CN117630649A

CN117630649A - PCBA testing method, device, equipment and storage medium

Info

Publication number: CN117630649A
Application number: CN202410110673.6A
Authority: CN
Inventors: 余海滨; 赵光礼
Original assignee: Shenzhen Evision Semiconductor Technology Co ltd
Current assignee: Shenzhen Evision Semiconductor Technology Co ltd
Priority date: 2024-01-26
Filing date: 2024-01-26
Publication date: 2024-03-01
Anticipated expiration: 2044-01-26
Also published as: CN117630649B

Abstract

The application relates to the technical field of PCBA testing and discloses a PCBA testing method, device, equipment and storage medium. The method comprises the following steps: acquiring material information of a target PCBA and heat sensitivity of an electronic component, and creating a first reflow heating curve; carrying out repeated reflux heating test and detection to obtain target temperature distribution data and target electromagnetic characteristic data; performing temperature change characteristic analysis and electromagnetic characteristic change analysis; vector conversion and feature fusion are carried out, so that a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector are obtained; PCBA welding fault detection is carried out through the PCBA welding fault detection model, and a PCBA welding fault detection result is obtained; the first reflow heating curve is subjected to curve optimization to obtain a second reflow heating curve, and the self-adaptive optimization of the reflow heating of the PCBA is realized and the accuracy of the welding test of the PCBA is improved.

Description

PCBA testing method, device, equipment and storage medium

Technical Field

The application relates to the technical field of PCBA testing, in particular to a PCBA testing method, a PCBA testing device, PCBA testing equipment and a PCBA storage medium.

Background

With the continuous development and intelligent trend of electronic products, PCBA is increasingly widely applied in various industries.

In the current PCBA production, the traditional testing method is difficult to comprehensively evaluate welding quality, such as insufficient detailed analysis on PCBA materials and thermal sensitivity, and challenges of how to accurately acquire temperature distribution and electromagnetic characteristic data in the testing process, so that the accuracy of the prior art is low.

Disclosure of Invention

The application provides a method, a device, equipment and a storage medium for testing PCBA, which realize self-adaptive optimization of PCBA reflow heating and improve the welding test accuracy of PCBA.

In a first aspect, the present application provides a method for testing a PCBA, the method for testing a PCBA comprising:

acquiring material information of a target PCBA and heat sensitivity of an electronic component, and creating a first reflow heating curve of the target PCBA according to the material information and the heat sensitivity;

performing multiple reflow heating tests and temperature distribution detection on the target PCBA according to the first reflow heating curve to obtain target temperature distribution data of each reflow heating test, and performing electromagnetic characteristic measurement on the target PCBA to obtain target electromagnetic characteristic data of each reflow heating test;

Performing temperature change characteristic analysis on the target temperature distribution data of each reflux heating test based on the first Wiener process model to obtain a temperature change characteristic set, and performing electromagnetic characteristic change analysis on the target electromagnetic characteristic data of each reflux heating test based on the second Wiener process model to obtain an electromagnetic change characteristic set;

carrying out correlation coefficient analysis on the temperature change feature set and the electromagnetic change feature set respectively to obtain feature correlation coefficients, and carrying out vector conversion and feature fusion on the temperature change feature set and the electromagnetic change feature set according to the feature correlation coefficients to obtain a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector;

inputting the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector into a preset PCBA welding fault detection model to detect PCBA welding faults, and obtaining a PCBA welding fault detection result;

and performing curve optimization on the first reflow heating curve according to the PCBA welding fault detection result by a preset double-layer programming solving algorithm to obtain a second reflow heating curve of the target PCBA.

In a second aspect, the present application provides a test device of a PCBA, the test device of a PCBA including:

the acquisition module is used for acquiring the material information of the target PCBA and the heat sensitivity of the electronic component, and creating a first reflow heating curve of the target PCBA according to the material information and the heat sensitivity;

the test module is used for carrying out multiple reflow heating tests and temperature distribution detection on the target PCBA according to the first reflow heating curve to obtain target temperature distribution data of each reflow heating test, and carrying out electromagnetic characteristic measurement on the target PCBA to obtain target electromagnetic characteristic data of each reflow heating test;

the analysis module is used for carrying out temperature change characteristic analysis on the target temperature distribution data of each reflow heating test based on the first Wiener process model to obtain a temperature change characteristic set, and carrying out electromagnetic characteristic change analysis on the target electromagnetic characteristic data of each reflow heating test based on the second Wiener process model to obtain an electromagnetic change characteristic set;

the conversion module is used for respectively carrying out correlation coefficient analysis on the temperature change feature set and the electromagnetic change feature set to obtain feature correlation coefficients, and carrying out vector conversion and feature fusion on the temperature change feature set and the electromagnetic change feature set according to the feature correlation coefficients to obtain temperature change feature vectors, electromagnetic change feature vectors and fusion change feature vectors;

The detection module is used for inputting the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector into a preset PCBA welding fault detection model to detect PCBA welding faults, and obtaining a PCBA welding fault detection result;

and the optimization module is used for carrying out curve optimization on the first reflow heating curve according to the PCBA welding fault detection result by a preset double-layer planning solving algorithm to obtain a second reflow heating curve of the target PCBA.

A third aspect of the present application provides a computer device comprising: a memory and at least one processor, the memory having instructions stored therein; the at least one processor invokes the instructions in the memory to cause the computer device to perform the test method of PCBA described above.

A fourth aspect of the present application provides a computer readable storage medium having instructions stored therein which, when run on a computer, cause the computer to perform the method of testing a PCBA as described above.

In the technical scheme provided by the application, the material information of the PCBA and the heat sensitivity of the electronic component are analyzed, so that the reflow soldering temperature point and the temperature range are predicted, and the potential soldering quality problem can be recognized in advance. And target temperature distribution data and electromagnetic characteristic data are obtained through repeated reflow heating tests and temperature distribution detection, so that accurate data support is provided for the prediction and control of welding quality. The temperature distribution data is analyzed by using the first Wiener process model, so that drift and diffusion characteristics of temperature change can be captured, and deeper temperature information is provided for subsequent fault detection. And the electromagnetic characteristic data is analyzed based on the second Wiener process model, so that the drift and diffusion characteristics of electromagnetic change can be obtained, and the understanding of the electromagnetic characteristic change is improved. By carrying out correlation coefficient analysis on the temperature change characteristic set and the electromagnetic change characteristic set, the association degree between the temperature change characteristic set and the electromagnetic change characteristic set can be quantified, and the understanding of the relation between different characteristics is improved. And vector conversion and feature fusion are carried out to obtain a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector, so that the influence of a plurality of features on welding quality is comprehensively considered, and the accuracy of fault detection is improved. And inputting the extracted feature vector into the model by using a preset PCBA welding fault detection model, so as to realize comprehensive detection of PCBA welding quality. By combining the detection results of the temperature influence factors, the electromagnetic influence factors and the comprehensive influence factors, the comprehensive judging capability of welding faults is improved. And the first reflow heating curve is subjected to curve optimization through a double-layer planning algorithm, so that the fine control of the welding process is realized, and the welding quality and the production efficiency are improved. Multiple optimization targets, such as improvement of welding quality, reduction of energy consumption and the like, can be considered simultaneously, so that multi-target optimization is realized. By comprehensively analyzing and optimizing the welding process, the time of trial and error and adjustment is reduced, and the production efficiency is improved. By optimizing the heating curve, the resources can be more effectively utilized, the energy consumption is reduced, the production cost is reduced, the self-adaptive optimization of the reflow heating of the PCBA is realized, and the welding test accuracy of the PCBA is improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and other drawings may be obtained based on these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of one embodiment of a method for testing PCBA in an embodiment of the present application;

fig. 2 is a schematic diagram of an embodiment of a test device for PCBA in an embodiment of the present application.

Detailed Description

The embodiment of the application provides a PCBA test method, device and equipment and a storage medium. The terms "first," "second," "third," "fourth" and the like in the description and in the claims of this application and in the above-described figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.

For ease of understanding, the following describes a specific flow of an embodiment of the present application, referring to fig. 1, and one embodiment of a method for testing PCBA in an embodiment of the present application includes:

step 101, acquiring material information of a target PCBA and heat sensitivity of an electronic component, and creating a first reflow heating curve of the target PCBA according to the material information and the heat sensitivity;

it can be understood that the execution body of the application may be a testing device of the PCBA, and may also be a terminal or a server, which is not limited herein. The embodiment of the present application will be described by taking a server as an execution body.

Specifically, the material information of the target PCBA is obtained, including the material of the circuit board itself and the material of the electronic component mounted on the circuit board. The materials of the circuit board include various types of insulating substrates such as FR4 or metal core PCBs, which have different resistance and conduction characteristics to heat. Likewise, electronic components also have their own thermal sensitivity, for example, some components deform or break at higher temperatures, while others are relatively stable to temperature changes. The target PCBA is hierarchically analyzed to identify and understand the physical locations of the various components on the PCBA and their role in the circuit, to ascertain the thermal sensitivity of each component, and to predict the thermal impact they are subjected to during the reflow soldering process. For example, components near the center of the PCBA may experience higher temperatures than components at the edge portions. Critical temperature points required during reflow soldering are predicted. This includes determining the minimum temperature required to achieve a good weld, and ensuring that the maximum temperature of any sensitive components is not compromised. Determination of the temperature range requires a combination of thermal sensitivity of all components and thermal properties of the board material. And carrying out reflow soldering time prediction on the target PCBA. The time required to heat the PCBA to the desired temperature is estimated, as well as the duration required to maintain this temperature, to ensure the weld quality. Too long a heating time results in damage to the assembly, while too short a time results in poor welding. A first reflow heating profile of the target PCBA is created from the temperature range data and the time data of the reflow soldering. This profile not only ensures that the proper welding temperature is achieved, but also maintains the safety of the assembly throughout the welding process.

102, performing multiple reflow heating tests and temperature distribution detection on a target PCBA according to a first reflow heating curve to obtain target temperature distribution data of each reflow heating test, and performing electromagnetic characteristic measurement on the target PCBA to obtain target electromagnetic characteristic data of each reflow heating test;

specifically, multiple reflow heating tests are performed on the target PCBA according to the first reflow heating profile. In the testing process, temperature distribution detection is carried out on the target PCBA through a preset infrared temperature sensor, and temperature changes of the PCBA surface and different components in the heating process are captured, so that initial temperature distribution data of each heating test are obtained. And carrying out data cleaning and data standardization treatment on the initial temperature distribution data obtained by each reflux heating test. The data cleaning aims to remove abnormal data caused by sensor errors, environmental interference or data transmission errors and the like, and ensure the accuracy of subsequent analysis. The data normalization process is to convert the temperature data into a common format, thereby facilitating comparison of data differences between different tests. And measuring electromagnetic characteristics of the target PCBA. Electromagnetic responses of the PCBA at different heating stages are detected to obtain initial electromagnetic characteristic data of each reflow heating test. These data include magnetic field strength data and frequency distribution data, which facilitate evaluation of the electromagnetic compatibility and functional performance of the PCBA. Because PCBA may exhibit different electromagnetic properties at different heating stages, these data provide insight into the electromagnetic behavior of PCBA. Performing scale normalization processing on the initial electromagnetic characteristic data obtained by each reflow heating test to obtain target electromagnetic characteristic data of each reflow heating test, wherein the target electromagnetic characteristic data comprises: magnetic field strength data and frequency distribution data. The data generated by the different tests are converted into a standard format with comparability, so that electromagnetic characteristic data between the different tests can be effectively compared.

Step 103, performing temperature change characteristic analysis on target temperature distribution data of each reflow heating test based on a first Wiener process model to obtain a temperature change characteristic set, and performing electromagnetic characteristic change analysis on target electromagnetic characteristic data of each reflow heating test based on a second Wiener process model to obtain an electromagnetic change characteristic set;

specifically, estimation of a drift coefficient and a diffusion coefficient is performed on target temperature distribution data of each reflow heating test based on a first Wiener process model. The temperature drift coefficient represents the trend of temperature over time, while the temperature diffusion coefficient describes the randomness or uncertainty of this change. And executing conditional probability calculation of a Wiener process on the target temperature distribution data according to the obtained temperature drift coefficient and diffusion coefficient. Based on current and past observation data, the probability distribution of the temperature at a certain moment in the future is predicted through a Wiener process model. The conditional probability data reflects the probability that the future temperature reaches a certain value given the current temperature state. And extracting temperature change characteristics of the target temperature distribution data according to the temperature condition probability data. A critical pattern of change is identified from the complex temperature data, forming a set of temperature change features. These features help understand the heating response of the PCBA at different stages, providing an important basis for optimizing the soldering process. And estimating the drift coefficient and the diffusion coefficient of the target electromagnetic characteristic data by adopting a second Wiener process model. The electromagnetic drift coefficient represents the trend of the electromagnetic characteristic over time, while the electromagnetic diffusion coefficient describes the randomness or uncertainty of this change. By calculating the coefficients, the dynamic change of the electromagnetic property of the PCBA in the heating process can be revealed. And performing conditional probability calculation of the Wiener process by using the electromagnetic drift coefficient and the diffusion coefficient to provide prediction for the future state of the electromagnetic characteristic. The electromagnetic conditional probability data provides a probability distribution that the electromagnetic property reaches a particular value at a future time in the current electromagnetic state. And carrying out electromagnetic change feature extraction on the target electromagnetic characteristic data according to the electromagnetic condition probability data. And identifying a key change mode from the electromagnetic characteristic data to form an electromagnetic change characteristic set. These electromagnetic features help understand the electromagnetic behavior of the PCBA during heating, providing key information for improving electromagnetic compatibility and optimizing electromagnetic characteristics.

104, respectively carrying out correlation coefficient analysis on the temperature change feature set and the electromagnetic change feature set to obtain feature correlation coefficients, and carrying out vector conversion and feature fusion on the temperature change feature set and the electromagnetic change feature set according to the feature correlation coefficients to obtain temperature change feature vectors, electromagnetic change feature vectors and fusion change feature vectors;

specifically, statistical analysis is performed on the temperature change feature set, and the average value and standard deviation of the temperature change feature set are calculated to obtain the general trend and the dispersion degree of the temperature change in the whole set. Likewise, the general variation pattern and variability of the electromagnetic properties are obtained by calculating the mean and standard deviation of the electromagnetic variation characteristics. And calculating and analyzing the correlation between the mean value and the standard deviation of the temperature change characteristic and the mean value and the standard deviation of the electromagnetic change characteristic through the Pearson correlation coefficient. The pearson correlation coefficient is a statistic that measures the degree of linear correlation between two sets of data, with values ranging from-1 to 1, with a near 1 or-1 indicating a strong correlation between the variables and a near 0 indicating no significant correlation. The temperature change feature set is subjected to normalized feature mapping and vector conversion, and the temperature change feature is converted into a series of numerical vectors, and the vectors can more effectively represent the features of the original data. Similarly, the electromagnetic change feature set is converted into an electromagnetic change feature vector through similar normalized feature mapping and vector conversion processes. And carrying out feature fusion on the temperature change feature vector and the electromagnetic change feature vector according to the obtained feature correlation coefficient. Based on the correlation between the two sets of features, these feature vectors are mathematically combined into a comprehensive fusion variant feature vector.

Step 105, inputting the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector into a preset PCBA welding fault detection model to detect PCBA welding faults, and obtaining a PCBA welding fault detection result;

specifically, the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector are input into a preset PCBA welding fault detection model. The model contains three different networks of GRUs (gated loop units) and one output layer, each network performing a specific function for a different feature set. The first GRU network in the model uses bi-directional GRU units to process the temperature change feature vectors. The bidirectional GRU unit can learn time sequence characteristics of temperature data from two directions, so that dynamic characteristics of temperature change are more comprehensively captured, and a temperature hidden characteristic vector is obtained, wherein the temperature hidden characteristic vector comprises deep characteristics of the temperature data. The unidirectional GRU units in the first GRU network further analyze the hidden features to detect temperature dependent PCBA weld faults, generating a first weld fault detection result. The second GRU network performs similar operations on the electromagnetic change feature vector. And extracting hidden features of electromagnetic data by the bidirectional GRU unit to obtain electromagnetic hidden feature vectors. This step enables the model to understand the intrinsic mode of electromagnetic property variation. The unidirectional GRU unit analyzes the hidden features, identifies welding faults related to electromagnetic characteristics, and generates a second welding fault detection result. And processing the fusion change feature vector through a third GRU network. And the bidirectional GRU unit extracts deep hidden features from the fusion feature vector to obtain the fusion hidden feature vector. These hidden features integrate temperature and electromagnetic property information and provide more abundant data for unidirectional GRU units. The unidirectional GRU unit then analyzes the fused hidden features, detects comprehensive welding faults, and outputs a third welding fault detection result. And integrating welding fault detection results generated by the three GRU networks through an output layer of the model. The output layer is used for fusing information from different sources to generate a final PCBA welding fault detection result. The result comprehensively considers temperature, electromagnetism and comprehensive influence factors thereof, and provides a comprehensive welding fault diagnosis.

And 106, performing curve optimization on the first reflow heating curve according to the PCBA welding fault detection result by a preset double-layer programming solving algorithm to obtain a second reflow heating curve of the target PCBA.

Specifically, a plurality of optimization targets are extracted from the PCBA welding fault detection result. These optimization objectives are determined based on the type and nature of the fault detected and are intended to address specific problems identified in the first reflow heating profile, such as reducing soldering defects, optimizing heating time, or improving temperature profile, etc. A preset double-layer planning solving algorithm is obtained, and the algorithm consists of two main parts: the first layer is a plurality of genetic algorithms and the second layer is a linear programming algorithm. Genetic algorithms are optimization algorithms that simulate natural selection and genetic principles, finding optimal solutions by the process of generating, evaluating, selecting and modifying a series of solutions. In this embodiment, each genetic algorithm is responsible for exploring different optimization directions, and finding an optimal heating curve that satisfies a specific constraint condition. And according to the constraint conditions of each genetic algorithm in the determined multiple optimization target double-layer planning solving algorithms. These constraints guide the genetic algorithm in what factors are considered during the search process, such as a particular temperature range, heating rate, or temperature profile for a particular region. Each genetic algorithm optimizes the first reflow heating profile according to these constraints, generating a series of profile optimization parameters. These parameters represent reflow soldering conditions in different schemes, such as temperature variations, heating and cooling rates, etc. The curve optimization parameters are input into the linear programming algorithm of the second layer. Linear programming is a mathematical method for finding the optimal solution under a given linear constraint. By comprehensively considering the optimization parameters from each genetic algorithm, and determining the final curve shape. The output of the linear programming algorithm is a second reflow heating profile for the target PCBA, which combines all optimization objectives and constraints, providing more optimal soldering conditions than the first time.

In the embodiment of the application, the material information of the PCBA and the heat sensitivity of the electronic component are analyzed, and the reflow soldering temperature point and the temperature range are predicted, so that potential soldering quality problems can be recognized in advance. And target temperature distribution data and electromagnetic characteristic data are obtained through repeated reflow heating tests and temperature distribution detection, so that accurate data support is provided for the prediction and control of welding quality. The temperature distribution data is analyzed by using the first Wiener process model, so that drift and diffusion characteristics of temperature change can be captured, and deeper temperature information is provided for subsequent fault detection. And the electromagnetic characteristic data is analyzed based on the second Wiener process model, so that the drift and diffusion characteristics of electromagnetic change can be obtained, and the understanding of the electromagnetic characteristic change is improved. By carrying out correlation coefficient analysis on the temperature change characteristic set and the electromagnetic change characteristic set, the association degree between the temperature change characteristic set and the electromagnetic change characteristic set can be quantified, and the understanding of the relation between different characteristics is improved. And vector conversion and feature fusion are carried out to obtain a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector, so that the influence of a plurality of features on welding quality is comprehensively considered, and the accuracy of fault detection is improved. And inputting the extracted feature vector into the model by using a preset PCBA welding fault detection model, so as to realize comprehensive detection of PCBA welding quality. By combining the detection results of the temperature influence factors, the electromagnetic influence factors and the comprehensive influence factors, the comprehensive judging capability of welding faults is improved. And the first reflow heating curve is subjected to curve optimization through a double-layer planning algorithm, so that the fine control of the welding process is realized, and the welding quality and the production efficiency are improved. Multiple optimization targets, such as improvement of welding quality, reduction of energy consumption and the like, can be considered simultaneously, so that multi-target optimization is realized. By comprehensively analyzing and optimizing the welding process, the time of trial and error and adjustment is reduced, and the production efficiency is improved. By optimizing the heating curve, the resources can be more effectively utilized, the energy consumption is reduced, the production cost is reduced, the self-adaptive optimization of the reflow heating of the PCBA is realized, and the welding test accuracy of the PCBA is improved.

In a specific embodiment, the process of executing step 101 may specifically include the following steps:

(1) Acquiring material information of a target PCBA, wherein the material information comprises materials of a circuit board and materials of an electronic component mounted on the circuit board;

(2) Performing hierarchical structure analysis on the target PCBA to obtain target hierarchical structure information, and performing thermal sensitivity analysis on the electronic component according to the material of the electronic component to obtain thermal sensitivity of the electronic component;

(3) Carrying out reflow soldering temperature point prediction on the target PCBA according to the target hierarchical structure information and the heat sensitivity of the electronic component to obtain reflow soldering temperature point data;

(4) Carrying out reflow soldering temperature range prediction on the target PCBA according to the reflow soldering temperature point data to obtain reflow soldering temperature range data, and carrying out reflow soldering time prediction on the target PCBA to obtain reflow soldering duration data;

(5) A first reflow heating profile of the target PCBA is created from the reflow soldering temperature range data and the reflow soldering duration data.

Specifically, the material information of the target PCBA is obtained, which includes the material of the circuit board itself and the materials of various electronic components mounted on the circuit board. For example, circuit boards are made of FR4 fiberglass composite, while electronic components include copper conductive paths, silicon-based semiconductor devices, and connectors and packages made of various plastics and metal alloys. Each material responds differently to heat, for example silicon-based semiconductors are damaged at high temperatures, while metal alloys are able to withstand higher temperatures. And carrying out hierarchical structure analysis on the target PCBA. The physical location, function, and interrelationship between the various electronic components on the circuit board are identified and understood. For example, a processor is located in a central area of the circuit board, surrounded by memory, power management modules, and various sensors. This hierarchical analysis helps determine which areas are affected by high temperatures during the welding process and which components require special attention because of thermal sensitivity. And carrying out heat sensitivity analysis on the electronic component on the basis of the hierarchical structure analysis. Each component material was evaluated for its performance during heating, such as physical or chemical changes due to temperature increases. For example, plastic encapsulated components deform when a certain temperature is reached, while the copper conductive path remains stable at higher temperatures. This thermal sensitivity analysis helps to take into account the specific requirements of each component in the subsequent reflow soldering temperature point predictions. And predicting reflow soldering temperature points of the target PCBA according to the hierarchical structure information and the heat sensitivity of the electronic component, and determining key temperature points which need to be reached in the soldering process. These temperature points are based on the principle of ensuring the quality of the weld while preventing damage to the heat-sensitive components. For example, if particularly thermally sensitive components are contained in the circuit board, the predicted temperature point may be relatively low to avoid damaging these components. Based on the temperature point data, the reflow soldering temperature range prediction is further performed on the target PCBA to determine the range in which the temperature should be maintained during the soldering process. This range is required to ensure that the solder joints are formed uniformly across the board while preventing thermal damage to the electronic assembly. Reflow soldering time predictions are made for the target PCBA to estimate the duration of heating and cooling to ensure quality and efficiency of soldering. A first reflow heating profile of the target PCBA is created based on the reflow soldering temperature range data and the duration data. This curve represents the temperature change over time throughout the welding process, taking into account all of the above factors. For example, if the circuit board contains a large number of heat-sensitive components, the heating curve can be designed to rise gently to avoid damage caused by rapid temperature rise; on a PCBA composed primarily of a high temperature resistant material, the heating profile is steeper to achieve the desired soldering temperature quickly.

In a specific embodiment, the process of executing step 102 may specifically include the following steps:

(1) Carrying out reflow heating test on the target PCBA for a plurality of times according to the first reflow heating curve, and carrying out temperature distribution detection on the target PCBA through a preset infrared temperature sensor to obtain initial temperature distribution data of each reflow heating test;

(2) Respectively carrying out data cleaning and data standardization treatment on the initial temperature distribution data of each reflux heating test to obtain target temperature distribution data of each reflux heating test;

(3) Performing electromagnetic characteristic measurement on the target PCBA to obtain initial electromagnetic characteristic data of each reflux heating test;

(4) Respectively carrying out scale normalization processing on the initial electromagnetic characteristic data of each reflow heating test to obtain target electromagnetic characteristic data of each reflow heating test, wherein the target electromagnetic characteristic data comprises: magnetic field strength data and frequency distribution data.

Specifically, a series of heating tests were performed on the PCBA according to a first reflow heating profile. These experiments were intended to simulate the temperature changes experienced by the PCBA during soldering. For example, if the first reflow heating profile indicates that the temperature should gradually rise during the first half of the soldering process, it should also be ensured in experiments that the temperature varies in this pattern. In each heating test, the temperature distribution of the PCBA is detected in detail by a preset infrared temperature sensor. The infrared temperature sensor is capable of contactlessly measuring the temperature of the surface of the circuit board, providing accurate temperature readings for each test point. For example, the infrared sensor can capture temperature changes during heating in different parts of the circuit board, such as near the processor, the power management module, or the connector area. These initial temperature profile data show the spatial distribution of the temperature of the PCBA during heating. And performing data cleaning and normalization processing on the initial temperature distribution data. The purpose of the data cleaning is to remove outliers that may be generated due to sensor errors, environmental factors, or other disturbances. For example, if a temperature reading of a certain area is abnormally high or low and has a significant difference from the readings of surrounding areas, this indicates that the reading is erroneous and should be excluded from analysis. And then, carrying out data standardization processing to convert all temperature data into a uniform format so as to facilitate analysis and comparison. For example, all temperature readings may be converted to differences relative to ambient temperature, or normalized to a fixed range, such as between 0 and 1. And measuring electromagnetic characteristics of the target PCBA. In each reflow heating test, a specialized electromagnetic property measurement device is used to capture the electromagnetic response of the PCBA. For example, the magnetic field strength generated by the PCBA during heating and its electromagnetic radiation at different frequencies can be measured. These initial electromagnetic property data provide a preliminary understanding of the electromagnetic behavior of the PCBA under heating conditions, such as which portions of the electromagnetic radiation are enhanced and which portions of the magnetic field strength vary significantly. And carrying out scale normalization processing on the initial electromagnetic characteristic data to ensure the consistency and comparability of the data. Scale normalization involves converting electromagnetic data into a standard scale so that data from different tests can be effectively compared. For example, all magnetic field strength data may be converted to a percentage relative to a certain standard reference value, or the frequency distribution data may be adjusted to an offset relative to a certain reference frequency.

In a specific embodiment, the process of executing step 103 may specifically include the following steps:

(1) Estimating a drift coefficient of target temperature distribution data of each reflux heating test based on a first Wiener process model to obtain a temperature drift coefficient, and estimating a diffusion coefficient of the target temperature distribution data of each reflux heating test to obtain a temperature diffusion coefficient;

(2) Carrying out conditional probability calculation of a Wiener process on target temperature distribution data of each reflow heating test according to the temperature drift coefficient and the temperature diffusion coefficient to obtain temperature conditional probability data;

(3) Extracting temperature change characteristics of target temperature distribution data of each reflux heating test according to the temperature condition probability data to obtain a temperature change characteristic set;

(4) Estimating a drift coefficient of the target electromagnetic characteristic data of each reflow heating test based on the second Wiener process model to obtain an electromagnetic drift coefficient, and estimating a diffusion coefficient of the target electromagnetic characteristic data of each reflow heating test to obtain an electromagnetic diffusion coefficient;

(5) Carrying out conditional probability calculation of a Wiener process on the target electromagnetic characteristic data of each reflow heating test according to the electromagnetic drift coefficient and the electromagnetic diffusion coefficient to obtain electromagnetic conditional probability data;

(6) And carrying out electromagnetic change feature extraction on the target electromagnetic characteristic data of each reflux heating test according to the electromagnetic condition probability data to obtain an electromagnetic change feature set.

Specifically, based on the estimation of the drift coefficient and the diffusion coefficient of the target temperature distribution data in each reflow heating test by the first Wiener process model, the coefficients are key parameters in the Wiener process model and are used for describing the dynamic behavior of variables in a random process. The estimation of the drift coefficient is to determine the average trend of the temperature change over time. For example, if a certain heating test shows that the temperature in the central area of the circuit board has a steady rising trend over time, the drift coefficient of the area is a positive value, reflecting this continuous rising trend of temperature. Conversely, if the temperature trend is decreasing, the drift coefficient is negative. Calculation of the drift coefficient requires analysis of temperature data at different time points and extraction of such average trend of variation therefrom. The estimation of the diffusion coefficient is concerned with the randomness of the temperature variation, or the degree to which the temperature deviates from its average trend. This can be seen as the "intensity" of the random fluctuations in temperature. The diffusion coefficient reflects the fluctuation of the temperature data over time. For example, if during a test a region of the circuit board has a large temperature fluctuation during heating, this indicates that the diffusion coefficient of that region is high, indicating a large randomness. And (3) carrying out conditional probability calculation of a Wiener process on target temperature distribution data of each reflow heating test according to the drift coefficient and the diffusion coefficient, and predicting the probability that the temperature reaches a certain specific value at a certain moment in the future under the given current temperature and time conditions. For example, if a particular region of a circuit board is rapidly raised in temperature during an initial stage of a heating test, the Wiener process model may help predict the probability that this region will reach a particular temperature threshold at some point in the future. And extracting temperature change characteristics according to the temperature condition probability data. A critical pattern of variation is identified from the complex temperature profile data to form a set comprising temperature variation characteristics. These features help to provide a deeper understanding of the thermal behavior of the PCBA under different heating conditions, thereby providing important information for optimizing the soldering process. Also, estimation of drift coefficient and diffusion coefficient is performed on the target electromagnetic characteristic data in each reflow heating test. The electromagnetic drift coefficient reflects the overall trend of the magnetic field strength or electromagnetic radiation level during heating, while the diffusion coefficient describes the random fluctuations of these electromagnetic properties over time. Based on the drift coefficient and the diffusion coefficient of the electromagnetic characteristic, the conditional probability calculation of the Wiener process is carried out, and the probability that the electromagnetic characteristic reaches a specific value at a certain moment in the future is predicted. For example, the probability that the electromagnetic radiation level reaches or exceeds a certain threshold at a certain stage of the heating process may be predicted. And carrying out electromagnetic change feature extraction on the target electromagnetic characteristic data according to the electromagnetic condition probability data. And extracting a key change mode from the electromagnetic characteristic data to form an electromagnetic change characteristic set. These features reflect the electromagnetic behavior of the PCBA during heating, and help understand the impact of heating on PCBA function.

In a specific embodiment, the process of executing step 104 may specifically include the following steps:

(1) Calculating the mean value and the standard deviation of the temperature change characteristic set to obtain the mean value and the standard deviation of the temperature change characteristic set, and calculating the mean value and the standard deviation of the electromagnetic change characteristic set to obtain the mean value and the standard deviation of the electromagnetic change characteristic set;

(2) Carrying out pearson correlation coefficient calculation on the temperature change characteristic mean value, the temperature change characteristic standard deviation, the electromagnetic change characteristic mean value and the electromagnetic change characteristic standard deviation to obtain a characteristic correlation coefficient;

(3) Carrying out normalized feature mapping and vector conversion on the temperature change feature set to obtain a temperature change feature vector, and carrying out normalized feature mapping and vector conversion on the electromagnetic change feature set to obtain an electromagnetic change feature vector;

(4) And carrying out feature vector fusion on the temperature change feature vector and the electromagnetic change feature vector according to the feature correlation coefficient to obtain a fusion change feature vector.

Specifically, the average value and the standard deviation of the temperature change feature set are calculated, and the central trend and the discrete degree quantitative description of the temperature data are obtained. Such statistical analysis helps understand the distribution characteristics of temperature on the PCBA, such as whether certain areas are more likely to heat or cool than others. And similarly, calculating the mean value and the standard deviation of the electromagnetic change characteristic set to obtain the mean value and the standard deviation of the electromagnetic change characteristic, and quantitatively describing the overall behavior and the change range of the electromagnetic characteristic in the heating process. And calculating the pearson correlation coefficient of the mean value and the standard deviation of the temperature change characteristic and the mean value and the standard deviation of the electromagnetic change characteristic. The pearson correlation coefficient is a statistic that measures the strength of a linear relationship between two sets of data, and its value varies between-1 and 1. Calculation of the correlation coefficient may reveal whether there is some linear correlation between the temperature change and the electromagnetic property. For example, if the correlation coefficient is close to 1 or-1, this indicates that the change in temperature is closely related to the change in electromagnetic characteristics under a specific heating condition; and if the correlation coefficient is close to 0, it indicates that there is no significant linear relationship between the two. And (3) carrying out normalized feature mapping and vector conversion on the temperature change feature set, and converting the statistical data into a format which is more suitable for machine learning and other advanced analysis technologies. Normalization refers to converting data into a common scale, typically in the range of 0 to 1, to eliminate the effects of different scale levels. For example, converting temperature change data for different areas of the PCBA under different test conditions into a standardized range of values makes it easier to compare and analyze the data. Similarly, the electromagnetic change feature set is subjected to similar normalization and vector conversion processing, so that a set of numerical vectors capable of representing electromagnetic characteristic changes is obtained. And carrying out feature vector fusion on the temperature change feature vector and the electromagnetic change feature vector according to the feature correlation coefficient. Based on the degree of correlation between temperature and electromagnetic properties, the two sets of vectors are combined into a comprehensive fusion variation feature vector. For example, if there is a strong correlation between temperature and electromagnetic properties, the two sets of vectors may be combined by a weighting method to form a composite eigenvector that contains both temperature change information and electromagnetic property information.

In a specific embodiment, the process of executing step 105 may specifically include the following steps:

(1) Inputting the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector into a preset PCBA welding fault detection model, wherein the PCBA welding fault detection model comprises a first GRU network, a second GRU network, a third GRU network and an output layer;

(2) Extracting hidden features of the temperature change feature vector through bidirectional GRU units in the first GRU network to obtain a temperature hidden feature vector, and detecting PCBA welding faults of the temperature hidden feature vector through unidirectional GRU units in the first GRU network to obtain a first welding fault detection result of temperature influence factors;

(3) Extracting hidden features of the electromagnetic change feature vector through a bidirectional GRU unit in the second GRU network to obtain an electromagnetic hidden feature vector, and detecting PCBA welding faults of the electromagnetic hidden feature vector through a unidirectional GRU unit in the second GRU network to obtain a second welding fault detection result of electromagnetic influence factors;

(4) Extracting hidden features of the fusion change feature vector through a bidirectional GRU unit in a third GRU network to obtain a fusion hidden feature vector, and detecting PCBA welding faults of the fusion hidden feature vector through a unidirectional GRU unit in the third GRU network to obtain a third welding fault detection result of comprehensive influence factors;

(5) And fusing the first welding fault detection result of the temperature influence factor, the second welding fault detection result of the electromagnetic influence factor and the third welding fault detection result of the comprehensive influence factor by the output layer to obtain the PCBA welding fault detection result.

Specifically, a temperature change feature vector is input to a first GRU network. The network comprises bi-directional GRU units, which function to extract hidden features from time series data. For example, when a PCBA is heat tested, variations in temperature at different points in time can suggest potential problems in the soldering process. The bidirectional GRU unit can capture dynamic characteristics of temperature change from past and future time points, so as to generate a temperature hiding characteristic vector containing deep information of temperature change along with time. The unidirectional GRU units in the network use these hidden features to perform the actual fault detection, generating welding fault detection results for temperature influencing factors. The electromagnetic change feature vector is input to a second GRU network. This network is similar in structure to the first network, and includes both bi-directional and uni-directional GRU units. The bidirectional GRU unit processes the electromagnetic change feature vector, extracts the hidden feature of the electromagnetic characteristic changing along with time, and the change of the electromagnetic characteristic is closely related to welding quality. For example, an abnormal increase in electromagnetic radiation indicates an undesirable current or hot spot generated during the welding process. The extracted electromagnetic concealed feature vector is then used to perform welding fault detection, generating welding fault detection results of electromagnetic influencing factors. And inputting the fusion change feature vector into a third GRU network. The network also uses bi-directional GRU units to extract hidden features in the fused feature vector that combine information of temperature and electromagnetic data. For example, if the temperature of a region increases significantly during heating, with an accompanying increase in electromagnetic radiation, this indicates that the region has welding defects. By analyzing the fused hidden features, the unidirectional GRU unit can more comprehensively detect potential welding faults and generate welding fault detection results of comprehensive influence factors. And integrating welding fault detection results generated by the three GRU networks by an output layer of the model. The output layer is used for fusing information from different sources to generate a final PCBA welding fault detection result. For example, if the temperature-influencing factor detection and the electromagnetic-influencing factor detection both indicate potential problems in the same area, the output layer may combine these information to provide a more deterministic weld failure signal.

In a specific embodiment, the process of executing step 106 may specifically include the following steps:

(1) Determining a plurality of optimization targets of a first reflow heating curve according to the PCBA welding fault detection result;

(2) Acquiring a preset double-layer programming solution algorithm, wherein the double-layer programming solution algorithm comprises a first-layer programming solution algorithm and a second-layer programming solution algorithm, the first-layer programming solution algorithm is a plurality of genetic algorithms, and the second-layer programming solution algorithm is a linear programming algorithm;

(3) Defining constraint conditions of each genetic algorithm in the double-layer planning and solving algorithm according to a plurality of optimization targets, and performing curve optimization on the first reflow heating curve through a plurality of genetic algorithms based on the constraint conditions to obtain curve optimization parameters corresponding to each genetic algorithm;

(4) And linearly programming curve optimization parameters corresponding to each genetic algorithm through a linear programming algorithm in the second-layer programming solving algorithm, and outputting a second reflow heating curve of the target PCBA.

Specifically, information is extracted from the PCBA welding fault detection result, and an optimization target of the first reflow heating curve is determined. These objectives include reducing solder joint defects, optimizing heating rates to reduce thermal stresses, or adjusting the temperature profile in specific areas to avoid damage to sensitive components. For example, if the fault detection results show that some areas are overheated, resulting in damage to the electronic components, one of the optimization objectives is to lower the maximum temperature limits of these areas. And acquiring a preset double-layer planning solution algorithm, wherein the algorithm structure comprises two layers of planning solutions. The first layer of programming solution algorithm uses a plurality of genetic algorithms, while the second layer uses a linear programming algorithm. Genetic algorithms are heuristic search algorithms that solve optimization problems by modeling natural choices and genetic mechanisms in biological evolution. For example, in adjusting the heating profile, each genetic algorithm may explore different heating strategies, such as changing the heating rate or adjusting the temperature at a particular stage. Constraint conditions are defined for each genetic algorithm according to the optimization objective. These constraints guide the algorithm in what factors are considered during the search process, ensuring that the solution meets both the weld quality requirements and does not cause damage to the PCBA. For example, if one of the optimization objectives is to reduce the temperature in a region, the corresponding genetic algorithm will look for a cooling strategy in that region. Each genetic algorithm optimizes the first reflow heating profile according to these constraints, generating a series of profile optimization parameters. These parameters represent heating conditions in different strategies, such as temperature change, heating and cooling rates, etc. The curve optimization parameters are input into the linear programming algorithm of the second layer. Linear programming is a method to find the optimal solution given the linear constraints. From which the final second reflow heating profile is determined by comprehensively considering the optimization parameters from the various genetic algorithms. For example, the linear programming algorithm may analyze the different heating strategies proposed by each genetic algorithm to select a heating curve that meets all welding quality requirements while minimizing failure probability.

The foregoing describes a method for testing a PCBA in an embodiment of the present application, and the following describes a device for testing a PCBA in an embodiment of the present application, referring to fig. 2, one embodiment of the device for testing a PCBA in an embodiment of the present application includes:

an obtaining module 201, configured to obtain material information of a target PCBA and thermal sensitivity of an electronic component, and create a first reflow heating curve of the target PCBA according to the material information and the thermal sensitivity;

the test module 202 is configured to perform multiple reflow heating tests and temperature distribution detection on the target PCBA according to the first reflow heating curve, obtain target temperature distribution data of each reflow heating test, and perform electromagnetic characteristic measurement on the target PCBA, so as to obtain target electromagnetic characteristic data of each reflow heating test;

the analysis module 203 is configured to perform a temperature change feature analysis on the target temperature distribution data of each reflow heating test based on the first Wiener process model to obtain a temperature change feature set, and perform an electromagnetic feature change analysis on the target electromagnetic feature data of each reflow heating test based on the second Wiener process model to obtain an electromagnetic change feature set;

The conversion module 204 is configured to perform correlation coefficient analysis on the temperature change feature set and the electromagnetic change feature set to obtain feature correlation coefficients, and perform vector conversion and feature fusion on the temperature change feature set and the electromagnetic change feature set according to the feature correlation coefficients to obtain a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector;

the detection module 205 is configured to input the temperature variation feature vector, the electromagnetic variation feature vector, and the fusion variation feature vector into a preset PCBA welding fault detection model to perform PCBA welding fault detection, so as to obtain a PCBA welding fault detection result;

and the optimizing module 206 is configured to perform curve optimization on the first reflow heating curve according to the PCBA welding fault detection result by using a preset double-layer planning solution algorithm, so as to obtain a second reflow heating curve of the target PCBA.

Through the cooperation of the components, the material information of the PCBA and the heat sensitivity of the electronic component are analyzed, and the reflow soldering temperature point and the temperature range are predicted, so that the potential soldering quality problem can be identified in advance. And target temperature distribution data and electromagnetic characteristic data are obtained through repeated reflow heating tests and temperature distribution detection, so that accurate data support is provided for the prediction and control of welding quality. The temperature distribution data is analyzed by using the first Wiener process model, so that drift and diffusion characteristics of temperature change can be captured, and deeper temperature information is provided for subsequent fault detection. And the electromagnetic characteristic data is analyzed based on the second Wiener process model, so that the drift and diffusion characteristics of electromagnetic change can be obtained, and the understanding of the electromagnetic characteristic change is improved. By carrying out correlation coefficient analysis on the temperature change characteristic set and the electromagnetic change characteristic set, the association degree between the temperature change characteristic set and the electromagnetic change characteristic set can be quantified, and the understanding of the relation between different characteristics is improved. And vector conversion and feature fusion are carried out to obtain a temperature change feature vector, an electromagnetic change feature vector and a fusion change feature vector, so that the influence of a plurality of features on welding quality is comprehensively considered, and the accuracy of fault detection is improved. And inputting the extracted feature vector into the model by using a preset PCBA welding fault detection model, so as to realize comprehensive detection of PCBA welding quality. By combining the detection results of the temperature influence factors, the electromagnetic influence factors and the comprehensive influence factors, the comprehensive judging capability of welding faults is improved. And the first reflow heating curve is subjected to curve optimization through a double-layer planning algorithm, so that the fine control of the welding process is realized, and the welding quality and the production efficiency are improved. Multiple optimization targets, such as improvement of welding quality, reduction of energy consumption and the like, can be considered simultaneously, so that multi-target optimization is realized. By comprehensively analyzing and optimizing the welding process, the time of trial and error and adjustment is reduced, and the production efficiency is improved. By optimizing the heating curve, the resources can be more effectively utilized, the energy consumption is reduced, the production cost is reduced, the self-adaptive optimization of the reflow heating of the PCBA is realized, and the welding test accuracy of the PCBA is improved.

The present application also provides a computer device, including a memory and a processor, where the memory stores computer readable instructions that, when executed by the processor, cause the processor to perform the steps of the method for testing PCBA in the foregoing embodiments.

The present application also provides a computer readable storage medium, which may be a non-volatile computer readable storage medium, or may be a volatile computer readable storage medium, where instructions are stored in the computer readable storage medium, when the instructions are executed on a computer, cause the computer to perform the steps of the method for testing a PCBA.

It will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, systems and units may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein.

The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.

The above embodiments are merely for illustrating the technical solution of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the corresponding technical solutions.

Claims

1. The test method of the PCBA is characterized by comprising the following steps of:

2. The method of testing a PCBA of claim 1, wherein the obtaining the material information of the target PCBA and the thermal sensitivity of the electronic component, and creating the first reflow heating profile of the target PCBA based on the material information and the thermal sensitivity, comprises:

acquiring material information of a target PCBA, wherein the material information comprises materials of a circuit board and materials of an electronic component mounted on the circuit board;

Performing hierarchical structure analysis on the target PCBA to obtain target hierarchical structure information, and performing heat sensitivity analysis on the electronic component according to the material of the electronic component to obtain heat sensitivity of the electronic component;

carrying out reflow soldering temperature point prediction on the target PCBA according to the target hierarchical structure information and the heat sensitivity of the electronic component to obtain reflow soldering temperature point data;

carrying out reflow soldering temperature range prediction on the target PCBA according to the reflow soldering temperature point data to obtain reflow soldering temperature range data, and carrying out reflow soldering time prediction on the target PCBA to obtain reflow soldering duration data;

and creating a first reflow heating curve of the target PCBA according to the reflow soldering temperature range data and the reflow soldering duration data.

3. The method for testing a PCBA according to claim 1, wherein the performing multiple reflow heating tests and temperature distribution detection on the target PCBA according to the first reflow heating curve to obtain target temperature distribution data of each reflow heating test, and performing electromagnetic property measurement on the target PCBA to obtain target electromagnetic property data of each reflow heating test, includes:

Carrying out multiple reflux heating tests on the target PCBA according to the first reflux heating curve, and carrying out temperature distribution detection on the target PCBA through a preset infrared temperature sensor to obtain initial temperature distribution data of each reflux heating test;

respectively carrying out data cleaning and data standardization treatment on the initial temperature distribution data of each reflux heating test to obtain target temperature distribution data of each reflux heating test;

performing electromagnetic characteristic measurement on the target PCBA to obtain initial electromagnetic characteristic data of each reflux heating test;

respectively carrying out scale normalization processing on initial electromagnetic characteristic data of each reflow heating test to obtain target electromagnetic characteristic data of each reflow heating test, wherein the target electromagnetic characteristic data comprises: magnetic field strength data and frequency distribution data.

4. The method for testing the PCBA according to claim 1, wherein the performing a temperature change feature analysis on the target temperature distribution data of each reflow heating test based on the first Wiener process model to obtain a temperature change feature set, and performing an electromagnetic feature change analysis on the target electromagnetic feature data of each reflow heating test based on the second Wiener process model to obtain an electromagnetic change feature set includes:

Estimating a drift coefficient of target temperature distribution data of each reflux heating test based on a first Wiener process model to obtain a temperature drift coefficient, and estimating a diffusion coefficient of the target temperature distribution data of each reflux heating test to obtain a temperature diffusion coefficient;

carrying out conditional probability calculation of a Wiener process on target temperature distribution data of each reflux heating test according to the temperature drift coefficient and the temperature diffusion coefficient to obtain temperature conditional probability data;

extracting temperature change characteristics of target temperature distribution data of each reflux heating test according to the temperature condition probability data to obtain a temperature change characteristic set;

estimating a drift coefficient of the target electromagnetic characteristic data of each reflow heating test based on the second Wiener process model to obtain an electromagnetic drift coefficient, and estimating a diffusion coefficient of the target electromagnetic characteristic data of each reflow heating test to obtain an electromagnetic diffusion coefficient;

carrying out conditional probability calculation of a Wiener process on the target electromagnetic characteristic data of each reflux heating test according to the electromagnetic drift coefficient and the electromagnetic diffusion coefficient to obtain electromagnetic conditional probability data;

And carrying out electromagnetic change feature extraction on the target electromagnetic characteristic data of each reflux heating test according to the electromagnetic condition probability data to obtain an electromagnetic change feature set.

5. The method for testing PCBA according to claim 1, wherein the performing correlation coefficient analysis on the temperature change feature set and the electromagnetic change feature set to obtain feature correlation coefficients, and performing vector conversion and feature fusion on the temperature change feature set and the electromagnetic change feature set according to the feature correlation coefficients to obtain a temperature change feature vector, an electromagnetic change feature vector, and a fusion change feature vector, includes:

calculating the mean value and the standard deviation of the temperature change characteristic set to obtain a temperature change characteristic mean value and a temperature change characteristic standard deviation, and calculating the mean value and the standard deviation of the electromagnetic change characteristic set to obtain an electromagnetic change characteristic mean value and an electromagnetic change characteristic standard deviation;

carrying out pearson correlation coefficient calculation on the temperature change characteristic mean value, the temperature change characteristic standard deviation, the electromagnetic change characteristic mean value and the electromagnetic change characteristic standard deviation to obtain a characteristic correlation coefficient;

Carrying out normalized feature mapping and vector conversion on the temperature change feature set to obtain a temperature change feature vector, and carrying out normalized feature mapping and vector conversion on the electromagnetic change feature set to obtain an electromagnetic change feature vector;

and carrying out feature vector fusion on the temperature change feature vector and the electromagnetic change feature vector according to the feature correlation coefficient to obtain a fusion change feature vector.

6. The method for testing a PCBA according to claim 1, wherein inputting the temperature change feature vector, the electromagnetic change feature vector, and the fusion change feature vector into a preset PCBA welding fault detection model to perform PCBA welding fault detection, to obtain a PCBA welding fault detection result, comprises:

inputting the temperature change feature vector, the electromagnetic change feature vector and the fusion change feature vector into a preset PCBA welding fault detection model, wherein the PCBA welding fault detection model comprises a first GRU network, a second GRU network, a third GRU network and an output layer;

extracting hidden features of the temperature change feature vector through bidirectional GRU units in the first GRU network to obtain a temperature hidden feature vector, and detecting PCBA welding faults of the temperature hidden feature vector through unidirectional GRU units in the first GRU network to obtain a first welding fault detection result of temperature influence factors;

Extracting hidden features of the electromagnetic change feature vector through a bidirectional GRU unit in the second GRU network to obtain an electromagnetic hidden feature vector, and detecting PCBA welding faults of the electromagnetic hidden feature vector through a unidirectional GRU unit in the second GRU network to obtain a second welding fault detection result of electromagnetic influence factors;

extracting hidden features of the fusion change feature vector through a bidirectional GRU unit in the third GRU network to obtain a fusion hidden feature vector, and detecting PCBA welding faults of the fusion hidden feature vector through a unidirectional GRU unit in the third GRU network to obtain a third welding fault detection result of comprehensive influence factors;

and fusing the first welding fault detection result of the temperature influence factor, the second welding fault detection result of the electromagnetic influence factor and the third welding fault detection result of the comprehensive influence factor by the output layer to obtain a PCBA welding fault detection result.

7. The method for testing a PCBA according to claim 1, wherein the performing curve optimization on the first reflow heating curve according to the PCBA welding fault detection result by a preset double-layer programming solution algorithm to obtain a second reflow heating curve of the target PCBA includes:

Determining a plurality of optimization targets of the first reflow heating curve according to the PCBA welding fault detection result;

acquiring a preset double-layer programming solution algorithm, wherein the double-layer programming solution algorithm comprises a first-layer programming solution algorithm and a second-layer programming solution algorithm, the first-layer programming solution algorithm is a plurality of genetic algorithms, and the second-layer programming solution algorithm is a linear programming algorithm;

defining constraint conditions of each genetic algorithm in a double-layer planning and solving algorithm according to the plurality of optimization targets, and performing curve optimization on the first reflow heating curve through the plurality of genetic algorithms based on the constraint conditions to obtain curve optimization parameters corresponding to each genetic algorithm;

and linearly programming curve optimization parameters corresponding to each genetic algorithm through a linear programming algorithm in the second-layer programming solving algorithm, and outputting a second reflux heating curve of the target PCBA.

8. A testing arrangement of PCBA, characterized in that the testing arrangement of PCBA includes:

9. A computer device, the computer device comprising: a memory and at least one processor, the memory having instructions stored therein;

the at least one processor invoking the instructions in the memory to cause the computer device to perform the method of testing a PCBA as recited in any one of claims 1-7.

10. A computer readable storage medium having instructions stored thereon, which when executed by a processor, implement a method of testing a PCBA according to any one of claims 1 to 7.