Circuit board detection method and system, electronic equipment and storage medium

Abstract

The invention relates to the technical field of circuit board detection, and discloses a circuit board detection method and system, electronic equipment and a storage medium, and the method comprises the steps: collecting historical design parameters and simulation results of a circuit board, processing the design parameters and the simulation results, and forming a data set # imgabs0 #; executing a prediction model training strategy according to the data set, and generating an output prediction model of the circuit board; obtaining current design parameters of the circuit board, and applying the design parameters to the output prediction model to obtain a model result; executing the simulation dynamic adjustment strategy, and dynamically adjusting the simulation process of the simulation tool; historical fault design parameters of the circuit board are collected, a fault model training strategy is executed, and a fault prediction model of the circuit board is generated; applying the current design parameters of the circuit board to the fault prediction model to obtain a fault probability; setting a fault probability threshold; if the fault probability is not less than the fault probability threshold. And performance prediction, simulation optimization and fault early warning and correction in the circuit board design and manufacturing process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of circuit board detection, and specifically to a circuit board detection method, system, electronic device and storage medium. Background Art

[0002] With the increasing complexity of circuit board design, traditional design and verification methods can no longer meet the requirements of high efficiency and accuracy. Therefore, combining intelligent analysis with adaptive simulation has become the key to improving the performance of circuit boards. First, circuit board design and simulation technology provides a basic framework for circuit design. By using simulation tools to simulate circuit behavior and predict its performance, it helps designers discover problems in the design stage. Second, machine learning technology, especially regression analysis, has been widely used in circuit board prediction. It can model based on the relationship between design parameters and simulation results, thus accurately predicting the output behavior of circuit boards. At the same time, data collection and preprocessing extract key features and perform normalization processing, eliminating the dimensional differences between data and improving the efficiency and accuracy of model training. Adaptive simulation technology dynamically adjusts the simulation process based on the prediction model, optimizes the simulation step size and accuracy, and improves the simulation efficiency and accuracy. Finally, fault prediction technology uses historical fault data and machine learning models to identify potential defects in circuit design in advance, avoiding failures in actual applications.

[0003] The present invention proposes a circuit board detection method, system, electronic device and storage medium, which combines modern intelligent technology with traditional circuit design methods, promoting the intelligent process of circuit board design and optimization. Summary of the Invention

[0004] The present invention provides a circuit board detection method, system, electronic device and storage medium, which helps to solve the problems mentioned in the above background art.

[0005] In a first aspect, the present application provides a circuit board detection method, adopting the following technical solution: A circuit board detection method includes: S1. Data collection and preprocessing; Collect the historical design parameters and simulation results of the circuit board, process the design parameters and simulation results to form a data set ; S2. Model training and verification; According to the data set, execute a prediction model training strategy to generate an output prediction model of the circuit board; S3. Adaptive simulation; Obtain the current design parameters of the circuit board, apply them to the output prediction model to obtain a model result; According to the model result, execute a simulation dynamic adjustment strategy to dynamically adjust the simulation process of the simulation tool; S4. Fault prediction optimization; Collect the failed design parameters of the circuit board history, execute the failure model training strategy, and generate a failure prediction model for the circuit board; Apply the current design parameters of the circuit board to the failure prediction model to obtain the failure probability; Set the failure probability threshold; If the failure probability ≥ the failure probability threshold, it is prompted that the design parameters of the circuit board are unqualified.

[0006] Preferably, collecting the historical design parameters and simulation results of the circuit board and processing the design parameters and simulation results includes: For any pair of design parameters and simulation results ; Record the collected historical design parameters as , where a is the number of elements in the design parameters; Record the collected historical simulation results as , where b is the number of elements in the simulation results; Perform normalization processing on each historical design parameter: Calculate the norm of the design parameter: ; Calculate the normalized design parameter ; Calculate the norm of the simulation result: ; Calculate the normalized simulation result .

[0007] By collecting historical design parameters and simulation results, the comprehensiveness and accuracy of the data source can be ensured. Extracting features and normalizing the data can eliminate the differences between different dimensions, making each feature equally important during the training process, thereby improving the accuracy of the model. After normalization, the range of the data is restricted to a certain interval, which helps to accelerate model training and optimization, improve computational efficiency, and at the same time avoid certain features having too much impact on model training. Especially when the parameters and simulation results of circuit design involve multi-dimensional data, normalization enables these data to be processed on the same scale, which is beneficial to model training. In addition, feature extraction can extract key information from a large amount of raw data, provide effective input for the subsequent prediction model, and help reduce waste of computing resources. Finally, through these steps, a training set with high quality and suitable for training and testing can be constructed, laying a foundation for the accuracy and generalization ability of the subsequent machine learning model.

[0008] Preferably, according to the data set, executing the prediction model training strategy to generate an output prediction model for the circuit board includes: Use a linear regression model to train the output prediction model: Set the predicted output result ; Wherein, is the regression coefficient of the linear regression model, c is the bias term of the linear regression model, 1 ≦ i ≦ b; Calculate the weighted loss function: , wherein, is a constant, is the weight of the i-th element of the simulation result, is the i-th element in the simulation result, is the value of the i-th element of the predicted simulation result; Adjust the regression coefficient and the bias term to reduce the weighted loss function: For any regression coefficient , wherein, is the j-th element of the design parameter; For the bias term c;

[0009] Set the deviation threshold ; When , then stop adjusting the regression coefficient and the bias term.

[0010] By training an appropriate prediction model, the relationship between the input design parameters and the circuit board output can be learned from historical data, and then the prediction of unknown data can be realized. By validating the model through methods such as cross-validation, the overfitting problem can be effectively avoided, ensuring that the model not only performs well on the training set but also has good generalization ability on new data. In addition, during the training process, by adjusting the regression coefficient and the bias term, the prediction ability of the model can be finely optimized to make it more in line with the actual circuit board behavior. Calculating the weighted loss function and optimizing it can further improve the accuracy of the model under specific conditions, especially when some features in the simulation result are more important. Through the weighted loss function, the model can pay more attention to the prediction results of these important features. By adjusting the regression coefficient and the bias term, not only can the fitting effect of the model be improved, but also the prediction accuracy of the model can be increased. Finally, after model training and validation, an accurate circuit board output prediction model can be generated, which can provide effective support for subsequent simulation and fault prediction.

[0011] Preferably, according to the model result, execute a simulation dynamic adjustment strategy to dynamically adjust the simulation process of the simulation tool, including: Dynamically adjust the step size: Obtain the current simulation step size ; Calculate the simulation step size for the next simulation ; , where is the stability threshold set for the i-th element of the simulation result, is the coefficient for step size adjustment; Dynamically adjust the simulation accuracy: Obtain the current simulation accuracy ; Calculate the simulation accuracy for the next simulation ; , where is the coefficient for accuracy adjustment, is the target value of the i-th element of the simulation result; Dynamically adjust the design parameters: For any one in the simulation result ; Obtain the influence coefficient of each element pair in the predicted design parameters on ; ; Obtain the maximum change range of each element in the design parameters ; Dynamically adjust the change range of each element in the design parameters: , 1 ≤ i ≤ a.

[0012] Adaptive simulation is an important part of the circuit board detection method. Its purpose is to dynamically adjust the simulation process according to the design parameters and prediction model of the circuit board to ensure the accuracy and efficiency of the simulation results. By applying the output prediction model, the performance of the circuit board can be predicted based on the current design parameters, and the simulation process can be adjusted on this basis. For example, by adjusting the simulation step size, the calculation accuracy and efficiency in the simulation process can be optimized. If the error in the simulation process is large, the step size can be appropriately reduced to ensure higher accuracy of the simulation results; on the contrary, if the accuracy requirement is not high, the step size can be increased to accelerate the simulation process. Dynamically adjusting the simulation accuracy can flexibly optimize the simulation accuracy according to the results and accuracy requirements of the model, thereby improving the calculation efficiency of the simulation process. By dynamically adjusting the change range of the design parameters, the design of the circuit board can be gradually optimized during the simulation process to ensure that the circuit board can work properly under various conditions. Adaptive simulation can not only improve the flexibility of the simulation process, but also greatly improve the simulation efficiency and accuracy, thereby effectively saving computing resources and shortening the product development cycle.

[0013] Preferably, collecting the historical fault design parameters of the circuit board, implementing the fault model training strategy, and generating the fault prediction model of the circuit board includes: Record the historical fault design parameters as ; Calculate similarity , where is a positive number used to adjust the sensitivity of the distance; is the current input value of the i-th element in the design parameters.

[0014] By generating a fault prediction model, potential problems in circuit board design can be identified in advance, avoiding faults during actual production and use. By collecting historical fault design parameters and training, an accurate fault prediction model can be generated, which can predict the fault probability of a circuit board based on its design parameters. If the fault probability of the current design parameters exceeds the set threshold, a warning can be issued in a timely manner, indicating that the circuit board design is unqualified, thus avoiding the occurrence of faults. This process can effectively improve the reliability and quality of the circuit board, reducing losses caused by faults. At the same time, fault prediction optimization can not only detect problems in advance, but also provide effective feedback to designers, helping them make design adjustments in the early stage to improve the stability and safety of the circuit board. By dynamically adjusting the parameters in the fault prediction model, the system can adapt to new design requirements and environmental changes, ensuring the accuracy and reliability of fault prediction, and thus providing a highly predictable circuit board design optimization solution.

[0015] Preferably, the method of collecting historical fault design parameters of the circuit board, executing a fault model training strategy, and generating a fault prediction model of the circuit board includes: Dynamically adjust the p value: Obtain u candidate p values set; ; Use cross-validation to calculate the performance metric function set on its validation set ; Substitute each candidate p value into the performance metric function set; For any candidate p value, calculate the target metric value , where is the weight of each performance metric function; Obtain the p value with the largest target metric value as the p of the fault prediction model.

[0016] By dynamically adjusting the p-value, the fault prediction model can be optimized according to different training data and problem requirements, ensuring that the model performs best in different scenarios. By adjusting the p-value, different types of distance metrics can be flexibly selected, thereby optimizing the prediction ability of the model. Methods such as Bayesian optimization can help automate the selection of the most suitable p-value without manual setting, thus improving the flexibility and automation level of the model. By calculating the performance metric function set through cross-validation, the effect of each candidate p-value can be quickly evaluated to find the optimal p-value. The process of dynamically adjusting the p-value can ensure that the model is optimized according to the actual situation, avoiding overfitting and underfitting problems, and ensuring the efficiency and accuracy of the fault prediction model in practical applications. Through these optimizations, the system can dynamically adapt to different design parameters and provide more accurate fault prediction results, thereby improving the quality and reliability of circuit board design.

[0017] In a second aspect, the present application provides a circuit board detection system, adopting the following technical solution: A circuit board detection system includes: A data collection and preprocessing module: collecting and processing historical design parameters and simulation data; A model training and verification module: training a prediction model and conducting verification; An adaptive simulation module: dynamically adjusting the simulation process based on the output of the prediction model; A fault prediction and optimization module: predicting the fault probability and making threshold judgments; A fault model training module: constructing and training a fault prediction model; A user interface module: interacting with the user and displaying simulation results and fault warnings; A data storage and management module: managing and storing historical data, models, and simulation results.

[0018] In a third aspect, the present application provides a circuit board detection device, adopting the following technical solution: A circuit board detection device includes: circuit modules such as an MCU, 2.4G, an LED indication module, a 12V power supply, a 5V power supply, a 2.5V power supply, current detection, voltage detection, and charging control; Among them, the LED indication module is used to display the charging status, battery status, and cost status; The current and voltage detection modules are used to detect the charging current and charging voltage; The 2.5V power supply is used to supply power to each circuit module; The 12V power supply and the LED lamp output control module are used for power supply and controlling external LED lamps; The 5V power supply and the USB output control module are used for charging external USB devices and controlling them; The memory is used to store the device ID and the usage status of the product; The 2.4G module is used for local communication and receives recharge passwords for recharge. Temperature detection is used to detect the temperature changes of the internal battery during charging and discharging. Charge control is used to manage and control the charging of the battery by the charger or solar panel.

[0019] In a fourth aspect, the present application provides a circuit board detection medium, adopting the following technical solution: a circuit board detection medium, including: a battery simulator, a DC power supply, a 5V test electronic load instrument, a 5V test electronic multimeter, a 12V test electronic load instrument, a 12V test electronic multimeter, a NEOV3.0 tooling fixture, and a PC.

[0020] The present invention has the following beneficial effects: 1. In this circuit board detection method, data collection and preprocessing are important steps in the circuit board detection method, which provide high-quality and accurate basic data for subsequent model training and prediction. First, by collecting the historical design parameters and simulation results of the circuit board, the comprehensiveness and authenticity of the data can be ensured. These historical data not only cover the performance of the circuit under different design conditions but also include the results in various simulation environments, thus providing a wide range of samples for the learning of the model. Normalization processing is particularly important. By adjusting the design parameters and simulation results to a unified scale range, it avoids the model being biased towards certain features due to the difference in the order of magnitude of the features. This process not only reduces the bias during model training but also speeds up the convergence rate of the model and improves the training efficiency. Through these preprocessing steps, the noise and inconsistency of the data can be eliminated, and the model's ability to learn key features can be enhanced. Finally, the preprocessed data can provide reliable input for subsequent model training, enabling the model to extract effective rules from the data during the training process and improving the accuracy and generalization ability of the prediction.

[0021] 2. In this circuit board detection method, model training and verification are the core links in the circuit board detection method. Through this process, an accurate and generalizable prediction model can be ensured. First, when using historical data for training, the model can learn the relationship between the design parameters and the simulation results, thereby predicting the output behavior of the circuit board. The use of the linear regression model enables the model to fit the output of the circuit by adjusting the regression coefficients and bias terms, ensuring that the prediction results are as close as possible to the actual values. Through the weighted loss function, the model can pay more attention to certain important simulation results and optimize the prediction accuracy under specific design parameters, which is particularly important for circuit boards that need to process different types of simulation data. The introduction of cross-validation ensures the generalization ability of the model, effectively preventing overfitting, so that the model is not only effective on the training set but also maintains good prediction performance on new unseen data.

[0022] 3. For this circuit board detection method, by adjusting the regression coefficients and bias terms, the model can be continuously optimized to reduce the prediction error. The training and validation processes not only improve the accuracy of the model but also provide a solid foundation for subsequent adaptive simulation and fault prediction, ensuring the reliability and efficiency of circuit board design and performance prediction. In summary, model training and validation are key steps to ensure the accuracy of circuit board prediction.

[0023] 4. For this circuit board detection method, adaptive simulation can significantly improve the simulation efficiency and accuracy by dynamically adjusting the simulation process according to the results of the prediction model. First, when the design parameters of the circuit board are obtained and input into the output prediction model, the model can predict the behavior of the circuit board under these design conditions. This process provides accurate initial conditions for the simulation, enabling the simulation process to more precisely simulate the working state of the circuit. Based on these prediction results, the simulation tool can dynamically adjust the simulation step size according to actual needs. If the simulation result has a large error, the step size can be appropriately reduced to improve the simulation accuracy; on the contrary, when the simulation result is relatively stable, the step size can be increased to accelerate the simulation process and save computing resources. The dynamic adjustment of the simulation accuracy can also improve the computing efficiency while meeting the simulation requirements. If it is found during the simulation process that certain parameters have a greater impact on the results, the accuracy can be appropriately increased to ensure the accuracy of the simulation results. In addition, the dynamic adjustment of the design parameters can optimize the circuit design according to the simulation results. By fine-tuning the range of design parameters, the reliability and performance of the circuit board can be further improved. Adaptive simulation not only enhances the flexibility and accuracy of the simulation process but also can respond to different simulation requirements through dynamic adjustment, thus providing efficient and accurate support for design optimization and product verification.

[0024] 5. For this circuit board detection method, fault prediction optimization is an essential part of the circuit board detection method. Through the fault prediction model, potential design problems can be identified in advance to avoid faults during the actual production and use of the circuit board. By collecting historical fault design parameters and analyzing them, the system can establish a fault prediction model for specific design conditions. This model can predict the fault probability of the circuit board based on new design parameters. When the fault probability of the circuit board exceeds the set threshold, the system can issue a warning in a timely manner, prompting the designer that there may be defects in the circuit board design, so as to take measures to correct the problem in advance. This preventive detection method can greatly reduce the risk of faults and improve the reliability and safety of the circuit board. By dynamically adjusting the parameters in the fault prediction model, the system can adapt to changes in different design conditions and environments, thus ensuring the accuracy and practicality of fault prediction. Fault prediction can not only help designers optimize the circuit board design and improve the overall quality of the product but also effectively reduce the economic losses and time costs caused by faults.

[0025] 6. The circuit board detection method. Dynamically adjusting the p value is an important part of optimizing the fault prediction model, which can automatically select the most appropriate p value according to the characteristics of the data and the actual performance of the model. By selecting the appropriate p value, the model can flexibly adjust the distance measurement method, thereby improving the accuracy of the prediction results. Through the cross-validation calculation of candidate p values, the performance of each candidate p on the validation set can be quickly evaluated, and then the optimal p value can be selected. Cross-validation can effectively avoid the bias of a single data set and ensure the stable performance of the model on different data sets. Dynamically adjusting the p value can not only improve the accuracy of the fault prediction model, but also make it adapt to different types of fault data and optimize the performance of the model under different design conditions. By calculating the target metric value and selecting the p with the best performance, the performance of the model can be further improved, enabling it to more accurately predict the fault probability of the circuit board in practical applications. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic flow chart of the method of the present invention.

[0027] Figure 2 It is a schematic diagram of the system modules of the present invention.

[0028] Figure 3 It is a schematic diagram of the composition relationship of the circuit board detection medium of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0029] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0030] Embodiment 1. Refer to Figure 1 , a circuit board detection method, including: S1. Data collection and preprocessing; Collect the historical design parameters and simulation results of the circuit board, and process the design parameters and simulation results to form a data set ; S2. Model training and verification; According to the data set, execute the prediction model training strategy to generate the output prediction model of the circuit board; S3. Adaptive simulation; Obtain the current design parameters of the circuit board, apply them to the output prediction model to obtain the model result; According to the model result, execute the simulation dynamic adjustment strategy to dynamically adjust the simulation process of the simulation tool; S4. Fault prediction optimization; Collect the failed design parameters of the circuit board's history, execute the failure model training strategy, and generate a failure prediction model for the circuit board; Apply the current design parameters of the circuit board to the failure prediction model to obtain the failure probability; Set the failure probability threshold; If the failure probability ≥ the failure probability threshold, it is prompted that the design parameters of the circuit board are unqualified.

[0031] The design parameters and simulation results of the circuit board's history are collected and processed, including: For any pair of design parameters and simulation results ; Record the collected historical design parameters as , where a is the number of elements in the design parameters; Record the collected historical simulation results as , where b is the number of elements in the simulation results; Perform normalization processing on each historical design parameter: Calculate the modulus of the design parameters: ; Calculate the normalized design parameters ; Calculate the modulus of the simulation results: ; Calculate the normalized simulation results .

[0032] Through data collection and preprocessing, the data quality and consistency in the circuit board detection method can be ensured. First, collecting the design parameters and simulation results of the circuit board's history provides a comprehensive basic dataset, which helps with the subsequent training and validation of the model. By extracting the features of the design parameters and simulation results, key features related to the prediction target can be refined, thereby reducing the interference of irrelevant information and improving the effectiveness of the data. Normalization processing is crucial for eliminating the dimensional differences between data. By unifying the numerical scales of different features to the same range, each feature is equally treated in model training, thus avoiding certain features from dominating the learning process of the model due to overly large scales. By means of standardization and denoising, outliers and noise in the data can be effectively removed, further improving the data quality. This process can provide clean and normalized input data for model training, thereby improving the accuracy and convergence speed of the model, reducing the waste of computing resources, and ultimately improving the prediction performance. Through data collection and preprocessing, a solid foundation can be provided for subsequent simulation, failure prediction, and optimization, making the overall operation of the system more efficient and reliable.

[0033] Performing a prediction model training strategy based on the data set to generate an output prediction model for the circuit board, including: Training the output prediction model using a linear regression model: Setting the predicted output result ; Among them, is the regression coefficient of the linear regression model, c is the bias term of the linear regression model, 1 ≦ i ≦ b; Calculating the weighted loss function: Among them, is a constant, is the weight of the i-th element of the simulation result, is the i-th element in the simulation result, is the value of the i-th element of the predicted simulation result; Adjusting the regression coefficient and the bias term to reduce the weighted loss function: For any regression coefficient Among them, is the j-th element of the design parameter; For the bias term c;

[0034] Setting the deviation threshold ; When then stop adjusting the regression coefficient and the bias term.

[0035] Through model training and validation, the prediction accuracy and generalization ability of the circuit board detection method can be ensured. First, historical data is used for model training, enabling the model to learn the relationship between circuit board design parameters and simulation results and generating a model suitable for actual prediction. Through cross-validation, the performance of the model on different data sets can be effectively evaluated, overfitting problems can be avoided, and the prediction ability of the model on unknown data can be improved. Through the weighted loss function, the model can optimize the more important parts of the simulation result, making the training process more accurate and able to specifically focus on the parameters that have a greater impact on the circuit board performance. By dynamically adjusting the regression coefficient and the bias term, the model can be continuously optimized to reduce the prediction error and ensure that its prediction results are more accurate. Through these steps, the model can not only improve the training accuracy but also enhance its reliability in practical applications, especially when facing unknown data. Model training and validation provide a precise tool for circuit board design optimization, enabling designers to predict and adjust the performance of the circuit board, thereby improving the quality and reliability of circuit board design.

[0036] Performing a simulation dynamic adjustment strategy according to the model result to dynamically adjust the simulation process of the simulation tool, including: Dynamically adjusting the step size: Obtain the current simulation step size ; Calculate the simulation step size for the next simulation ; , where is the stability threshold set for the i-th element of the simulation result, is the coefficient for step size adjustment; Dynamically adjust the simulation accuracy: Obtain the current simulation accuracy ; Calculate the simulation accuracy for the next simulation ; , where is the coefficient for accuracy adjustment, is the target value of the i-th element of the simulation result; Dynamically adjust the design parameters: For any one in the simulation result ; Obtain the influence coefficient of each element in the predicted design parameters on ; Obtain the maximum change range of each element in the design parameters ; Dynamically adjust the change range of each element in the design parameters: , 1 ≤ i ≤ a.

[0037] Collect the historical fault design parameters of the circuit board, execute the fault model training strategy, and generate the fault prediction model of the circuit board, including: Record the historical fault design parameters as ; Calculate the similarity , where is a positive number used to adjust the sensitivity of the distance; is the current input value of the i-th element in the design parameters.

[0038] ​Through adaptive simulation, it is possible to dynamically optimize the simulation process in circuit board detection, improving simulation efficiency and accuracy. First of all, based on the current design parameters and prediction model, the simulation tool can quickly generate simulation results to ensure that the simulation process meets the requirements of the actual circuit board design. By dynamically adjusting the simulation step size, the system can flexibly adjust the step size according to the accuracy requirements of the simulation results, ensuring that the simulation process improves the calculation efficiency while guaranteeing accuracy. When the error of the simulation results is large, reducing the step size can improve the simulation accuracy; when the simulation results are relatively stable, increasing the step size helps to accelerate the simulation process and save computing resources. Dynamically adjusting the simulation accuracy can also optimize the calculation efficiency while meeting the simulation requirements, avoiding unnecessary calculation waste. By dynamically adjusting the range of changes in design parameters, the simulation tool can optimize the circuit board design to ensure that the circuit board can achieve an ideal working state under different design conditions. This flexible and dynamic simulation adjustment method makes the simulation process not only more efficient, but also provides accurate simulation results for different scenarios and requirements, helping to accelerate the design verification and optimization of circuit boards Collect the historical fault design parameters of the circuit board, execute the fault model training strategy, and generate the fault prediction model of the circuit board, including: Dynamically adjust the p value: Obtain u candidate p values set ; Use cross-validation to calculate the performance metric function set on its validation set ; Substitute each candidate p value into the performance metric function set; For any candidate p value, calculate the target metric value , where is the weight of each performance metric function; Obtain the p value with the largest target metric value as the p of the fault prediction model.

[0039] Through fault prediction optimization, potential problems in circuit board design can be identified in advance, avoiding failures during actual use. First, the fault prediction model can be trained based on historical fault data to generate a prediction model adapted to the current design parameters. By inputting the current design parameters of the circuit board into the fault prediction model, the fault probability under this design condition can be calculated. When the fault probability exceeds the set threshold, the system can issue a warning in a timely manner, indicating that there may be design defects to the designers, avoiding the risk of failures during the actual production process. This process can not only improve the reliability of the circuit board but also provide targeted feedback to the designers, helping them make adjustments and optimizations during the design stage, thus reducing the later repair costs. In addition, the fault prediction model can be flexibly adjusted according to different design parameters and working conditions to ensure the prediction accuracy of the model under different situations. Through fault prediction optimization, the stability and safety of the circuit board can be effectively improved, the probability of failures can be reduced, and the overall design quality can be enhanced.

[0040] By performing cross-validation on multiple candidate p-values, the performance of different p-values on the validation set can be evaluated, and thus the optimal p-value can be selected. This adjustment method can ensure that the fault prediction model performs best under different data conditions, avoiding errors caused by fixed p-values. By calculating the target index value and dynamically adjusting p, the sensitivity of the model can be adjusted according to different fault data, further improving the accuracy of the prediction results. This flexible adjustment method enables the fault prediction model to still provide efficient and accurate predictions when facing various complex situations, enhancing the adaptability and stability of the circuit board detection system.

[0041] Example Two, referring to Figure 2 , a circuit board detection system, including: Data collection and preprocessing module: Collect and process historical design parameters and simulation data; Model training and verification module: Train the prediction model and conduct verification; Adaptive simulation module: Dynamically adjust the simulation process based on the output of the prediction model; Fault prediction and optimization module: Conduct fault probability prediction and threshold judgment; Fault model training module: Construct and train the fault prediction model; User interface module: Interact with users and display simulation results and fault warnings; Data storage and management module: Manage and store historical data, models, and simulation results.

[0042] Example Three, a circuit board detection device, including: circuit modules such as MCU, 2.4G, LED indication module, 12V power supply, 5V power supply, 2.5V power supply, current detection, voltage detection, and charging control. Among them, the LED indication module is used to display the charging status, power status, and cost status; The current and voltage detection module is used to detect the charging current and charging voltage; The 2.5V power supply is used to supply power to each circuit module; The 12V power supply and the LED lamp output control module are used to supply power and control external LED lamps; The 5V power supply and the USB output control module are used to charge and control external USB devices; The memory is used to store the device ID and the usage status of the product; The 2.4G module is used for local communication and receiving recharge passwords for recharge; Temperature detection is used to detect the temperature change of the internal battery during charge and discharge; Charge control is used to manage and control the charging of the battery by the charger or solar panel.

[0043] Example 4, referring to Figure 3 , a circuit board detection medium, including: Battery simulator, DC power supply, 5V test electronic load instrument, 5V test electronic multimeter, 12V test electronic load instrument, 12V test electronic multimeter, NEOV3.0 tooling fixture, and PC.

[0044] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising", or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article, or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or elements inherent to such process, method, article, or device.

[0045] The above are only the preferred embodiments of the present invention. It should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A circuit board detection method, characterized in that, Including: S1. Data collection and preprocessing; Collect the design parameters and simulation results of the circuit board history, process the design parameters and simulation results, and form a data set ; S2. Model training and validation; According to the data set, execute the prediction model training strategy to generate the output prediction model of the circuit board; S3. Adaptive simulation; Obtain the design parameters of the circuit board this time, apply them to the output prediction model, and obtain the model result; According to the model result, execute the simulation dynamic adjustment strategy to dynamically adjust the simulation process of the simulation tool; S4. Fault prediction optimization; Collect the historical fault design parameters of the circuit board, execute the fault model training strategy, and generate the fault prediction model of the circuit board; Apply the design parameters of the circuit board this time to the fault prediction model to obtain the fault probability; Set the fault probability threshold; If the fault probability ≥ the fault probability threshold, it is prompted that the design parameters of the circuit board are unqualified.

2. The circuit board detection method according to claim 1, wherein The historical design parameters and simulation results of the circuit board are collected, and the design parameters and simulation results are processed, including: For any pair of design parameters and simulation results ; Denote the collected historical design parameters as , where a is the number of elements in the design parameters; Denote the collected historical simulation results as , where b is the number of elements in the simulation results; Perform normalization processing on each historical design parameter: Calculate the modulus of the design parameter: ; Calculate the normalized design parameters ; Calculate the modulus of the simulation result: ; Calculate the normalized simulation results .

3. The circuit board detection method according to claim 1, characterized in that, The execution of the prediction model training strategy according to the data set to generate the output prediction model of the circuit board includes: Use a linear regression model to train the output prediction model: Set the predicted output result ; Among them, is the regression coefficient of the linear regression model, c is the bias term of the linear regression model, and 1 ≤ i ≤ b; Calculate the weighted loss function: , where is a constant is the weight of the i-th element of the simulation result is the i-th element in the simulation result is the value of the predicted i-th element of the simulation result; Adjust the regression coefficient and bias term to reduce the weighted loss function: Update any one of the regression coefficients , where is the j-th element of the design parameter; For the bias term c; 4. Set the deviation threshold ; When , stop adjusting the regression coefficient and the bias term.

5. The circuit board detection method according to claim 1, wherein, The execution of the simulation dynamic adjustment strategy according to the model result to dynamically adjust the simulation process of the simulation tool includes: Dynamically adjust the step size: Obtain the current simulation step size ; Calculate the simulation step size for the next simulation ; , where is the stability threshold set for the i-th element of the simulation result, is the coefficient for step size adjustment; Dynamically adjust the simulation accuracy: Obtain the current simulation accuracy ; Calculate the simulation accuracy for the next simulation ; , where is the coefficient for precision adjustment, is the target value of the i-th element of the simulation result; Dynamically adjust the design parameters: For any one of the simulation results ; Obtain the influence coefficient of each element pair in the predicted design parameters for ; Obtain the maximum variation range of each element in the design parameters ; Dynamically adjust the change range of each element in the design parameters: , where 1 ≦ i ≦ a.

6. The circuit board detection method according to claim 1, wherein The collection of the historical fault design parameters of the circuit board, the execution of the fault model training strategy, and the generation of the fault prediction model of the circuit board include: Record the historical failure design parameters as ; Calculate similarity , where is a positive number used to adjust the sensitivity of the distance; is the current input value of the i-th element in the design parameters.

7. The circuit board detection method according to claim 5, wherein The collection of the historical fault design parameters of the circuit board, the execution of the fault model training strategy, and the generation of the fault prediction model of the circuit board include: Dynamically adjust the p value: Obtain the set u candidate p-values ; Calculate the performance metric function set on its validation set using cross-validation ; Substitute each candidate p value into the performance index function set; For any candidate p value, calculate the target metric value , where is the weight of each performance metric function; Obtain the p value with the largest target index value as the p of the fault prediction model.

8. A circuit board detection system, characterized in that, Including: Data collection and preprocessing module: Collect and process historical design parameters and simulation data; Model training and validation module: Train the prediction model and perform validation; Adaptive simulation module: Dynamically adjust the simulation process based on the output of the prediction model; Fault prediction and optimization module: Perform fault probability prediction and threshold judgment; Fault model training module: Construct and train the fault prediction model; User interface module: Interact with the user and display the simulation results and fault warnings; Data storage and management module: Manage and store historical data, models, and simulation results.

9. A circuit board detection device, characterized in that, Including: Composed of circuit modules such as MCU, 2.4G, LED indication module, 12V power supply, 5V power supply, 2.5V power supply, current detection, voltage detection, and charging control; Among them, the LED indication module is used to display the charging status, power status, and cost status; The current and voltage detection modules are used to detect the charging current and charging voltage; The 2.5V power supply is used to supply power to each circuit module; The 12V power supply and LED lamp output control module are used for power supply and control of external LED lamps; The 5V power supply and USB output control module are used for charging and control of external USB devices; The memory is used to store the ID of the device and the usage status of the product. The 2.4G module is used for local communication and receives recharge passwords for recharge; Temperature detection is used to detect the temperature changes of the internal battery during charge and discharge; Charge control is used to manage and control the charging of the battery by the charger or solar panel.

10. A circuit board detection medium, characterized in that, It includes: Battery simulator, DC power supply, 5V test electronic load meter, 5V test electronic multimeter, 12V test electronic load meter, 12V test electronic multimeter, NEOV3.0 tooling fixture and PC.

Images