Liquid crystal module hot-pressing binding method based on intelligent temperature control

Abstract

The invention relates to the technical field of display panel manufacturing, in particular to a liquid crystal module hot-pressing binding method based on intelligent temperature control, which comprises the following steps: step 1, hot-pressing area partitioning and dynamic temperature calibration: dividing a contact surface of a hot-pressing head into a plurality of independently controlled temperature areas, each area is provided with at least one temperature sensor, and the temperature sensors in all the areas are calibrated through thermal imaging equipment, so that the deviation between the sensors and the real temperature is eliminated; 2, dynamic temperature compensation and pressure cooperative control are carried out; step 3, iteratively optimizing hot pressing process parameters, and training the deep learning model based on historical binding data to predict an optimal temperature-pressure parameter combination of a next binding period; and automatically correcting the initial temperature threshold in the step 1 and the target pressure in the step 2 according to a prediction result. Through the intelligent temperature control and pressure cooperative control technology, the temperature and pressure in the hot pressing process of the liquid crystal module are accurately adjusted, and the bonding quality and the production efficiency are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of display panel manufacturing, and in particular to a liquid crystal module hot pressing binding method based on intelligent temperature control. Background Art

[0002] The hot pressing bonding process for liquid crystal display modules (LCD / OLED) is a core step in display panel manufacturing. It uses hot pressing equipment to precisely bond driver chips (such as COF / FPC) to LCD panel electrodes. With the rapid development of flexible display technologies (such as foldable and curved screens) and ultra-thin LCD modules, the traditional hot pressing process has gradually exposed the following technical defects:

[0003] Existing hot pressing devices often rely on a single temperature sensor for global temperature control and are unable to dynamically respond to temperature differences in different areas of the hot pressing head. For example, the edge areas of special-shaped screens are prone to underheating due to faster heat dissipation, while the center area may overheat due to heat accumulation, causing thermal degradation of the adhesive layer or poor soldering of the electrodes. In addition, traditional temperature control algorithms (such as fixed PID parameters) are difficult to adapt to the dynamic heat conduction characteristics of complex geometric structures, resulting in temperature fluctuations exceeding the process tolerance range of ±3°C.

[0004] During hot pressing, adhesive layer viscosity is strongly correlated with temperature and pressure, but existing equipment lacks a dynamic temperature control mechanism with integrated pressure feedback. For example, during hot pressing of flexible substrates (such as PI film), uneven pressure distribution can lead to insufficient flow in some areas of the adhesive layer. Conventional equipment cannot adjust the temperature threshold in real time based on pressure to optimize the rheological behavior of the adhesive layer, ultimately affecting bonding uniformity. Summary of the Invention

[0005] Based on the above objectives, the present invention provides a liquid crystal module hot pressing binding method based on intelligent temperature control, comprising the following steps:

[0006] Step 1: Hot-pressing area zoning and dynamic temperature calibration: The contact surface of the hot-pressing head is divided into multiple independently controlled temperature zones, each equipped with at least one temperature sensor. Based on the material type and geometry of the liquid crystal module, the initial temperature threshold of each zone is determined through a pre-calibration experiment. The pre-calibration experiment includes performing multiple sets of hot-pressing experiments on test pieces of the target material, recording the electrode adhesion strength and adhesive layer morphology at different temperatures. Based on the relationship curve between adhesion strength and temperature, the equilibrium temperature between the adhesive layer fluidity and thermal degradation threshold is selected as the initial threshold. The temperature sensors in each zone are calibrated using thermal imaging equipment to eliminate deviations between the sensors and the actual temperature.

[0007] Step 2: Dynamic temperature compensation and pressure coordinated control, real-time collection of temperature data from each area calibrated in step 1, and calculation of the deviation from the target temperature; dynamic adjustment of heating power based on the deviation, specifically using a fuzzy logic algorithm to adjust the heating power when the deviation is greater than a first set range, and switching to a linear proportional integral algorithm to adjust the heating power when the deviation is less than or equal to the first set range; synchronous collection of pressure distribution data from each area during the hot pressing process, the pressure distribution data being acquired in real time by a pressure sensor array disposed below the hot pressing table; adjustment of the temperature threshold of the corresponding area based on the real-time pressure value, including increasing the temperature threshold to reduce the viscosity of the adhesive layer when the pressure is below the target pressure range, and decreasing the temperature threshold to prevent excessive extrusion of the adhesive layer when the pressure is above the target pressure range;

[0008] Step 3: Iteratively optimize the hot pressing process parameters. Input the temperature data and pressure distribution data collected in Step 2 into the deep learning model. The deep learning model is trained based on the historical binding data to predict the optimal temperature-pressure parameter combination for the next binding cycle. The initial temperature threshold in Step 1 and the target pressure in Step 2 are automatically corrected based on the prediction results.

[0009] Preferably, the calibration method of the temperature sensor in step 1 is specifically as follows: when the thermal press head is not in operation, the surface temperature distribution of the thermal press head is scanned by an external thermal imager, the two-dimensional temperature field data collected by the thermal imager is compared with the single-point measurement data of each temperature sensor, and the offset of each sensor at different temperatures of the thermal press head is calculated; the offset is quadratically compensated according to the thermal expansion coefficient of the thermal press head material, and the compensation method includes calculating the influence of the deformation variable of the thermal press head during the heating process on the sensor installation position through finite element simulation, establishing a correlation function between the deformation variable and the temperature change, and dynamically adjusting the correction coefficient of the offset according to the current temperature; the calibrated sensor data is eliminated from random noise interference by the Kalman filter algorithm, the input of the Kalman filter is the raw data of the sensor and the rate of change of the heating power of the thermal press head, and the output is the corrected temperature feedback signal.

[0010] Preferably, the adjustment strategy of the fuzzy logic algorithm in step 2 is specifically as follows: the input variables are the temperature deviation value and the deviation change rate, the temperature deviation value is the difference between the current temperature and the target temperature, and the deviation change rate is the temperature deviation difference between two adjacent samples divided by the time interval; the output variable is the adjustment amplitude and direction of the heating power, the adjustment amplitude is determined according to the temperature deviation level defined by the membership function, the membership function divides the deviation value into three nonlinear intervals of high, medium and low, and assigns a weight coefficient to each interval; the adjustment instruction is generated by an empirical rule library, and the empirical rule library constructs a mapping relationship between pressure and temperature based on the temperature deviation and the heating power adjustment amount in the historical hot pressing data; the fuzzy logic algorithm introduces a delay compensation mechanism in the adjustment process, and the delay time is dynamically calculated according to the thermal conductivity of the material of the hot pressing head and the current heating power change rate to avoid overshoot caused by temperature lag.

[0011] Preferably, the method for establishing the mapping relationship between pressure and temperature in step 2 is specifically as follows: in the pre-calibration stage, different pressure combinations are applied to the test piece of the same material, the diffusion process of the adhesive layer during the hot pressing process is captured by a high-speed camera device, the movement trajectory of the edge of the adhesive layer is extracted and the displacement per unit time is calculated as the flow rate; at the same time, the pixel ratio of the contact area between the adhesive layer and the electrode is identified as the coverage area through an image processing algorithm; the coordinated adjustment ratio of pressure and temperature is determined according to the ratio of flow rate to coverage area, and the coordinated adjustment ratio is established by fitting the experimental data to establish a three-dimensional relationship surface of pressure-temperature-adhesive layer morphology, and the corresponding temperature compensation amount is obtained by interpolation on the surface according to the current pressure value.

[0012] Preferably, the training method of the deep learning model in step 3 is specifically as follows: the input data includes the temperature data and pressure distribution data collected in step 2, and the output data is the temperature threshold and target pressure recommended for the next cycle; the model uses a convolutional neural network to extract the temperature-pressure spatiotemporal characteristics, and predicts the parameter optimization direction through a long short-term memory network; the input data is normalized during the training process, and the normalization method includes scaling the temperature data to the range of 0-1 according to the maximum temperature resistance value of the thermal pressure head, and scaling the pressure distribution data to the range of 0-1 according to the upper limit of the sensor range; the loss function of the model is defined as the mean square error between the predicted parameters and the actual optimal parameters, and the optimizer uses an adaptive learning rate algorithm to dynamically adjust the gradient descent step size; after the training is completed, the generalization ability of the model is evaluated by cross-validation, and the overfitting phenomenon is suppressed by the discard method.

[0013] Preferably, the method for collecting pressure distribution data in step 2 is specifically as follows: a pressure sensor array is set under the hot pressing table, and the array density matches the number of hot pressing head partitions; the analog signal of the pressure sensor is converted into a digital quantity through a signal amplifier, and the gain of the signal amplifier is dynamically adjusted according to the sensitivity of the pressure sensor; the converted digital signal is filtered out of high-frequency noise interference through a digital filtering algorithm, and the filtering algorithm includes Fourier transforming the original signal and then cutting off the components higher than the set frequency, and then reconstructing the pressure signal through inverse transformation; the reconstructed pressure signal is used to generate a continuous pressure distribution thermodynamic map through a spatial interpolation algorithm, and the pressure abnormality area is identified according to the thermodynamic map.

[0014] Preferably, the pre-calibration experiment in step 1 also includes dividing the priority areas of the thermal head according to the geometric structure of the liquid crystal module; for the edge area of the curved surface or special-shaped structure, adding a compensation temperature value based on the initial temperature threshold; the method for determining the compensation temperature value includes predicting the heat dissipation path of the contact area between the thermal head and the module through finite element thermal conduction simulation, and calculating the temperature loss in the edge area due to heat dissipation; the simulation input parameters include the thermal conductivity of the thermal head material, the environmental convection coefficient and the specific heat capacity of the module substrate, and the boundary conditions are set to a constant temperature of the thermal head heating surface and the ambient temperature; the error between the simulation result and the actual measurement data is corrected by iteratively correcting the thermal conductivity parameters until convergence.

[0015] Preferably, the method for controlling the viscosity of the adhesive layer in step 2 is specifically as follows: when the pressure is lower than the target range, the temperature is increased to a first range higher than the glass transition temperature of the adhesive layer; the glass transition temperature is determined by measuring the phase change curve of the adhesive layer material with a differential scanning calorimeter and extracting the starting point corresponding to the endothermic peak; when the pressure is higher than the target range, the temperature is decreased to a second range lower than the thermal degradation temperature of the adhesive layer; the thermal degradation temperature is determined by the starting point of the exothermic peak of the phase change curve; the temperature ranges are set in a pre-calibration stage to verify the rheological properties of the adhesive layer at different temperatures, to ensure that the decrease in the viscosity of the adhesive layer after the temperature is increased meets the flow and filling requirements.

[0016] Preferably, the triggering conditions for the iterative optimization of parameters in step 3 include detecting that the continuous binding yield is lower than a set threshold or that the ambient temperature and humidity changes exceed a preset tolerance range; the statistical method for the yield is to automatically analyze the electrode morphology after binding through optical detection equipment to identify defects such as cold solder joints, short circuits, and incomplete adhesive coverage; the monitoring of the ambient temperature and humidity changes is carried out by real-time data collection through high-precision sensors, and parameter optimization is triggered when the temperature and humidity fluctuations exceed the standard deviation of the historical average; the optimization process includes reloading pre-calibrated experimental data and updating the input feature weights of the deep learning model, giving priority to adjusting parameters that are sensitive to the current environment.

[0017] Preferably, it also includes a step of automatic defect repair after hot pressing: identifying the cold solder joint or short circuit defects in the binding area through optical detection equipment, locating the defective area and triggering local secondary hot pressing; the specific method of the secondary hot pressing is to increase the temperature to a percentage range of the original set value, and reduce the pressure to a percentage range of the original set value; the basis for determining the percentage range is the size and position of the defective area, and the correspondence between the defect size and the temperature / pressure adjustment amount is established through preliminary experiments; during the secondary hot pressing, the temperature control signal of the normal area is shielded, and only the adjusted parameters are applied to the defective area.

[0018] Beneficial effects of the present invention:

[0019] Through intelligent temperature control and pressure coordination, the accuracy and consistency of hot-press bonding of LCD modules are effectively improved. By zoning the hot-pressing area and dynamically calibrating the temperature, the temperature of each area can be precisely adjusted, avoiding the temperature unevenness and excessive extrusion of the adhesive layer common in traditional hot-pressing processes. Dynamic temperature compensation and coordinated pressure control enable heating power and temperature thresholds to be adjusted based on real-time data, further optimizing bonding quality. On this basis, a deep learning model predicts the optimal temperature-pressure combination based on historical data, enabling automatic iterative optimization of process parameters, reducing manual intervention, improving production efficiency and product quality, and demonstrating high adaptability and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work. Figure 1 is a flow chart of the steps of the method of the present invention;

[0021] Figure 2 A flowchart of the method for establishing a mapping relationship between pressure and temperature in step 2 of the method of the present invention;

[0022] Figure 3 This is a flow chart of the steps of the method for collecting pressure distribution data in step 2 of the method of the present invention. DETAILED DESCRIPTION

[0023] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0024] See Figure 1-Figure 3 , an embodiment of the present invention provides a method for hot pressing and binding liquid crystal modules based on intelligent temperature control. In step 1, first, by dividing the hot pressing head into multiple independent temperature control areas, each area is equipped with at least one temperature sensor to ensure accurate monitoring of the temperature of each area. According to the different materials and geometric structures of the liquid crystal module, the pre-calibration experiment determines the appropriate initial temperature threshold by testing the electrode bonding strength and glue layer morphology at different temperatures. This process not only optimizes the temperature distribution during the hot pressing process, but also provides an accurate benchmark for subsequent temperature control. In addition, the temperature sensor is calibrated using thermal imaging equipment to eliminate the error between it and the actual temperature, thereby further improving the accuracy of temperature control.

[0025] Furthermore, during the hot pressing process in step 2, the temperature and pressure distribution data of each area are collected and monitored in real time to calculate the deviation between the current and target temperatures. According to the temperature deviation, the system will dynamically adjust the heating power, use the fuzzy logic algorithm to adjust the heating rate, and switch to the linear proportional integral algorithm when the temperature deviation is less than or equal to the set range. This dynamic adjustment not only ensures the stability of the temperature, but also adjusts the temperature threshold in real time to ensure the optimal viscosity of the adhesive layer by synchronously collecting pressure distribution data. For example, when the pressure is lower than the target range, the temperature is increased to reduce the viscosity of the adhesive layer. When the pressure is too high, the temperature is appropriately lowered to avoid excessive extrusion of the adhesive layer. This step ensures the synergistic effect of temperature and pressure, optimizes the fluidity of the adhesive layer, and avoids the negative impact of overpressure or overtemperature on the quality of the module.

[0026] In step 3, the data collected in step 2 is fed into a deep learning model. Based on historical binding data, the model predicts the optimal temperature and pressure combination for the next binding cycle. The advantage of this process is that the model automatically optimizes temperature and pressure settings by learning from the regularities of historical data, improving binding efficiency while ensuring consistent product quality. By automatically adjusting the initial temperature threshold in step 1 and the target pressure in step 2, the entire system can intelligently adjust to the actual production environment.

[0027] Through precise temperature and pressure control, combined with intelligent data analysis and optimization, the accuracy, efficiency and quality control level of the LCD module hot pressing bonding process have been greatly improved, which has important practical application value.

[0028] In one possible implementation, a thermal imager is used to scan the surface of the thermal head while it is not in use, acquiring two-dimensional temperature field data. The temperature field data provided by the thermal imager is highly accurate and fully covers the temperature distribution on the surface of the thermal head. The data collected by the thermal imager is then compared with the single-point measurement data from each temperature sensor to calculate the offset of each sensor at different temperatures. This comparison process can identify sensor reading errors caused by installation position, material, or other factors.

[0029] To eliminate the deviation caused by thermal expansion of the thermal head material, the next step involves performing secondary compensation based on the thermal expansion coefficient of the thermal head. This compensation method uses finite element simulation to simulate the deformation of the thermal head during heating and analyze the impact of this deformation on the sensor installation position. This simulation generates a correlation function between deformation and temperature change. Based on this function, the offset correction factor is dynamically adjusted during actual operation to ensure that the measurement data of each sensor reflects the accurate temperature.

[0030] Even after offset compensation, the sensor data may still be subject to random noise. To further improve the stability of temperature measurement, a Kalman filter algorithm is used to correct the raw sensor data. The Kalman filter inputs are the raw sensor data and the rate of change of the thermal head's heating power, and the output is a corrected temperature feedback signal. Through recursive processing, the Kalman filter effectively removes noise and provides a more accurate temperature feedback signal.

[0031] Through multiple compensation and filtering technologies, the temperature control accuracy and system stability during the hot pressing and bonding process of the LCD module are ensured, which optimizes production efficiency and product quality and has obvious technical advantages.

[0032] In one possible implementation, the fuzzy logic algorithm adjustment strategy in step 2 is designed to intelligently adjust the heating power based on the temperature deviation and its rate of change to achieve precise temperature control. This embodiment of the present invention addresses the issue of how to dynamically adjust the heating power based on temperature information fed back by the sensor to ensure accurate and stable temperature control during the hot pressing process.

[0033] Specifically, the fuzzy logic algorithm first requires two key input variables: the temperature deviation value and the deviation change rate. The temperature deviation value represents the difference between the current temperature and the target temperature, while the deviation change rate is the temperature deviation difference between two consecutive sampling times divided by the time interval. The temperature deviation value and the deviation change rate together reflect the current performance of the temperature control system and serve as the basis for subsequent adjustment decisions.

[0034] Based on the input temperature deviation value, a membership function is used to classify the deviation into three nonlinear intervals: high, medium, and low. Each interval has a different weight coefficient, and the membership function can be used to assign different membership values to temperature deviations. This allows the algorithm to select the appropriate regulation strategy based on the size of the temperature deviation, allowing the system to adjust power in a targeted manner to avoid excessive or insufficient heating.

[0035] The algorithm combines the membership function and the deviation level to generate the adjustment range and direction. The adjustment range of the heating power is determined by the temperature deviation level defined by the membership function, while the adjustment direction (increase or decrease heating power) is determined by the deviation trend. This strategy enables the system to respond appropriately to the dynamic characteristics of temperature changes, ensuring temperature control accuracy.

[0036] The control instructions within the fuzzy logic algorithm are generated based on an empirical rule base. This rule base maps pressure and temperature based on temperature deviations and heating power adjustments from historical hot-pressing data. By accumulating historical data, this empirical rule base provides the algorithm with a precise control framework, enabling the system to generate optimized control strategies based on varying temperature deviations, ensuring rational and efficient heating power adjustments during the hot-pressing process.

[0037] During the adjustment process, the temperature control system may be affected by the thermal conductivity of the hot-press material and the rate of change of the heating power, resulting in a certain amount of delay. To avoid overshoot (i.e., excessive adjustment of the heating power) caused by temperature lag, the algorithm introduces a delay compensation mechanism. This mechanism dynamically calculates the delay time based on the thermal conductivity of the hot-press material and the current rate of change of the heating power, thereby preventing overshoot of the temperature control and ensuring a smooth system response.

[0038] The fuzzy logic algorithm's regulation strategy ensures the stability, accuracy, and efficiency of temperature control during the hot-press bonding process of LCD modules through precise adjustment amplitude, dynamic compensation, and an empirical rule library, thereby improving production efficiency and product quality.

[0039] In one possible embodiment, the method for establishing the mapping relationship between pressure and temperature in step 2 is to achieve more precise intelligent temperature control and ensure that the flow and coverage area of the adhesive layer during the hot pressing bonding process of the liquid crystal module can be reasonably controlled under different pressure and temperature conditions.

[0040] Specifically, during the pre-calibration phase, to establish a mapping relationship between pressure and temperature, different pressure combinations are first applied to test pieces of the same material. This process fully tests the effects of temperature and pressure during the hot pressing process, allowing for the subsequent establishment of an effective mapping relationship.

[0041] While applying varying pressures, high-speed video equipment monitors the adhesive layer's diffusion process in real time during the hot pressing process. The captured data on the adhesive layer's edge movement is used to calculate displacement per unit time, thereby deriving the adhesive layer's flow rate. This flow rate reflects the adhesive layer's flow characteristics under pressure and temperature, providing critical data for subsequent adjustments.

[0042] At the same time, an image processing algorithm analyzes the contact area between the adhesive layer and the electrode, identifying the pixel ratio of the contact area. This ratio is considered an effective indicator of the adhesive layer's coverage area, reflecting the distribution of the adhesive layer during the hot pressing process. This data is crucial for the coordinated regulation of temperature and pressure.

[0043] The ratio of flow rate to coverage area is a key factor in the coordinated adjustment of pressure and temperature. By analyzing the relationship between flow rate and coverage area, we can determine their ratio, providing a basis for the coordinated adjustment of pressure and temperature. This ratio becomes the basis for the subsequent mapping surface construction.

[0044] By collecting experimental data under different pressure and temperature conditions, a three-dimensional surface representing the relationship between pressure, temperature, and adhesive layer morphology is fitted. This surface clearly depicts the flow behavior and morphological changes of the adhesive layer under different pressures and temperatures. Based on the currently applied pressure, interpolation calculations can be performed on the surface to obtain the corresponding temperature compensation.

[0045] Finally, by interpolating the fitted 3D relationship surface, the corresponding temperature compensation is calculated based on the current pressure value. This compensation reflects the temperature change that the temperature control system needs to adjust to achieve optimal adhesive flow and coverage under a given pressure.

[0046] A method based on the mapping relationship between pressure and temperature uses high-speed camera technology, image processing algorithms, and 3D surface fitting to precisely adjust temperature control parameters during the hot pressing process. This method enables high-precision control of the adhesive layer flow, effectively improving the quality and efficiency of the hot pressing bonding process for LCD modules.

[0047] In one possible implementation, the deep learning model training method in step 3 utilizes temperature and pressure distribution data to automatically predict the temperature threshold and target pressure for the next hot pressing cycle. The specific training method involves combining a convolutional neural network with a long-short-term memory network. The goal is to learn the spatiotemporal characteristics of temperature and pressure from historical data and optimize the temperature control process using a predictive model.

[0048] Specifically, during the training process, we first need to collect temperature data, pressure distribution data, and ambient humidity data. This data, already acquired and processed in step 2, represents the key factors in the hot pressing process. This data is then fed into the deep learning model to train it to learn the laws governing the temperature control process.

[0049] Convolutional neural networks (CNNs) are a key tool for feature extraction in deep learning. They can extract spatiotemporal features from input temperature, pressure, and humidity data. During the hot pressing process, temperature and pressure vary over time and are affected by their spatial distribution. CNNs, through multi-layer convolution operations, automatically extract high-level features relevant to hot pressing quality.

[0050] Long Short-Term Memory (LSTM) is a recursive neural network suitable for time series data. It can effectively process temperature and pressure sequence data that changes over time. Using the LSTM model, the training system can predict temperature and pressure trends at a specific moment in the future and provide optimization directions for the next cycle, helping the system automatically adjust temperature control parameters.

[0051] To ensure data stability and efficient training, the input data needs to be normalized. Specifically, the temperature data is scaled to the maximum temperature resistance of the thermal head, while the pressure distribution data is scaled to the range of 0-1 according to the upper limit of the sensor's range. This processing method helps accelerate model convergence and prevent the impact of different data scales.

[0052] The model's loss function is defined as the mean squared error (MSE) between the predicted temperature and pressure values and the actual optimal parameters. By minimizing this loss function, the model gradually improves prediction accuracy. The optimizer uses an adaptive learning rate algorithm that dynamically adjusts the gradient descent step size, improving training efficiency and avoiding oscillations caused by excessively large step sizes during training.

[0053] During training, cross-validation is used to assess the model's generalization capabilities, ensuring that the model can adapt to different thermal stress conditions and not simply overfit to the training data. At the same time, dropout is used to suppress overfitting and enhance the model's robustness, allowing the trained model to better generalize to unknown data.

[0054] The deep learning model training method that combines convolutional neural networks and long short-term memory networks can achieve intelligent temperature control optimization, accurately predict the set values of temperature and pressure, and significantly improve the quality and efficiency of the LCD module hot press bonding process.

[0055] In one possible embodiment, in step 2, the method for collecting pressure distribution data involves using a pressure sensor array to accurately monitor the pressure during the hot pressing process, and obtaining continuous and accurate pressure distribution data through a series of signal processing and algorithm processing.

[0056] Specifically, an array of pressure sensors is installed beneath the hot press platform, with the number and density of sensors matching the number of zones within the hot press head. This means that each pressure sensor's acquisition area covers every area of the hot press head, ensuring the accuracy of the pressure distribution data by precisely measuring pressure changes in each zone.

[0057] The raw signals generated by pressure sensors are typically analog. For subsequent processing, these signals need to be converted to digital signals using a signal amplifier. The gain setting of the signal amplifier is dynamic, adjusting the gain in real time based on the sensor's sensitivity and the measurement environment to ensure signal quality and accuracy.

[0058] The converted digital signal may contain high-frequency noise, which can affect data accuracy. Therefore, a digital filtering algorithm is used to process the signal. Specifically, a Fourier transform is used to convert the signal from the time domain to the frequency domain, removing noise components above a set frequency. Then, an inverse Fourier transform is used to reconstruct a clean pressure signal. This process effectively removes interference and improves data reliability.

[0059] The reconstructed pressure signal is then processed using a spatial interpolation algorithm to generate a continuous pressure distribution heat map. This heat map visually displays pressure variations across different regions and helps identify areas of abnormal pressure. In this process, the spatial interpolation algorithm smooths the data and enhances signal accuracy, making the resulting pressure distribution map more valuable for reference.

[0060] On the generated pressure heat map, the system uses algorithms to detect abnormal pressure areas. These abnormal areas may indicate uneven pressure distribution or other potential production issues, such as poor contact between the hot press head and the mold, or a faulty pressure sensor. Promptly identifying these abnormal areas helps operators make adjustments and avoid quality issues.

[0061] Through this series of steps including data acquisition, signal processing, spatial interpolation, and abnormal area identification, this method can provide high-precision pressure distribution data, offer reliable support for intelligent temperature control, and significantly improve the quality and stability of the LCD module hot-press bonding process.

[0062] In one possible implementation, during a pre-calibration experiment, the hot-head priority areas are first divided according to the LCM geometry. Different areas have different heat transfer characteristics, so different pressures and temperatures need to be applied to the hot-head based on the module's structural characteristics. Especially for curved or irregularly shaped LCM structures, the hot-head pressure and temperature distribution in these areas need to be adjusted.

[0063] The edge areas of curved or irregularly shaped structures may experience uneven temperatures due to variations in geometry and contact area. To address this, a compensation temperature value is added to the initial temperature threshold during the pre-calibration experiment. This compensation temperature value compensates for temperature loss caused by heat dissipation at the edge areas, ensuring consistent temperature across the entire LCM surface.

[0064] Furthermore, the compensation temperature value is determined through finite element heat conduction simulation. This simulation models the heat transfer process in the contact area between the thermal head and the module, predicts the heat dissipation path in this area, and calculates the temperature loss in the edge area due to heat dissipation. This process accurately predicts the temperature deviation that may occur in the edge area during the hot pressing process, allowing appropriate temperature compensation to be performed.

[0065] When performing finite element heat conduction simulation, some key input parameters need to be set. These parameters include:

[0066] Thermal conductivity of the thermal head material: determines the heat conduction efficiency between the thermal head and the module.

[0067] Environmental convection coefficient: affects the loss of heat during heat conduction.

[0068] Specific heat capacity of the module substrate: determines the heat storage capacity of the module material.

[0069] Boundary conditions: Assume that the temperature of the heating surface of the thermal press head is constant and the ambient temperature is a known value.

[0070] There may be some error between the simulation results and the actual measured data. To improve the accuracy of the simulation results, the thermal conductivity parameters are iteratively corrected until the error between the simulation results and the actual measured data converges. This process ensures the accuracy of the simulation model and makes subsequent temperature compensation more precise.

[0071] Precise temperature compensation and heat conduction simulation not only optimize temperature control during hot pressing, but also effectively improve the quality and production efficiency of LCD modules. This precise pre-calibration and compensation technology ensures temperature uniformity during hot pressing, especially in edge areas, reducing temperature deviations during hot pressing and improving the intelligence and automation level of the entire production process.

[0072] In one possible implementation, this example proposes a specific method for controlling adhesive layer viscosity through intelligent temperature regulation. This method addresses issues such as poor bonding or over-compaction caused by abnormal pressure during hot-press bonding. By dynamically controlling the temperature, the rheological properties of the adhesive layer are adjusted to ensure that it maintains good filling and bonding conditions under varying pressures.

[0073] Specifically, during the hot pressing process, the pressure applied by the hot pressing head to the LCD module is monitored in real time and compared with a preset target pressure range. If the actual pressure is detected to be lower than the target range, it means there is a gap or poor contact between the hot pressing head and the module, and the adhesive layer may not be able to fully flow and fill the gap.

[0074] To improve the fluidity of the adhesive layer, the temperature is raised to a level above the first range of the adhesive layer's glass transition temperature. The glass transition temperature is the critical temperature at which the adhesive layer material changes from a hard and brittle state to a soft state with a certain degree of fluidity. The determination method is:

[0075] Differential scanning calorimetry (DSC) was used to perform thermal analysis of the adhesive layer material.

[0076] Obtain its phase change curve and identify the endothermic peak region.

[0077] The starting point of the endothermic peak was taken as the glass transition temperature.

[0078] After increasing the temperature, the viscosity of the adhesive layer decreases, which enhances its fluidity and promotes the colloid to fully flow under lower pressure and fill the gap between the module and the hot pressing head, thereby achieving good bonding.

[0079] If the pressure detected is higher than the target range, it means that the force applied by the indenter is too great, which may cause excessive extrusion or thermal degradation of the adhesive layer. In this case, the temperature should be lowered to a second range below the thermal degradation temperature to slow the flow of the adhesive and maintain its structural stability.

[0080] The thermal degradation temperature was also obtained by DSC:

[0081] Identify the exothermic peak in the phase transition curve.

[0082] The starting point of the exothermic peak is taken as the thermal degradation temperature.

[0083] The temperature is controlled in a range slightly lower than this temperature to avoid material degradation.

[0084] Before formal production, the rheological properties of the adhesive layer material are tested under different temperature conditions through pre-calibration experiments to clarify the viscosity change law at different temperatures.

[0085] Determine that the viscosity of the adhesive layer decreases significantly above the glass transition point;

[0086] Maintains a certain viscosity before thermal degradation to avoid adhesive failure or decomposition.

[0087] Finally, a set of temperature range data tables for the intelligent temperature control system is formed for dynamic adjustment in the subsequent binding process.

[0088] By incorporating differential scanning calorimetry test results into the temperature control logic, we were able to scientifically identify the two key temperature points of the adhesive layer—glass transition and thermal degradation. Based on this, we set a dynamic temperature control strategy, coupled with pressure feedback for intelligent regulation. This approach effectively enhances the intelligence of hot-press bonding, enhancing the stability and bonding quality of the adhesive layer during the hot-pressing process of LCD modules, and providing reliable support for large-scale, high-quality LCD module production.

[0089] In one possible implementation, during the hot-press bonding process, the intelligent control system initiates iterative parameter optimization through two triggering conditions:

[0090] The first is that the yield rate is lower than the set threshold: If the bonding yield rate of multiple consecutive production batches is lower than the predetermined standard value, it indicates that the current process settings may not be able to adapt to actual production needs, and it is necessary to optimize the parameters to improve the bonding quality and consistency.

[0091] The second is excessive fluctuations in ambient temperature and humidity: These changes can affect the rheological properties and bonding quality of the adhesive layer. By monitoring environmental changes in real time, if the fluctuations exceed the standard deviation of the historical average, the system deems the change to be beyond tolerance and triggers parameter optimization.

[0092] Optical inspection equipment is used to perform automated image analysis on each batch of bonded LCD module electrodes to identify defects, including common problems such as cold solder joints, short circuits, and incomplete adhesive coverage.

[0093] The optical inspection system uses high-precision image recognition technology to detect the morphology of each electrode and calculate the yield rate of each batch, which serves as a basis for determining whether the process needs to be adjusted.

[0094] High-precision sensors collect real-time temperature and humidity data. The system compares the standard deviation of the real-time data with the historical average to determine whether it exceeds the preset tolerance range.

[0095] Once the temperature and humidity fluctuate too much, the system will determine that the external environment changes have affected the performance of the adhesive layer, and then trigger the parameter optimization process.

[0096] Specifically, when the trigger conditions are met, the system enters the parameter optimization phase. First, the pre-calibrated experimental data is reloaded to ensure that the data used matches the current production environment and material properties.

[0097] Update the input feature weights of the deep learning model so that the model can be adjusted according to real-time production data, giving priority to optimizing and adjusting key parameters that are more sensitive to environmental changes (such as temperature, pressure, adhesive layer thickness, etc.).

[0098] Through this iterative optimization, the system continuously updates and optimizes the thermal compression bonding process to ensure a high yield rate at all times under different environmental conditions.

[0099] Through the intelligent temperature-controlled hot-press bonding process for LCD modules, real-time monitoring of yield and ambient temperature and humidity changes triggers iterative parameter optimization. Combining optical inspection with deep learning model optimization, the system automatically adapts to changing production environments, ensuring high quality and stability in the bonding process. This significantly improves product yield and production efficiency, while reducing manual intervention and operational errors, and enhancing overall production automation.

[0100] In one possible implementation, after the initial hot-press bonding of the LCD module, optical inspection equipment comprehensively scans the bonding area to automatically identify electrode defects such as cold solder joints or short circuits. The optical inspection system uses high-precision image processing technology to analyze the electrode's topography and electrical properties to locate and mark defective areas. This automated and efficient process can promptly identify defects that were not fully repaired during the hot-press bonding process.

[0101] Once the optical inspection equipment locates the defective area, the system automatically triggers the local secondary hot pressing step. The purpose of the secondary hot pressing is to improve the bonding quality of the defective area by adjusting the temperature and pressure.

[0102] The specific method of secondary hot pressing: The system will increase the temperature to a percentage of the original set value according to a pre-set ratio, while simultaneously reducing the pressure to a percentage of the original set value. The percentage range is selected based on the size and location of the defect area. In other words, defects of different sizes and locations require different temperature and pressure adjustments. Preliminary experiments determine the corresponding relationship between defect size and temperature / pressure adjustment, ensuring that the secondary hot pressing process is effective for the defect area.

[0103] During the secondary hot pressing process, the temperature control signal in the normal area is shielded, preventing additional heat and pressure from being applied to the already bonded area, thereby minimizing the impact on areas without defects. The temperature control system applies adjusted temperature and pressure parameters only to the defective area, ensuring that the hot pressing process affects only the problem area, without affecting the quality of other parts of the LCD module.

[0104] By introducing an automated post-press defect repair step, combined with optical inspection and secondary hot pressing technology, this system effectively addresses defects such as cold solder joints and short circuits that can occur during the hot pressing process. This localized secondary hot pressing process precisely controls temperature and pressure, enabling targeted repair of defective areas. This not only improves product quality but also saves production resources and time. This intelligent repair mechanism significantly improves the production quality and efficiency of LCD modules while avoiding the waste and complexity of traditional repair methods.

[0105] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0106] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A liquid crystal module hot pressing binding method based on intelligent temperature control, characterized in that: The following steps are involved: Step 1: Hot-pressing area zoning and dynamic temperature calibration: The contact surface of the hot-pressing head is divided into multiple independently controlled temperature zones, each equipped with at least one temperature sensor. Based on the material type and geometry of the liquid crystal module, the initial temperature threshold of each zone is determined through a pre-calibration experiment. The pre-calibration experiment includes performing multiple sets of hot-pressing experiments on test pieces of the target material, recording the electrode adhesion strength and adhesive layer morphology at different temperatures. Based on the relationship curve between adhesion strength and temperature, the equilibrium temperature between the adhesive layer fluidity and thermal degradation threshold is selected as the initial threshold. The temperature sensors in each zone are calibrated using thermal imaging equipment to eliminate deviations between the sensors and the actual temperature. Step 2: Dynamic temperature compensation and pressure coordinated control, real-time collection of temperature data from each area calibrated in step 1, and calculation of the deviation from the target temperature; dynamic adjustment of heating power based on the deviation, specifically using a fuzzy logic algorithm to adjust the heating power when the deviation is greater than a first set range, and switching to a linear proportional integral algorithm to adjust the heating power when the deviation is less than or equal to the first set range; synchronous collection of pressure distribution data from each area during the hot pressing process, the pressure distribution data being acquired in real time by a pressure sensor array disposed below the hot pressing table; adjustment of the temperature threshold of the corresponding area based on the real-time pressure value, including increasing the temperature threshold to reduce the viscosity of the adhesive layer when the pressure is below the target pressure range, and decreasing the temperature threshold to prevent excessive extrusion of the adhesive layer when the pressure is above the target pressure range; Step 3: Iteratively optimize the hot pressing process parameters. Input the temperature data and pressure distribution data collected in Step 2 into the deep learning model. The deep learning model is trained based on the historical binding data to predict the optimal temperature-pressure parameter combination for the next binding cycle. The initial temperature threshold in Step 1 and the target pressure in Step 2 are automatically corrected based on the prediction results.

2. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 1, characterized in that: The calibration method of the temperature sensor in step 1 is specifically as follows: when the thermal head is not in operation, the surface temperature distribution of the thermal head is scanned by an external thermal imager, the two-dimensional temperature field data collected by the thermal imager is compared with the single-point measurement data of each temperature sensor, and the offset of each sensor at different temperatures of the thermal head is calculated; the offset is quadratically compensated according to the thermal expansion coefficient of the thermal head material, and the compensation method includes calculating the influence of the deformation variable of the thermal head during the heating process on the sensor installation position through finite element simulation, establishing a correlation function between the deformation variable and the temperature change, and dynamically adjusting the correction coefficient of the offset according to the current temperature; the calibrated sensor data is eliminated from random noise interference through the Kalman filter algorithm, the input of the Kalman filter is the sensor raw data and the rate of change of the thermal head heating power, and the output is the corrected temperature feedback signal.

3. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 2, characterized in that: The adjustment strategy of the fuzzy logic algorithm in step 2 is specifically as follows: the input variables are the temperature deviation value and the deviation change rate, the temperature deviation value is the difference between the current temperature and the target temperature, and the deviation change rate is the temperature deviation difference between two adjacent samples divided by the time interval; the output variable is the adjustment amplitude and direction of the heating power, the adjustment amplitude is determined according to the temperature deviation level defined by the membership function, the membership function divides the deviation value into three nonlinear intervals of high, medium and low, and assigns a weight coefficient to each interval; the adjustment instruction is generated by an empirical rule library, and the empirical rule library constructs a mapping relationship between pressure and temperature based on the temperature deviation and the heating power adjustment amount in the historical hot pressing data; the fuzzy logic algorithm introduces a delay compensation mechanism in the adjustment process, and the delay time is dynamically calculated according to the thermal conductivity of the material of the hot pressing head and the current heating power change rate to avoid overshoot caused by temperature lag.

4. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 3, characterized in that: The method for establishing the mapping relationship between pressure and temperature in step 2 is specifically as follows: in the pre-calibration stage, different pressure combinations are applied to the same material test piece, the diffusion process of the adhesive layer during the hot pressing process is captured by a high-speed camera device, the movement trajectory of the adhesive layer edge is extracted, and the displacement per unit time is calculated as the flow rate; at the same time, the pixel ratio of the contact area between the adhesive layer and the electrode is identified as the coverage area through an image processing algorithm; the coordinated adjustment ratio of pressure and temperature is determined based on the ratio of flow rate to coverage area, and the coordinated adjustment ratio is established by fitting experimental data to establish a three-dimensional relationship surface of pressure-temperature-adhesive layer morphology, and the corresponding temperature compensation amount is interpolated on the surface according to the current pressure value.

5. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 4, characterized in that: The training method of the deep learning model in step 3 is specifically as follows: the input data includes the temperature data and pressure distribution data collected in step 2, and the output data is the temperature threshold and target pressure recommended for the next cycle; the model uses a convolutional neural network to extract the temperature-pressure spatiotemporal characteristics, and predicts the parameter optimization direction through a long short-term memory network; the input data is normalized during the training process, and the normalization method includes scaling the temperature data to a range of 0-1 according to the maximum temperature resistance value of the thermal pressure head, and scaling the pressure distribution data to a range of 0-1 according to the upper limit of the sensor range; the loss function of the model is defined as the mean square error between the predicted parameters and the actual optimal parameters, and the optimizer uses an adaptive learning rate algorithm to dynamically adjust the gradient descent step size; after the training is completed, the generalization ability of the model is evaluated through cross-validation, and the overfitting phenomenon is suppressed by the discard method.

6. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 5, characterized in that: The method for collecting pressure distribution data in step 2 is specifically as follows: a pressure sensor array is set under the hot pressing table, and the array density matches the number of hot pressing head partitions; the analog signal of the pressure sensor is converted into a digital quantity through a signal amplifier, and the gain of the signal amplifier is dynamically adjusted according to the sensitivity of the pressure sensor; the converted digital signal is filtered out of high-frequency noise interference through a digital filtering algorithm, and the filtering algorithm includes performing Fourier transform on the original signal and then cutting off the components higher than the set frequency, and then reconstructing the pressure signal through inverse transformation; the reconstructed pressure signal is used to generate a continuous pressure distribution thermodynamic map through a spatial interpolation algorithm, and the pressure abnormality area is identified based on the thermodynamic map.

7. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 6, characterized in that: The pre-calibration experiment in step 1 also includes dividing the priority areas of the thermal head according to the geometric structure of the liquid crystal module; for the edge areas of curved surfaces or special-shaped structures, adding a compensation temperature value based on the initial temperature threshold; The method for determining the compensation temperature value includes predicting the heat dissipation path of the contact area between the thermal head and the module through finite element heat conduction simulation, and calculating the temperature loss in the edge area caused by heat dissipation; the simulation input parameters include the thermal conductivity of the thermal head material, the environmental convection coefficient and the specific heat capacity of the module substrate, and the boundary conditions are set to a constant temperature of the thermal head heating surface and the ambient temperature; the error between the simulation result and the actual measurement data is corrected by iteratively correcting the thermal conductivity parameter until convergence.

8. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 7, characterized in that: The method for controlling the viscosity of the adhesive layer in step 2 is specifically as follows: when the pressure is lower than the target range, the temperature is increased to a first range higher than the glass transition temperature of the adhesive layer; the glass transition temperature is determined by measuring the phase change curve of the adhesive layer material with a differential scanning calorimeter and extracting the starting point corresponding to the endothermic peak; when the pressure is higher than the target range, the temperature is reduced to a second range lower than the thermal degradation temperature of the adhesive layer; the thermal degradation temperature is determined by the starting point of the exothermic peak of the phase change curve; the temperature ranges are set in a pre-calibration stage to verify the rheological properties of the adhesive layer at different temperatures, ensuring that the decrease in the viscosity of the adhesive layer after the temperature is increased meets the flow and filling requirements.

9. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 8, characterized in that: The triggering conditions for the iterative optimization of parameters in step 3 include detecting that the continuous binding yield rate is lower than a set threshold or that the ambient temperature and humidity changes exceed a preset tolerance range; the statistical method for the yield rate is to automatically analyze the electrode morphology after binding through optical detection equipment to identify defects such as cold solder joints, short circuits, and incomplete adhesive coverage; the monitoring of the ambient temperature and humidity changes is carried out by real-time data collection through high-precision sensors, and parameter optimization is triggered when the temperature and humidity fluctuations exceed the standard deviation of the historical average value; the optimization process includes reloading the pre-calibrated experimental data and updating the input feature weights of the deep learning model, giving priority to adjusting parameters that are sensitive to the current environment.

10. The method for hot-pressing and bonding liquid crystal modules based on intelligent temperature control according to claim 9, characterized in that: It also includes a step of automatic defect repair after hot pressing: using optical detection equipment to identify cold solder joints or short circuit defects in the binding area, locating the defective area and triggering local secondary hot pressing; the specific method of the secondary hot pressing is to increase the temperature to a percentage range of the original set value and reduce the pressure to a percentage range of the original set value; the basis for determining the percentage range is the size and position of the defective area, and a correspondence between the defect size and the temperature / pressure adjustment amount is established through preliminary experiments; during the secondary hot pressing, the temperature control signal of the normal area is shielded, and only the adjusted parameters are applied to the defective area.

Images