A method and apparatus for manufacturing a flexible circuit board

Abstract

The application provides a flexible circuit board manufacturing method and device. The method comprises the following steps: obtaining a first preset pressing parameter; obtaining a bubble prediction result according to the first preset pressing parameter and a trained bubble prediction model; adjusting the first preset pressing parameter according to the bubble prediction result to obtain a second preset pressing parameter; obtaining a first flexible circuit board according to the second preset pressing parameter, a cover film and a flexible circuit board substrate; adjusting the second preset pressing parameter in real time according to the first flexible circuit board to obtain a third preset pressing parameter; and obtaining a target flexible circuit board according to the third preset pressing parameter, the cover film and the flexible circuit board substrate. The application is beneficial to reducing the bubble rate of the flexible circuit board and improving the use safety of the flexible circuit board.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of flexible circuit board manufacturing, and further relates to a flexible circuit board manufacturing method and device. BACKGROUND

[0002] In the production process of a flexible circuit board (FPC), after etching to form a circuit pattern, a polyimide film and a glue layer are attached to the circuit pattern as a cover layer to protect the circuit, enhance the reliability of the flexible circuit board, and ensure the realization of flexible functions. However, in the process of pressing the cover layer and the circuit, due to incomplete gas discharge or insufficient flowability of the glue layer, air bubbles are left between the cover layer and the substrate, and the air bubble area is prone to cause copper foil oxidation and circuit rupture, and accelerate delamination failure when bending, which greatly affects the safety of the flexible circuit board. SUMMARY

[0003] The present application aims to solve the technical problem of providing a flexible circuit board manufacturing method and device to reduce air bubbles and improve the safety of the flexible circuit board.

[0004] To solve the above technical problems, the technical solution of the present application is as follows:

[0005] In a first aspect, the present application provides a flexible circuit board manufacturing method, comprising:

[0006] Obtaining a first preset pressing parameter;

[0007] According to the first preset pressing parameter and the trained air bubble prediction model, an air bubble prediction result is obtained;

[0008] According to the air bubble prediction result, the first preset pressing parameter is adjusted to obtain a second preset pressing parameter;

[0009] According to the second preset pressing parameter, a cover film and a flexible circuit board substrate, a first flexible circuit board is obtained; the cover film includes a polyimide film and a glue layer, and the flexible circuit board substrate has etched circuit lines thereon;

[0010] According to the first flexible circuit board, the second preset pressing parameter is adjusted in real time to obtain a third preset pressing parameter;

[0011] According to the third preset pressing parameter, a cover film and a flexible circuit board substrate, a target flexible circuit board is obtained.

[0012] Optionally, the first preset pressing parameter includes:

[0013] The preset pressing temperature, the preset pressing pressure, the preset pressing time and the preset glue layer thickness.

[0014] Optionally, according to the first preset lamination parameter and the trained bubble prediction model, a bubble prediction result is obtained, including:

[0015] The first preset lamination parameter is preprocessed to obtain preprocessed data;

[0016] The preprocessed data is format-converted to obtain a feature vector;

[0017] The feature vector is input into the trained bubble prediction model to obtain a bubble prediction result.

[0018] Optionally, the training process of the bubble prediction model includes:

[0019] Historical data is obtained; the historical data includes historical lamination parameters and corresponding bubble area proportions;

[0020] The historical data is processed to obtain sample data;

[0021] The sample data is divided to obtain a training set and a test set;

[0022] The preset network model is trained according to the training set to obtain a first network model;

[0023] The first network model is tested according to the test set to obtain a test result;

[0024] The first network model is adjusted according to the test result and a preset evaluation index to obtain a bubble prediction model.

[0025] Optionally, the first preset lamination parameter is adjusted according to the bubble prediction result to obtain a second preset lamination parameter, including:

[0026] The first preset lamination parameter is adjusted according to the bubble prediction result and a preset bubble rate to obtain a second preset lamination parameter.

[0027] Optionally, a first flexible circuit board is obtained according to the second preset lamination parameter, a cover film and a flexible circuit board substrate, including:

[0028] The second preset lamination parameter is input into a laminator, so that the laminator performs lamination processing on the cover film and the flexible circuit board substrate according to the second preset lamination parameter to obtain a first flexible circuit board.

[0029] Optionally, the second preset lamination parameter is adjusted in real time according to the first flexible circuit board to obtain a third preset lamination parameter, including:

[0030] According to the first flexible circuit board, bubble data is obtained; the bubble data includes bubble position data and bubble area data;

[0031] According to the bubble data and a preset bubble rate, the second preset pressing parameter is adjusted in real time to obtain a third preset pressing parameter.

[0032] In a second aspect of the present application, a flexible circuit board manufacturing device is provided, comprising:

[0033] The acquisition module is configured to acquire a first preset pressing parameter.

[0034] The processing module is configured to obtain a bubble prediction result according to the first preset pressing parameter and a trained bubble prediction model, adjust the first preset pressing parameter according to the bubble prediction result to obtain a second preset pressing parameter, obtain a first flexible circuit board according to the second preset pressing parameter, a cover film and a flexible circuit board substrate, wherein the cover film comprises a polyimide film and a glue layer, and the flexible circuit board substrate has a circuit etched thereon, adjust the second preset pressing parameter in real time according to the first flexible circuit board to obtain a third preset pressing parameter, and obtain a target flexible circuit board according to the third preset pressing parameter, the cover film and the flexible circuit board substrate.

[0035] In a third aspect of the present application, a computing device is provided, comprising a processor and a memory storing a computer program, wherein the computer program is executed by the processor to perform the method of the first aspect.

[0036] In a fourth aspect of the present application, a computer readable storage medium is provided, storing instructions, wherein the instructions are executed on a computer to cause the computer to perform the method of the first aspect.

[0037] The above-mentioned scheme of the present application has at least the following beneficial effects:

[0038] The above-mentioned scheme of the present application predicts bubbles through a bubble prediction model and the acquired first preset pressing parameter, then adjusts the first preset pressing parameter according to the prediction result, then manufactures a first flexible circuit board according to the second preset pressing parameter, the cover film and the flexible circuit board substrate, and then manufactures a target flexible circuit board by adjusting the second preset pressing parameter in real time according to the bubble data on the first flexible circuit board, which is beneficial to reduce the bubble rate on the flexible circuit board and improve the use safety of the flexible circuit board. BRIEF DESCRIPTION OF DRAWINGS

[0039] Figure 1 is a flowchart of the manufacturing method of the flexible circuit board in the embodiment of the present application;

[0040] Figure 2It is a structure schematic view of a manufacturing device of a flexible circuit board in the embodiment of the present application. DETAILED DESCRIPTION

[0041] Exemplary embodiments of the present application will be described herein below with reference to the accompanying drawings. While exemplary embodiments of the present application are shown in the drawings, it is understood that the present application can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this application will be thorough and complete, and will fully convey the scope of the application to those skilled in the art.

[0042] As shown in Figure 1 the embodiment of the present application proposes a manufacturing method of a flexible circuit board, comprising the following steps:

[0043] Step 101, obtaining a first preset pressing parameter;

[0044] Step 102, obtaining a bubble prediction result according to the first preset pressing parameter and a trained bubble prediction model;

[0045] Step 103, adjusting the first preset pressing parameter according to the bubble prediction result to obtain a second preset pressing parameter;

[0046] Step 104, obtaining a first flexible circuit board according to the second preset pressing parameter, a cover film and a flexible circuit board substrate; the cover film comprises a polyimide film and a glue layer, and the flexible circuit board substrate has etched lines thereon;

[0047] Step 105, adjusting the second preset pressing parameter in real time according to the first flexible circuit board to obtain a third preset pressing parameter;

[0048] Step 106, obtaining a target flexible circuit board according to the third preset pressing parameter, the cover film and the flexible circuit board substrate.

[0049] The manufacturing method of the flexible circuit board in the embodiment of the present application performs bubble prediction through the bubble prediction model and the obtained first preset pressing parameter, then adjusts the first preset pressing parameter according to the prediction result, then manufactures the first flexible circuit board according to the second preset pressing parameter, the cover film and the flexible circuit board substrate, then adjusts the second preset pressing parameter in real time according to the bubble data on the first flexible circuit board to manufacture the target flexible circuit board, which is conducive to reducing the bubble rate on the flexible circuit board and improving the use safety of the flexible circuit board.

[0050] In an optional embodiment of the present application, the first preset pressing parameter in step 101 comprises:

[0051] a preset pressing temperature, a preset pressing pressure, a preset pressing time and a preset glue layer thickness.

[0052] Specifically, the temperature, the pressure, the time and the thickness of the adhesive layer are key parameters for pressing the cover layer (such as the adhesive layer of the polyimide film) and the flexible circuit board substrate with the circuit (or the circuit pattern, etc.), in a specific embodiment, the preset pressing temperature can be 160℃, the preset pressing pressure can be 1.5MPa, the preset pressing time can be 30 seconds, and the preset thickness of the adhesive layer can be 25μm, the first preset pressing parameter can be set according to the type of the cover layer and the material of the flexible circuit board substrate, etc., and the subsequent bubble prediction and parameter adjustment are performed according to the first preset pressing parameter.

[0053] In an optional embodiment of the present application, step 102 comprises:

[0054] Step 10211, preprocessing the first preset pressing parameter to obtain preprocessing data;

[0055] Specifically, preprocessing the first preset pressing parameter can include: verifying the first preset pressing parameter to obtain the verified parameter, so as to ensure the rationality and effectiveness of the first preset pressing parameter; generating the target interaction parameter according to the verified parameter, using the target interaction parameter can effectively capture the nonlinear relationship in the flexible circuit board cover layer film pasting process, and improve the prediction accuracy of the subsequent prediction model for the bubble area ratio; the preprocessing data includes the verified parameter and the target interaction parameter.

[0056] Here, the interaction parameter can include a first interaction parameter and a second interaction parameter, wherein the first interaction parameter is obtained by A1=T×P, wherein A1 is the first interaction parameter, T is the preset pressing temperature, and P is the preset pressing pressure; the first interaction parameter reflects the synergistic effect of temperature and pressure, such as high temperature enhancing the flowability of the adhesive layer, and high pressure promoting the filling of the adhesive layer gap, and the product of the two can represent the comprehensive effect; the second interaction parameter is obtained by A2=t / d, wherein A2 is the second interaction parameter, t is the preset pressing time, and d is the preset thickness of the adhesive layer; the second interaction parameter represents the processing time required by the unit thickness of the adhesive layer, such as thinner adhesive layer needs shorter time to complete the gas discharge, and the ratio of time and adhesive layer thickness can quantify this relationship; the greater the ratio of the preset pressing time and the preset thickness of the adhesive layer, the longer the processing time of the unit thickness of the adhesive layer, which will affect the bubble residual rate.

[0057] Step 10212, format conversion is performed on the preprocessing data to obtain a feature vector;

[0058] Specifically, the temperature, pressure, time, adhesive layer thickness, first interaction parameter and second interaction parameter in the preprocessed data can be sorted in a preset order, and then arranged into an array, i.e., a feature vector. For example, the feature vector of a specific embodiment can be represented as [T, P, t, d, A1, A2]. This facilitates subsequent processing of the data by the bubble prediction model and improves processing efficiency.

[0059] At step 10213, the feature vector is input into the trained bubble prediction model to obtain a bubble prediction result.

[0060] Specifically, the bubble prediction model receives the feature vector as input, and each tree in the bubble prediction model independently predicts the input feature vector to obtain a prediction result. The prediction results of all trees are aggregated (which can be averaging or majority voting) to obtain the final bubble area ratio prediction value as the output result of the model. In addition, the bubble prediction model can also output the importance of each feature in the feature vector, such as temperature, pressure, time, and adhesive layer thickness, so as to optimize the first preset pressing parameter according to the importance. The importance of each feature is calculated as follows:

[0061]

[0062] where D is the feature set contained by the node, Entropy(D) is the information entropy, c is the number of classes, p i is the probability that the feature belongs to the i-th class, Entropy split (D, feature, threshold) is the weighted information entropy after splitting, |D left | is the number of features in the left child node after splitting, |D right | is the number of features in the right child node after splitting, Entropy(D left ) is the information entropy of the left child node, Entropy(D right ) is the information entropy of the right child node, Gain(D, feature, threshold) is the information gain, TotalGain f is the total information gain of the f-th feature, F is the total number of features, N is the number of trees in the bubble prediction model, and Importance f is the importance of the f-th feature.

[0063] It should be noted that the bubble prediction result can include a bubble area ratio prediction value, an importance degree of temperature, an importance degree of pressure, an importance degree of time, and an importance degree of glue layer thickness. For example, in a specific embodiment, the bubble prediction result includes a bubble area ratio prediction value of 0.5%, an importance degree of temperature of 35%, an importance degree of pressure of 28%, an importance degree of time of 20%, and an importance degree of glue layer thickness of 12%. Since the bubble area ratio prediction value is greater than the preset bubble rate of 0.2%, the first preset pressing parameter needs to be adjusted. When the first preset pressing parameter is adjusted, the temperature and the pressure with higher importance degrees can be adjusted preferentially.

[0064] In an optional embodiment of the present application, the training process of the bubble prediction model in step 102 includes:

[0065] In step 10221, historical data is obtained. The historical data includes historical pressing parameters and corresponding bubble area ratios.

[0066] Specifically, a preset number of historical data can be obtained as needed, such as 500 groups of historical data from a manufacturer. The historical data includes historical pressing parameters and corresponding bubble area ratios. The historical pressing parameters include historical pressing temperature, historical pressing pressure, historical pressing time, and historical glue layer thickness. In addition, the historical pressing parameters need to cover the first preset pressing parameter. For example, the first preset pressing parameter includes a preset pressing temperature of 160°C, a preset pressing pressure of 1.5 MPa, a preset pressing time of 30 seconds, and a preset glue layer thickness of 25 μm. The historical pressing temperature in the historical pressing parameter can range from 150 to 200°C, the historical pressing pressure can range from 1 to 4 MPa, the historical pressing time can range from 20 to 80 seconds, and the historical glue layer thickness can range from 15 to 50 μm.

[0067] In step 10222, the historical data is processed to obtain sample data.

[0068] Specifically, the historical data is processed for missing values to obtain first historical data. The missing value processing method can be filling (such as mean interpolation) or deleting incomplete data.

[0069] The first historical data is processed to remove outliers to obtain second historical data. The outlier removal processing method can be to identify and remove outliers (such as removing abnormal batches with a bubble area ratio > 10%) through a box plot or a 3σ principle.

[0070] The second historical data is processed for standardization to obtain third historical data. The standardization processing method can be to process the second historical data through The second historical data is standardized, wherein X2 is the third historical data, X1 is the second historical data, μ is the mean of the second historical data, and σ is the standard deviation of the second historical data.

[0071] The historical interaction parameters are generated according to the third historical data, and the historical interaction parameters include first historical interaction parameters and second historical interaction parameters; wherein the first interaction parameter is obtained by A1 ′ =T ′ ×P ′ , wherein A1 ′ is the first historical interaction parameter, T ′ is the historical pressing temperature, and P ′ is the historical pressing pressure; the second interaction parameter is obtained by A2 ′ =t ′ / d ′ , wherein A2 ′ is the second historical interaction parameter, t ′ is the historical pressing time, and d ′ is the historical glue layer thickness. The purpose of calculating the historical interaction parameters is to enhance the model's ability to capture nonlinear relationships and improve the accuracy of the model.

[0072] The sample data is obtained according to the third historical data and the historical interaction parameters; wherein the sample data includes the third historical data and the historical interaction parameters.

[0073] In step 10223, the sample data is divided to obtain a training set and a test set;

[0074] Specifically, the sample data can be divided according to a ratio of 8:2, with the training set being 80% and the test set being 20%.

[0075] In step 10224, a first network model is obtained by training a preset network model according to the training set.

[0076] Specifically, the number of numbers in the preset network model is 100 to 500, the maximum depth of the tree is 5 to 20, the minimum sample number for node splitting is 2 to 10, and the maximum number of features for each tree is sqrt(the number of features that each tree can consider when splitting nodes is the square root of the total number of features) or log2(the number of features that each tree can consider when splitting nodes is the logarithm value of the total number of features with 2 as the base). After the training set is converted into a feature matrix X=[T, P, t, d, A1, A2] and input into the preset network model, the preset network model is trained using the least mean square error as the objective function to obtain the first network model. The objective function is wherein MSE is the mean square error, M is the number of training sets, is the predicted value of the i-th training data, The actual value of the i-th training data.

[0077] Step 10225, testing the first network model according to the test set to obtain a test result;

[0078] Specifically, the first network model is tested using the test set to obtain a test result, and the test result includes an output result of the model and a corresponding bubble area proportion in the test set.

[0079] Step 10226, adjusting the first network model according to the test result and a preset evaluation index to obtain a bubble prediction model.

[0080] Specifically, the preset evaluation index can include a determination coefficient and a mean absolute error, the determination coefficient can measure the proportion of the model to explain the data variation, and the mean absolute error can directly reflect the absolute value of the prediction error, and by obtaining the determination coefficient, and by obtaining the mean absolute error, wherein R 2 is the determination coefficient, n is the data amount in the test set, y i is the i-th actual observation value, is the i-th predicted value of the model, is the average value of all actual observation values, and MAE is the mean absolute error. If the determination coefficient is greater than a first preset value and the mean absolute error is less than a second preset value, the first network model at this time is the bubble prediction model; if the determination coefficient is less than the first preset value or the mean absolute error is greater than the second preset value, the number of numbers in the model, the maximum depth of the tree, the minimum sample number of node splitting, and the maximum feature number of each tree are adjusted to make the determination coefficient greater than the first preset value and the mean absolute error less than the second preset value, so as to obtain the bubble prediction model.

[0081] In an optional embodiment of the application, step 103 includes:

[0082] According to the bubble prediction result and a preset bubble rate, the first preset pressing parameter is adjusted to obtain a second preset pressing parameter.

[0083] Specifically, if the bubble area proportion prediction value in the bubble prediction result is greater than the preset bubble rate, an adjustment order is determined according to the importance of each parameter in the bubble prediction result, and the first preset pressing parameter is adjusted according to the adjustment order. In a specific embodiment, the bubble area proportion prediction value is 0.5%, the importance of temperature is 35%, the importance of pressure is 28%, the importance of time is 20%, and the importance of the adhesive layer thickness is 12%. Since the bubble area proportion prediction value is greater than the preset bubble rate 0.2%, the first preset pressing parameter needs to be adjusted. Since the importance of temperature and pressure is higher, temperature and pressure are adjusted first to obtain the second preset pressing parameter, and the second preset pressing parameter includes the adjusted temperature, the adjusted pressure, the adjusted time and the adjusted adhesive layer thickness.

[0084] In a specific embodiment, the manner of adjusting the first preset pressing parameter is:

[0085] The preset bubble rate and the constraint condition are obtained, such as the preset bubble rate is 0.2%, and the constraint condition is that the adjusted temperature is within the actual feasible temperature range, the adjusted pressure is within the actual feasible pressure range, the adjusted time is within the actual feasible time range, the adjusted adhesive layer thickness is within the actual feasible thickness range, and the target bubble rate is within the actual feasible bubble rate range.

[0086] The adjustment target function is determined according to the preset bubble rate, such as minimizing the absolute value of the difference between the target bubble rate and the preset bubble rate.

[0087] A group of pressing parameters are randomly generated as an initial population. The target function value of each individual (i.e., each group of parameter combinations) is calculated, including the absolute value of the difference between the target bubble rate and the preset bubble rate. The better individual is selected according to the fitness value for breeding. The offspring individual is generated by exchanging part of the genes of the parent individual. The genes of the offspring individual are randomly changed with a preset probability to increase the diversity of the population. The steps of calculating the target function value of each individual are repeated until the step of randomly changing the genes of the offspring individual with a preset probability until a predetermined iteration number or a convergence condition is reached.

[0088] After each iteration, the non-dominated sorting is used to extract the Pareto front according to the target function value of the individual.

[0089] One or more final schemes are selected from the Pareto front as the second preset pressing parameter according to the preset bubble rate.

[0090] In an optional embodiment of the present application, step 104 comprises:

[0091] The second preset pressing parameter is input into the laminator, so that the laminator performs pressing processing on the cover film and the flexible circuit board substrate according to the second preset pressing parameter to obtain the first flexible circuit board.

[0092] Specifically, the second preset pressing parameter can be converted into a control parameter recognizable by the control system of the laminator, and then the control parameter is input into the laminator, so that the laminator performs pressing treatment on the cover film and the flexible circuit board substrate according to the control parameter, and a first flexible circuit board is obtained.

[0093] In an optional embodiment of the present application, step 105 comprises:

[0094] Step 1051, obtaining bubble data according to the first flexible circuit board; the bubble data comprises bubble position data and bubble area data;

[0095] Step 1052, adjusting the second preset pressing parameter in real time according to the bubble data and a preset bubble rate, to obtain a third preset pressing parameter.

[0096] Specifically, the AOI (Automatic Optical Inspection) system is used to perform high-resolution scanning on the first flexible circuit board to obtain a clear image; the scanned image is preprocessed, including denoising, contrast enhancement and other operations, to improve the accuracy of image segmentation; a preset image segmentation algorithm (such as UNet image segmentation algorithm) is applied to segment the preprocessed image to obtain a segmentation result; the position (such as the center of mass coordinates) and area of each bubble are calculated according to the segmentation result; the position and area information of the bubble are converted into a heat map to intuitively show the distribution of the bubble on the cover layer, and the heat map can use different colors or brightness to represent the density or size of the bubble; the real-time bubble rate is calculated according to the total area of the bubble and the total area of the cover layer, and the bubble rate is the percentage of the total area of the bubble to the total area of the cover layer; if the real-time bubble rate is greater than the preset bubble rate, the second preset pressing parameter needs to be adjusted in real time.

[0097] Specifically, the temperature in the second preset pressing parameter can be increased (for example, from 160°C to 180°C) to enhance the flowability of the glue layer; the pressure can also be increased (from 1.5 MPa to 2.5 MPa) to promote the uniform spreading of the glue layer; the pressing time can also be extended (from 30 seconds to 60 seconds) to ensure that the gas is fully discharged. It should be noted that the adjustment of the second preset pressing parameter can also be performed in the manner of step 103 described above.

[0098] In an optional embodiment of the present application, step 106 comprises:

[0099] The third preset pressing parameter is input into the laminator, so that the laminator performs pressing treatment on the cover film and the flexible circuit board substrate according to the third preset pressing parameter, and a target flexible circuit board is obtained.

[0100] Specifically, the third preset pressing parameter can be converted into a control parameter recognizable by the control system of the laminator, and then the control parameter is input into the laminator, so that the laminator performs pressing processing on the cover film and the flexible circuit board substrate according to the control parameter, and obtains the target flexible circuit board.

[0101] One specific embodiment of the flexible circuit board manufacturing method of the embodiment of the present application includes:

[0102] Step 111, obtaining a first preset pressing parameter;

[0103] The first preset pressing parameter includes a preset pressing temperature, a preset pressing pressure, a preset pressing time, and a preset glue layer thickness.

[0104] Step 112, bubble prediction;

[0105] Before manufacturing the flexible circuit board, the bubble prediction is performed according to the trained bubble prediction model, thereby saving the manufacturing cost.

[0106] Step 113, adjusting the first preset pressing parameter;

[0107] If the bubble area ratio prediction value in the bubble prediction result is greater than the preset bubble rate, it indicates that the flexible circuit board manufactured according to the first preset pressing parameter has more bubbles and cannot meet the expected demand, and the first preset pressing parameter needs to be adjusted to obtain a flexible circuit board with fewer bubbles.

[0108] Step 114, manufacturing a first flexible circuit board;

[0109] According to the adjusted preset pressing parameter and the laminator, the cover film is pressed on the flexible circuit board substrate to obtain the first flexible circuit board, and the number of the first flexible circuit board can be one or multiple. One sample can be manufactured for subsequent parameter adjustment, or a batch can be produced, and the bubble rate and product quality are reduced through real-time adjustment of the parameters subsequently.

[0110] Step 115, real-time adjustment of a second preset pressing parameter;

[0111] The real-time bubble rate of the first flexible circuit board is calculated according to the scanned image of the first flexible circuit board, and if it is greater than the preset bubble rate, the temperature, pressure, and time in the second preset pressing parameter need to be increased to reduce the bubble rate.

[0112] Step 116, manufacturing a target flexible circuit board.

[0113] The twice-adjusted pressing parameter is input into the laminator, so that the laminator performs pressing processing on the cover film and the flexible circuit board substrate according to the twice-adjusted pressing parameter, and obtains the target flexible circuit board.

[0114] The manufacturing method of the flexible circuit board of the embodiment of the present application reduces the bubble rate on the target flexible circuit board by adjusting the pressing parameters multiple times, greatly improves the quality and product qualification rate of the target flexible circuit board, and is beneficial to reducing the production cost and improving the use safety of the flexible circuit board.

[0115] As shown in Figure 2 The embodiment of the present application proposes a manufacturing device 200 of a flexible circuit board, which comprises:

[0116] The acquisition module 201 is configured to acquire a first preset pressing parameter.

[0117] The processing module 202 is configured to obtain a bubble prediction result according to the first preset pressing parameter and a trained bubble prediction model, adjust the first preset pressing parameter according to the bubble prediction result to obtain a second preset pressing parameter, obtain a first flexible circuit board according to the second preset pressing parameter, a cover film and a flexible circuit board substrate, wherein the cover film comprises a polyimide film and a glue layer, and the flexible circuit board substrate has a circuit etched thereon, adjust the second preset pressing parameter in real time according to the first flexible circuit board to obtain a third preset pressing parameter, and obtain a target flexible circuit board according to the third preset pressing parameter, the cover film and the flexible circuit board substrate.

[0118] Optionally, the first preset pressing parameter comprises:

[0119] a preset pressing temperature, a preset pressing pressure, a preset pressing time and a preset glue layer thickness.

[0120] Optionally, obtaining a bubble prediction result according to the first preset pressing parameter and a trained bubble prediction model comprises:

[0121] preprocessing the first preset pressing parameter to obtain preprocessing data;

[0122] performing format conversion on the preprocessing data to obtain a feature vector;

[0123] inputting the feature vector into the trained bubble prediction model to obtain the bubble prediction result.

[0124] Optionally, the training process of the bubble prediction model comprises:

[0125] acquiring historical data; the historical data comprises historical pressing parameters and corresponding bubble area proportions;

[0126] performing data processing on the historical data to obtain sample data;

[0127] dividing the sample data to obtain a training set and a test set;

[0128] training a preset network model according to the training set, to obtain a first network model;

[0129] testing the first network model according to the test set, to obtain a test result;

[0130] adjusting the first network model according to the test result and a preset evaluation index, to obtain a bubble prediction model.

[0131] Optionally, the first preset pressing parameter is adjusted according to the bubble prediction result, to obtain a second preset pressing parameter, including:

[0132] The first preset pressing parameter is adjusted according to the bubble prediction result and a preset bubble rate, to obtain a second preset pressing parameter.

[0133] Optionally, a first flexible circuit board is obtained according to the second preset pressing parameter, a cover film and a flexible circuit board substrate, including:

[0134] The second preset pressing parameter is input into a laminator, so that the laminator performs pressing processing on the cover film and the flexible circuit board substrate according to the second preset pressing parameter, to obtain the first flexible circuit board.

[0135] Optionally, the second preset pressing parameter is adjusted in real time according to the first flexible circuit board, to obtain a third preset pressing parameter, including:

[0136] Bubble data is obtained according to the first flexible circuit board; the bubble data includes bubble position data and bubble area data;

[0137] The second preset pressing parameter is adjusted in real time according to the bubble data and a preset bubble rate, to obtain a third preset pressing parameter.

[0138] The manufacturing device of the flexible circuit board in the embodiment of the application performs bubble prediction through the bubble prediction model and the obtained first preset pressing parameter, then adjusts the first preset pressing parameter according to the prediction result, and then manufactures the first flexible circuit board according to the second preset pressing parameter, the cover film and the flexible circuit board substrate, and finally adjusts the second preset pressing parameter in real time according to the bubble data on the first flexible circuit board to manufacture the target flexible circuit board, which is conducive to reducing the bubble rate on the flexible circuit board and improving the use safety of the flexible circuit board.

[0139] It should be noted that the device corresponds to the above method, and all implementation manners in the above method embodiments are applicable to the embodiments of the device and can achieve the same technical effects. The embodiments will not be described again.

[0140] The embodiment of the present application also provides a computing device, comprising a processor and a memory storing a computer program, when the computer program is run by the processor, the method in any of the above embodiments is executed. All implementation manners in the above method embodiments are applicable to the embodiment of the device, and the same technical effects can also be achieved. The embodiment will not be described here.

[0141] The embodiment of the present application also provides a computer readable storage medium, which stores instructions, when the instructions are run on a computer, the computer executes the method in any of the above embodiments. All implementation manners in the above method embodiments are applicable to the embodiment of the device, and the same technical effects can also be achieved. The embodiment will not be described here.

[0142] It should be noted that in the device and method of the present application, each component or step can be decomposed and / or recombined. These decompositions and / or recombination should be considered as equivalent solutions of the present application. Moreover, the steps of performing the above series of processes can be executed in time sequence according to the order of description, but it is not necessary to be executed in time sequence. Some steps can be executed in parallel, cross or independently of each other.

[0143] It should be noted that in the above embodiments, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitation, the element defined by the sentence "including a…" does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the method and device in the implementation of the above embodiments is not limited to the order of executing the functions shown or discussed, but can also include executing the functions in a substantially simultaneous manner or in reverse order according to the functions involved, for example, the described method can be executed in an order different from the described order, and various steps can also be added, omitted or combined. In addition, the features described with reference to some examples can be combined in other examples.

[0144] The above is the preferred embodiment of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should also be considered as the protection scope of the present application.

Claims

1. A method for manufacturing a flexible circuit board, characterized in that, include: Obtain the first preset pressing parameters; Based on the first preset pressing parameters and the trained bubble prediction model, the bubble prediction result is obtained; The first preset pressing parameters are adjusted based on the bubble prediction results to obtain the second preset pressing parameters; A first flexible circuit board is obtained based on the second preset pressing parameters, the cover film, and the flexible circuit board substrate; the cover film includes a polyimide film and an adhesive layer, and the flexible circuit board substrate has been etched with circuits. The second preset pressing parameters are adjusted in real time based on the first flexible circuit board to obtain the third preset pressing parameters; The target flexible circuit board is obtained based on the third preset pressing parameters, the cover film, and the flexible circuit board substrate; The bubble prediction result is obtained based on the first preset pressing parameters and the trained bubble prediction model, including: The first preset pressing parameters are verified to obtain the verified parameters; the target interaction parameters are generated based on the verified parameters; the preprocessed data includes the verified parameters and the target interaction parameters; the target interaction parameters include the first interaction parameter and the second interaction parameter, wherein the first interaction parameter is obtained by A1=T×P, where A1 is the first interaction parameter, T is the preset pressing temperature, and P is the preset pressing pressure; the second interaction parameter is obtained by A2=t / d, where A2 is the second interaction parameter, t is the preset pressing time, and d is the preset adhesive layer thickness; The preprocessed data is format-converted to obtain feature vectors; The bubble prediction model receives feature vectors as input. Each tree in the model independently predicts the input feature vector, obtaining a prediction result. The prediction results of all trees are aggregated to obtain the final predicted bubble area ratio, which is the output of the model. The bubble prediction result includes the predicted bubble area ratio, the importance of temperature, the importance of pressure, the importance of time, and the importance of the adhesive layer thickness. The importance of each feature is calculated as follows: ； ； ； ； ； Where D is the feature set contained in the node. It is information entropy, and c is the number of categories. It represents the probability that a feature belongs to the i-th category. It is the weighted information entropy after splitting. It is the number of features of the left child node after the split. It is the number of features of the right child node after the split. It is the information entropy of the left child node. It is the information entropy of the right child node. It is information gain. This represents the total information gain of the f-th feature, where F is the total number of features and N is the number of trees in the bubble prediction model. It is the importance of the f-th feature; The second preset pressing parameters are obtained by adjusting the first preset pressing parameters based on the bubble prediction results, including: The first preset pressing parameters are adjusted based on the bubble prediction results and the preset bubble rate to obtain the second preset pressing parameters. If the predicted bubble area ratio in the bubble prediction results is greater than the preset bubble rate, the adjustment order is determined according to the importance of each parameter in the bubble prediction results, and the first preset pressing parameters are adjusted according to this adjustment order. The method for adjusting the first preset pressing parameters is as follows: Obtain the preset bubble ratio and constraints; The adjustment objective function is determined based on the preset bubble rate. The adjustment objective function is to minimize the absolute value of the difference between the target bubble rate and the preset bubble rate. A set of compression parameters is randomly generated as the initial population; the objective function value is calculated for each individual, including the absolute value of the difference between the target bubble rate and the preset bubble rate; individuals with better fitness values ​​are selected for breeding; offspring individuals are generated by exchanging some genes of parent individuals; the genes of offspring individuals are randomly modified with a preset probability; the steps of calculating the objective function value for each individual and randomly modifying the genes of offspring individuals with a preset probability are repeated until the predetermined number of iterations or convergence conditions are reached. After each iteration, the Pareto front is extracted using non-dominated sorting based on the individual objective function values; Based on the preset bubble rate, one or more final schemes are selected from the Pareto front as the second preset pressing parameters.

2. The method for manufacturing a flexible circuit board according to claim 1, characterized in that, The first preset pressing parameters include: Preset pressing temperature, preset pressing pressure, preset pressing time, and preset adhesive layer thickness.

3. The method for manufacturing a flexible circuit board according to claim 1, characterized in that, The training process of the bubble prediction model includes: Acquire historical data; the historical data includes historical pressing parameters and the corresponding bubble area percentage; The historical data is processed to obtain sample data; The sample data is divided into a training set and a test set; The preset network model is trained based on the training set to obtain the first network model; The first network model is tested according to the test set to obtain test results; The first network model is adjusted based on the test results and preset evaluation indicators to obtain the bubble prediction model.

4. The method for manufacturing a flexible circuit board according to claim 1, characterized in that, Based on the second preset pressing parameters, the cover film, and the flexible circuit board substrate, a first flexible circuit board is obtained, comprising: The second preset pressing parameters are input into the laminator, so that the laminator presses the cover film and the flexible circuit board substrate according to the second preset pressing parameters to obtain the first flexible circuit board.

5. The method for manufacturing a flexible circuit board according to claim 1, characterized in that, The third preset pressing parameters are obtained by adjusting the second preset pressing parameters in real time based on the first flexible circuit board, including: Based on the first flexible circuit board, bubble data is obtained; the bubble data includes bubble position data and bubble area data; The second preset pressing parameter is adjusted in real time based on the bubble data and the preset bubble rate to obtain the third preset pressing parameter.

6. An apparatus for manufacturing flexible circuit boards, characterized in that, include: The acquisition module is used to acquire the first preset pressing parameters; The processing module is configured to: obtain a bubble prediction result based on the first preset pressing parameters and a trained bubble prediction model; adjust the first preset pressing parameters based on the bubble prediction result to obtain a second preset pressing parameter; obtain a first flexible circuit board based on the second preset pressing parameter, a cover film, and a flexible circuit board substrate; the cover film includes a polyimide film and an adhesive layer, and the flexible circuit board substrate has etched circuitry; adjust the second preset pressing parameters in real time based on the first flexible circuit board to obtain a third preset pressing parameter; and obtain a target flexible circuit board based on the third preset pressing parameter, the cover film, and the flexible circuit board substrate. The bubble prediction result is obtained based on the first preset pressing parameters and the trained bubble prediction model, including: The first preset pressing parameters are verified to obtain the verified parameters; the target interaction parameters are generated based on the verified parameters; the preprocessed data includes the verified parameters and the target interaction parameters; the target interaction parameters include the first interaction parameter and the second interaction parameter, wherein the first interaction parameter is obtained by A1=T×P, where A1 is the first interaction parameter, T is the preset pressing temperature, and P is the preset pressing pressure; the second interaction parameter is obtained by A2=t / d, where A2 is the second interaction parameter, t is the preset pressing time, and d is the preset adhesive layer thickness; The preprocessed data is format-converted to obtain feature vectors; The bubble prediction model receives feature vectors as input. Each tree in the model independently predicts the input feature vector, obtaining a prediction result. The prediction results of all trees are aggregated to obtain the final predicted bubble area ratio, which is the output of the model. The bubble prediction result includes the predicted bubble area ratio, the importance of temperature, the importance of pressure, the importance of time, and the importance of the adhesive layer thickness. The importance of each feature is calculated as follows: ； ； ； ； ； Where D is the feature set contained in the node. It is information entropy, and c is the number of categories. It represents the probability that a feature belongs to the i-th category. It is the weighted information entropy after splitting. It is the number of features of the left child node after the split. It is the number of features of the right child node after the split. It is the information entropy of the left child node. It is the information entropy of the right child node. It is information gain. This represents the total information gain of the f-th feature, where F is the total number of features and N is the number of trees in the bubble prediction model. It is the importance of the f-th feature; The second preset pressing parameters are obtained by adjusting the first preset pressing parameters based on the bubble prediction results, including: The first preset pressing parameters are adjusted based on the bubble prediction results and the preset bubble rate to obtain the second preset pressing parameters. If the predicted bubble area ratio in the bubble prediction results is greater than the preset bubble rate, the adjustment order is determined according to the importance of each parameter in the bubble prediction results, and the first preset pressing parameters are adjusted according to this adjustment order. The method for adjusting the first preset pressing parameters is as follows: Obtain the preset bubble ratio and constraints; The adjustment objective function is determined based on the preset bubble rate. The adjustment objective function is to minimize the absolute value of the difference between the target bubble rate and the preset bubble rate. A set of compression parameters is randomly generated as the initial population; the objective function value is calculated for each individual, including the absolute value of the difference between the target bubble rate and the preset bubble rate; individuals with better fitness values ​​are selected for breeding; offspring individuals are generated by exchanging some genes of parent individuals; the genes of offspring individuals are randomly modified with a preset probability; the steps of calculating the objective function value for each individual and randomly modifying the genes of offspring individuals with a preset probability are repeated until the predetermined number of iterations or convergence conditions are reached. After each iteration, the Pareto front is extracted using non-dominated sorting based on the individual objective function values; Based on the preset bubble rate, one or more final schemes are selected from the Pareto front as the second preset pressing parameters.

7. A computing device, characterized in that, include: A processor, a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The system stores instructions that, when executed on a computer, cause the computer to perform the method as described in any one of claims 1 to 5.

Images