OLED substrate pixel trench positioning methods, systems, devices and storage media

Abstract

This invention relates to the field of inkjet printing technology, and discloses a method, system, device, and storage medium for OLED substrate pixel slot positioning. The OLED substrate pixel slot positioning method includes the following steps: acquiring temperature data from all sensors and images of the glass substrate; calculating a two-dimensional temperature field function and the actual displacement of a reference marker point; constructing a substrate stage simulation model; calculating temperature prediction data and first displacement prediction data for multiple points based on the simulation model; training a preset convolutional neural network model to obtain an OLED substrate pixel slot positioning model; and predicting and calculating the coordinates of the reference marker point at the next moment using the OLED substrate pixel slot positioning model. This invention compensates for thermally induced deformation errors through temperature field and convolutional neural network model, reducing noise and delay caused by relying solely on real-time visual detection, and is suitable for high-precision pixel slot positioning of large-size glass substrates during processing.

Description

Technical Field

[0001] This invention relates to the field of inkjet printing technology, and in particular to a method, system, device, and storage medium for OLED substrate pixel slot positioning. Background Technology

[0002] Existing inkjet printing alignment methods can only measure deformation after it has occurred, using optical images – a passive, reactive detection method. The root cause of deformation is often thermal stress. During the manufacturing process, uneven heating of the substrate creates temperature gradients, leading to uneven thermal expansion and deformation. As OLED substrates become larger, they are more susceptible to complex deformations during transport and printing due to temperature changes (potentially causing thermal expansion and contraction) or mechanical stress. Relying solely on a few discrete alignment markers for position estimation is like trying to guess the overall distortion of a large sheet of paper with a few points; it cannot detect and compensate for the complex distortions between the marker points, resulting in localized misalignment. Error compensation also exhibits lag. For high-speed continuous printing, especially when the system has significant inertia or mechanical delay, this lag compensation may not be able to keep up with the speed at which errors occur, leading to printing deviations. Summary of the Invention

[0003] The present invention aims to improve at least one technical problem in the prior art.

[0004] The first aspect of this invention provides a pixel slot positioning method for an OLED substrate, applied to an OLED substrate pixel slot positioning system. The OLED substrate pixel slot positioning system includes: a control device, a stage, and multiple temperature sensors. The stage is used to support a glass substrate, and the multiple temperature sensors are mounted on the stage. The control device is electrically connected to the multiple temperature sensors. The OLED substrate pixel slot positioning method includes the following steps:

[0005] Acquire temperature data from all sensors and obtain images of the glass substrate;

[0006] Interpolate all the temperature data to calculate the two-dimensional temperature field function;

[0007] Determine the coordinates of the reference marker points based on the image of the glass substrate;

[0008] The actual displacement of the reference marker point is calculated based on its coordinates.

[0009] Based on the actual displacement of the reference markers and the two-dimensional temperature field function, a simulation model of the substrate stage is constructed.

[0010] Based on the substrate stage simulation model, temperature prediction data and first displacement prediction data at multiple points are calculated. The temperature prediction data and first displacement prediction data are used to train a preset convolutional neural network model to obtain the OLED substrate pixel slot positioning model.

[0011] The two-dimensional temperature field function is input into the OLED substrate pixel slot positioning model, and the predicted displacement of the reference marker point is output.

[0012] Based on the coordinates of the reference marker and the predicted displacement, the coordinates of the reference marker at the next moment are calculated.

[0013] Optionally, in a first implementation of the first aspect of the present invention, a simulation model of the substrate stage is constructed based on the actual displacement of the reference marker point and the two-dimensional temperature field function, including:

[0014] A thermo-coupled finite element model was constructed based on the glass substrate and the stage.

[0015] The thermo-coupled finite element model is given thermal properties, mechanical properties, and coefficient of thermal expansion to obtain the first model;

[0016] By applying boundary conditions and loads to the first model, the second model is obtained.

[0017] The second model was solved by sequential coupling analysis to obtain the second displacement prediction data of the reference marker point.

[0018] Based on the second displacement prediction data and the actual displacement, the second model is iteratively optimized until it meets the preset conditions, thus obtaining the simulation model of the substrate stage.

[0019] Optionally, in a second implementation of the first aspect of the present invention, applying boundary conditions and loads to the first model includes:

[0020] A two-dimensional temperature field function is applied as a load to the bottom surface of the glass substrate;

[0021] Convection and radiation heat transfer are applied as boundary conditions to the top and side surfaces of the glass substrate and the surface of the stage.

[0022] Contact thermal conductivity and contact stiffness are applied as boundary conditions to the interface between the glass substrate and the stage.

[0023] Optionally, in a third implementation of the first aspect of the present invention, the step of using sequential coupling to solve the thermo-mechanical coupling analysis of the second model and calculating the second displacement prediction data of the reference marker points includes:

[0024] Steady-state thermal analysis was performed on the second model to calculate its temperature distribution.

[0025] Temperature distribution is used as a volume load input for static structural analysis to calculate the deformation of the glass substrate.

[0026] Based on the deformation of the glass substrate, the second displacement prediction data of the reference mark point is obtained.

[0027] Optionally, in a fourth implementation of the first aspect of the present invention, based on a substrate stage simulation model, temperature prediction data and first displacement prediction data for multiple points are calculated, and a preset convolutional neural network model is trained using the temperature prediction data and the first displacement prediction data to obtain an OLED substrate pixel slot positioning model, including:

[0028] Each temperature prediction data is converted into a first two-dimensional array, and each first displacement prediction data is converted into a second two-dimensional array;

[0029] All the first two-dimensional arrays and their corresponding second two-dimensional arrays are combined into multiple data pairs, and all data pairs are randomly divided to obtain the training set.

[0030] Construct a U-Net network structure, input the first two-dimensional array of the training set into the U-Net network structure, and output a predicted two-dimensional array;

[0031] Calculate the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first input two-dimensional array;

[0032] Based on the loss function, update the network weights of the U-Net network structure, and return to the step of calculating the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first two-dimensional array of input, until the training set is traversed to obtain the OLED substrate pixel slot localization model.

[0033] A second aspect of the present invention provides an OLED substrate pixel slot positioning system, comprising: a control device, a frame, a stage, and a plurality of temperature sensors, wherein the stage is mounted on the frame and is used to support a glass substrate, the temperature sensors are mounted on the stage, and the glass substrate is provided with reference marking points, and the control device is electrically connected to the plurality of temperature sensors; the OLED substrate pixel slot positioning system is used to perform the OLED substrate pixel slot positioning method as described above.

[0034] As a further improvement to the above technical solution, the OLED substrate pixel slot positioning system further includes an X-axis moving mechanism, a Y-axis moving mechanism, and a CCD camera. The fixed part of the Y-axis moving mechanism is mounted on the frame, and the movable part of the Y-axis moving mechanism is connected to the stage for transmission. The Y-axis moving mechanism drives the stage to move along the Y-axis direction. The fixed part of the X-axis moving mechanism is mounted on the frame, and the CCD camera is mounted on the movable part of the X-axis moving mechanism. The X-axis moving mechanism drives the CCD camera to move along the X-axis direction. The CCD camera is located above the glass substrate and is used to photograph the glass substrate. The control device is electrically connected to the Y-axis moving mechanism, the X-axis moving mechanism, and the CCD camera.

[0035] As a further improvement to the above technical solution, the OLED substrate pixel slot positioning system further includes a first laser interferometer and a second laser interferometer. The first laser interferometer is mounted on the frame and is used to measure its distance from the stage along the Y-axis. The second laser interferometer is mounted on the fixed part of the X-axis moving mechanism and is used to measure its distance from the CCD camera along the X-axis. The control device is electrically connected to the first laser interferometer and the second laser interferometer.

[0036] A third aspect of the present invention provides an OLED substrate pixel slot positioning device, the OLED substrate pixel slot positioning device comprising: a memory and at least one processor, wherein the memory stores instructions;

[0037] At least one of the processors invokes the instructions in the memory to cause the OLED substrate pixel slot positioning device to perform the steps of the OLED substrate pixel slot positioning method as described above.

[0038] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the steps of the OLED substrate pixel slot positioning method described above.

[0039] The beneficial effects of this invention are as follows: By acquiring temperature data from multiple temperature sensors on the stage and interpolating to construct a two-dimensional temperature field function, an image of the glass substrate is obtained to determine the coordinates of the reference marker points. The actual displacement of the reference marker points is calculated, and a simulation model of the glass substrate stage is established based on the actual displacement and the two-dimensional temperature field function. This model is then used to train and call an OLED substrate pixel slot positioning model to predict the displacement of the reference marker points. Existing inkjet printing alignment methods can only measure the deformation result through optical images after deformation has occurred. Uneven heating of the glass substrate generates temperature gradients, leading to uneven thermal expansion and thus deformation. This invention, by monitoring the temperature at different locations on the substrate in real time, can predict the trend and magnitude of deformation before or during deformation. By compensating for thermally induced deformation errors through the temperature field and a convolutional neural network model, it significantly improves positioning accuracy, reduces noise and delay caused by relying solely on real-time visual detection, and improves positioning stability. It is suitable for high-precision pixel slot positioning of large-size glass substrates during processing. Furthermore, the temperature sensing system is relatively simple, low-cost, and easier to maintain. Attached Figure Description

[0040] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which:

[0041] Figure 1 This is a first flowchart of an OLED substrate pixel slot positioning method provided in an embodiment of the present invention;

[0042] Figure 2 This is a second flowchart of the OLED substrate pixel slot positioning method provided in an embodiment of the present invention;

[0043] Figure 3 This is a third flowchart of the OLED substrate pixel slot positioning method provided in the embodiments of the present invention;

[0044] Figure 4 This is a fourth flowchart of the OLED substrate pixel slot positioning method provided in the embodiments of the present invention;

[0045] Figure 5 This is a fifth flowchart of the OLED substrate pixel groove positioning method provided in the embodiments of the present invention;

[0046] Figure 6 An isometric view of the ink supply control system provided in an embodiment of the present invention;

[0047] Figure 7 A schematic diagram of a glass substrate provided in an embodiment of the present invention;

[0048] Figure 8 This is a schematic diagram of the structure of the ink supply control device provided in an embodiment of the present invention.

[0049] In the attached diagram: 1-stage; 2-temperature sensor; 3-CCD camera; 4-second laser interferometer; 5-first laser interferometer; 6-Y-axis moving mechanism; 7-X-axis moving mechanism; 8-frame. Detailed Implementation

[0050] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0051] The following is combined with Figures 1 to 8 Embodiments of the present invention will be described.

[0052] Please refer to Figure 1 and Figure 7 An embodiment of the OLED substrate pixel slot positioning method of the present invention includes: an OLED substrate pixel slot positioning system, the OLED substrate pixel slot positioning system including: a control device, a stage, and multiple temperature sensors, the stage being used to support a glass substrate, the glass substrate having reference marking points, the temperature sensors being mounted on the stage, and the control device being electrically connected to the multiple temperature sensors; the OLED substrate pixel slot positioning method includes the following steps:

[0053] 101. Acquire temperature data from all sensors and obtain an image of the glass substrate;

[0054] In this embodiment, temperature data from multiple temperature sensors mounted on the stage are collected in real time, and image information of the glass substrate is also acquired, which can be done by setting up a CCD camera.

[0055] 102. Interpolate all the temperature data to calculate the two-dimensional temperature field function;

[0056] In this embodiment, interpolation is performed on the acquired discrete temperature data to construct a two-dimensional temperature field function covering the stage and glass substrate region, which is used to characterize the current temperature distribution state on the stage surface. Discrete sensor readings are then used to... Transformed into a continuous two-dimensional temperature field function Interpolation method: The coordinates of each sensor are known. and its temperature readings , Representing the A sensor is used to estimate the temperature at any point on the stage surface using a spatial interpolation algorithm.

[0057] 103. Determine the coordinates of the reference marker points based on the image of the glass substrate;

[0058] In this embodiment, image processing is performed on the glass substrate image to identify and extract reference marker points on the glass substrate, and the coordinate information of each reference marker point at the current time is obtained. (Refer to...) Figure 7 The reference markers can be set along the loop.

[0059] 104. Calculate the actual displacement of the reference marker point based on its coordinates;

[0060] In this embodiment, the coordinates of the reference marker point at the current moment are compared with the coordinates under the reference state to calculate the actual displacement of the reference marker point.

[0061] 105. Based on the actual displacement of the reference markers and the two-dimensional temperature field function, construct a simulation model of the substrate stage;

[0062] In this embodiment, a simulation model of the substrate stage is constructed based on the actual displacement of the reference marker points and the two-dimensional temperature field function to describe the thermal deformation relationship between the stage and the glass substrate.

[0063] 106. Based on the substrate stage simulation model, calculate the temperature prediction data and the first displacement prediction data at multiple points. Use the temperature prediction data and the first displacement prediction data to train the preset convolutional neural network model to obtain the OLED substrate pixel slot positioning model.

[0064] In this embodiment, the simulation model of the control substrate stage outputs temperature data and corresponding displacement prediction data at multiple different spatial points, forming a temperature-displacement prediction dataset, which serves as the basis for subsequent model training. A convolutional neural network model is trained using this temperature-displacement prediction data to obtain the OLED substrate pixel slot positioning model.

[0065] 107. Input the two-dimensional temperature field function into the OLED substrate pixel slot positioning model and output the predicted displacement of the reference marker point;

[0066] In this embodiment, the currently constructed two-dimensional temperature field function is input into the OLED substrate pixel slot positioning model, and the predicted displacement of the reference marker point is output.

[0067] 108. Based on the coordinates of the reference marker and the predicted displacement, calculate the coordinates of the reference marker at the next moment.

[0068] In this embodiment, the coordinates of the reference marker at the next moment are calculated based on the current coordinates and predicted displacement of the reference marker, thereby realizing real-time prediction and positioning of the pixel slot position of the OLED substrate.

[0069] In this embodiment of the invention, temperature data from multiple temperature sensors on the stage are acquired and interpolated to construct a two-dimensional temperature field function. An image of the glass substrate is then obtained to determine the coordinates of a reference marker point. The actual displacement of the reference marker point is calculated, and a simulation model of the glass substrate stage is established based on the actual displacement and the two-dimensional temperature field function. This model is then trained and used to predict the displacement of the reference marker point using an OLED substrate pixel slot positioning model. Existing inkjet printing alignment methods can only measure the deformation result through optical images after deformation has occurred. Uneven heating of the glass substrate generates temperature gradients, leading to uneven thermal expansion and thus deformation. In this embodiment of the invention, by monitoring the temperature at different locations on the substrate in real time, the trend and magnitude of deformation can be predicted before or during deformation. The temperature field and convolutional neural network model compensate for thermally induced deformation errors, significantly improving positioning accuracy, reducing noise and delay caused by relying solely on real-time visual detection, and improving positioning stability. This method is suitable for high-precision pixel slot positioning of large-size glass substrates during processing. Furthermore, the temperature sensing system is relatively simple, low-cost, and easier to maintain.

[0070] Please refer to Figure 2 The second embodiment of the OLED substrate pixel slot positioning method of the present invention includes:

[0071] Based on the actual displacement of the reference marker points and the two-dimensional temperature field function, a simulation model of the substrate stage is constructed, including:

[0072] 201. Construct a thermo-coupled finite element model based on the glass substrate and the stage;

[0073] In this embodiment, 3D geometric models of the glass substrate and marble stage can be created using CAD software (such as SolidWorks, CATIA) or the modeling tools built into the finite element analysis software. Finite element analysis software options include ANSYS, Abaqus, or COMSOL Multiphysics. The glass substrate and marble stage are meshed using a hexahedral-dominated swept mesh. In critical areas, such as substrate edges and near marker points, mesh refinement is necessary to accurately capture stress concentration and displacement gradients. The mesh is continuously refined and recalculated until the change in maximum displacement is less than a threshold (1%), ensuring that the calculation results are independent of mesh density.

[0074] 202. Assign thermal properties, mechanical properties, and coefficient of thermal expansion to the thermo-coupled finite element model to obtain the first model;

[0075] In this embodiment, the properties of the glass substrate and the stage are as follows:

[0076] Thermal properties: thermal conductivity (k), specific heat capacity (c), density (ρ);

[0077] Mechanical properties: Young's modulus (E), Poisson's ratio (ν);

[0078] Thermodynamic coupling property: coefficient of thermal expansion.

[0079] Initial parameter values ​​can be obtained from material handbooks or data provided by suppliers, and subsequent models calibrate and adjust these nominal values.

[0080] 203. Apply boundary conditions and loads to the first model to obtain the second model;

[0081] In this embodiment, boundary conditions and loads are applied to the first model. These can include two-dimensional temperature field functions, convective heat transfer and radiative heat transfer, as well as boundary conditions at the interface between the glass substrate and the stage, making the model more consistent with real data.

[0082] 204. The thermo-mechanical coupling analysis of the second model is solved using a sequential coupling method to calculate the second displacement prediction data of the reference marker points;

[0083] In this embodiment, the deformation of the glass substrate has a minimal impact on macroscopic heat conduction; therefore, a more computationally efficient sequential coupling method is employed. First, a steady-state thermal analysis is performed to calculate the temperature distribution of the entire model. Then, this temperature distribution is used as a volume load and input into the subsequent static structural analysis to calculate thermally induced deformation.

[0084] 205. Based on the second displacement prediction data and the actual displacement, iteratively optimize the second model until it meets the preset conditions to obtain the substrate stage simulation model.

[0085] In this embodiment, the uncertain input parameters in the model are adjusted to minimize the difference between the model output (second displacement prediction data) and the actual displacement.

[0086] The optimization objective is to minimize the residual between the simulated second displacement prediction data and the actual displacement measured by the CCD camera.

[0087]

[0088] in, Represents the residual; It is the parameter vector to be calibrated; Iterate through all steady-state experimental data points (under different temperature conditions). Iterate through all CCD measurement markers; , It is the reference marker point of the finite element model. Temperature operating conditions and parameters Below, the second displacement prediction data in the X and Y axis directions; , It is a reference marker point Temperature operating conditions The actual displacement in the X and Y axis directions.

[0089] Parameter vector Candidate parameters may include the following:

[0090] P1: Coefficient of thermal expansion (CTE) of the glass substrate: It can be optimized as a constant, or if the temperature difference is large, it can be optimized as a first or second function of temperature.

[0091] P2: Contact thermal conductivity (TCC) at the glass-marble interface: determines the efficiency of heat transfer from the stage to the glass substrate, thus affecting the actual temperature distribution inside the glass and indirectly affecting deformation.

[0092] P3: Contact stiffness of the glass-marble interface: Under vacuum adsorption, the contact behavior of micro-irregularities can affect local displacement.

[0093] P4: Young's modulus (E) of glass: affects the ability of glass substrate materials to resist deformation.

[0094] P5: Convection heat transfer coefficient: Affects heat dissipation from the upper surface and sides of the glass substrate, thus affecting the overall temperature distribution.

[0095] First, a sensitivity analysis is performed to identify the parameters that have the greatest impact on the final displacement result as the main calibration targets, in order to reduce the dimensionality and complexity of the optimization problem.

[0096] The parameter optimization algorithm adopts a hybrid optimization strategy. First, a genetic algorithm is used to perform a global search to find a region close to the global optimum in the entire parameter space. The optimal solution obtained from the global search is used as the initial value, and the Levenberg-Marquardt local optimization algorithm is used for fine-tuning until it converges to the final solution.

[0097] The iterative calibration process is as follows:

[0098] Define the vector of parameters to be calibrated And its value range (constraints); select the optimization algorithm and set its parameters (such as the population size and crossover mutation rate of GA); set the iteration stopping condition; set the initial parameters. (Based on experience-based material manuals or data provided by suppliers).

[0099] Iteration loop begins (current iteration step) Based on historical evaluation results, a new candidate parameter vector is generated. For each steady-state experimental condition ( ), using parameters Update the input to the finite element model and run the solution to obtain the predicted displacement field for each working condition. Extract the values ​​of all reference points from the simulation results file. The second displacement prediction data corresponding to the position is then used. The extracted second displacement prediction data and the corresponding actual displacement are then substituted into the formula to calculate the current parameters. The objective function value .Will The value is returned to the optimization algorithm, which checks whether the preset iteration stopping condition is met. If it is met, the loop ends, and the current optimal value is returned. This is the calibration result; if it does not meet the requirements, then... Then, begin the next iteration.

[0100] The iteration terminates when any of the following conditions are met:

[0101] (1) The objective function converges: the relative change in the objective function value is less than a very small threshold in several consecutive steps.

[0102]

[0103] (2) Parameter convergence: The relative change of the parameter vector is less than a threshold in several consecutive steps.

[0104]

[0105] (3) Reaching the maximum number of iterations: To prevent the algorithm from failing to converge and taking too long, a maximum number of iterations N_max (200 times) is set.

[0106] In this embodiment of the invention, a thermo-mechanically coupled finite element model of the glass substrate and the stage is established. The model is assigned thermal properties, mechanical properties, and a coefficient of thermal expansion. A two-dimensional temperature field function is applied as a thermal load. The second displacement prediction data of the reference marker points is solved through sequential coupling. Based on the second displacement prediction data and the actual displacement, the model is iteratively optimized to obtain a simulation model of the substrate and stage. The thermo-induced deformation mechanism of the glass substrate and stage is introduced into the positioning process in the form of a finite element model, improving the physical consistency of the model. Iterative optimization makes the simulation model match the actual working conditions, reducing errors caused by model parameter uncertainties. This provides a highly reliable model foundation for subsequent displacement prediction and neural network training, improving the accuracy and repeatability of the temperature-displacement relationship description.

[0107] Please refer to Figure 3 The third embodiment of the OLED substrate pixel slot positioning method of the present invention includes:

[0108] The application of boundary conditions and loads to the first model includes:

[0109] 301. Apply a two-dimensional temperature field function as a load to the bottom surface of the glass substrate;

[0110] In this embodiment, the two-dimensional temperature field function As a thermal load applied to the model, As a specified temperature boundary condition, it is applied to all nodes on the bottom surface of the glass substrate (the surface in contact with the stage).

[0111] 302. Convection heat transfer and radiation heat transfer are applied as boundary conditions to the top and side surfaces of the glass substrate and the surface of the stage.

[0112] In this embodiment, convective and radiative heat transfer boundary conditions are applied to the upper and side surfaces of the glass substrate and the exposed surface of the marble stage to simulate heat exchange with the surrounding ambient air, and the ambient temperature, convective heat transfer coefficient and surface emissivity are defined.

[0113] 303. Contact thermal conductivity and contact stiffness are applied as boundary conditions to the interface between the glass substrate and the stage.

[0114] In this embodiment, the contact thermal conductivity (TCC) of the glass-marble interface determines the efficiency of heat transfer from the stage to the glass substrate, thereby affecting the actual temperature distribution inside the glass substrate and indirectly affecting deformation. The contact stiffness of the glass-marble interface, under vacuum adsorption, is affected by the contact behavior of microscopic irregularities, which can influence local displacement.

[0115] Additionally, structural loads can be assigned, with options including: gravity; vacuum adsorption; contact definition: defining a contact pair between the bottom surface of the glass substrate and the top surface of the marble platform; normal behavior: set to hard contact, allowing separation but not penetration; tangential behavior: can be set to frictional contact, with a friction coefficient set; displacement constraint: to prevent rigid body displacement during the solution process, appropriate fixed constraints need to be applied to the bottom or sides of the marble platform.

[0116] In this embodiment of the invention, convective and radiative heat transfer boundary conditions are applied to the glass substrate and the stage in the thermo-coupled finite element model, and contact thermal conductivity and contact stiffness are set at the contact interface between the glass substrate and the stage. This enables a more realistic simulation of the heat exchange and mechanical constraint state between the glass substrate and the stage during actual operation, reduces the deviation in temperature field and deformation calculations caused by the simplification of boundary conditions, improves the applicability of the simulation model under different ambient temperatures and heat dissipation conditions, and provides a more accurate training dataset for predicting the displacement of the reference marker point.

[0117] Please refer to Figure 4 The fourth embodiment of the OLED substrate pixel slot positioning method of the present invention includes:

[0118] The method of solving the thermo-mechanical coupling analysis of the second model using sequential coupling to calculate the second displacement prediction data of the reference marker points includes:

[0119] 401. Perform steady-state thermal analysis on the second model and calculate the temperature distribution of the second model;

[0120] 402. Input the temperature distribution as a volume load into the static structural analysis and calculate the deformation of the glass substrate;

[0121] In this embodiment, a steady-state thermal analysis is first performed to calculate the temperature distribution of the entire model. This temperature distribution is then used as a volume load and input into the subsequent static structural analysis to calculate thermally induced deformation. Since the deformation of the glass substrate is at the micrometer level, its impact on macroscopic heat conduction is minimal. Therefore, a sequential coupling method with higher computational efficiency is adopted.

[0122] 403. Based on the deformation of the glass substrate, the second displacement prediction data of the reference mark point is obtained.

[0123] In this embodiment, the second displacement prediction data of the reference marker points is extracted based on the deformation results of the glass substrate.

[0124] In this embodiment of the invention, a sequential coupling approach is used to solve the thermo-coupled finite element model. First, steady-state thermal analysis is performed to obtain the temperature distribution. Then, this temperature distribution is used as a volume load input to static structural analysis to calculate the deformation of the glass substrate, thereby obtaining the predicted second displacement data of the reference marker point. While ensuring the accuracy of the thermo-analysis, this approach reduces the overall computational complexity, minimizes numerical instability issues in fully coupled analysis, improves the reliability of the model solution, facilitates model parameter adjustment and multiple iterative calculations, and enhances modeling efficiency.

[0125] Please refer to Figure 5 The fifth embodiment of the OLED substrate pixel slot positioning method of the present invention includes:

[0126] The process involves calculating temperature prediction data and first displacement prediction data for multiple points based on the substrate stage simulation model, and then training a pre-defined convolutional neural network model using the temperature prediction data and the first displacement prediction data to obtain an OLED substrate pixel slot positioning model, including:

[0127] 501. Convert each temperature prediction data into a first two-dimensional array, and convert each first displacement prediction data into a second two-dimensional array;

[0128] In this embodiment, for the input features, the input temperature field Discretize the image into an N×M two-dimensional array (the first two-dimensional array), where each pixel value represents the temperature at that location. For the output target: output the deformation field. and The input and output are also discretized into two N×M two-dimensional arrays (images), i.e., the predicted two-dimensional array. The actual displacement in the X and Y axes is discretized into two N×M two-dimensional arrays (images), i.e., the second two-dimensional array. Transforming the input and output into two-dimensional arrays is suitable for processing by convolutional neural networks (CNNs), transforming the problem into an image-to-image regression task.

[0129] 502. Combine all the first two-dimensional arrays and their corresponding second two-dimensional arrays into multiple data pairs, and randomly divide all the data pairs to obtain the training set;

[0130] In this embodiment, all generated data pairs are randomly divided into a training set, a validation set, and a test set in an 8:1:1 ratio. The training set is used to train the model and adjust its weights. The validation set is used during training to monitor model performance, perform hyperparameter tuning, and prevent overfitting. The test set is not involved in the training process; it is only used to evaluate the model's generalization ability and performance after training is complete.

[0131] 503. Construct the U-Net network structure, input the first two-dimensional array of the training set into the U-Net network structure, and output the predicted two-dimensional array;

[0132] 504. Calculate the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first input two-dimensional array;

[0133] 505. Update the network weights of the U-Net network structure according to the loss function, and return to the step of calculating the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first two-dimensional array of input, until the training set is traversed to obtain the OLED substrate pixel slot localization model.

[0134] In this embodiment, data preprocessing involves normalizing the input (temperature field) and output (deformation field) data (subtracting the mean and dividing by the standard deviation, or scaling to the [0, 1] interval) to accelerate model convergence and improve performance.

[0135] A U-Net network structure was constructed, using mean squared error as the loss function.

[0136]

[0137] in, The model outputs a two-dimensional array of predictions. This is a second two-dimensional array transformed from the input first displacement prediction data. Based on the prediction two-dimensional array, it can be converted into predicted displacement.

[0138] Training loop: Data is drawn from the training set in mini-batches and fed into the U-Net model for forward propagation to obtain the predicted deformation field. The loss between the predicted displacement and the actual displacement is calculated. Then, backpropagation is performed to calculate the gradient of the loss function with respect to each weight in the network. The Adam optimizer is used to update the network weights based on the gradient. The above steps are repeated until the entire training set has been traversed.

[0139] After each training cycle, the model performance is evaluated on the validation set. If the loss on the validation set no longer decreases, an early stopping mechanism can be triggered to stop training and prevent overfitting.

[0140] In this embodiment of the invention, a dataset is constructed and a convolutional neural network model is trained based on multiple temperature prediction data and first displacement prediction data output from the substrate stage simulation model. The convolutional neural network model adopts a U-Net network structure and is used to quickly predict the displacement of the reference marker point based on the two-dimensional temperature field. By mapping the complex finite element simulation calculation results to an efficient neural network prediction model, rapid online displacement prediction is achieved. The U-Net network's ability to extract spatial features of the two-dimensional temperature field improves prediction accuracy, significantly reduces computational resource consumption during real-time positioning, and enhances the overall response speed of the OLED substrate pixel slot positioning method.

[0141] The OLED substrate pixel slot positioning method of the present invention has been described above. The OLED substrate pixel slot positioning system in the embodiments of the present invention will be described below, with reference to... Figure 6 One embodiment of the OLED substrate pixel slot positioning system of the present invention includes:

[0142] The system includes a control device, a frame 8, a stage 1, and multiple temperature sensors 2. The stage 1 is mounted on the frame 8 and is used to support a glass substrate. The temperature sensors 2 are mounted on the stage 1. The glass substrate has reference marking points. The control device is electrically connected to the multiple temperature sensors 2. The OLED substrate pixel slot positioning system is used to perform the OLED substrate pixel slot positioning method as described above.

[0143] In this embodiment, multiple temperature sensors 2 are set on the stage 1. The structure is simple and can stably acquire the temperature information of the stage 1, providing a reliable data source for establishing a two-dimensional temperature field function and simulation model. Specifically, the temperature sensors 2 are embedded in the surface of the stage 1 and arranged at key locations on the stage 1, including the edges, center and key support points of the stage 1, so as to collect the temperature distribution of the entire stage 1 and the glass substrate in real time and synchronously.

[0144] In a second embodiment of the OLED substrate pixel slot positioning system of the present invention, the OLED substrate pixel slot positioning system further includes an X-axis moving mechanism 7, a Y-axis moving mechanism 6, and a CCD camera 3. The fixed part of the Y-axis moving mechanism 6 is mounted on the frame, and the movable part of the Y-axis moving mechanism 6 is connected to the stage 1 via a transmission connection. The Y-axis moving mechanism 6 drives the stage 1 to move along the Y-axis direction. The fixed part of the X-axis moving mechanism 7 is mounted on the frame, and the CCD camera 3 is mounted on the movable part of the X-axis moving mechanism 7. The X-axis moving mechanism 7 drives the CCD camera 3 to move along the X-axis direction. The CCD camera 3 is located above the glass substrate and is used to photograph the glass substrate. The control device is electrically connected to the Y-axis moving mechanism 6, the X-axis moving mechanism 7, and the CCD camera 3.

[0145] In this embodiment, an X-axis moving mechanism 7, a Y-axis moving mechanism 6, and a CCD camera 3 are configured to acquire images of the glass substrate and obtain the coordinates of reference marker points, thereby improving the spatial coverage of reference marker point detection, meeting the positioning requirements of large-size glass substrates, and providing image data for reference marker point displacement calculation and model correction. Specifically, the CCD camera 3 identifies reference marker points on the glass substrate and determines the actual positions of these reference marker points in the coordinate systems of the X-axis moving mechanism 7 and the Y-axis moving mechanism 6.

[0146] In a third embodiment of the OLED substrate pixel slot positioning system of the present invention, the OLED substrate pixel slot positioning system further includes a first laser interferometer 5 and a second laser interferometer 4. The first laser interferometer 5 is mounted on the frame 8 and is used to measure the distance between itself and the stage 1 along the Y-axis. The second laser interferometer 4 is mounted on the fixed part of the X-axis moving mechanism 7 and is used to measure the distance between itself and the CCD camera 3 along the X-axis. The control device is electrically connected to the first laser interferometer 5 and the second laser interferometer 4.

[0147] In this embodiment, laser interferometers are respectively set in the movement directions of the stage 1 and the CCD camera 3 to measure the displacement information of the stage 1 and the CCD camera 3 and to link with the control device, thereby further improving the overall positioning accuracy and stability of the system. Specifically, the first laser interferometer 5 measures the movement distance of the stage 1 in the Y-axis direction, and the second laser interferometer 4 and the CCD camera 3 measure the movement distance in the X-axis direction.

[0148] Figure 8This is a schematic diagram of the structure of an OLED substrate pixel slot positioning device 800 provided in an embodiment of the present invention. The OLED substrate pixel slot positioning device 800 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 810 (e.g., one or more processors) and a memory 820, and one or more storage media 830 (e.g., one or more mass storage devices) storing application programs 833 or data 832. The memory 820 and storage media 830 can be temporary or persistent storage. The program stored in the storage media 830 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the OLED substrate pixel slot positioning device 800. Furthermore, the processor 810 may be configured to communicate with the storage media 830 and execute the series of instruction operations in the storage media 830 on the OLED substrate pixel slot positioning device 800 to implement the steps of the OLED substrate pixel slot positioning method provided in the above-described method embodiments.

[0149] The OLED substrate pixel slot positioning device 800 may also include one or more power supplies 840, one or more wired or wireless network interfaces 850, one or more input / output interfaces 860, and / or one or more operating systems 831, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 8 The illustrated OLED substrate pixel slot positioning device structure does not constitute a limitation on the OLED substrate pixel slot positioning device, and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0150] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the OLED substrate pixel slot positioning method.

[0151] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0152] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0153] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An OLED substrate pixel slot positioning method, characterized in that, An OLED substrate pixel slot positioning system is applied, comprising: a control device, a stage, and multiple temperature sensors. The stage supports a glass substrate, and the multiple temperature sensors are mounted on the stage. The control device is electrically connected to the multiple temperature sensors. The OLED substrate pixel slot positioning method includes the following steps: Acquire temperature data from all sensors and obtain images of the glass substrate; Interpolate all the temperature data to calculate the two-dimensional temperature field function; Determine the coordinates of the reference marker points based on the image of the glass substrate; The actual displacement of the reference marker point is calculated based on its coordinates. Based on the actual displacement of the reference markers and the two-dimensional temperature field function, a simulation model of the substrate stage is constructed. Based on the substrate stage simulation model, temperature prediction data and first displacement prediction data at multiple points are calculated. The temperature prediction data and first displacement prediction data are used to train a preset convolutional neural network model to obtain the OLED substrate pixel slot positioning model. The two-dimensional temperature field function is input into the OLED substrate pixel slot positioning model, and the predicted displacement of the reference marker point is output. Based on the coordinates of the reference marker and the predicted displacement, the coordinates of the reference marker at the next moment are calculated.

2. The method of claim 1, wherein, Based on the actual displacement of the reference marker points and the two-dimensional temperature field function, a simulation model of the substrate stage is constructed, including: A thermo-coupled finite element model was constructed based on the glass substrate and the stage. The thermo-coupled finite element model is given thermal properties, mechanical properties, and coefficient of thermal expansion to obtain the first model; By applying boundary conditions and loads to the first model, the second model is obtained. The second model was solved by sequential coupling analysis to obtain the second displacement prediction data of the reference marker point. Based on the second displacement prediction data and the actual displacement, the second model is iteratively optimized until it meets the preset conditions, thus obtaining the simulation model of the substrate stage.

3. The method of claim 2, wherein, The application of boundary conditions and loads to the first model includes: A two-dimensional temperature field function is applied as a load to the bottom surface of the glass substrate; Convection and radiation heat transfer are applied as boundary conditions to the top and side surfaces of the glass substrate and the surface of the stage. Contact thermal conductivity and contact stiffness are applied as boundary conditions to the interface between the glass substrate and the stage.

4. The method of claim 2, wherein, The method of solving the thermo-mechanical coupling analysis of the second model using sequential coupling to calculate the second displacement prediction data of the reference marker points includes: Steady-state thermal analysis was performed on the second model to calculate its temperature distribution. Temperature distribution is used as a volume load input for static structural analysis to calculate the deformation of the glass substrate. Based on the deformation of the glass substrate, the second displacement prediction data of the reference mark point is obtained.

5. The method of claim 1, wherein: Based on the substrate stage simulation model, temperature prediction data and first displacement prediction data for multiple points are calculated. A pre-set convolutional neural network model is trained using the temperature prediction data and the first displacement prediction data to obtain the OLED substrate pixel slot positioning model, including: Each temperature prediction data is converted into a first two-dimensional array, and each first displacement prediction data is converted into a second two-dimensional array; All the first two-dimensional arrays and their corresponding second two-dimensional arrays are combined into multiple data pairs, and all data pairs are randomly divided to obtain the training set. Construct a U-Net network structure, input the first two-dimensional array of the training set into the U-Net network structure, and output a predicted two-dimensional array; Calculate the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first input two-dimensional array; Based on the loss function, update the network weights of the U-Net network structure, and return to the step of calculating the loss function based on the predicted two-dimensional array and the second two-dimensional array corresponding to the first two-dimensional array of input, until the training set is traversed to obtain the OLED substrate pixel slot localization model.

6. An OLED substrate pixel slot positioning system, characterized in that, include: The system comprises a control device, a frame (8), a stage (1), and multiple temperature sensors (2). The stage (1) is mounted on the frame (8) and is used to support a glass substrate. The temperature sensors (2) are mounted on the stage (1). The glass substrate has reference markings. The control device is electrically connected to the multiple temperature sensors (2). The OLED substrate pixel slot positioning system is used to perform the OLED substrate pixel slot positioning method as described in any one of claims 1-5.

7. The OLED substrate pixel slot positioning system of claim 6, wherein, The OLED substrate pixel slot positioning system further includes an X-axis moving mechanism (7), a Y-axis moving mechanism (6), and a CCD camera (3). The fixed part of the Y-axis moving mechanism (6) is mounted on the frame (8), and the movable part of the Y-axis moving mechanism (6) is connected to the stage (1) for transmission. The Y-axis moving mechanism (6) drives the stage (1) to move along the Y-axis direction. The fixed part of the X-axis moving mechanism (7) is mounted on the frame (8), and the CCD camera (3) is mounted on the movable part of the X-axis moving mechanism (7). The X-axis moving mechanism (7) drives the CCD camera (3) to move along the X-axis direction. The CCD camera (3) is located above the glass substrate and is used to photograph the glass substrate. The control device is electrically connected to the Y-axis moving mechanism (6), the X-axis moving mechanism (7), and the CCD camera (3).

8. The OLED substrate pixel slot positioning system of claim 7, wherein: The OLED substrate pixel slot positioning system further includes a first laser interferometer (5) and a second laser interferometer (4). The first laser interferometer (5) is mounted on the frame (8) and is used to measure the distance between it and the stage (1) along the Y-axis. The second laser interferometer (4) is mounted on the fixed part of the X-axis moving mechanism (7) and is used to measure the distance between it and the CCD camera (3) along the X-axis. The control device is electrically connected to the first laser interferometer (5) and the second laser interferometer (4).

9. An OLED substrate pixel slot positioning apparatus, characterized in that, The OLED substrate pixel slot positioning device includes: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause the OLED substrate pixel slot positioning device to perform the steps of the OLED substrate pixel slot positioning method as described in any one of claims 1-5.

10. A computer-readable storage medium having stored thereon instructions, the instructions comprising, When the instructions are executed by the processor, they implement each step of the OLED substrate pixel slot positioning method as described in any one of claims 1-5.

Images