Automatic fault prediction method and system for PCBA board

Abstract

The invention discloses an automatic fault prediction method and system for a PCBA board, and relates to the technical field of fault prediction, and the automatic fault prediction method comprises the steps: carrying out the electrical test of the PCBA board through an electrical test module, recording test data, and carrying out the analysis and processing of the test data through a data analysis module; pHM feature vectors are constructed based on test data through a PHM analysis module, a prediction model is input, the prediction model predicts the future fault probability of the PCBA board based on historical data, if the output future fault probability of the PCBA board is larger than a set threshold value, the test board gives an alarm and waits for manual taking out for processing, and otherwise, the PCBA board continues to be tested; by adopting the method, the future fault probability can be predicted by testing and analyzing the PCBA board, and more potential fault products are prevented from flowing into the market.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of failure prediction, in particular to an automatic failure prediction method and system for a PCBA board. BACKGROUND

[0002] In the field of intelligent wear and health devices, the massage waist support is an integrated product that combines precise mechanical control, multi-modal sensor feedback, and complex circuit driving. The performance stability of its core PCBA board directly determines the accuracy of user experience and the safety of device operation. As consumers increasingly demand control accuracy, low-noise operation characteristics, and long-term reliability of massage intensity, PCBA failure prediction technology has become a core link in the quality control of massage waist support products.

[0003] However, existing PCBA failure identification methods still rely heavily on manual operation mode, which has significant technical bottlenecks. On the one hand, massage waist support PCBA generally uses customized flexible printed circuit boards (FPC) and ultra-miniature components (such as 0.4mm pitch surface mount pressure sensors and 0.3mm line width flexible bending area circuits). Manual visual inspection cannot reliably identify hidden defects such as FPC micro-cracks and miniature component pin virtual soldering. Additionally, the flexible material is prone to hidden damage during manual handling due to improper handling. SUMMARY

[0004] The present application provides an automatic failure prediction method and system for a PCBA board to address the problems of existing technology.

[0005] To achieve the above-mentioned purposes, the technical solutions adopted by the present application are as follows: An automatic failure prediction method for a PCBA board, comprising: Step one: Place the PCBA board to be tested on the test table through the transportation module. Perform appearance detection on the PCBA board through the image detection module. If the detection is qualified, continue testing. If the detection is unqualified, the test table alarms and waits for manual removal for processing. Step two: Perform electrical testing on the PCBA board through the electrical testing module. Record the test data and analyze and process it through the data analysis module. Construct a PHM feature vector based on the test data through the PHM analysis module. Input the prediction model. The prediction model predicts the future failure probability of the PCBA board based on historical data. If the output future failure probability of the PCBA board is greater than the set threshold, the test table alarms and waits for manual removal for processing. Otherwise, the PCBA board continues testing. Step three: temperature rise test is performed on the PCBA board. In a preset time, the PCBA board is heated to a preset temperature under a given rated power voltage. In this process, the temperature rise voltage data of each circuit on the PCBA board is continuously monitored. If there is voltage abnormality on the PCBA board, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing. Otherwise, it is determined that the PCBA board is a qualified product.

[0006] Preferably, step one acquires images of the PCBA board through the image detection module, outputs a test image group, compares the test image group with the qualified image group and the unqualified image group in the database, and if there is an unqualified image in the test image group, the unqualified image is labeled as unqualified, and it is determined that the PCBA board is unqualified. If all the images in the PCBA board test image group are qualified images, the PCBA board continues to be tested. If the images in the PCBA board test image group do not match the images in the qualified image group or the unqualified image group, it is determined that the PCBA board is manually reviewed. If the manual review result is unqualified, it is determined that the PCBA board is unqualified. If the manual review result is qualified, it is determined that the PCBA board is qualified, and the PCBA board continues to be tested.

[0007] Preferably, the test image group includes an overall image of the PCBA board, an element image of all elements on the PCBA board, and a solder point image of all solder points on the PCBA board.

[0008] Preferably, when the PCBA board is manually reviewed, if the manual review result is qualified, all images in the test image group corresponding to the PCBA board are stored in the qualified image group of the database. If the manual review result is unqualified, the unqualified images in the test image group corresponding to the PCBA board are stored in the unqualified image group of the database.

[0009] Preferably, the data analysis module in step two stores standard voltage data tables of different models of PCBA boards. The standard voltage data table includes a qualified parameter range of the voltage of each circuit on the corresponding model PCBA board. If the test data of the circuit voltage does not belong to the qualified parameter range of the corresponding circuit on the standard voltage data table, it is determined that the circuit voltage is abnormal. Otherwise, it is determined that the circuit voltage is normal.

[0010] Preferably, step two includes: The pin communication test is performed on the PCBA board by the electrical test module to test whether the communication of each component is good. If the test is qualified, the test continues. If the test is unqualified, the test bench alarms, and waits for manual removal for processing.

[0011] Preferably, the second step comprises: performing a normal voltage test on the PCBA board by the electrical test module, testing normal voltage data of each circuit under a given rated power supply voltage, and sending the normal voltage data to the data analysis module for analysis and processing; The data analysis module performs a preliminary analysis on the normal voltage data, and if the data analysis module preliminarily analyzes that there is a voltage anomaly on the PCBA board, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing; otherwise, the normal voltage data of the PCBA board is subjected to a one-time future failure probability prediction; A PHM feature vector is constructed based on the normal voltage data of the PCBA board, and input into a prediction model, and the prediction model predicts the one-time future failure probability of the PCBA board based on historical data; The set threshold of the future failure probability includes a first threshold and a second threshold, and the first threshold is greater than the second threshold; If the one-time future failure probability of the PCBA board is not greater than the second threshold, the PCBA board is subjected to a temperature rise test; If the one-time future failure probability of the PCBA board is greater than the first threshold, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing; If the one-time future failure probability of the PCBA board is not greater than the first threshold and greater than the second threshold, the PCBA board is subjected to a deep analysis; The PCBA board subjected to the deep analysis is subjected to an irregular electrical test, testing irregular voltage data of each circuit under a given over-limit power supply voltage, and sending the irregular voltage data to a PHM analysis module for a two-time future failure probability prediction; A PHM feature vector is constructed based on the irregular voltage data of the PCBA board, and input into a prediction model, and the prediction model predicts the two-time future failure probability of the PCBA board based on historical data; If the two-time future failure probability of the PCBA board is not greater than the first threshold, the PCBA board continues to be detected; otherwise, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing.

[0012] Preferably, if the two-time future failure probability of the PCBA board is not greater than the first threshold, the PCBA board is subjected to a two-time appearance detection; The PCBA board is subjected to a two-time appearance detection, and if the detection is unqualified, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing, otherwise, the PCBA board is marked, and the PCBA board is subjected to a temperature rise test.

[0013] Preferably, if the PCBA board is determined to be qualified after the temperature rise test, and the PCBA board is not marked, the PCBA board is determined to be a first-class product among the qualified products; if the PCBA board is determined to be qualified after the temperature rise test, and the PCBA board is marked, the PCBA board is determined to be a second-class product among the qualified products.

[0014] An automatic fault prediction system of a PCBA board, which is matched with the automatic fault prediction method described above, the automatic fault prediction system comprises: A test table, which can be integrated with a central control display screen, an alarm unit, a control system and an output power supply unit; the central control display screen is used for convenient viewing and control by relevant personnel, the alarm unit is used for alarm prompt; the control system is used for setting and adjusting electrical test parameters; the output power supply unit is used to provide the power voltage required for PCBA board testing; A transportation module, which comprises a mechanical arm and a positioning unit, the mechanical arm is used to transport the PCBA board to the test table, and the positioning unit is used to position the PCBA board; An image detection module, which is used for image detection of the PCBA board; A database, which plays a role of storing records; An electrical test module, which is used for electrical test of the PCBA board; A data analysis module, which is used for analysis and processing of electrical test data; A PHM analysis module, which is used for future fault probability prediction; A temperature rise module, which is used for temperature rise and temperature measurement of the PCBA board.

[0015] One or more technical solutions provided in the application have at least the following technical effects or advantages: The electrical test module performs regular voltage test (based on standard voltage table matched with model) and over-limit voltage test in stages, synchronously extracts dynamic characteristics such as voltage mean value and variance by the PHM analysis module, constructs a characteristic vector to input an XGBoost survival analysis model, and predicts future fault probability. Finally, the stability is verified through the temperature rise test, the qualified products are divided into first-class products (directly qualified) and second-class products (qualified after marking) according to the prediction result, the long-term reliability of the PCBA board is predicted, and many potential fault products are avoided to flow into the market. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 It is a flowchart of the automatic fault prediction method of the PCBA board of the application; Figure 2 It is a flowchart of step two of the automatic fault prediction method of the PCBA board of the application. DETAILED DESCRIPTION

[0017] To facilitate understanding of the present invention, a more complete description of this application will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the present invention.

[0018] It should be noted that the terms "vertical," "horizontal," "up," "down," "left," "right," and similar expressions used in this article are for illustrative purposes only and do not represent the only possible implementation.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains; the terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention; the term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0020] Example 1: As Figure 1 As shown, the automated fault prediction method for PCBA boards in this application includes: Step 1: Place the PCBA board to be tested on the test bench using the transport module. The PCBA board is visually inspected using the image detection module. If the inspection is successful, the test continues. If the inspection fails, the test bench will sound an alarm and wait for manual removal and processing.

[0021] Specifically, the image detection module acquires images of the PCBA board and outputs a test image set. This test image set is compared with qualified and unqualified image sets in the database. If any unqualified image exists in the test image set, it is labeled as unqualified, and the PCBA board is deemed unqualified. If all images in the PCBA board's test image set are qualified, the PCBA board continues testing. If the images in the PCBA board's test image set do not match either the qualified or unqualified image sets, the PCBA board undergoes manual review. If the manual review result is unqualified, the PCBA board is deemed unqualified; if the manual review result is qualified, the PCBA board is deemed qualified, and testing continues.

[0022] The test image set includes an overall view of the PCBA board, component views of all components on the PCBA board, and solder joint views of all solder points on the PCBA board.

[0023] The database stores a qualified image group and an unqualified image group, wherein the qualified image group includes an overall qualified image of the PCBA board, an element qualified image of all elements on the PCBA board, and a welding point qualified image of all welding points on the PCBA board; and the unqualified image group includes an overall unqualified image of a plurality of common PCBA boards, an element unqualified image of all elements on the PCBA board, and a welding point unqualified image of all welding points on the PCBA board.

[0024] Meanwhile, the database also stores test data and market feedback data of different models of PCBA boards, wherein the test data at least includes batch number, test date, test result (qualified product quantity, unqualified product quantity) and detailed test data; and the market feedback data at least includes batch number, failure type, failure time and failure cause; and the test data and the market feedback data of the different models of PCBA boards can be updated in real time by manual operation.

[0025] In addition, when the PCBA board is manually reviewed, if the manual review result is qualified, all images in the test image group corresponding to the PCBA board are stored in the qualified image group of the database; and if the manual review result is unqualified, the unqualified image in the test image group corresponding to the PCBA board is stored in the unqualified image group of the database.

[0026] It should be noted that the qualified image group can be a batch of qualified PCBA products screened by manual inspection and microscope inspection by relevant personnel, and then all test images output by appearance test of the batch of qualified products by the image detection module are used as the qualified image group; similarly, the unqualified image group can be a batch of unqualified PCBA products screened by manual inspection and microscope inspection by relevant personnel, and then all test images output by appearance test of the batch of unqualified products by the image detection module are used as the unqualified image group.

[0027] It should be noted that after the image detection module performs image acquisition, image preprocessing is performed, including denoising, enhancement and correction, then the OpenCV tool is used to extract features from the preprocessed image, identify key features (such as shape features, color features, texture features, etc.) in the image, use SIFT algorithm or ORB algorithm to extract descriptors, match the database template (qualified image group, unqualified image group) through FLANN, calculate the matching similarity, and output the matching result (such as matching similarity greater than 80% is considered as matching success).

[0028] Step 2: Perform electrical tests on the PCBA board using the electrical testing module, record the test data, and analyze and process it using the data analysis module. Construct a PHM feature vector based on the test data using the PHM analysis module, input it into the prediction model, and the prediction model predicts the future failure probability of the PCBA board based on historical data. If the output future failure probability of the PCBA board is greater than the set threshold, the test bench will sound an alarm and wait for manual removal and processing. Otherwise, the PCBA board will continue to be tested.

[0029] The data analysis module can store standard voltage data tables for different PCBA board models. The standard voltage data tables include the acceptable parameter ranges for each circuit voltage on the corresponding PCBA board model. If the test data of the circuit voltage (including conventional voltage data and unconventional voltage data) does not fall within the acceptable parameter range for the corresponding circuit on the standard voltage data sheet... If the voltage of the circuit is normal, it is determined that the circuit voltage is abnormal; otherwise, it is determined that the circuit voltage is normal. If there is a circuit with abnormal voltage on the PCBA board, it is determined that there is an abnormal voltage on the PCBA board.

[0030] in, and All are positive numbers.

[0031] It should be noted that the acceptable parameter ranges for each circuit voltage in the standard voltage data sheet are as follows. The voltage can be limited according to the design parameters of the PCBA board, and the standard voltage data sheet can be manually modified.

[0032] Step 3: Perform a temperature rise test on the PCBA board. Under a continuously applied rated power supply voltage for a preset time, the temperature rise module will heat the PCBA board to the preset temperature. During this process, the temperature rise voltage data of each circuit on the PCBA board will be continuously monitored. If there is an abnormal voltage on the PCBA board, the PCBA board will be deemed unqualified, the test bench will sound an alarm, and the board will be manually removed for processing. Otherwise, the PCBA board will be deemed qualified.

[0033] In this embodiment, step two specifically includes: The PCBA board is tested for pin connectivity using an electrical testing module to check the connectivity of each component. If the test is successful, the test continues; if the test fails, the test bench will sound an alarm and the board will need to be manually removed for further processing. The PCBA board is subjected to routine voltage tests through the electrical testing module. Under a given rated power supply voltage, the routine voltage data of each circuit is tested and sent to the data analysis module for analysis and processing. The data analysis module performs preliminary analysis on the normal voltage data. If the data analysis module preliminarily analyzes that there is voltage abnormality on the PCBA board, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing. Otherwise, the normal voltage data of the PCBA board is subjected to a future failure probability prediction. It should be noted that in the normal voltage test, if the normal voltage data of any circuit on the PCBA board exceeds the voltage qualified parameter range of the corresponding circuit on the standard voltage data table, it is determined that there is voltage abnormality on the PCBA board. If the normal voltage data of all circuits on the PCBA board does not exceed the voltage qualified parameter range of the corresponding circuit on the standard voltage data table, it is determined that there is no voltage abnormality on the PCBA board, and the normal voltage data of the PCBA board is subjected to a future failure probability prediction.

[0034] The future failure probability prediction is performed, the PHM feature vector is constructed based on the normal voltage data of the PCBA board, and the prediction model is input. The prediction model predicts the future failure probability of the PCBA board based on historical data. The set threshold of the future failure probability includes a first threshold and a second threshold, and the first threshold is greater than the second threshold. The first threshold and the second threshold can be set by manual, for example, the first threshold is set to 35%, and the second threshold is set to 20%.

[0035] If the future failure probability of the PCBA board is not greater than the second threshold, the PCBA board is subjected to a temperature rise test. If the future failure probability of the PCBA board is greater than the first threshold, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing. If the future failure probability of the PCBA board is not greater than the first threshold and is greater than the second threshold, the PCBA board is subjected to an abnormal electrical property test.

[0036] The screened PCBA board is subjected to an abnormal electrical property test. The abnormal voltage data of each circuit is tested under a given over-limit power supply voltage, and the abnormal voltage data is sent to the PHM analysis module for a secondary future failure probability prediction. The PHM feature vector is constructed based on the abnormal voltage data of the PCBA board, and the prediction model is input. The prediction model predicts the secondary future failure probability of the PCBA board based on historical data (test data and market feedback data of the PCBA board of the same type stored in the database). If the secondary future failure probability of the PCBA board is not greater than the first threshold, the PCBA board continues to be detected. Otherwise, it is determined that the PCBA board is unqualified, the test bench alarms, and waits for manual removal for processing.

[0037] It should be noted that in the future failure probability process in the present fact example, the prediction model can be an ensemble learning model, and the PHM feature vector is constructed, including extracting key indicators from the test data, which at least include voltage mean, voltage variance, transient overshoot, etc. The feature vector is input into the trained prediction model to match similar cases in the historical data. The model calculates the probability of failure of the PCBA board in the future T period according to the matching result. The T period can be within 60 days or within 180 days. The specific length of time can be freely selected according to the demand.

[0038] For example, the prediction model uses the XGBoost survival analysis algorithm (XGBSE) to construct a gradient boosting decision tree based on the Cox proportional hazards model. The static attributes (model, component density) of the historical PCBA board, dynamic electrical characteristics (voltage mean, variance, drift rate) and environmental data are used to learn the non-linear relationship between failure risk and features. The output result is the failure probability of the board within a specified time window (such as 30 days) in the future; the failure probability calculation formula of the prediction model is.

[0039] Among them, the first future failure probability represents the probability of failure of the PCBA board in a period of time in the future under the normal use voltage; the second future failure probability represents the probability of failure of the PCBA board in a period of time in the future under the over-limit use voltage.

[0040] Alternatively, in another embodiment of the present application, if the second future failure probability of the PCBA board is not greater than the first threshold value, the PCBA board is subjected to secondary appearance detection, and if the detection is unqualified, the PCBA board is determined to be unqualified, and the test bench alarms, waiting for manual removal for processing, otherwise, the PCBA board is marked, and the PCBA board is subjected to temperature rise test.

[0041] Among them, the over-limit power supply voltage is a preset voltage, and the over-limit power supply voltage is 1.1 to 1.2 times the rated power supply voltage.

[0042] In addition, if the PCBA board is determined to be qualified after temperature rise test, and the PCBA board is not marked, the PCBA board is determined to be a first-class product among qualified products; if the PCBA board is determined to be qualified after temperature rise test, and the PCBA board is marked, the PCBA board is determined to be a second-class product among qualified products.

[0043] It should be noted that the first-class product and the second-class product are both qualified products, and the quality evaluation of the first-class product is higher than that of the second-class product.

[0044] It should be noted that each circuit voltage on the PCBA board corresponds to a qualified parameter range ; the qualified parameter range corresponding to different circuit voltages and The value of the voltage can be different.

[0045] In addition, the data analysis module can store standard voltage data tables of different models of PCBA boards. Specifically, there can be multiple standard voltage data tables, and each standard voltage data table can correspond to a model of PCBA board. It can be understood that the circuit components on different models of PCBA boards are generally not completely the same, so the circuit design parameters of different models of PCBA boards are different, and therefore the qualified value range of the circuit voltage on different models of PCBA boards is different. Therefore, it is necessary to set the corresponding standard voltage data table according to the model of the PCBA board, and the qualified range of each circuit on the standard voltage data table is selected according to the circuit design parameters. Specifically, if a worker wants to test a PCBA board of model A, the corresponding standard voltage data table of model A needs to be called out in the data analysis module. After selecting the standard voltage data table of model A, the system automatically compares and analyzes the detected normal voltage data or abnormal voltage data with the corresponding value range on the standard voltage data table of model A during the normal voltage test or abnormal voltage test of the PCBA board of model A. If a worker wants to test a PCBA board of model B, the corresponding standard voltage data table of model B needs to be called out in the data analysis module. After selecting the standard voltage data table of model B, the system automatically compares and analyzes the detected normal voltage data or abnormal voltage data with the corresponding value range on the standard voltage data table of model B during the normal voltage test or abnormal voltage test of the PCBA board of model B.

[0046] The circuit design parameters can refer to the recommended operating voltage range in the data manual of the components and the chip specification book, and the like design guidance documents. The qualified parameter range of each circuit on the above standard voltage data table can be the circuit design parameter range set during the design of the PCBA circuit board, or can be the circuit qualified parameter range redefined by relevant technical personnel through production measurement or failure analysis. The parameter setting of the standard voltage data table is a prior art, which will not be described here.

[0047] Embodiment two: The application also provides an automatic fault prediction system for a PCBA board, which is applied to the automatic fault prediction method of embodiment one. The fault prediction system comprises: A test bench, which can be integrated with a central control display screen, an alarm unit, a control system, and an output power supply unit. The central control display screen is used for convenient viewing and control by relevant personnel, and the alarm unit is used for alarm prompt. The control system is used for setting and adjusting electrical test parameters, and the output power supply unit is used to provide the power voltage required for PCBA board testing. The transport module includes a mechanical arm for transporting the PCBA board to the test table and a positioning unit for positioning the PCBA board. The image detection module is configured to detect the PCBA board. The database is configured to store records. The electrical test module is configured to test the electrical properties of the PCBA board. The data analysis module is configured to analyze the electrical test data. The PHM analysis module is configured to predict future failure probabilities. The temperature rise module is configured to heat and measure the temperature of the PCBA board.

[0048] The image detection module can be an AOI optical detector and its components.

[0049] It should be noted that the electrical test module includes a test probe and a data acquisition unit. The test probe is configured to form an electrical connection with the test points on the PCBA board to perform electrical tests. The data acquisition unit records the electrical test data.

[0050] It should be noted that the temperature rise module can be a hot air blower and its components, and the temperature rise module further integrates a temperature sensor to test the temperature of the PCBA board.

[0051] The technical solutions of the present application have at least the following technical effects or advantages: 1. Image detection is used to compare with the pre-stored qualified image group and unqualified image group in the database, combined with manual review, to improve the accuracy of appearance inspection while ensuring test efficiency. 2. The manual review results are fed back to the database (qualified images are stored in the qualified group, and unqualified images are stored in the unqualified group), which continuously optimizes the detection model and solves the misjudgment problem caused by the fixed traditional image library. 3. Two-stage voltage analysis: preliminary analysis for rapid screening of obvious abnormalities, and deep analysis for identifying potential risks through the center qualified range, which can screen products with voltage fluctuations close to the critical value and improve the accuracy of testing. 4. Fault prediction for the PCBA board, quality control of the PCBA board based on the fault prediction results, and reduction of potential fault products flowing into the market. 5. Simulating abnormal working conditions (voltage fluctuations exceeding the rated value) to test products with voltage fluctuations close to the critical value, detecting the reliability of the PCBA board under extreme conditions, and improving the safety and reliability of the test. 6. Secondary appearance detection after unconventional electrical tests to prevent hidden appearance problems such as solder joint cracking and component damage caused by high-voltage tests. 7. Based on the test results, qualified products are further divided into first-class products (no abnormal records) and second-class products (previously triggered non-routine tests but ultimately qualified), providing a more accurate basis for quality grading and facilitating subsequent differentiated processing.

[0052] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automated fault prediction method for PCBA boards, characterized in that, include: Step 1: Place the PCBA board to be tested on the test bench using the transport module. The PCBA board is visually inspected using the image detection module. If the inspection is successful, the test continues. If the inspection fails, the test bench will sound an alarm and wait for manual removal and processing. Step 2: Perform electrical tests on the PCBA board using the electrical testing module, record the test data, and analyze and process the data using the data analysis module; The PHM analysis module constructs PHM feature vectors based on test data and inputs them into the prediction model. The prediction model predicts the future failure probability of the PCBA board based on historical data. If the output future failure probability of the PCBA board is greater than the set threshold, the test bench will sound an alarm and wait for manual removal and processing. Otherwise, the PCBA board will continue to be tested. Step 3: Perform a temperature rise test on the PCBA board. Under a continuously applied rated power supply voltage for a preset time, the temperature rise module will heat the PCBA board to the preset temperature. During this process, the temperature rise voltage data of each circuit on the PCBA board will be continuously monitored. If there is an abnormal voltage on the PCBA board, the PCBA board will be deemed unqualified, the test bench will sound an alarm, and the board will be manually removed for processing. Otherwise, the PCBA board will be deemed qualified.

2. The automated fault prediction method for PCBA boards as described in claim 1, characterized in that, Step one involves acquiring images of the PCBA board using an image detection module, outputting a test image set, and comparing the test image set with qualified and unqualified image sets in the database. If an unqualified image is found in the test image set, it is labeled as unqualified, and the PCBA board is deemed unqualified. If all images in the PCBA board's test image set are qualified, the PCBA board continues testing. If the images in the PCBA board's test image set do not match either the qualified or unqualified image sets, the PCBA board undergoes manual review. If the manual review result is unqualified, the PCBA board is deemed unqualified; if the manual review result is qualified, the PCBA board is deemed qualified, and testing continues.

3. The automated fault prediction method for PCBA boards as described in claim 2, characterized in that, The test image set includes an overall view of the PCBA board, component views of all components on the PCBA board, and solder joint views of all solder points on the PCBA board.

4. The automated fault prediction method for PCBA boards as described in claim 2, characterized in that, When a PCBA board is manually reviewed, if the review result is qualified, all images in the test image group corresponding to the PCBA board are stored in the qualified image group in the database; if the review result is unqualified, the unqualified images in the test image group corresponding to the PCBA board are stored in the unqualified image group in the database.

5. The automated fault prediction method for PCBA boards as described in claim 1, characterized in that, The data analysis module in step two stores standard voltage data tables for different types of PCBA boards. The standard voltage data sheet includes the acceptable parameter range for the voltage of each circuit on the corresponding PCBA board. If the test data of the circuit voltage does not fall within the acceptable parameter range of the corresponding circuit on the standard voltage data sheet, the circuit voltage is determined to be abnormal. Conversely, if the voltage is normal, the circuit is considered to be functioning correctly.

6. The automated fault prediction method for PCBA boards as described in claim 5, characterized in that, Step two includes: The PCBA board is tested for pin connectivity using an electrical testing module to check the connectivity of each component. If the test is successful, the test continues; if the test fails, the test bench will sound an alarm and the board will need to be manually removed for further processing.

7. The automated fault prediction method for PCBA boards as described in claim 5, characterized in that, Step two includes: performing conventional voltage tests on the PCBA board through an electrical testing module, testing the conventional voltage data of each circuit under a given rated power supply voltage, and sending the conventional voltage data to a data analysis module for analysis and processing; The data analysis module performs a preliminary analysis of the routine voltage data. If the preliminary analysis shows that there is an abnormal voltage on the PCBA board, the PCBA board is determined to be unqualified, the test bench will sound an alarm, and the board will be manually removed for processing. Otherwise, the routine voltage data of the PCBA board will be used to predict the probability of future failures. The PHM feature vector is constructed based on the conventional voltage data of the PCBA board and input into the prediction model. The prediction model predicts the probability of a future failure of the PCBA board based on historical data. The threshold for setting the probability of future failures includes a first threshold and a second threshold, and the first threshold is greater than the second threshold; If the probability of a future failure of the PCBA board is not greater than the second threshold, then the PCBA board shall undergo a temperature rise test. If the probability of a future failure of a PCBA board is greater than the first threshold, the PCBA board is deemed unqualified, the test bench will sound an alarm, and the board will be manually removed for processing. If the probability of a future failure of the PCBA board is not greater than the first threshold and is greater than the second threshold, then the PCBA board shall undergo unconventional electrical testing. The PCBA board undergoes unconventional electrical testing. Under a given over-limit power supply voltage, the unconventional voltage data of each circuit is tested, and the unconventional voltage data is sent to the PHM analysis module for secondary future failure probability prediction. A PHM feature vector is constructed based on unconventional voltage data of PCBA boards and input into a prediction model. The prediction model predicts the probability of secondary future failures of PCBA boards based on historical data. If the probability of a secondary future failure of the PCBA board is not greater than the first threshold, the PCBA board will continue to be tested; otherwise, the PCBA board will be deemed unqualified, the test bench will sound an alarm, and the board will be manually removed for processing.

8. The automated fault prediction method for PCBA boards as described in claim 7, characterized in that, If the probability of secondary future failure of the PCBA board is not greater than the first threshold, then the PCBA board shall undergo secondary appearance inspection. A secondary visual inspection is performed on the PCBA board. If the inspection fails, the PCBA board is deemed unqualified, the test bench will sound an alarm, and the board will be manually removed for processing. Otherwise, the PCBA board will be marked and subjected to a temperature rise test.

9. The automated fault prediction method for PCBA boards as described in claim 7, characterized in that, If a PCBA board passes the temperature rise test and is not marked, it is considered a first-class product among qualified products. If a PCBA board passes the temperature rise test and is marked, it is considered a second-class product among qualified products.

10. An automated fault prediction system for PCBA boards, characterized in that, The automated fault prediction system, used in conjunction with the automated fault prediction method as described in any one of claims 4-8, comprises: The test bench integrates a central control display screen, an alarm unit, a control system, and an output power supply unit. The central control display screen is for easy viewing and operation by relevant personnel, the alarm unit is for alarm prompts, the control system is for setting and adjusting electrical test parameters, and the output power supply unit is for providing the power voltage required for PCBA board testing. The transport module includes a robotic arm and a positioning unit. The robotic arm is used to transport PCBA boards to the test bench, and the positioning unit is used to position the PCBA boards. Image detection module, used for image detection of PCBA boards; Databases serve to store records; The electrical testing module is used to perform electrical tests on PCBA boards. The data analysis module is used to analyze and process electrical test data; The PHM analysis module is used to predict the probability of future failures. The temperature rise module is used to heat and measure the temperature of the PCBA board.

Images