance convergence change rate is calculated based on the time window to generate the corresponding convergence error vector. The convergence error vector is compared with the calibrated consistency threshold to construct a partition deviation judgment identifier matrix; The pressure relief duration parameter of each extraction zone is incrementally corrected based on the judgment identifier matrix, and the pressure relief control timing of the corresponding zone is updated. The rate of change of local interface impedance is synchronously acquired and mapped to a displacement correction coefficient according to the piecewise linear mapping rule; The local advancing displacement amplitude of the cover plate is adjusted in stages based on the displacement correction coefficient.

[0014] The process of correcting process control parameters based on pressure relief cycle time and migration impedance parameters is as follows: After bonding is completed, the entire screen is subjected to multi-point optical scanning to obtain the optical adhesive layer thickness distribution matrix and edge alignment deviation data. Perform regional segmentation analysis on pixel mapping errors to locate the corresponding bonding response units in the concentrated error areas; The abnormal fit response unit is associated with the corresponding pressure relief cycle parameters and migration impedance parameters for tracing the source. Based on the tracing results, the partition control parameters in the vacuum bonding fixture controller are updated and written.

[0015] The beneficial effects of this invention are as follows: By introducing a migration impedance modeling mechanism based on the coupling relationship between the depressurization characteristics of the pumping zones and the local deflection response of the cover plate, the vacuum bonding process is transformed from traditional empirical control to a zoned dynamic control system with physical mapping constraints. By aligning and mapping the depressurization time constant and the deflection displacement time sequence and dividing the bonding response into bonding response units, a refined zoning expression of the bonding interface from macroscopic regions to local units is achieved, thereby improving the identifiability and controllability of gas path damping differences in different regions. During the pressing process, the optical adhesive leading edge trajectory is used as a unified timing benchmark to perform spatial consistency correction on the gap closing rate and local deformation, giving the interface migration impedance evolution process a comparable timing consistency basis, thus improving the synchronous evaluation capability of the pressing state in different regions. Furthermore, by introducing a second-order local deflection... The system identifies the trigger node for the elastic release of the cover plate by identifying its features, and combines this with the differential distribution of the unidirectional restricted flow path of the colloid to constrain and analyze the interface flow behavior. This transforms the optical adhesive expansion process from free flow to controlled path evolution, thereby reducing the risks of localized adhesive buildup, bubble retention, and uneven flow. Furthermore, based on the constraints of path differences and impedance convergence consistency, the system collaboratively optimizes the depressurization cycle and local propulsion displacement sequence of each evacuation zone, and implements zoned dynamic correction control, enabling the bonding process to achieve adaptive equilibrium convergence in both spatial and temporal dimensions. Finally, through closed-loop detection and parameter write-back of thickness distribution, edge alignment deviation, and pixel mapping error, the system achieves closed-loop control of bonding quality from post-detection to process-driven correction, effectively improving the consistency, interface stability, and yield of large-size LCD bonding. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 This is a flowchart of the LCD screen bonding manufacturing method provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the process framework of the LCD liquid crystal display screen bonding manufacturing method provided in Embodiment 1 of the present invention. Detailed Implementation

[0018] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0019] Example 1: Please refer to Figure 1-2 As shown, the LCD screen bonding manufacturing method according to an embodiment of the present invention includes the following steps: Step S10: Collect the graded decompression time constant and the local flexural response displacement sequence of each pumping zone of the vacuum bonding fixture, align and map the decompression rate with the flexural response time sequence, establish the migration impedance parameter characterizing the air path damping characteristics of each pumping zone, and divide the bonding response unit according to the changing trend of the migration impedance parameter between adjacent pumping zones.

[0020] In step S10, the process of establishing the migration impedance parameters characterizing the gas path damping characteristics of each extraction zone is as follows: The pressure decay curves of each evacuation zone of the vacuum bonding fixture during the staged depressurization process were obtained, and the corresponding depressurization time constant was extracted using an exponential fitting method. The vacuum bonding fixture was first divided into multiple independent evacuation zones according to process requirements. For large-size LCD bonding scenarios, 8 to 16 zones were typically used. Each zone was equipped with a high-precision vacuum pressure sensor, and a fixed sampling accuracy and frequency were set to ensure accurate and stable pressure data acquisition. After the cover plate and LCD module were aligned and the vacuum chamber was completely sealed, the staged depressurization program was initiated, using a stepped gradient depressurization method to gradually depressurize from high vacuum steady-state pressure to atmospheric pressure. Real-time pressure data from each independent evacuation zone was continuously collected throughout the process, and the raw pressure decay data for each zone was organized in chronological order to create a complete pressure decay curve. By leveraging the inherent inertial decay characteristics of vacuum decompression, a standard exponential fitting method is employed to fit the global pressure curve. Through iterative optimization, abnormal data caused by equipment vibration and environmental interference are eliminated to ensure the fitting accuracy meets the standards. Ultimately, the decompression time constant corresponding to each pumping zone is accurately extracted. This time constant can intuitively reflect the differentiated characteristics of air passage resistance, pipeline patency, and cavity volume in each zone, providing reliable basic time-series data for subsequent quantification of air passage damping characteristics.

[0021] The system synchronously collects the local deflection displacement sequence of the cover plate under the action of each evacuation zone, and timestamps the pressure decay sequence and displacement response sequence based on a unified sampling frequency. High-precision laser displacement sensors are deployed in the cover plate contact area corresponding to each evacuation zone of the fixture to accurately collect the microscopic deflection deformation of the cover plate during the depressurization process. The sensor data acquisition is synchronized with the vacuum pressure acquisition system throughout the entire process. During the entire staged depressurization process, the system continuously captures the vertical local deflection displacement changes of the cover plate corresponding to each zone, and organizes them in chronological order to generate a complete cover plate deformation time sequence. The sequence fully records the entire process of the cover plate's changes from its initial flat state and depressurization deformation to its later elastic recovery. To address the timing misalignment caused by inconsistent sampling periods of different acquisition hardware, the sampling frequencies of pressure and displacement data are uniformly calibrated. Using the depressurization start moment as a unified zero-time reference, the two sets of timing data are matched, completed, and trimmed frame by frame to remove invalid data with timing offsets. This achieves complete timing synchronization between pressure change data and cover plate deformation data, completely eliminating analysis errors caused by hardware sampling time differences and establishing a unified timing reference for subsequent coupled analysis.

[0022] The aligned pressure change gradient and flexural response amplitude are normalized, and the dynamic response ratio coefficient between them is calculated. For a single zone, the maximum and minimum values ​​are first found from all pressure data during the entire depressurization process of that zone, and then the maximum and minimum values ​​are found from the flexural displacement data of the same time period. Subsequently, the minimum pressure value of that zone is subtracted from the real-time pressure value at each sampling moment, and the difference is divided by the difference between the maximum and minimum pressure values ​​of that zone to obtain the normalized pressure value at that moment. Simultaneously, the minimum flexural displacement value of that zone is subtracted from the real-time flexural displacement value at each sampling moment, and the difference is divided by the maximum flexural displacement value of that zone. The difference between the normalized and minimum values ​​is the normalized flexural displacement value at that moment. After normalization, the normalized pressure values ​​at two adjacent sampling moments are subtracted and the absolute value is taken to obtain the pressure change gradient at that moment. The normalized flexural displacement values ​​at two adjacent sampling moments are subtracted and the absolute value is taken to obtain the flexural change amplitude at that moment. Finally, the flexural change amplitude at the same moment is divided by the pressure change gradient and a very small positive number. The quotient is the dynamic response ratio coefficient of that partition at that moment. This coefficient is a dimensionless value, and each sampling moment corresponds to an independent coefficient value. The coefficients at all moments are arranged in chronological order to form a dynamic response ratio coefficient time series curve.

[0023] Using the phase lag mapping relationship between the local deflection response of the cover plate and the pressure decay process as a constraint, a pressure and deformation coupling consistency correction term is introduced between regions to correct the spatiotemporal consistency of the dynamic response ratio coefficient. The cover glass is an elastic thin-plate structure, and the response speed of vacuum pressure change is faster than that of cover plate deflection deformation, resulting in a natural time difference in response. Furthermore, the difference in air path damping between different pumping zones further exacerbates this deviation, leading to spatiotemporal distortion in the original response coefficient. Therefore, a time consistency correction is performed as follows: For each pumping zone, the normalized pressure time-series curve and the normalized deflection displacement time-series curve are compared frame-by-frame with misalignment, ranging from -50 frames to +50 frames. For each of the ten frames, after each misalignment, the two sequences are multiplied frame by frame and summed to obtain the correlation coefficient for that number of misaligned frames. After traversing all misaligned frames, the number of misaligned frames corresponding to the maximum correlation coefficient is found. This number of frames is multiplied by the time length corresponding to the single-frame sampling period to obtain the lag time of the relative pressure decay of the flexural response in that partition, in seconds. After performing the same operation on all partitions, the lag time of each partition is obtained. The lag time of each partition is used as the quantitative expression of the phase lag mapping relationship to form the temporal constraint boundary for subsequent spatial correction. That is, the adjustment amount of the local advance displacement amplitude of the partition must not destroy the temporal sequence relationship corresponding to the lag time, ensuring the corrected flexural response. The temporal sequence should still follow the physical driving order of pressure decay on deformation. Based on this, spatial consistency correction is performed as follows: Extract the final value of the normalized flexural displacement time-series curve of each partition at the current moment to construct a global flexural displacement spatial distribution vector; calculate the global arithmetic mean of this vector; for each partition, subtract the global average from the final flexural displacement value of that partition to obtain the positional deviation value of that partition; multiply the positional deviation value by a spatial equilibrium coefficient pre-calibrated using five sets of good-quality bonding data, and the resulting product is defined as the inter-regional pressure and deformation coupling consistency correction term for that partition; the calibration method for the spatial equilibrium coefficient is as follows: select five sets... For the qualified bonding samples, the final value of the flexural displacement of each zone is extracted and the corresponding optimal correction amount is calculated to minimize the sum of squares of the flexural displacement differences between adjacent zones after correction. The arithmetic mean of the ratios of the five optimal correction amounts to the corresponding positional deviation values ​​is taken to obtain the spatial balance coefficient. Once determined, this coefficient is stored in the process database and used permanently under the current product specifications, only to be recalibrated when the product structure or bonding material changes. After the correction item calculation is completed, the correction item for that zone is applied to the normalized flexural displacement value after time-compensated in a weighted superposition manner. The superposition formula is: final corrected displacement value of that zone = displacement value after time-compensated + correction item × spatial balance coefficient.After correction, the final corrected displacement values ​​of each partition are used to replace the original flexural displacement data. The dynamic response ratio coefficient is calculated according to a predefined formula. This ratio coefficient is physically the ratio of the change in flexural displacement to the change in pressure during the same period. It is obtained by dividing the change in flexural displacement at each sampling time by the change in pressure during the same period. The calculation is repeated frame by frame to obtain the corrected dynamic response ratio coefficient time series. After correction, the standard deviation of the corrected displacement values ​​for all partitions is calculated. If the standard deviation is greater than 20% of the original standard deviation, the spatial equalization coefficient is finely adjusted and incremented, the correction term is recalculated, and the correction is iterated again until the standard deviation is reduced to within 20% of the original standard deviation. At this point, the spatiotemporal consistency correction is considered complete. Through the above two-stage correction of temporal alignment and spatial equalization, the corrected coefficients retain the dynamic change characteristics of a single partition while meeting the requirements for overall spatial consistency.

[0024] Based on the dynamic response ratio coefficient, a gas path damping characterization function is constructed for each extraction zone. The coupling relationship between pressure decay rate and structural deflection response is parameterized to form a migration impedance parameter. For a single zone, the coefficient values ​​at all sampling moments during the entire depressurization process in the time series of the modified dynamic response ratio coefficient for that zone are accumulated item by item, and then divided by the total number of sampling frames to obtain the full-time average value of the dynamic response ratio coefficient for that zone. This full-time average value is then multiplied by the depressurization time constant for that zone. The product is then multiplied by the dimension conversion constant, and the value is assigned the unit "Pascal-second". The final value obtained is the migration impedance parameter for that zone. The migration impedance parameter determined in this way is... The migration impedance parameter is a single value, with its physical unit being Pascal. This value comprehensively reflects the coupling effect of the zoned gas path damping characteristics, vacuum pressure response characteristics, and cover plate structural deformation response characteristics. Each pumping zone independently calculates a migration impedance parameter value. The difference in values ​​between different zones directly reflects the differences in their respective pressure relief response characteristics, providing a core quantitative basis for subsequent fitting response unit division and interface state analysis. It should be emphasized that all the above calculation steps are completed based on frame-by-frame discrete data at the same sampling frequency. Each operation has a clear execution object, calculation method, and parameter value, and all results can be independently reproduced according to the same process.

[0025] In step S10, the process of dividing the fitting response unit according to the changing trend of the migration impedance parameter between adjacent pumping zones is as follows: The migration impedance parameters corresponding to each evacuation zone are serialized and rearranged according to spatial topological adjacency relationships to construct a continuous impedance distribution sequence. The calibrated migration impedance parameters of all evacuation zones of the vacuum bonding fixture are taken, and combined with the actual physical layout of the fixture, the spatial topological adjacency relationships of all evacuation zones are analyzed to clarify the adjacent order of each zone vertically and horizontally, discarding the original random numbering method of the equipment. According to the actual spatial distribution pattern of the bonding screen's horizontal extension and vertical arrangement, the migration impedance parameters of all zones are sequentially rearranged to ensure that the parameter arrangement order completely corresponds to the actual physical spatial position of the fixture. Based on this, the discretely distributed impedance parameters of each zone are spatially interpolated and made continuous, filling the transition impedance characteristic data between zones to eliminate data discontinuity caused by discrete zones. Finally, a global impedance distribution sequence covering the entire bonding area and with continuous spatial correspondence is constructed, providing a complete and continuous data base for subsequent global impedance gradient change analysis.

[0026] Based on the difference in migration impedance parameters between adjacent extraction zones, the local impedance gradient and its acceleration are calculated. Using a constructed continuous impedance distribution sequence, the impedance parameter differences between spatially adjacent extraction zones are compared and calculated one by one. The magnitude of the impedance difference in each adjacent region is statistically analyzed to characterize the impedance change amplitude in the local area. Combined with the spatial spacing parameters of each zone, the spatial rate of change of the adjacent impedance difference is converted to obtain the local impedance gradient at each location across the entire domain, intuitively reflecting the rate of spatial change of gas path damping in different regions. Simultaneously, multiple sets of time-series continuous impedance gradient data are continuously collected, and the gradient change differences at the same spatial location at different times are compared. The real-time rate of change of the gradient parameters is statistically analyzed, and the acceleration of the impedance gradient is calculated. This is used to accurately determine whether the impedance change is a gradual and stable change or a sudden change, providing a dual-layer judgment basis for subsequent boundary identification.

[0027] The impedance gradient change process is equated to the equivalent damping propagation process of the vacuum air path, and an impedance propagation delay factor is introduced to correct the difference between adjacent zones. During vacuum sealing and depressurization, the changes in damping of each zone's air path are not transmitted instantaneously and synchronously; both airflow propagation and cavity pressure transmission have inherent delays, and the spatial change of the impedance gradient is accompanied by propagation lag characteristics. Therefore, the spatial change law of the global impedance gradient is equivalently simulated as the propagation process of vacuum air path damping from near to far, conforming to the diffusion and transmission logic of the actual depressurization airflow. Based on the inherent hardware parameters such as the length of the fixture's air path, cavity volume, and zone spacing, the impedance propagation is set as follows: Propagation delay factor: First, each extraction zone is equivalent to a lumped parameter gas path unit. Its characteristic length is the actual length of the pipeline from the extraction port of that zone along the airflow direction to the center point of the mating cavity, in meters. Its characteristic volume is the volume of the mating cavity area covered by that zone, in cubic meters, calculated by multiplying the zone area by the mating gap height. The distance between adjacent zones is the straight-line distance between the center points of the extraction ports of the two zones, in meters. The delay factor is based on the propagation delay time, which is the time required for a pressure disturbance to propagate from one zone to an adjacent zone. The calculation method is: propagation delay time equals the center-to-center distance between zones divided by the gas path length. The effective sound velocity, determined by experimental fitting and correction of the ratio of pipe cross-sectional area to cavity volume, is set at 340 meters per second. This value is an empirical value of the air sound velocity under standard atmospheric pressure after pipe structure correction, applicable to the negative pressure air path propagation conditions of this fixture. The propagation delay time is used as the initial reference value of the impedance propagation delay factor, in seconds. Based on this, for each group of adjacent partitions, the center-to-center distance between the adjacent partitions is divided by the effective sound velocity to obtain the corresponding delay factor value, so that each group of adjacent partitions has an independently matched delay factor, rather than a uniform single value across the entire domain. This delay factor is then used to apply to all adjacent partitions. The impedance difference and gradient data are corrected by hysteresis compensation. The specific correction method is as follows: For each group of adjacent partitions, the impedance time series data of one partition is shifted backward along the time axis by the time frame corresponding to the delay factor of the group. The difference between the shifted impedance sequence and the original impedance sequence of the adjacent partition is then calculated. The shift direction is to shift the impedance data of the upstream partition of the airflow backward so that it is aligned with the time node when the downstream partition feels the pressure disturbance. This eliminates the problem of false difference being too large caused by airflow propagation delay, so that the calculated impedance difference data completely matches the real air path damping propagation state and improves the accuracy of subsequent boundary judgment.

[0028] When the local impedance gradient changes from a stable state to a discontinuous transition state, and the acceleration of the change exceeds a preset structural sensitivity threshold, the corresponding location is determined as the interface impedance fracture boundary point. The preset structural sensitivity threshold is determined as follows: During the bonding equipment debugging stage, three sets of cover plate and display module samples of the same specifications as the current production batch are selected. Each set of samples is bonded under the same vacuum pressing conditions. During the bonding process, the impedance gradient change acceleration time-series data of each bonding response unit is collected at a fixed sampling frequency throughout the process. After bonding, each set of samples is scanned in sequence through an ultrasonic scanning microscope to scan the entire bonding interface. The interface area with a sudden change in ultrasonic echo amplitude and a continuous length exceeding 0.5 mm is marked as the actual structural fracture boundary location. The impedance gradient corresponding to the actual fracture boundary location in each set of samples is extracted. The lower limit of the impedance gradient change acceleration value is taken, which is the minimum value of the impedance gradient change acceleration at the actual fracture boundary among the three sets of samples. This lower limit value is then multiplied by a safety margin coefficient of 0.8, and the product is used as the preset structural sensitivity threshold for this product specification. Once this threshold is determined, it is stored in the equipment process parameter database as the benchmark control parameter for the current product specification. Subsequent batches of products of the same specification directly call this threshold to perform boundary judgment without recalibrating before each batch. Only when the product structure design, cover plate thickness, or adhesive layer material changes, the threshold in the database is re-determined and updated according to the above three sets of sample calibration procedures. When the local impedance gradient changes from a stable change to a discontinuous transition state, and the change acceleration exceeds the above preset structural sensitivity threshold, the corresponding position is determined as the interface impedance fracture boundary point.

[0029] Using boundary points as constraints, the bonding area of ​​the display module is structurally segmented and reconstructed to form multiple bonding response units. All interface impedance breakage boundary points identified across the entire domain are aggregated, and discrete boundary points are connected by lines according to their spatial positions to form continuous, closed regional boundaries, which serve as hard constraints for region segmentation. Combining the effective display range of the LCD screen bonding area and the boundaries of non-effective bonding areas at the edges, the entire module bonding area is structurally and finely segmented and reconstructed, eliminating invalid and redundant partitions, integrating continuous regions with consistent characteristics, and splitting abruptly different regions with significant characteristic differences. Finally, the entire bonding area is divided into several independent bonding response units with uniform internal impedance characteristics, consistent deformation response patterns, and significant differences in characteristics between regions. This achieves a shift from overall macroscopic control of the bonding interface to refined control of local units, providing standardized analysis units for subsequent precise regional control and colloidal flow analysis.

[0030] Step S20: During the controlled pressing stage of the cover plate relative to the display module, the gap closing rate, local deflection change and optical adhesive front advancement displacement trajectory of each bonding response unit are collected. The continuous advancement trajectory of the optical adhesive front is used as a unified timing reference to perform spatial consistency correction on the gap closing rate and local deflection change, and interface migration impedance evolution data of each bonding response unit are generated.

[0031] In step S20, the process of generating the interface migration impedance evolution data for each bonding response unit is as follows: During the controlled pressing process, real-time gap change sequences and local deflection response sequences of each bonding response unit are collected. During the controlled vacuum pressing stage between the LCD display cover and the liquid crystal display module, the vacuum cavity is kept sealed, and the fixture's adsorption state is stabilized. A high-precision online monitoring system is activated throughout the process to continuously collect data at fixed points for each bonding response unit that has been pre-defined. Using a high-resolution gap detection device, dynamic change data of the bonding gap between the cover and the LCD panel are collected unit by unit in real time, recording the continuous change in gap from large to small throughout the pressing process. This process generates a unique time sequence of gap changes for each unit. Simultaneously, a high-precision laser detection component is used to collect the local flexural deformation state of the cover plate within each bonding response unit area. This captures the dynamic response process of the cover plate under pressure indentation and micro-deformation rebound during the pressing process, generating a corresponding time sequence of local flexural response of the cover plate. A uniform and fixed sampling frequency is maintained throughout the process to ensure that the time sequence density of each data set is consistent and the data is complete and without missing data. This process fully records the dynamic change characteristics of the interface gap and structural deformation of each unit throughout the pressing process, providing original measured data support for subsequent data reconstruction and analysis.

[0032] Using the spatial position of the optical adhesive leading edge trajectory as the primary reference coordinate, the sampling data of different bonding response units are resampled and reconstructed in spatiotemporal time. During the vacuum pressing and spreading of the adhesive, a visual alignment detection device captures the leading edge propagation state of the liquid optical adhesive from the dispensing area to the surrounding areas in real time, continuously recording the spatial propagation position, spreading range, and diffusion path of the optical adhesive leading edge in the bonding plane, forming a complete spatial propagation trajectory of the adhesive leading edge. This propagation trajectory is represented by a two-dimensional spatial coordinate sequence in the bonding plane, with each set of spatial coordinates corresponding to the arrival time of the adhesive leading edge. Using this spatial coordinate sequence as the globally unified primary reference coordinate, the following spatial resampling operation is performed on the original sampling data: First, for the geometric center of each bonding response unit, its fixed two-dimensional coordinates in the bonding plane are calculated; second, all leading edges in the spatial propagation trajectory of the adhesive leading edge are resampled. Along the spatial coordinate points, find the leading edge spatial coordinate point closest to the geometric center of the unit and the corresponding leading edge arrival time; then, using this arrival time as the spatiotemporal reference matching point of the unit, extract the original gap data and deflection data within one sampling period before and after this time; finally, using the colloid leading edge propagation velocity as a weighting factor, perform linear interpolation on the extracted original data to obtain the gap value and deflection value at the geometric center of the unit that precisely correspond to the arrival time of the colloid leading edge, completing the coordinate mapping transformation of the original time series data from "based on the device clock" to "based on the spatial position of the colloid leading edge"; by performing the above spatial coordinate matching and interpolation operations for each unit one by one, the monitoring data of all bonding response units are based on the actual spatial position of the colloid spread, realizing the comparability and linkage analysis of multi-region data in the spatial coordinate system.

[0033] Phase alignment correction was performed on the response data from different regions. Due to the different spatial positions of each bonding response unit, there were differences in the timing of optical adhesive spreading progress, interface contact time, and cover plate deformation under stress. This resulted in temporal phase shifts in the gap and deflection data of different units, making them unsuitable for direct coupling analysis. After completing the interpolation resampling based on the spatial coordinates of the colloid front, all data had a unified physical reference. However, residual phase deviations still existed because the center coordinates of each unit were not completely equal to the nearest neighbor distance in the trajectory of the colloid front. Therefore, the theoretical calculation time when the optical adhesive front advanced to the geometric center of each unit was used as the standard phase reference. The residual time offset between this theoretical time and the actual data recording time of each unit after interpolation resampling was calculated. Based on this offset, all data points of each unit were shifted temporally to ensure that the zero time of each unit's data uniformly corresponded to the physical time when the colloid front advanced to its respective geometric center. This completely eliminated the residual phase deviations caused by the differences in the geometrical distribution of the units, ensuring the synchronization and accuracy of subsequent global interface state analysis.

[0034] The phase alignment correction uses the activation moment of local interface contact caused by the advancement of the optical adhesive front as the trigger reference. Specifically, the geometric center coordinates of each bonding response unit are substituted into the spatial advancement trajectory of the adhesive front. The arrival time of the optical adhesive front advancing to the geometric center of the unit is accurately calculated through coordinate matching and linear interpolation. This moment is defined as the valid starting trigger point for the data record of the corresponding unit. Using the interface contact activation moment of each unit as an independent zero point, the timing starting point of all gap change data and flexural deformation data of the unit is recalibrated. This accurately eliminates the inherent phase difference caused by the asynchronous arrival time of the adhesive in each unit, so that all response data correspond to the physical process of actual interface pressure, contact, and deformation, ensuring that the phase alignment result fully matches the actual bonding conditions.

[0035] Based on the corrected gap closing rate and deflection change, the interface compression response evolution coefficient per unit time is calculated to form interface migration impedance evolution data. After completing spatiotemporal reconstruction and precise phase alignment correction, stable and synchronous gap closing dynamic data and cover plate local deflection change data of each bonding response unit are extracted, and the gap closing advancement speed and local deformation change amplitude of each unit per unit time are statistically analyzed. Combining the dynamic change law of interface compression during the pressing process, and integrating the gap closing progress and cover plate deflection deformation state, the interface compression response degree of each bonding response unit at different pressing times is quantitatively calculated to obtain the interface compression response evolution coefficient per unit time. The specific calculation method is as follows: at each sampling time, the current measured value of the gap closing rate is divided by the initial gap value of the bonding response unit to obtain the gap closing progress ratio; the current measured value of the deflection change is divided by the local thickness value of the cover plate of the unit to obtain the deflection strain ratio; the gap closing progress ratio and the deflection strain ratio are calculated according to the pre-stored data in the bonding process database. The weighted sums are calculated using a weighted fusion method, with the gap closure progress ratio having the dominant weight (60% to 80%) and the flexural strain ratio having the auxiliary weight (20% to 40%), the sum of which is 100%. The resulting weighted sum is then multiplied by the interface contact area correction coefficient for that unit at the current moment to obtain the interface compression response evolution coefficient. This interface contact area correction coefficient is determined based on the ratio of the actual widening area of ​​the optical adhesive layer to the area of ​​the bonding response unit during the pressing process. This ratio is calculated by a high-speed vision system that measures the wetting front position of the colloid in real time. The dynamic change trend of this coefficient is continuously recorded throughout the pressing process, fully reflecting the real-time evolution of interface damping, stress state, and deformation coupling relationships of each unit. This data is then compiled to form temporally continuous and spatially corresponding interface migration impedance evolution data covering all bonding response units, providing a precise data foundation for subsequent flexural characteristic analysis, elastic release node identification, and colloid flow analysis.

[0036] Step S30: Perform second-order flexural change analysis on the interface migration impedance evolution data. When the local second-order flexural change rate changes from negative to positive and continues to decay, identify the corresponding bonding response unit as the local elastic release trigger node of the cover plate. Use the advance hysteresis of the optical adhesive front edge between adjacent bonding response units as the unidirectional flow boundary constraint condition, and analyze the difference distribution of the unidirectional restricted flow path of the colloid between each bonding response unit.

[0037] In step S30, the process of identifying the corresponding bonding response unit as the cover plate local elastic release trigger node is as follows: Second-order difference calculations were performed on the interface migration impedance evolution data to obtain the local flexural acceleration change sequence. Complete interface migration impedance evolution data of each bonding response unit after spatiotemporal correction and phase alignment were retrieved. This set of data includes the continuous gap closure state and local flexural evolution information of the cover plate throughout the entire pressing process. For the time-series evolution data of each bonding response unit, continuous second-order difference operations were performed segment by segment. First, the original flexural change time-series data was solved by first-order difference to obtain the flexural change rate sequence within a unit period. Then, the rate sequence was differentially processed again to obtain the corresponding flexural change acceleration sequence. The entire process adopted a global continuous sampling calculation method without data truncation or simplified fitting, which completely preserved the dynamic change characteristics of flexural change in each sampling period and accurately restored the increase and decrease of deformation rate of the cover plate during the pressing process. Finally, a local flexural acceleration change sequence covering the entire pressing cycle, with continuous time sequence and corresponding to each independent bonding response unit was generated, providing original analysis data for subsequent state inflection point identification.

[0038] The sign transition points of the second-order rate of change of flexure are analyzed, and the inflection point region from negative to positive values ​​is identified. The sign and state analysis of the local flexure acceleration change sequence corresponding to each unit are performed frame-by-frame to distinguish the positive and negative change attributes at different times. Negative change states represent the compression-dominated stage where the cover plate is under continuous pressure and deformation, with the degree of compression constantly intensifying. Positive change states represent the recovery stage where the rate of deformation slows down and a rebound trend begins. The acceleration sign states are compared sequentially to accurately capture the moment when the sequence completely switches from negative to positive values, marking all transition points where positive and negative states switch. The sign transition points within continuous intervals are summarized to form a complete inflection point region. This region corresponds to the critical interval where the bonding interface transitions from forced compression to elastic rebound, and is the core characteristic interval for determining the start of elastic release of the cover plate.

[0039] Duration window analysis is performed on the inflection point region. When the positive rate of change shows a decreasing trend within the continuous sampling period, the inflection point region is determined to have transitioned from a compression-dominated state to an elastic release state. For each identified inflection point region, a fixed-duration sliding time window is used to continuously verify the positive flexural change characteristics before and after the inflection point, avoiding misjudgments caused by instantaneous data fluctuations. Within the continuous sampling period covered by the time window, the fluctuation status and evolution trend of the positive flexural change rate are continuously monitored. If, after the inflection point, multiple sets of sampling data show a stable characteristic of gradually decreasing positive change amplitude and a continuously decreasing growth trend, it indicates that the cover plate is no longer continuously subjected to pressure deformation, the interface compression force gradually subsides, and the cover plate's own elastic recovery force begins to dominate. At this point, it can be accurately determined that the inflection point region has completely left the compression-dominated state and officially entered a stable elastic release state, effectively filtering out false inflection point interference caused by single jitters and local pressure fluctuations, and improving the stability and accuracy of state determination.

[0040] The bonding response units that meet the above conditions are marked as local elastic release trigger nodes. All bonding response units that have undergone second-order difference analysis, sign inflection point identification, and continuous time window verification are summarized, and target units that simultaneously possess the characteristics of a negative-to-positive inflection point and a continuously decreasing positive rate of change are selected. These qualified units are marked with exclusive features and defined as local elastic release trigger nodes of the cover plate in the bonding process. This represents that the local area is the first to complete compression energy storage and initiate elastic rebound, and is a key feature point where the flow resistance of the optical adhesive changes and the interface stress state switches. Through full-domain unit screening and marking, all elastic release trigger positions in the entire screen bonding area are completely located, providing accurate regional feature basis for subsequent analysis of the unidirectional restricted flow path of the adhesive and coordinated control of the zoned pressure relief rhythm.

[0041] In step S30, the process of analyzing the difference distribution of the unidirectional confined flow path of the colloid between each bonding response unit is as follows: The advancement time difference and spatial displacement difference of the optical adhesive front edge between adjacent bonding response units are extracted to calculate the advancement lag. Throughout the entire process of vacuum pressing and colloid spreading, high-definition high-speed vision acquisition equipment is used to continuously track the global spreading dynamics of the optical adhesive wetting front edge, and record the accurate time and spatial position of the colloid front edge reaching the boundary of each bonding response unit in real time. For each group of spatially adjacent bonding response units, the time nodes of the colloid front edge entering the preceding and following units are compared one by one to obtain the colloid advancement time difference between adjacent units. At the same time, combined with the coordinate spacing of the bonding plane, the spatial displacement difference of the colloid front edge during the spreading process of adjacent units is measured. Combining the dual deviation characteristics of time lag and spatial lag, the colloid advancement lag at the interface of each adjacent unit is comprehensively evaluated, which intuitively represents the difference in the speed and degree of hindrance of colloid spreading in different regions, providing real working condition data support for subsequent flow constraint modeling.

[0042] Based on the spatial distribution of trigger nodes, spatial weighting coefficients are assigned to the propulsion hysteresis between adjacent bonding response units. During the cover plate pressing process, when a bonding response unit is detected to have a local deflection second-order rate of change from negative to positive and continuously decaying, the coordinates of the center point of that unit are recorded as the spatial location of a trigger node. After completing the identification of trigger nodes for all bonding response units, the distribution points of all trigger nodes are marked in the bonding plane coordinate system. For each group of adjacent bonding response units, the midpoint coordinates of their common boundary line segment are extracted, and the straight-line distances from the midpoint coordinates to the coordinates of each surrounding trigger node are calculated. The shortest distance value is taken as the proximity index between the boundary and the nearest trigger node. Spatial weighting coefficients are assigned according to the proximity: when the shortest distance is less than one-third of the characteristic side length of the bonding response unit, a high point weight is assigned; when the shortest distance is greater than one-half of the characteristic side length, a low point weight is assigned; and for distances in between, a median weight is assigned according to an arithmetic progression. Through the above operations, the boundaries of adjacent units closer to the elastic release area receive a larger spatial weight, while the boundaries farther from the trigger nodes receive a smaller spatial weight. The spatial weight coefficient is negatively correlated with the distance from the trigger node to the boundary of the corresponding fitting response unit, so as to perform spatial weight correction on the initial propulsion lag to obtain the corrected propulsion lag. The spatial weight coefficient assigned to each group of adjacent fitting response units in step two is multiplied by the initial propulsion lag calculated in step one to obtain the spatial weight corrected propulsion lag. In specific operation, if there are multiple trigger nodes near the boundary of an adjacent unit, the maximum value of the corresponding weight of each trigger node is used in the correction calculation to ensure that the node with the strongest elastic release influence dominates the correction direction. After correction, the initial propulsion lag and the corrected propulsion lag at the boundary of the same group of adjacent units are stored together. For areas with dense distribution of trigger nodes, the corrected propulsion lag is significantly greater than the initial value, reflecting the additional flow obstruction effect caused by elastic release in this area. For areas far from the trigger nodes, the corrected propulsion lag is basically the same as the initial value, reflecting no additional obstruction influence.

[0043] Using the corrected propulsion hysteresis as a flow hindrance constraint factor, a flow priority matrix of the colloid among different bonding response units is constructed. On the topological grid diagram of the bonding response units, the row and column order of the grid is used as the row and column labels of the matrix to establish a flow priority data table corresponding one-to-one with the boundaries of adjacent units. Before determining the priority of each group of adjacent bonding response units, the specific value of the system flow passage threshold is first determined through three sets of parallel calibration tests: three sets of cover plate and display module samples of the same specifications as the current bonding batch are selected, and three different depressurization rates are set for trial bonding under the same vacuum pressing conditions. During the bonding process, the aforementioned high-speed vision system is used to record the time and spatial position of the colloid front actually crossing the boundaries of each group of adjacent units. After bonding, each set of samples is cut open along the bonding plane, and the actual wetting and broadening of the colloid at the boundaries of each group of adjacent units is measured by optical microscope. The flow priority threshold is determined by taking the minimum flow priority value corresponding to the moment when the colloid can stably cross the boundary and expand for more than 30 seconds without retraction. This value is used as the calibration value for the flow passage threshold. For each group of adjacent bonding response units, the corrected propulsion hysteresis is compared with the system reference value: unit boundaries with a corrected propulsion hysteresis less than half of the system reference value are marked as high flow priority; unit boundaries with a corrected propulsion hysteresis more than twice the system reference value are marked as low flow priority; and those in between are marked as medium flow priority. The system reference value is the median value of the corrected propulsion hysteresis of all adjacent units in the three calibration tests. The high, medium, and low priority levels are assigned numerical flow priority coefficients to form a complete flow priority matrix data table. Each value in this data table represents the ease with which the colloid can propagate from one unit to the other at the corresponding boundary.

[0044] After constructing the flow priority matrix, a correction based on real physical field measurement data is applied to the matrix. Six miniature pressure sensors are evenly distributed around the bonding plane within the bonding cavity to collect the actual pressure values ​​at the boundaries of each extraction zone in real time. These values ​​are then compared with preset target pressure values ​​to obtain the actual air path damping deviation at each boundary. Simultaneously, twelve laser displacement sensors are arrayed on the back of the cover plate to measure the off-surface displacement of the cover plate surface at a sampling frequency of 2,000 times per second. For each bonding response unit boundary, off-surface displacement data from two sampling points on each side of that boundary are extracted. Second-order difference calculations are performed to obtain the change in the curvature of the cover plate at the boundary. The air path damping deviation at the same boundary is added to the change in the curvature of the cover plate to generate the comprehensive interface damping coefficient at the boundary. This coefficient is a dimensionless value, ranging from negative one to positive one. A positive value indicates that the air path resistance at the boundary is too high or the cover plate bulges and bends towards the bonding surface, hindering the propagation of the colloid in that direction. A negative value indicates that the air path resistance at the boundary is too low or the cover plate is concave and bends away from the bonding surface, which is conducive to the propagation of the colloid in that direction. The comprehensive interface damping coefficient is used to analyze the flow obtained in step four. The flow priority matrix is ​​adjusted point by point: Each boundary in the flow priority data table is checked. If the comprehensive interface damping coefficient of the boundary is greater than +0.5, the flow priority coefficient at that boundary is adjusted down one level. Specifically, if the current priority is high, it is reduced to medium priority; if the current priority is medium, it is reduced to low priority; if the current priority is low, it remains unchanged. If the comprehensive interface damping coefficient is less than -0.5, the flow priority coefficient at that boundary is adjusted up one level, i.e., low priority is upgraded to medium priority, medium priority is upgraded to low priority, and so on. High priority is maintained; if the overall interface damping coefficient is between -0.5 and +0.5, the flow priority coefficient of that boundary remains unchanged; where +0.5 and -0.5 are preset damping influence thresholds, which are set based on bonding process experience and represent the minimum limit at which the overall interface damping coefficient has a substantial impact on flow priority; after completing the correction of all boundaries, the flow priority matrix after double correction by the above measured air pressure and measured deformation data is obtained, which simultaneously couples the actual air path damping and the constraint effect of the cover plate morphology on the colloid spreading.

[0045] Constraint propagation analysis of the colloidal propagation path is performed based on the modified flow priority matrix to determine the dominant and restricted unidirectional flow paths, and the distribution of differences in colloidal flow paths among each bonding response unit is output. Starting from the bonding response unit where the initial injection point of the optical adhesive is located, the search expands outwards level by level along the modified flow priority matrix from step five. Specifically, starting from the starting unit, the values ​​corresponding to the four adjacent boundaries of that unit in the modified flow priority matrix are read. These values ​​are compared one by one with the flow passage thresholds obtained from the aforementioned calibration experiments. Adjacent units with priority values ​​greater than the thresholds are included in the first-level candidate set. The process begins by marking the boundary of a direction as passable and marking directions with priority values ​​less than or equal to a threshold as blocked boundaries, thus terminating the expansion in that direction. Then, starting from each unit in the first-level candidate set, the process checks the boundaries of its adjacent, unvisited units. New units corresponding to boundaries with priority values ​​greater than the threshold are added to the second-level candidate set, and so on, expanding outwards layer by layer until no new candidate units are generated during the expansion of a certain layer, at which point the entire search process terminates. After the search terminates, each visited fitting response unit recorded during the path search process is labeled with its level number, and this level number is then assigned to the appropriate level. The search results are verified against the timestamps of the colloidal front reaching the cell boundary as recorded by the high-speed vision system. If the layer number of a cell does not match the order of its actual arrival times, the priority values ​​of that cell are reordered according to the actual arrival times, and the search process is repeated until the layer numbers completely correspond to the measured time sequence, ensuring the physical authenticity of the search results. During the search, all actually traversed boundaries are marked as flow-connected boundaries, and all boundaries not selected due to low priority coefficients are marked as flow-blocking boundaries. After the search is completed, the timestamps of all connected and blocked boundaries are statistically analyzed. Spatial Distribution: Path chains passing through the top 40% of units are designated as flow-dominant paths, meaning the main spreading channels where the colloid propagates fastest and experiences least flow resistance during bonding. Path chains passing through the bottom 30% of units are designated as restricted flow paths, meaning localized channels where the colloid propagates slowest and is subject to both elastic release damping and interfacial damping. Path chains passing through the middle 30% of units are designated as conventional flow paths, meaning transitional channels where the colloid spreading speed is moderate and not significantly affected by resistance effects. The basis for these classification ratios is as follows: Pre-analyzing experimental data from 100 sets of bonding processes of the same specification revealed that the colloid spreading speed distribution follows a normal distribution. The standard deviation limit at the faster end of the speed distribution corresponds to the top 40% quantile, and the standard deviation limit at the slower end corresponds to the bottom 30% quantile. This statistical law has been entered into the bonding process database as a fixed benchmark for path classification.The boundary of complete blockage is marked separately as the blocking interface; finally, the above three types of paths are marked with different legends on the topology mesh diagram of the fitting response unit, and the differential distribution of colloidal flow paths is output as the basis for the subsequent coordinated configuration of the depressurization cycle of the pumping zone and the displacement sequence of the cover plate.

[0046] Step S40: Based on the differential distribution of the unidirectional confined flow path, the execution sequence of the depressurization cycle and the local propulsion displacement of the cover plate in each extraction zone is configured in a coordinated manner. The impedance convergence consistency threshold of each fitting response unit is used as the process constraint condition to perform dynamic correction control of the depressurization duration and local propulsion displacement amplitude of each extraction zone.

[0047] In step S40, the process of coordinating the execution sequence of the depressurization cycle and the local advance displacement of the cover plate in each extraction zone is as follows: Based on the differential distribution of colloidal flow paths, the flow resistance coefficients corresponding to each bonding response unit are calculated, and a topological connection matrix of flow dependence relationships between each bonding response unit is constructed. The results of the differential distribution of colloidal flow paths across the entire domain, obtained from the previous analysis, are retrieved. Combined with the dominant flow state, restricted resistance state, and propulsion lag degree of each bonding response unit, the flow resistance characteristics of each bonding response unit are quantitatively evaluated. Considering multiple operating conditions such as the magnitude of unit interface damping, the degree of colloidal spreading lag, and the strength of elastic release node constraints, the flow resistance coefficients corresponding to each unit are uniformly calculated, accurately characterizing the degree of obstruction of optical adhesive flow in different regions. Using all bonding response units as topological nodes, and the colloidal flow transmission relationship and the difference in resistance coefficients between adjacent units as connection weights, the correlation logic of positive flow dependence and reverse resistance constraints between units is sorted out, and a topological connection matrix of flow dependence relationships covering the entire bonding area is constructed. This matrix can completely record the mutual flow constraints between units across the entire domain, clarify the colloidal conduction priority and resistance constraint level in local areas, and provide a structured data foundation for subsequent pressure relief priority ranking.

[0048] Based on the weight distribution of the hindrance coefficient of each bonding response unit in the topology connection matrix, a pressure relief priority sequence is generated. The weight distribution of the hindrance coefficient of all units within the flow dependency topology connection matrix is ​​traversed to uniformly sort and classify the flow hindrance levels of the bonding response units across the entire domain. For regions with high hindrance coefficients, severely restricted colloid spreading, and prone to colloid buildup and uneven filling, a higher pressure relief priority is set to relieve the deformation constraints caused by negative pressure adsorption at the interface through early and sufficient pressure relief. For regions with low hindrance coefficients, smooth colloid flow, and sensitive filling response, a relatively lower pressure relief priority is set to avoid premature pressure relief causing colloid turbulence due to sudden changes in interface micro-gap. According to the global weight ranking rules, a partitioned pressure relief priority sequence adapted to the entire screen bonding state is generated, achieving differentiated matching of pressure relief timing based on colloid flow characteristics.

[0049] Based on the pressure relief priority sequence, the pressure relief duration and rate of each extraction zone are configured in a graded manner to form zone pressure relief cycle parameters. Strictly following the generated pressure relief priority sequence, differentiated and graded pressure relief parameter configurations are implemented for each corresponding extraction zone of the vacuum fixture. For zones with high priority and severe flow obstruction, the pressure relief duration is appropriately extended and the pressure relief rate is slowed down to ensure stable release of local vacuum pressure, avoiding sudden pressure changes that could cause instantaneous deformation of the cover plate, and providing a stable interface environment for the slow and uniform spread of colloid in confined areas. For zones with low priority and smooth and stable flow, conventional pressure relief duration and standard pressure relief rate are adopted to ensure synchronized progress of the overall bonding process. Through targeted graded parameter tuning, the pressure relief start time, pressure relief duration, and pressure relief termination rhythm of each zone are standardized, ultimately forming standardized zone pressure relief cycle parameters adapted to the flow characteristics of each area, achieving refined zone control of the pressure relief process.

[0050] Based on the flow direction vector distribution of each bonding response unit, a sequence of local propulsion displacement directions of the cover plate is generated. Combining the distribution of differences in the global colloid flow path, the mainstream optical adhesive spreading direction, lateral diffusion trend, and restricted sealing direction of each bonding response unit are statistically analyzed, and the colloid flow direction distribution characteristics of each unit are summarized. According to the mainstream direction of natural colloid spreading, the displacement guidance direction of local micro-propulsion of the cover plate is matched and set, so that the micro-pressing and propulsion action of the cover plate conforms to the natural wetting and expansion law of the colloid, avoiding the problems of shear turbulence and uneven colloid thickness caused by the artificial forced pressing direction contradicting the colloid flow direction. According to the spatial arrangement order and flow conduction logic of each unit, the sequential orientation relationship of local propulsion in the global domain is sorted out in turn, and a continuous and orderly sequence of local propulsion displacement directions of the cover plate is generated, providing a directional reference for the timing arrangement of mechanical pressing action.

[0051] Based on the temporal correspondence between the pressure relief cycle parameters and the propulsion displacement direction sequence, the execution sequence of the cover plate's local propulsion displacement is determined, and the pressure relief execution cycle of each extraction zone is synchronously matched and configured. The temporal rhythm of the zone pressure relief cycle parameters is matched one-to-one with the action logic of the cover plate's local propulsion displacement direction sequence, establishing a temporal linkage mechanism between pressure relief release and mechanical micro-propulsion. Taking the zone pressure relief progress as a prerequisite, after the high-resistance zone completes the initial stable pressure relief and the interface constraint is initially released, the corresponding area's cover plate local propulsion action is executed synchronously. The step-by-step execution sequence of the cover plate's micro-propulsion in each area is determined according to the order of pressure relief, realizing the linkage logic of pressure relief unlocking first and displacement compaction last. At the same time, according to the execution sequence of the local propulsion action, the pressure relief closing cycle of the corresponding extraction zone is finely adjusted in reverse, so that the pressure relief release rhythm is completely synchronized with the mechanical propulsion amplitude and propulsion timing, thoroughly solving the problem of colloidal flow imbalance caused by asynchronous zone pressure relief and mismatched compression propulsion, and completing the coordinated and integrated configuration of the entire domain pressure relief cycle and the cover plate's local propulsion action.

[0052] In step S40, the process of dynamically correcting and controlling the depressurization duration and local propulsion displacement amplitude of each extraction zone is as follows: The system collects the migration impedance sequence data of each bonding response unit in real time, calculates the impedance convergence rate of change based on a time window, and generates the corresponding convergence error vector. During the dynamic process of continuous bonding and pressing of the entire machine and the depressurization of the zones, the system collects continuous migration impedance time-series data of each bonding response unit online in real time, maintaining a fixed sampling frequency throughout to ensure synchronous, continuous, and uninterrupted updates of data from each unit. A sliding time window approach is used to segment and analyze the real-time impedance sequence, with each time window covering a fixed number of continuous sampled data points to stably characterize the local interface impedance at the current moment. Convergence status: By comparing the amplitude and stability of impedance data changes before and after the window, the real-time impedance convergence rate of each bonding response unit is calculated to determine whether the coupling state between gas path damping and interface deformation in the current region is stabilizing. The actual convergence status of each unit is compared with the standard stable convergence status item by item, and the real-time convergence deviation of each unit is extracted. The deviations are then integrated according to the spatial arrangement of the entire domain to form a complete global convergence error vector, which centrally reflects the magnitude and trend of the current impedance convergence deviation in each bonding region, providing original data support for subsequent partition deviation determination.

[0053] The convergence error vector is compared with the calibrated consistency threshold to construct a partition deviation judgment identifier matrix; the impedance convergence consistency threshold stored in the equipment system is retrieved; the threshold is determined as follows: before the start of batch bonding production, five sets of standard cover plates and display module samples with the same specifications as the current product are selected. Each set of samples uses the benchmark bonding formula recorded in the equipment process parameter library, including standard vacuum degree, standard pressing speed and standard pressure relief curve, for trial bonding. During the bonding process, the migration impedance time series data of each bonding response unit is collected throughout the process at the same sampling frequency as in step S40; after bonding, the actual impedance value of each bonding response unit interface of the five sets of samples is measured sequentially through an offline impedance analyzer. The offline measurement value is compared and fitted with the final value of the time series impedance collected online. The final impedance value of all bonding response units in the five sets of samples whose impedance value is within ±5% of the stable fluctuation zone at the time of bonding completion is taken as the arithmetic mean. This average value is determined as the impedance convergence consistency threshold; once this threshold is determined... The threshold is stored in the equipment process parameter database as the benchmark control parameter for the current product specification. Subsequent batches of products of the same specification can directly call this threshold for comparison and judgment without recalibrating before each batch. The threshold is only updated according to the above five calibration procedures when the product specification or bonding adhesive material changes. The deviation data of each bonding response unit in the real-time generated global convergence error vector is compared with the consistency threshold obtained by calibration to distinguish the qualified area where the convergence deviation is within the allowable range and the abnormal area where the deviation exceeds the limit. The deviation degree, deviation type and positive and negative deviation trend of each unit are classified and marked. According to the spatial topological distribution relationship of each bonding response unit, the judgment results of all units are structured and integrated to form a partition deviation judgment identification matrix covering the entire bonding area. The matrix can intuitively distinguish the convergence qualified status and abnormal level of each area in the entire domain, accurately locate the problem areas of uneven impedance convergence and interface state imbalance, and provide a judgment basis for differential parameter correction.

[0054] Based on the judgment identifier matrix, the depressurization duration parameter of each extraction zone is incrementally corrected, and the depressurization control timing of the corresponding zone is updated. According to the overall judgment results of the zone deviation judgment identifier matrix, targeted incremental correction control is implemented for extraction zones in different states. For qualified zones with small convergence deviations and stable contact, the original depressurization duration parameter remains unchanged to maintain the current stable depressurization cycle. For abnormal zones with large convergence deviations, unstable impedance, and incomplete interface deformation, the depressurization duration is increased or finely adjusted according to the deviation level, using a small incremental gradual correction method to gradually compensate for local depressurization deficiencies, avoiding process fluctuations caused by large-scale parameter adjustments at once. After completing the incremental parameter correction, the complete depressurization control timing of the corresponding extraction zone is simultaneously refreshed, and the depressurization start node, pressure stabilization duration, and depressurization end node of the zone are recalibrated, so that the impedance state of abnormal zones gradually approaches the standard convergence range, achieving a consistent balance of the gas path damping state across all zones.

[0055] The local interface impedance change rate is synchronously acquired and mapped to a displacement correction coefficient according to the piecewise linear mapping rule. The measured impedance change rate of each bonding response unit is compared with the standard impedance change rate stored in the process database, and the rate deviation ratio between the two is calculated, which is the measured rate minus the standard rate and divided by the standard rate. If the rate deviation ratio is between -10% and +10%, it indicates that the impedance change rate in this area is in a stable range. In this case, the displacement correction coefficient is set to 1.0, that is, the advancing displacement amplitude in this area is not changed. If the rate deviation ratio is greater than +10%, it indicates that the impedance change in this area is too fast and the interface state tends to be unstable. In this case, the displacement correction coefficient is set to 1.0 plus the product of the rate deviation ratio and the conversion coefficient, where the conversion coefficient is... The coefficients are obtained by calibration based on the piezoelectric response characteristics of the displacement actuator. The calibration method is as follows: During the equipment commissioning phase, a set of standard voltage step signals are applied to the displacement actuator, and the correspondence between the actual output displacement and the control voltage is measured. The slope of the displacement relative to the voltage change is taken as the conversion coefficient. Once calibrated, this coefficient is used in the same type of equipment. If the rate deviation ratio is less than negative 10%, it indicates that the impedance change in this area is too slow and the interface is not sufficiently pressed. At this time, the displacement correction coefficient is set to 1.0 minus the product of the absolute value of the rate deviation ratio and the conversion coefficient, and the lower limit is not lower than 0.85 to prevent over-correction from causing local pressing failure of the cover plate. Through the above piecewise linear mapping rule, any measured impedance change rate is uniquely mapped to a displacement correction coefficient value.

[0056] The local advancing displacement amplitude of the cover plate is adjusted in stages based on the displacement correction coefficient; Regions with a correction coefficient greater than 1.10 are identified as fluctuating regions. The cover plate displacement amplitude in these regions is increased by one increment from the original step value, i.e., the minimum controllable step of one displacement actuator is increased. Regions with a correction coefficient between 0.90 and 1.10 are identified as stable regions, and the original displacement amplitude remains unchanged. Regions with a correction coefficient less than 0.90 are identified as lagging regions, and the displacement amplitude in these regions is decreased by one increment from the original step value, i.e., the minimum controllable step of one displacement actuator is reduced. Through dual dynamic correction of pressure relief timing and displacement amplitude, the closure progress of interface gaps, colloid spreading state, and deformation rebound state of each region are kept highly synchronized, ultimately achieving adaptive balance convergence of the entire bonding interface.

[0057] Step S50: Based on the bonding interface state after partition dynamic correction control, collect the optical adhesive layer thickness distribution, display area edge alignment deviation and pixel mapping error distribution. When any parameter exceeds the limit, locate the corresponding bonding response unit, and correct the process control parameters based on the pressure relief cycle and migration impedance parameters, and write them into the vacuum bonding fixture controller.

[0058] In step S50, the process of correcting the process control parameters based on the pressure relief cycle time and migration impedance parameters is as follows: After lamination, the entire screen is subjected to multi-point optical scanning to obtain the optical adhesive layer thickness distribution matrix and edge alignment deviation data. Before the LCD display is vacuum-pressed and laminated, and the optical adhesive has completed initial wetting and spreading but has not yet entered the curing stage, a high-precision optical scanning and inspection device is activated to perform full-coverage, high-density multi-point scanning and acquisition of the entire screen lamination area, avoiding data omissions caused by traditional local sampling inspections. The testing equipment follows a preset uniform dot matrix sampling strategy, simultaneously covering the effective display area in the center of the screen and the bonding areas at the four edges. It collects the actual thickness values ​​of the optical adhesive layer at each location, completely recording the differences in adhesive layer thickness, thinness, and local missing or piled adhesive in different areas. Through the integration and processing of the full-area point data, it generates an optical adhesive layer thickness distribution matrix corresponding to the entire screen area, intuitively reflecting the uniformity distribution characteristics of the adhesive layer. At the same time, it performs high-precision image detection on the four-sided bonding alignment position of the cover plate and the LCD module, collecting bonding offset and alignment gap deviation data of each edge, and summarizing them to form a complete edge alignment deviation dataset. This dataset comprehensively records the defect information such as misalignment and edge offset generated during the bonding process, providing a complete and accurate measured data foundation for subsequent error analysis and process traceability.

[0059] Pixel mapping errors are analyzed by regional segmentation to locate the bonding response units corresponding to areas of concentrated error. After bonding, the entire device's pixel imaging inspection data is retrieved. For common display errors in finished products, such as pixel mapping offset, uneven local brightness, fine line shift, and pixel misalignment, a refined regional segmentation analysis mode is adopted, abandoning the traditional coarse analysis method of average judgment across the entire screen. Combining the spatial boundaries of each bonding response unit already defined in the previous process flow, the pixel error data of the entire screen is divided and categorized according to the spatial range of each unit. The error distribution density, error offset amplitude, and frequency of anomalies within each bonding response unit are calculated one by one, accurately distinguishing between overall global errors and localized regional errors. Through horizontal comparison of data from each unit, abnormal areas with concentrated errors and exceeding error standards are screened out, and the corresponding bonding response units are accurately matched. This achieves precise location and correspondence between display defects and local bonding units, completely solving the problem that traditional inspection can only detect defects but cannot accurately locate abnormal process areas.

[0060] The abnormal bonding response unit is associated with the corresponding depressurization cycle parameters and migration impedance parameters for tracing. Using the located abnormal bonding response unit as the tracing target, the partition process log data stored throughout the bonding process is retrieved to accurately match the depressurization cycle parameters of the pumping partition corresponding to the abnormal unit with the unit's unique migration impedance parameters. The depressurization start-up sequence, depressurization duration, depressurization gradient rate, and other cycle execution states of the unit during the entire pressing and depressurization process are checked one by one. At the same time, the unit's gas path damping characteristics, pressure and deformation coupling response characteristics, impedance convergence and stabilization states, and other impedance characteristic parameters are reviewed. By comparing multi-dimensional data, we can distinguish the types of defects and determine whether the defects are caused by asynchronous depressurization timing, unbalanced depressurization rhythm leading to disordered colloid flow and uneven spreading, or by abnormal local migration impedance, delayed interface deformation response, and asynchronous elastic release leading to unbalanced bonding interface. We can establish a one-to-one correspondence between display error, adhesive layer thickness deviation, edge alignment offset and abnormal process parameters to accurately pinpoint the shortcomings of process parameters and the root causes of process disturbances.

[0061] Based on the source tracing results, the partition control parameters in the vacuum bonding fixture controller are updated and written. According to the defect causes and parameter deviation magnitudes obtained from the previous parameter tracing, a targeted partition parameter correction scheme is formulated. Precise parameter optimization is performed on the evacuation partition corresponding to the abnormal bonding response unit, while the process parameters of the normal bonding area remain unchanged to avoid process fluctuations in benign areas caused by uniform parameter adjustment across the entire area. For partitions with mismatched depressurization cycles, the depressurization duration, depressurization rate, and timing matching nodes are fine-tuned. For areas with unbalanced migration impedance and lag in deformation response, the local pressing advancement amplitude and depressurization matching constraint threshold are optimized. After completing the incremental parameter correction and adaptation verification, the optimized partition depressurization parameters, impedance convergence control parameters, and local displacement control parameters are packaged together and stably written into the controller storage module of the vacuum bonding fixture through the equipment communication protocol, overwriting the original old parameters and completing parameter solidification. Subsequent mass production processes will automatically call the updated optimized parameters to execute partition bonding control, realizing closed-loop correction of defects and process self-iterative optimization, continuously improving the overall bonding consistency and mass production yield.

[0062] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for manufacturing an LCD liquid crystal display screen, characterized in that: Includes the following steps: The graded decompression time constants and local flexural response displacement sequences of each evacuation zone of the vacuum bonding fixture were collected. The decompression rate and flexural response time sequence were aligned and mapped to establish the migration impedance parameters characterizing the air path damping characteristics of each evacuation zone. The bonding response units were divided according to the variation trend of the migration impedance parameters between adjacent evacuation zones. During the controlled pressing stage of the cover plate relative to the display module, the gap closing rate, local deflection change and optical adhesive front advancement displacement trajectory of each bonding response unit are collected. The continuous advancement trajectory of the optical adhesive front is used as a unified timing reference to perform spatial consistency correction on the gap closing rate and local deflection change, and generate interface migration impedance evolution data of each bonding response unit. Second-order flexural change analysis was performed on the interface migration impedance evolution data. When the local second-order flexural change rate changed from negative to positive and continued to decrease, the corresponding bonding response unit was identified as the local elastic release trigger node of the cover plate. Based on the spatial distribution of the trigger node, the propulsion hysteresis of the optical adhesive front edge between adjacent bonding response units was spatially weighted and corrected. The corrected propulsion hysteresis was used as the unidirectional flow boundary constraint condition, and the difference distribution of the unidirectional restricted flow path of the adhesive between each bonding response unit was analyzed. Based on the differential distribution of unidirectional confined flow paths, the execution sequence of depressurization cycle and local propulsion displacement of cover plate in each extraction zone is configured in a coordinated manner. The impedance convergence reference value obtained by calibration of each bonding response unit is used as the process constraint condition to perform dynamic correction control of depressurization duration and local propulsion displacement amplitude in each extraction zone. Based on the bonding interface state after partitioned dynamic correction control, the optical adhesive layer thickness distribution, display area edge alignment deviation and pixel mapping error distribution are collected. When any parameter exceeds the limit, the corresponding bonding response unit is located, and the process control parameters are corrected based on the pressure relief cycle and migration impedance parameters, and written into the vacuum bonding fixture controller.

2. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of establishing the migration impedance parameters characterizing the gas path damping properties of each extraction zone is as follows: The pressure decay curves of each evacuation zone of the vacuum bonding fixture during the staged depressurization process were obtained, and the corresponding depressurization time constants were extracted using an exponential fitting method. The local flexural displacement sequence of the cover plate under the action of each air extraction zone was collected synchronously, and the pressure decay sequence and displacement response sequence were time-stamped and aligned based on a unified sampling frequency. The aligned pressure change gradient and flexural response amplitude are normalized, and the dynamic response ratio coefficient between the two is calculated. Using the phase lag mapping relationship between the local flexural response of the cover plate and the pressure attenuation process as a constraint, an inter-regional pressure and deformation coupling consistency correction term is introduced to correct the spatiotemporal consistency of the dynamic response ratio coefficient. Based on the dynamic response ratio coefficient, the gas path damping characterization function of each extraction zone is constructed, and the coupling relationship between pressure decay rate and structural flexural response is parameterized to form migration impedance parameters.

3. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of dividing the fitted response unit based on the variation trend of the migration impedance parameter between adjacent pumping zones is as follows: The migration impedance parameters corresponding to each extraction zone are serialized and rearranged according to the spatial topological adjacency relationship to construct a continuous impedance distribution sequence. Based on the difference in migration impedance parameters between adjacent extraction zones, the local impedance gradient and the acceleration of change are calculated. The impedance gradient change process is equivalent to the vacuum path equivalent damped propagation process, and an impedance propagation delay factor is introduced to correct the difference between adjacent partitions. When the local impedance gradient changes from a steady change to a discontinuous transition state, and the acceleration of the change exceeds the preset structural sensitivity threshold, the corresponding location is determined as the interface impedance fracture boundary point. Using boundary points as constraints, the bonding area of ​​the display module is reconstructed in a structured segmentation to form multiple bonding response units.

4. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of generating the interface migration impedance evolution data for each mating response unit is as follows: During the controlled pressing process, the real-time gap change sequence of each bonding response unit and the local deflection response sequence of the cover plate are collected. Using the spatial position of the optical adhesive leading edge trajectory as the main reference coordinate, the sampling data of different bonding response units are resampled and reconstructed in a spatiotemporal manner. Phase alignment correction is performed on response data from different regions; Among them, the phase alignment correction uses the activation moment of local interface contact caused by the advancement of the optical adhesive front edge as the trigger reference; Based on the corrected gap closing rate and deflection change, the interface compression response evolution coefficient per unit time is calculated to form interface migration impedance evolution data.

5. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of identifying the corresponding bonding response unit as the cover plate local elastic release trigger node is as follows: Second-order difference calculations were performed on the interface migration impedance evolution data to obtain the local flexural acceleration variation sequence. Analyze the sign transition points of the second-order rate of change of flexure and identify the inflection point region where the value changes from negative to positive. A duration window analysis was performed on the inflection point region. When the positive rate of change showed a decreasing trend within the continuous sampling period, it was determined that the inflection point region changed from a compression-dominated state to an elastic release state. The bonding response unit that meets the above conditions is marked as a local elastic release trigger node.

6. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of analyzing the difference distribution of the colloidal unidirectional confined flow path between each bonding response unit is as follows: Extract the propulsion time difference and spatial displacement difference of the optical adhesive front edge between adjacent bonding response units, and calculate the propulsion hysteresis. Based on the spatial distribution of the triggering nodes, spatial weighting coefficients are assigned to the propulsion lag between each adjacent mating response unit; The spatial weight coefficient is negatively correlated with the distance from the trigger node to the boundary of the corresponding fitting response unit, so as to perform spatial weight correction on the initial propulsion lag and obtain the corrected propulsion lag. Using the corrected propulsion hysteresis as a flow retardation constraint factor, a flow priority matrix of colloids is constructed among different bonding response units. After completing the construction of the flow priority matrix, a correction based on real physical field measurement data is applied to the matrix. Constraint propagation analysis of the colloidal propagation path is performed based on the modified flow priority matrix to determine the dominant and restricted unidirectional flow paths and output the distribution of differences in colloidal flow paths among each bonding response unit.

7. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of coordinating the execution sequence of depressurization cycle and local advance displacement of cover plate in each extraction zone is as follows: Based on the differential distribution of colloidal flow paths, the flow retardation coefficient corresponding to each bonding response unit is calculated, and the topological connection matrix of flow dependence between each bonding response unit is constructed. Based on the weight distribution of the hysteresis coefficients of each fitted response unit in the topology connection matrix, a pressure relief priority sequence is generated. Based on the pressure relief priority sequence, the pressure relief duration and pressure relief rate of each extraction zone are configured in a graded manner to form the zone pressure relief cycle parameters; Based on the flow direction vector distribution of each bonding response unit, a sequence of local propulsion displacement directions of the cover plate is generated; Based on the temporal correspondence between the depressurization cycle parameters and the propulsion displacement direction sequence, the execution sequence of the local propulsion displacement of the cover plate is determined, and the depressurization execution cycle of each extraction zone is synchronously matched and configured.

8. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of dynamically correcting and controlling the depressurization duration and local propulsion displacement of each extraction zone is as follows: The migration impedance sequence data of each mating response unit is collected in real time, and the impedance convergence change rate is calculated based on the time window to generate the corresponding convergence error vector. The convergence error vector is compared with the calibrated consistency threshold to construct a partition deviation judgment identifier matrix; The pressure relief duration parameter of each extraction zone is incrementally corrected based on the judgment identifier matrix, and the pressure relief control timing of the corresponding zone is updated. The rate of change of local interface impedance is synchronously acquired and mapped to a displacement correction coefficient according to the piecewise linear mapping rule; The local advancing displacement amplitude of the cover plate is adjusted in stages based on the displacement correction coefficient.

9. The LCD liquid crystal display screen bonding manufacturing method according to claim 1, characterized in that, The process of correcting process control parameters based on pressure relief cycle time and migration impedance parameters is as follows: After bonding is completed, the entire screen is subjected to multi-point optical scanning to obtain the optical adhesive layer thickness distribution matrix and edge alignment deviation data. Perform regional segmentation analysis on pixel mapping errors to locate the corresponding bonding response units in the concentrated error areas; The abnormal fit response unit is associated with the corresponding pressure relief cycle parameters and migration impedance parameters for tracing the source. Based on the tracing results, the partition control parameters in the vacuum bonding fixture controller are updated and written.