Silver paste coating conductive uniformity test system and test method

Abstract

The invention discloses a silver paste coating conductive uniformity test system and a test method, particularly relates to the technical field of silver paste coating conductive test, and is used for solving the problem that the silver paste coating conductive uniformity test is inaccurate. Through a high-precision pressure sensor and a camera device, contact state data of a probe are monitored in real time at each measuring point, a contact state matrix is generated to support accurate correction of contact resistance, after collected data are subjected to standardization processing, a support vector regression model is used for capturing a nonlinear relation, and the accuracy of the contact resistance is improved. The method comprises the following steps: optimizing a compensation model through cross validation and parameter tuning, setting an optimal pressure range to stabilize a contact condition and smoothly process data, correcting time deviation by adopting a multi-channel data acquisition and distributed sampling synchronization algorithm, correcting an error in combination with a compensation coefficient, and finally calculating a conductivity uniformity index through a resistance data matrix. Heat map classification evaluation is generated, a low-degree uniform region is repaired, and the consistency and reliability of a conductive uniformity test are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of silver paste coating conductivity testing, and more specifically, to a silver paste coating conductivity uniformity testing system and testing method. Background Art

[0002] During the coating process, due to the influence of factors such as equipment, environment and material characteristics, the thickness and particle distribution of the silver paste coating may be uneven, resulting in deviations in conductivity in different areas. If the conductivity of the silver paste coating is uneven, it will directly affect the overall performance of the electronic device and may even cause problems such as unstable current distribution and abnormal local resistance in the later use of the product. Therefore, conductivity uniformity testing has become a key link in silver paste coating quality control.

[0003] Deficiencies of the existing technology: Due to the micro-roughness of the silver paste coating surface or poor probe contact, the contact resistance at different positions during the test will fluctuate significantly. This fluctuation not only affects the accuracy of the conductivity test in each area, but also easily leads to the accumulation of data deviations, reducing the accuracy of the overall uniformity assessment. The existing technology mostly uses a manual adjustment method with fixed contact pressure, which is difficult to adapt to the dynamically changing contact resistance requirements and lacks an effective real-time compensation mechanism. Especially in the multi-point test process, the unevenness of the contact resistance will be aggravated by factors such as ambient temperature and probe wear, making it difficult to achieve consistency in each measurement result, resulting in measurement errors that are difficult to control. Summary of the invention

[0004] In order to overcome the above defects of the prior art, the present invention provides a silver paste coating conductivity uniformity testing system and testing method to solve the problem of inaccurate silver paste coating conductivity uniformity testing in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for testing the conductivity uniformity of a silver paste coating comprises the following steps:

[0007] At each measuring point, a pressure sensor is used to monitor the contact pressure and area of ​​the probe, and the surface roughness characteristics are obtained through a camera device. The frequency domain information is analyzed using Fourier transform, and the contact resistance is corrected by combining the pressure and current voltage data and integrated into a contact state matrix.

[0008] The collected data is standardized, and nonlinear relationships are captured based on support vector regression. After cross-validation and parameter tuning, an optimized contact resistance compensation model is generated for real-time resistance data adjustment, improving conductivity consistency, and setting the optimal pressure range based on the compensation model;

[0009] Collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. Monitor the smoothed data and make adaptive adjustments to the contact resistance compensation model if the deviation exceeds the tolerance range.

[0010] Through multi-channel data acquisition and distributed sampling synchronization algorithm, measurement points are collected synchronously, time deviation is corrected, and error correction is performed in combination with compensation coefficients. The uniformity index is calculated based on the final resistance data matrix, a thermal map is generated, conductivity is evaluated by classification, and low uniformity areas are repaired.

[0011] In a preferred embodiment, a pressure sensor is used to monitor the contact pressure and area of ​​the probe at each measuring point, and the surface roughness characteristics are obtained by a camera device. The frequency domain information is analyzed by Fourier transform, and the contact resistance is corrected and integrated into a contact state matrix in combination with the pressure and current voltage data. The specific process is as follows:

[0012] At each measurement point, the pressure sensor collects the contact pressure data of the probe in real time to generate a pressure matrix;

[0013] The actual contact area at each measuring point is calculated based on the measured contact pressure matrix;

[0014] Use a micro camera device to perform microscopic scanning to generate high-resolution image data of the probe contact point, perform frequency domain analysis on the image through Fourier transform, extract surface roughness features, and generate a roughness matrix;

[0015] The contact resistance value is obtained by measuring the current and voltage, and the contact pressure, contact area and contact resistance are integrated to form a contact state matrix.

[0016] In a preferred embodiment, the collected data is standardized and nonlinear relationships are captured using support vector regression. The specific process is as follows:

[0017] The collected data are contact status data, which include contact pressure, contact area, surface roughness and preliminary contact resistance measurement values;

[0018] The contact status data were converted to a standard distribution with zero mean and unit variance through Z-score standardization;

[0019] After completing data normalization, multiple linear regression was used to preliminarily model the relationship between contact state data and contact resistance to determine the effects of contact pressure, contact area, and roughness on contact resistance.

[0020] Support vector regression is used to model the data. The data is mapped from the original space to a high-dimensional space through kernel function mapping, thereby determining the nonlinear relationship in the data.

[0021] In a preferred embodiment, after cross-validation and parameter tuning, an optimized contact resistance compensation model is generated for real-time resistance data adjustment to improve conductivity consistency, and an optimal pressure range is set according to the compensation model. The specific process is as follows:

[0022] After establishing the preliminary support vector regression model, the K-fold method was used for cross-validation, and the penalty coefficient and kernel function parameters of the support vector regression were tuned using the grid search method;

[0023] After obtaining the optimized support vector regression model, the corresponding contact resistance compensation coefficient is calculated based on the contact state data of each measurement point. The contact resistance compensation coefficient formula is defined as: in, is the contact resistance compensation value predicted by the model, R i，j is the measured value of initial contact resistance, K i，j The compensation coefficient for each measuring point characterizes the compensation relationship between contact state and resistance;

[0024] Integrate the contact resistance compensation coefficients of all measurement points to generate a compensation coefficient matrix, which is corrected using the measured values ​​in the conductivity uniformity test;

[0025] Determine the optimal contact pressure range for each measuring point based on the contact resistance compensation model;

[0026] Apply initial contact pressure through a piezoelectric regulator or a micro-electric device, and collect the contact resistance value for real-time feedback;

[0027] The feedback contact resistance value is used as the initial input of PID control;

[0028] The contact pressure of the probe is adjusted through PID, and the resistance feedback value is collected in real time. After each adjustment, the PID controller will recalculate the deviation of the contact pressure and adjust the pressure to the optimal pressure range until the target pressure is reached;

[0029] When the probe contact pressure is maintained within the optimal pressure range and the resistance is stable, the pressure adjustment is completed and the optimal measurement state is reached, and the final contact pressure and resistance values ​​are recorded as the acquisition data of the measurement point.

[0030] In a preferred embodiment, the contact pressure, area, roughness and resistance data are collected in real time, the resistance value is input into the compensation model, and the weighted moving average algorithm is used for smoothing, and the smoothed data is monitored. If the deviation exceeds the tolerance range, the contact resistance compensation model is adaptively adjusted. The specific process is as follows:

[0031] Collect contact pressure, contact area, roughness and contact resistance data in real time. After acquiring the real-time data, input it into the established contact resistance compensation model to calculate the real-time contact resistance compensation coefficient of each measuring point.

[0032] The contact resistance value measured in real time is corrected by the contact resistance compensation coefficient, and the corrected contact resistance value is smoothed by using the weighted moving average algorithm;

[0033] The tolerance range is set for the smoothed resistance data and the contact resistance compensation model is dynamically adjusted.

[0034] In a preferred embodiment, the measurement points are synchronously collected through multi-channel data acquisition and distributed sampling synchronization algorithm, the time deviation is corrected, and the error is corrected in combination with the compensation coefficient. The uniformity index is calculated based on the final resistance data matrix, a thermal map is generated, the conductivity is evaluated by classification, and the low uniformity area is repaired. The specific steps are as follows:

[0035] In the process of synchronous data collection, a distributed sampling synchronization algorithm is used to balance and correct the sampling deviation of each measurement point;

[0036] After the synchronous data is collected, the deviation between the resistance data of each measuring point is compared and the compensation coefficient matrix is ​​applied for correction;

[0037] After error compensation, the resistance difference between different measurement points is calculated to evaluate the uniformity of the data, and the measurement points are recalibrated using a consistency correction algorithm;

[0038] When the data of all measurement points have passed the consistency analysis and correction, the final corrected resistance data of each measurement point is recorded and a synchronously corrected resistance data matrix is ​​generated;

[0039] The resistance data matrix was analyzed, and the median, interquartile range and distribution density of the distribution characteristics were combined to generate the conductivity uniformity index;

[0040] Visualize the resistance data matrix to generate a resistance uniformity distribution heat map;

[0041] Based on the conductivity uniformity index and resistance distribution thermal map, a comprehensive judgment and repair is made on the conductivity uniformity of the silver paste coating.

[0042] A silver paste coating conductive uniformity testing system, used to implement the above-mentioned silver paste coating conductive uniformity testing method, comprising:

[0043] The measurement module is used to monitor the contact pressure and area of ​​the probe at each measurement point using a pressure sensor, obtain surface roughness characteristics through a camera, analyze frequency domain information using Fourier transform, and correct the contact resistance in combination with pressure and current-voltage data and integrate them into a contact state matrix;

[0044] The compensation determination module is used to standardize the collected data, capture nonlinear relationships based on support vector regression, generate an optimized contact resistance compensation model after cross-validation and parameter tuning, and is used for real-time resistance data adjustment to improve conductivity consistency and set the optimal pressure range based on the compensation model;

[0045] The model adjustment module is used to collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. It also monitors the smoothed data and performs adaptive adjustment of the contact resistance compensation model if the deviation exceeds the tolerance range.

[0046] The test and repair module is used to synchronize the measurement points through multi-channel data acquisition and distributed sampling synchronization algorithm, correct the time deviation, and perform error correction in combination with the compensation coefficient. It calculates the uniformity index based on the final resistance data matrix, generates a thermal map, classifies and evaluates the conductivity, and repairs low uniformity areas.

[0047] Technical effects and advantages of the present invention:

[0048] The present invention uses a high-precision pressure sensor and a camera device at each measuring point to monitor and collect data on the probe contact pressure, contact area and surface roughness in real time, and combines Fourier transform to analyze surface frequency domain information, generate a contact state matrix to support accurate calculation and correction of contact resistance, collect data and perform standardization, and use a support vector regression model to capture nonlinear relationships. An optimized compensation model is obtained through cross-validation and parameter tuning to achieve accurate adjustment of real-time resistance data, and set an optimal pressure range to stabilize contact conditions. At the same time, the real-time data is smoothed and deviations are monitored through a weighted moving average algorithm, and the contact resistance compensation model is adaptively adjusted to ensure data stability. The synchronous acquisition of measurement points is achieved according to multi-channel data acquisition and distributed sampling synchronization algorithms, the time deviation is corrected, and the error is further corrected in combination with the compensation coefficient. Finally, the conductivity uniformity index is calculated through the resistance data matrix and a heat map is generated, and classification evaluation and low-uniformity area repair are performed, which significantly improves the overall consistency and reliability of the silver paste coating conductivity uniformity test. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 The present invention is a flow chart of a method for testing the conductivity uniformity of a silver paste coating.

[0050] Figure 2The present invention is a schematic structural diagram of a silver paste coating conductivity uniformity testing system. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Example 1: Figure 1 As shown, a method for testing the conductivity uniformity of a silver paste coating comprises the following steps:

[0053] At each measuring point, a pressure sensor is used to monitor the contact pressure and area of ​​the probe, and the surface roughness characteristics are obtained through a camera device. The frequency domain information is analyzed using Fourier transform, and the contact resistance is corrected by combining the pressure and current voltage data and integrated into a contact state matrix.

[0054] The collected data is standardized, and nonlinear relationships are captured based on support vector regression. After cross-validation and parameter tuning, an optimized contact resistance compensation model is generated for real-time resistance data adjustment, improving conductivity consistency, and setting the optimal pressure range based on the compensation model;

[0055] Collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. Monitor the smoothed data and make adaptive adjustments to the contact resistance compensation model if the deviation exceeds the tolerance range.

[0056] Through multi-channel data acquisition and distributed sampling synchronization algorithm, measurement points are collected synchronously, time deviation is corrected, and error correction is performed in combination with compensation coefficients. The uniformity index is calculated based on the final resistance data matrix, a thermal map is generated, conductivity is evaluated by classification, and low uniformity areas are repaired.

[0057] Step 1: Real-time monitoring and data collection of the probe contact status. A high-precision pressure sensor is used to monitor the contact pressure and contact area between the probe and the silver paste coating surface in real time. Slight changes in the contact pressure and contact area will cause fluctuations in the contact resistance. Therefore, the probe contact status data is recorded at each measuring point. The contact status can provide a basic basis for the compensation of the contact resistance. At the same time, a micro-camera device is used to perform a microscopic scan of the probe contact area to identify the surface roughness and the probe contact condition.

[0058] At each measuring point, the pressure sensor collects the contact pressure data of the probe in real time and generates a pressure matrix P i，j, where i and j represent the horizontal and vertical coordinates of the measurement point, respectively. A small change in the probe contact pressure will significantly affect the contact resistance, so the pressure value at each point must be accurately collected. The pressure value is calculated using the following formula: Where α is the sensor sensitivity coefficient, which represents the pressure change under unit voltage output, and β is the offset correction factor, which is used to correct the initial pressure error caused by environmental conditions;

[0059] Calculate the actual contact area A at each measuring point based on the measured contact pressure matrix i，j , the contact area directly affects the resistance, so the contact area is corrected by the following formula: Among them, σ r represents the microscopic roughness of the silver paste coating, ∈ is a small correction term to avoid the denominator being zero, and γ is a nonlinear adjustment coefficient used to correct the nonlinear relationship of the contact area of ​​the rough surface;

[0060] Microscopic scanning and feature extraction of surface roughness are performed. In order to obtain the surface characteristics of the probe contact area, a micro camera device is used for microscopic scanning to generate high-resolution image data of the probe contact point. The image is analyzed in the frequency domain by Fourier transform to extract the surface roughness features and generate a roughness matrix: Σ i，j =F{I i，j}, where is the original image data at the measurement point (i, j), and F represents the Fourier transform operation, which is used to extract the frequency domain components of the image to calculate the local roughness characteristics;

[0061] Use a high-resolution camera to capture images of the silver paste coating surface, and then extract roughness information through image processing algorithms (such as edge detection and frequency domain analysis). Assuming that the height data is Z (x, y), its roughness frequency distribution is transformed by Fourier transform into And it is used to represent the frequency amplitude value after Fourier transformation, and the micro-roughness σ can be defined r for;

[0062] By measuring the current I i，j and voltage V i，j , get the contact resistance value R i，j , the contact resistance at each measurement point is calculated according to Ohm's law and is defined as follows: Among them, δ i，j It is the microscopic fluctuation factor of contact resistance, which indicates the resistance correction coefficient caused by tiny contact non-uniformity on the rough surface, and is used to correct the contact resistance deviation caused by surface roughness and pressure changes.

[0063] The contact pressure P i，j , contact area A i，j and contact resistance R i，jThe three are integrated to form the contact state matrix C i，j This matrix provides a full range of reference data for subsequent resistance compensation and data correction.

[0064] θ

[0065] The body expression is as follows: C i，j =(P i，j ·A i，j )·R i，j , where θ is the nonlinear combination factor during matrix integration, which is used to dynamically adjust the weight relationship between pressure, area and resistance; C i，j That is, the complete contact state matrix, which is used in the compensation model to correct the conductivity error caused by uneven probe contact in real time.

[0066] Step 2, establish and optimize the contact resistance compensation model, that is, standardize and preprocess the contact state data to ensure the consistency of the data dimension and provide a unified data scale for the subsequent model training. The contact state data includes contact pressure, contact area, surface roughness, and preliminary contact resistance measurement values. The variables vary greatly in magnitude. Directly inputting the model may cause uneven weights and cause the model to favor specific variables. Therefore, all variables are converted to a standard distribution with zero mean and unit variance through Z-score standardization: Among them, X is the original value of the variable, μ X and σ X Represent the mean and standard deviation of each variable respectively, X′ is the standardized data. After the unified scale, the data has a standard distribution with zero mean and unit variance;

[0067] The standardization process can avoid the influence of different variables due to dimensional differences, so that each variable is calculated with equal weight during model training. The standardized data is more conducive to the convergence and stability of subsequent machine learning algorithms (such as support vector regression SVR);

[0068] After completing the data standardization, multiple linear regression is used to preliminarily model the relationship between contact state data and contact resistance to determine the influence of variables such as contact pressure, contact area and roughness on contact resistance. The multiple regression model is as follows: R′ i，j =β 0 +β 1 P′ i,j + β 2A′ i，j +β 3 σ′ r +∈, where R′ i,j is the standardized contact resistance value, P′ i,j , A′ i，j ,σ′r are the standardized contact pressure, contact area and roughness, β 0 is the bias term of the model; β 1 , β 2 , β 3 is the regression coefficient to be determined, which represents the linear effect of each variable on the contact resistance; ∈ is the error term, which is used to represent the influence of random noise or unobserved factors;

[0069] Multiple linear regression can quickly build a preliminary relationship between contact state data and contact resistance, and determine the linear influence weight of each variable on contact resistance. The model can provide a basic prediction value, so that the subsequent support vector regression (SVR) model can perform more complex nonlinear fitting on this basis. The size of the regression coefficient can help determine the relative importance of the contact pressure, contact area and roughness on the contact resistance, which is helpful for further optimization and adjustment of the model structure.

[0070] A nonlinear support vector regression (SVR) model was constructed. After the preliminary multivariate regression model was established, in order to improve the nonlinear fitting ability of the model for contact resistance fluctuations, support vector regression (SVR) was used to further model the data. The data was mapped from the original space to a high-dimensional space through kernel function mapping, thereby capturing the nonlinear relationship in the data. The basic model form of SVR is as follows: Among them, K((P′ i，j , A′ i，j ,σ′ r ), X k ) represents the kernel function (RBF kernel function is selected here); α k is the weight coefficient of the support vector, which determines the importance of the support vector in nonlinear fitting; b is the bias term;

[0071] It should be noted that the RBF kernel function formula is selected based on its high efficiency in dealing with multivariate nonlinear relationships, and can capture the complex nonlinear relationship between contact resistance and contact state data;

[0072] After establishing the preliminary support vector regression (SVR) model, cross-validation is required to evaluate the generalization ability of the model and further optimize the model performance through hyperparameter tuning. The cross-validation adopts the K-fold method to divide the data set into K subsets, select one subset as the validation set in turn, and the remaining K-1 subsets as the training set. Repeat K times and calculate the average error of all rounds to obtain the stability and accuracy of the model. The score calculation formula of K-fold cross-validation is as follows: Among them, MSE k represents the mean square error of the samples in the k-fold validation set;

[0073] In the cross-validation process, in order to find the best hyperparameter combination that minimizes the validation error, the grid search method is used to tune the hyperparameters (penalty coefficient) and kernel function parameters of SVR. The grid search performs K-fold cross-validation on each combination by specifying a range of hyperparameter combinations to find the optimal hyperparameter combination that minimizes the validation error. S The smallest parameter combination of core;

[0074] You can also add L2 regularization terms during model training to suppress excessive regression coefficients in the model and prevent overfitting problems. L2 regularization constrains the size of model parameters by adding the sum of squares of regression coefficients to the loss function, thereby improving the stability of the model. The addition of L2 regularization terms prevents the model from tending to give too high weights to certain specific variables during training, especially the relative importance of variables such as contact pressure and contact area can be balanced.

[0075] The steps for generating the contact resistance compensation coefficient matrix are as follows: after obtaining the optimized support vector regression model, the corresponding contact resistance compensation coefficient is calculated based on the contact state data of each measurement point. The compensation coefficient is used to adjust the measured value of the contact resistance to more accurately reflect the actual conductive performance of the silver paste coating under different contact states. The specific calculation steps are as follows:

[0076] Input the contact state data, and input the standardized contact pressure, contact area, and roughness of each measurement point into the optimized SVR model;

[0077] Calculate the compensation coefficient and use the trained SVR model to predict the input data of each measurement point to obtain the predicted value of contact resistance after compensation, the contact resistance compensation coefficient K i，j Defined by the following formula: in, is the contact resistance compensation value predicted by the model, R i，j is the measured value of initial contact resistance, K i，j The compensation coefficient for each measurement point characterizes the compensation relationship between contact state and resistance;

[0078] K of all measurement points i，j The values ​​are integrated to generate the compensation coefficient matrix K, which is used to correct the measured values ​​in the subsequent conductivity uniformity test.

[0079] Step 3: Adaptive adjustment of the probe contact pressure. In order to ensure the stability of the measurement data, the contact pressure needs to be adaptively adjusted. According to the contact resistance compensation model established in step 2, the optimal pressure range of each measurement point is determined. The selection of the optimal pressure range is based on the model's minimization of the resistance fluctuation under different pressure values. That is, through the relationship diagram between contact pressure and contact resistance, find the pressure range that can effectively reduce the contact resistance fluctuation. The optimal pressure range can be expressed as: Popt =[P min , P max ], where P min and P max The minimum and maximum pressure values ​​predicted by the model ensure that the resistance fluctuation is minimal within this range;

[0080] The optimal pressure range ensures that the probe pressure is sufficiently stable, reduces the error caused by uneven contact, and provides a reference standard for adaptive adjustment, so that the probe can be kept in the appropriate pressure range in subsequent adjustments;

[0081] After the probe is moved to the measuring position, the initial contact pressure P is applied by a piezoelectric regulator or a micro-electric device. 0 , and immediately collect the real-time feedback contact resistance value P init , the initial contact pressure is based on the optimal pressure range P opt A median value is selected in the range to ensure that the initial pressure is close to the optimal range. Real-time collection of the initial contact resistance value is convenient for detecting whether the current contact state is stable, providing initial data for feedback control. By applying the initial pressure, it can quickly enter the state close to the target pressure range. The real-time feedback resistance value is used as the initial input for subsequent PID control, which helps to determine whether further adjustment is needed to achieve the ideal pressure.

[0082] Set the ratio of the PID controller (K p ), integral(K i ) and differential (K d ) parameters so that the controller can achieve fast response and stability in pressure regulation. The goal of the PID controller is to make the measured pressure P t Gradually approaching the set optimal range P opt , the formula is as follows: Among them, P t+1 is the pressure value at the next moment t+1, e(t)=P opt -P t is the deviation between the current pressure and the optimal pressure range; K p , K i , K d They are the proportional, integral and differential gain coefficients of the PID controller, which affect the response speed and stability of the system;

[0083] Under the action of PID control, the system continuously adjusts the contact pressure of the probe and collects the resistance feedback value in real time. After each adjustment, the PID controller recalculates the deviation of the contact pressure and adjusts the pressure to P. opt In the range, until the target pressure is reached and the contact resistance reaches a stable value R stable,In this process, the resistance value is used as a reference indicator of the pressure deviation. If the resistance change is no longer significant, the contact is considered to be stable;

[0084] When the probe contact pressure is maintained at P opt If the contact pressure and resistance are within the specified range and stable, the pressure adjustment is considered to be completed and the optimal measurement state is reached. The final contact pressure and resistance values ​​are recorded as the acquisition data of the measurement point for further use in subsequent uniformity analysis and compensation models.

[0085] Step 4: Real-time contact resistance compensation and data correction. While measuring, the pressure sensor, contact area and other data are input into the compensation model in real time, and the conductivity data is corrected instantly according to the contact resistance compensation coefficient. The error caused by poor contact is automatically corrected by jointly calculating the real-time resistance data with the compensation parameters in the model. The weighted moving average (WMA) algorithm is used to smooth the real-time resistance data to eliminate sudden noise interference, ensuring that the conductivity data is more stable and accurate, close to the actual conductivity performance. The specific steps are as follows:

[0086] First, data such as contact pressure, contact area, roughness and contact resistance are collected in real time and passed to the compensation model as input. The data collected in real time constitutes the contact state parameter set of the measurement point and provides basic data for contact resistance compensation. Real-time data collection and input ensure the timeliness of the measurement data, so that the compensation model can immediately respond to changes in the contact state.

[0087] After acquiring the real-time data, it is input into the previously established contact resistance compensation model to calculate the real-time contact resistance compensation coefficient K of each measuring point. i，j , the compensation coefficient is used to correct the currently measured resistance value: Among them, R i，j is the measured original contact resistance, is the corrected resistance value after applying the compensation coefficient;

[0088] The real-time resistance data after compensation may still be affected by instantaneous fluctuations or external noise. Therefore, the weighted moving average (WMA) algorithm is used to smooth the corrected resistance value to eliminate sudden noise interference. The formula of the weighted moving average algorithm is as follows: in, is the resistance value after smoothing; w t is the weighting coefficient, which makes the weight of recent data larger, w t The sum of is 1;

[0089] Real-time monitoring of smoothed resistance data The stability and consistency of the test data are tested. When the data deviation is found to exceed the set tolerance range, the error information is fed back to the compensation model so that the model can make adaptive adjustments. The specific tolerance range is set as follows; Among them, ∈ is the allowable error threshold, which is used to determine whether the data deviates from the normal range;

[0090] The error feedback mechanism ensures that the compensation model can be adjusted dynamically to prevent the contact resistance from being inaccurate due to environmental fluctuations or changes in contact status. By continuously monitoring and feeding back errors, the compensation model can be continuously optimized to make the measured values ​​more stable and reliable.

[0091] When the contact resistance data After the stability and consistency requirements are met, it is recorded as the final resistance value of the measurement point. This data is recorded in the uniformity analysis data set as the basis for subsequent conductivity uniformity evaluation. The final data format recorded is as follows: Among them, D i，j Represents the final data set for the measurement point (i, j), including contact pressure, contact area, roughness, and final contact resistance after smoothing.

[0092] Step 5, perform multi-point synchronous calibration and data consistency assessment, use a multi-channel data acquisition device, arrange measurement channels at each measurement point at the same time, ensure that all measurement points can synchronously collect contact resistance data, set a unique timestamp for each channel, and synchronize all channels in the acquisition device to ensure that data collection is performed at the same time node, and synchronize the recording of each measurement point so that the data at different measurement points have time consistency;

[0093] In the process of synchronous data collection, a distributed sampling synchronization algorithm is used to balance and correct the sampling deviation of each measurement point. The distributed sampling synchronization algorithm can perform time calibration on the collected raw resistance data to ensure that the data of all measurement points are on the same time scale. The calculation process of the algorithm is as follows: T adjusted，i，j =T i，j +ΔT sync , where T i，j is the original timestamp of the measurement point (i, j), ΔT sync The time correction value calculated by the synchronization algorithm ensures that the data of all measurement points are consistent at all times. adjusted，i，j is the timestamp after calibration;

[0094] The distributed sampling synchronization algorithm makes the data of each measurement point strictly aligned in sampling time. By correcting the time deviation, the error introduced by time asynchrony can be eliminated during data processing.

[0095] After the synchronous data is collected, the measurement error of each measurement point is calculated to evaluate the accuracy of the contact resistance data of each point. By comparing the deviations between the resistance data of each measurement point, the points with large errors are identified and the compensation coefficient matrix K is applied. i，j To perform error correction: in, is the original measured resistance value of the measuring point (i, j), is the resistance value after correction;

[0096] The error calculation and compensation steps ensure that the resistance data of each measurement point is consistent within the error range. Through the effect of the compensation coefficient matrix, the uncertainty in the contact resistance measurement is eliminated, so that the resistance data at different points can more accurately reflect the actual conductivity distribution.

[0097] After error compensation, the data consistency of each measurement point is further analyzed. The resistance difference between different measurement points is calculated to evaluate the uniformity of the data. The consistency correction algorithm is used to perform secondary correction on the measurement points with large deviations to ensure that the resistance values ​​of all measurement points meet the consistency requirements. The consistency calculation formula for resistance difference is as follows: Among them, D i，j is the average resistance of the measuring point (i, j) and all measuring points If D i，j If the allowed deviation threshold is exceeded, Perform secondary calibration to ensure that the resistance data of all measurement points have high consistency;

[0098] After the data of all measurement points have passed the consistency analysis and correction, the final corrected resistance data of each measurement point is recorded to generate a synchronously corrected resistance data matrix: This matrix serves as the data basis for uniformity analysis and is used for the overall assessment of the conductivity uniformity of the silver paste coating.

[0099] The final resistance data matrix R final Analyze and calculate its distribution characteristics, mainly including median, interquartile range and distribution density, to avoid the possible misleading of simple characteristics of mean and variance. The median represents the central value of conductivity distribution, the interquartile range is used to describe the degree of dispersion of resistance distribution, and the distribution density estimates the resistance value to observe the distribution of resistance values ​​in different intervals.

[0100] Based on the resistance distribution characteristics, a conductivity uniformity index is defined to quantify the conductivity consistency of the silver paste coating. This index combines the discreteness of the resistance distribution with the degree of deviation. The specific definition is as follows: Among them, M is the median resistance, IQR is the interquartile range, which indicates the discreteness of resistance; D is the variance of resistance density estimation, which indicates the discreteness of resistance distribution density; is the adjustment coefficient, which is used to control the proportion of IQR and D in CUI;

[0101] Visualize the resistance data matrix to generate a resistance uniformity distribution heat map, map the resistance values ​​to different colors, and the color depth indicates the resistance value. Through the color gradient, the conductivity distribution can be intuitively observed.

[0102] Perform heat map color mapping, that is, by using a continuous gradient color spectrum, lower resistance areas are represented by cold colors, higher resistance areas are represented by warm colors, and abnormal resistance measurement points are marked, especially those that deviate greatly from the median;

[0103] Based on the conductivity uniformity index (CUI) and resistance distribution heat map, the conductivity uniformity of the silver paste coating is comprehensively judged and divided into the following levels;

[0104] Setting a first threshold value and a second threshold value for uniform conductivity, a first threshold value and a second threshold value for heat map color, and the first threshold value is greater than the second threshold value;

[0105] Highly uniform (high CUI, uniform color of thermal map): indicates that the conductivity of the silver paste coating is consistent and the overall conductivity is stable, that is, the conductivity uniformity of the silver paste coating that exceeds the first threshold of conductivity uniformity and the first threshold of thermal map color is considered highly uniform;

[0106] Moderate uniformity (medium CUI, slight deviation in thermal map color): The conductivity is basically consistent, but there is a small range of resistance deviation in some areas, that is, the conductivity uniformity of the silver paste coating that exceeds the second threshold of conductivity uniformity and the second threshold of thermal map color is considered to be moderately uniform;

[0107] Low uniformity (low CUI, uneven color distribution of thermal map): The conductivity of the silver paste coating varies greatly in different areas, and there may be coating defects or discontinuity problems. That is, the conductivity uniformity of the silver paste coating that does not exceed the second threshold of conductivity uniformity and the second threshold of thermal map color is considered to be low uniformity;

[0108] The color distribution in the uniformity distribution heat map can be used to identify areas with high or low resistance, locate points with large conductivity differences, and take local repair measures for these areas, including increasing the thickness of the silver paste, improving the coating process, or reducing the physical damage of the coating. Precision spraying equipment can be used to evenly replenish the silver paste in areas with low conductivity to ensure that the thickness reaches the preset standard. Areas with higher conductivity are coated with polished or slightly trimmed.

[0109] It should be noted that the thresholds involved in the embodiments can be determined according to specific scenarios and requirements.

[0110] The present invention uses a high-precision pressure sensor and a camera device at each measuring point to monitor and collect data on the probe contact pressure, contact area and surface roughness in real time, and combines Fourier transform to analyze surface frequency domain information, generate a contact state matrix to support accurate calculation and correction of contact resistance, collect data and perform standardization, and use a support vector regression model to capture nonlinear relationships. An optimized compensation model is obtained through cross-validation and parameter tuning to achieve accurate adjustment of real-time resistance data, and set an optimal pressure range to stabilize contact conditions. At the same time, the real-time data is smoothed and deviations are monitored through a weighted moving average algorithm, and the contact resistance compensation model is adaptively adjusted to ensure data stability. The synchronous acquisition of measurement points is achieved according to multi-channel data acquisition and distributed sampling synchronization algorithms, the time deviation is corrected, and the error is further corrected in combination with the compensation coefficient. Finally, the conductivity uniformity index is calculated through the resistance data matrix and a heat map is generated, and classification evaluation and low-uniformity area repair are performed, which significantly improves the overall consistency and reliability of the silver paste coating conductivity uniformity test.

[0111] Embodiment 2: A silver paste coating conductive uniformity test system, such as Figure 2 As shown, specifically including:

[0112] The measurement module is used to monitor the contact pressure and area of ​​the probe at each measurement point using a pressure sensor, obtain surface roughness characteristics through a camera, analyze frequency domain information using Fourier transform, and correct the contact resistance in combination with pressure and current-voltage data and integrate them into a contact state matrix;

[0113] The compensation determination module is used to standardize the collected data, capture nonlinear relationships based on support vector regression, generate an optimized contact resistance compensation model after cross-validation and parameter tuning, and is used for real-time resistance data adjustment to improve conductivity consistency and set the optimal pressure range based on the compensation model;

[0114] The model adjustment module is used to collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. It also monitors the smoothed data and performs adaptive adjustment of the contact resistance compensation model if the deviation exceeds the tolerance range.

[0115] The test and repair module is used to synchronize the measurement points through multi-channel data acquisition and distributed sampling synchronization algorithm, correct the time deviation, and perform error correction in combination with the compensation coefficient. It calculates the uniformity index based on the final resistance data matrix, generates a thermal map, classifies and evaluates the conductivity, and repairs low uniformity areas.

[0116] The above formulas are all dimensionless and calculated numerically. Specific dimension removal can be achieved by various means such as standardization, which will not be elaborated here. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0117] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, an ATA hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0118] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0119] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0120] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0121] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0123] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for testing the conductivity uniformity of a silver paste coating, characterized in that: The steps include: At each measuring point, a pressure sensor is used to monitor the contact pressure and area of ​​the probe, and the surface roughness characteristics are obtained through a camera device. The frequency domain information is analyzed using Fourier transform, and the contact resistance is corrected by combining the pressure and current voltage data and integrated into a contact state matrix. The collected data is standardized, and nonlinear relationships are captured based on support vector regression. After cross-validation and parameter tuning, an optimized contact resistance compensation model is generated for real-time resistance data adjustment, improving conductivity consistency, and setting the optimal pressure range based on the compensation model; Collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. Monitor the smoothed data and make adaptive adjustments to the contact resistance compensation model if the deviation exceeds the tolerance range. Through multi-channel data acquisition and distributed sampling synchronization algorithm, measurement points are collected synchronously, time deviation is corrected, and error correction is performed in combination with compensation coefficients. The uniformity index is calculated based on the final resistance data matrix, a thermal map is generated, conductivity is evaluated by classification, and low uniformity areas are repaired.

2. A method for testing the conductivity uniformity of a silver paste coating according to claim 1, characterized in that: At each measuring point, a pressure sensor is used to monitor the contact pressure and area of ​​the probe, and the surface roughness characteristics are obtained through a camera device. The frequency domain information is analyzed using Fourier transform, and the contact resistance is corrected and integrated into a contact state matrix in combination with the pressure and current voltage data. The specific process is as follows: At each measurement point, the pressure sensor collects the contact pressure data of the probe in real time to generate a pressure matrix; The actual contact area at each measuring point is calculated based on the measured contact pressure matrix; Use a micro camera device to perform microscopic scanning to generate high-resolution image data of the probe contact point, perform frequency domain analysis on the image through Fourier transform, extract surface roughness features, and generate a roughness matrix; The contact resistance value is obtained by measuring the current and voltage, and the contact pressure, contact area and contact resistance are integrated to form a contact state matrix.

3. A method for testing the conductivity uniformity of a silver paste coating according to claim 2, characterized in that: The collected data is standardized and nonlinear relationships are captured based on support vector regression. The specific process is as follows: The collected data are contact status data, which include contact pressure, contact area, surface roughness and preliminary contact resistance measurement values; The contact status data were converted to a standard distribution with zero mean and unit variance through Z-score standardization; After completing data normalization, multiple linear regression was used to preliminarily model the relationship between contact state data and contact resistance to determine the effects of contact pressure, contact area, and roughness on contact resistance. Support vector regression is used to model the data. The data is mapped from the original space to a high-dimensional space through kernel function mapping, thereby determining the nonlinear relationship in the data.

4. A method for testing the conductivity uniformity of a silver paste coating according to claim 3, characterized in that: After cross-validation and parameter tuning, an optimized contact resistance compensation model is generated for real-time resistance data adjustment to improve conductivity consistency, and the optimal pressure range is set according to the compensation model. The specific process is as follows: After establishing the preliminary support vector regression model, the K-fold method was used for cross-validation, and the penalty coefficient and kernel function parameters of the support vector regression were tuned using the grid search method; After obtaining the optimized support vector regression model, the corresponding contact resistance compensation coefficient is calculated based on the contact state data of each measurement point. The contact resistance compensation coefficient formula is defined as: in, is the contact resistance compensation value predicted by the model, R i，j is the measured value of initial contact resistance, K i，j The compensation coefficient for each measuring point characterizes the compensation relationship between contact state and resistance; Integrate the contact resistance compensation coefficients of all measurement points to generate a compensation coefficient matrix, which is corrected using the measured values ​​in the conductivity uniformity test; Determine the optimal contact pressure range for each measuring point based on the contact resistance compensation model; Apply initial contact pressure through a piezoelectric regulator or a micro-electric device, and collect the contact resistance value for real-time feedback; The feedback contact resistance value is used as the initial input of PID control; The contact pressure of the probe is adjusted through PID, and the resistance feedback value is collected in real time. After each adjustment, the PID controller will recalculate the deviation of the contact pressure and adjust the pressure to the optimal pressure range until the target pressure is reached; When the probe contact pressure is maintained within the optimal pressure range and the resistance is stable, the pressure adjustment is completed and the optimal measurement state is reached, and the final contact pressure and resistance values ​​are recorded as the acquisition data of the measurement point.

5. A method for testing the conductivity uniformity of a silver paste coating according to claim 4, characterized in that: The contact pressure, area, roughness and resistance data are collected in real time, the compensation model is input to correct the resistance value, and the weighted moving average algorithm is used for smoothing. The smoothed data is monitored. If the deviation exceeds the tolerance range, the contact resistance compensation model is adaptively adjusted. The specific process is as follows: Collect contact pressure, contact area, roughness and contact resistance data in real time. After acquiring the real-time data, input it into the established contact resistance compensation model to calculate the real-time contact resistance compensation coefficient of each measuring point. The contact resistance value measured in real time is corrected by the contact resistance compensation coefficient, and the corrected contact resistance value is smoothed by using the weighted moving average algorithm; The tolerance range is set for the smoothed resistance data and the contact resistance compensation model is dynamically adjusted.

6. A method for testing the conductivity uniformity of a silver paste coating according to claim 5, characterized in that: Through multi-channel data acquisition and distributed sampling synchronization algorithm, the measurement points are collected synchronously, the time deviation is corrected, and the error is corrected in combination with the compensation coefficient. The uniformity index is calculated based on the final resistance data matrix, a thermal map is generated, the conductivity is evaluated by classification, and the low uniformity area is repaired. The specific steps are as follows: In the process of synchronous data collection, a distributed sampling synchronization algorithm is used to balance and correct the sampling deviation of each measurement point; After the synchronous data is collected, the deviation between the resistance data of each measuring point is compared and the compensation coefficient matrix is ​​applied for correction; After error compensation, the resistance difference between different measurement points is calculated to evaluate the uniformity of the data, and the measurement points are corrected again using the consistency correction algorithm; When the data of all measurement points have passed the consistency analysis and correction, the final corrected resistance data of each measurement point is recorded and a synchronously corrected resistance data matrix is ​​generated; The resistance data matrix was analyzed, and the median, interquartile range and distribution density of the distribution characteristics were combined to generate the conductivity uniformity index; Visualize the resistance data matrix to generate a resistance uniformity distribution heat map; Based on the conductivity uniformity index and resistance distribution thermal map, a comprehensive judgment and repair is made on the conductivity uniformity of the silver paste coating.

7. A silver paste coating conductive uniformity testing system, used to implement a silver paste coating conductive uniformity testing method according to any one of claims 1 to 6, characterized in that: include: The measurement module is used to monitor the contact pressure and area of ​​the probe at each measurement point using a pressure sensor, obtain surface roughness characteristics through a camera, analyze frequency domain information using Fourier transform, and correct the contact resistance in combination with pressure and current-voltage data and integrate them into a contact state matrix; The compensation determination module is used to standardize the collected data, capture nonlinear relationships based on support vector regression, generate an optimized contact resistance compensation model after cross-validation and parameter tuning, and is used for real-time resistance data adjustment to improve conductivity consistency and set the optimal pressure range based on the compensation model; The model adjustment module is used to collect contact pressure, area, roughness and resistance data in real time, input the compensation model to correct the resistance value, and smooth it through the weighted moving average algorithm. It also monitors the smoothed data and performs adaptive adjustment of the contact resistance compensation model if the deviation exceeds the tolerance range. The test and repair module is used to synchronize the measurement points through multi-channel data acquisition and distributed sampling synchronization algorithm, correct the time deviation, and perform error correction in combination with the compensation coefficient. It calculates the uniformity index based on the final resistance data matrix, generates a thermal map, classifies and evaluates the conductivity, and repairs low uniformity areas.

Images