Welding process and post-welding joint quality detection device and system

By designing the welding process and post-weld joint quality inspection system, and using multi-level cache and multi-layer error detection technology, the welding process is efficient, accurate and real-time inspection is achieved, the shortcomings of traditional detection methods are solved, the real-time and accuracy of detection are improved, and the scrap rate is reduced.

CN120404943APending Publication Date: 2025-08-01ZHENJIANG LIHAO PRECISION TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510484638.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-04-17
Publication Date
2025-08-01

AI Technical Summary

Technical Problem

Traditional welding detection methods have shortcomings in information acquisition, real-time, detection accuracy and data accuracy, and cannot meet the needs of modern industry for efficient, accurate and real-time inspection.

Method used

A welding process and post-weld joint quality detection device and system are designed, including data acquisition, preprocessing, transmission, analysis, evaluation and storage units. It adopts a multi-level cache architecture, a multi-layer error detection mechanism and automatic retransmission recovery technology, combined with electrical signals, arc light, temperature field and sound monitoring, and uses machine learning models for real-time analysis and quality evaluation.

Benefits of technology

It significantly improves the real-time and data accuracy of welding inspection, and reduces the data transmission bit error rate to below 0.001%, ensuring the reliability of welding quality evaluation, reducing the scrap rate, and improving the production efficiency and credibility of the test results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120404943A_ABST
    Figure CN120404943A_ABST
Patent Text Reader

Abstract

The invention provides a welding process and post-welding joint quality detection device and system, and relates to the technical field of welding detection. The welding process and post-welding joint quality detection device comprises a data acquisition system unit, a data preprocessing system unit, a data transmission system unit, a data analysis system unit, a quality evaluation system unit, a storage system unit, a control and reality system unit and a welding process and post-welding joint quality detection system, comprising a welding process real-time monitoring module, the welding process real-time monitoring module is connected with a signal preprocessing and feature extraction module, the signal preprocessing and feature extraction module is connected with a data fusion and analysis module, and the data fusion and analysis module is connected with a post-welding joint quality detection module. And the post-welding joint quality detection module is connected with a quality evaluation and report generation module. The method has accurate detection capability, guarantees the accuracy and completeness of data, and greatly improves the welding quality control and the production efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of welding detection, and particularly to a device and system for detecting the quality of welding process and post-weld joints. Background Art

[0002] In modern industrial production, welding, as a key joining process, is widely used in many fields such as aerospace, automotive manufacturing, and shipbuilding industry. The welding quality is directly related to the safety, reliability, and service life of products. Therefore, it is crucial to accurately detect the welding process and the quality of post-weld joints.

[0003] Traditional welding detection means have many limitations. In terms of welding process monitoring, in the early stage, only simple current and voltage measuring instruments could be relied on, and the obtained information was limited and could not reflect the welding state in real time and comprehensively, making it difficult to give early warnings for complex welding defects. For the quality detection of post-weld joints, conventional visual inspection is greatly affected by human factors and it is difficult to detect internal micro-defects; although radiographic inspection can detect internal defects, the equipment is expensive, the inspection process is cumbersome, and it is harmful to the human body; ultrasonic inspection is easily interfered by factors such as the shape and material of the welded parts, and the inspection accuracy and reliability need to be improved.

[0004] With the development of industrial automation and intelligence, the welding process has been continuously innovated, putting forward higher requirements for detection technology. On the one hand, the application of new welding materials and processes makes the physical phenomena in the welding process more complex, and traditional detection means cannot meet the detection needs of diverse welding scenarios. On the other hand, the real-time requirement for detection in the production process has been significantly improved. It is necessary to detect and correct problems in time during the welding process to avoid generating a large number of defective products and reduce production costs. The existing welding detection systems have deficiencies in the real-time data transmission and processing. For example, when the data volume is large, the data transmission delay causes the inability to respond to abnormal situations in the welding process in time; at the same time, there is a lack of effective error control and recovery mechanisms during data transmission, which easily causes data loss or errors, affecting the accuracy and reliability of the detection results. Therefore, there is an urgent practical need to develop a device and system for detecting the quality of welding process and post-weld joints that can overcome the above defects and achieve efficient, accurate, and real-time detection. Summary of the Invention

[0005] Technical Problems to be Solved

[0006] In view of the deficiencies of the prior art, the present invention provides a device and system for detecting the quality of welding process and post-weld joints, which solves the problems of traditional welding detection means in terms of monitoring information acquisition, real-time performance, detection accuracy, and data accuracy.

[0007] Technical Solutions

[0008] To achieve the above object, the present invention is realized through the following technical solutions: a welding process and post-weld joint quality detection device, and the detection device includes a data acquisition system unit, a data preprocessing system unit, a data transmission system unit, a data analysis system unit, a quality assessment system unit, a storage system unit, and a control and display system unit.

[0009] A welding process and post-weld joint quality detection system, including a real-time monitoring module for the welding process, characterized in that: the real-time monitoring module for the welding process is connected to a signal preprocessing and feature extraction module, the signal preprocessing and feature extraction module is connected to a data fusion and analysis module, the data fusion and analysis module is connected to a post-weld joint quality detection module, and the post-weld joint quality detection module is connected to a quality assessment and report generation module;

[0010] The signal preprocessing and feature extraction module includes data caching and prefetching, error control and data recovery, a signal preprocessing subsystem, and a feature extraction unit. In error control, let the data to be sent be D(x), the generating polynomial be G(x), and its highest power be r. Shift D(x) to the left by r bits to get D(x)×x^r, then divide D(x)×x^r by G(x) to get the remainder R(x). This remainder is the CRC check code. The data sent by the sending end is T(x) = D(x)×x^r + R(x). After the receiving end receives the data, divide T(x) by G(x). If the remainder is 0, it is considered that the data transmission is correct, otherwise it is considered that the data transmission is incorrect.

[0011] Preferably, the real-time monitoring module for the welding process includes an electrical signal acquisition unit, an arc light and temperature field sensing unit, and a sound monitoring unit.

[0012] Preferably, the data fusion and analysis module includes a data fusion system, a real-time analysis subsystem, and a historical data storage and mining subsystem.

[0013] Preferably, the post-weld joint quality detection module includes an ultrasonic phased array detection unit and an X-ray data program detection unit.

[0014] Preferably, the quality assessment and report generation module includes a quality assessment system and a report generation subsystem.

[0015] Advantageous Effects

[0016] The present invention provides a welding process and post-weld joint quality detection device and system. It has the following advantageous effects:

[0017] 1. The present invention provides a device and system for detecting the quality of welded joints during and after the welding process. Through the data caching and prefetching unit set in the system, a multi-level caching architecture is constructed at the front-end device and the edge node, and intelligent prefetching of data is driven based on historical data and real-time tasks. For example, during the welding process, the data relied on by the next process can be prefetched from the storage end to the edge node cache in advance, and a quick response can be made when actually needed, greatly reducing the data waiting time. Through actual tests, in complex welding tasks, the data acquisition time is shortened by an average of

[40] %, effectively avoiding the interruption of the detection process caused by data transmission delay, ensuring the continuity of the welding process monitoring and post-weld detection, greatly improving the overall real-time performance of the system, and meeting the strict requirements of modern industry for the high efficiency of welding detection.

[0018] 2. The present invention provides a device and system for detecting the quality of welded joints during and after the welding process. The error control and data recovery unit adopts a multi-layer error detection mechanism, from the CRC-32 checksum at the link layer, the checksums of the transport layer TCP and UDP, to the MD5 hash checksum and image feature point checksum for important data blocks at the application layer, to comprehensively ensure the accuracy of data transmission. Once an error is detected, the automatic retransmission recovery mechanism combines the exponential backoff algorithm to accurately retransmit the error data. At the same time, for some lost or damaged data, technologies such as interpolation algorithms and error correction codes are used for repair and reconstruction. Through testing, the data transmission error rate is reduced to less than 0.001%, effectively avoiding misjudgment of welding quality caused by data errors, providing a reliable data basis for welding quality assessment, and greatly improving the credibility and usability of the detection results. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a schematic diagram of the system flow of the present invention;

[0020] Figure 2 It is a schematic diagram of the device unit structure of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] As Figure 1 shown, the embodiment of the present invention provides a device for detecting the quality of welded joints during and after the welding process. The detection device includes a data acquisition system unit, a data preprocessing system unit, a data transmission system unit, a data analysis system unit, a quality assessment system unit, a storage system unit, and a control and display system unit;

[0023] Specifically, in the welding workshop of an automobile manufacturing enterprise, multiple sets of welding process and post-weld joint quality detection devices are deployed for multi-station and multi-type welding operations.

[0024] Data acquisition system unit:

[0025] At the connection between the welding power supply output terminal and the welding cable at each welding station, Honeywell's high-precision current and voltage sensors, model STG-A500, are installed. They can accurately measure currents from 0 to 500A and voltages from 0 to 1000V, with a sampling frequency as high as 1000Hz, ensuring that the subtle changes in welding current and voltage can be captured in real time.

[0026] The high-speed arc light camera acA2040-90um from Basler in Germany is selected, with a frame rate of up to 90fps. It is equipped with a narrow-band filter and is directed at the welding arc area to clearly capture the arc light dynamics. At the same time, the A325sc infrared thermal imager from FLIR is deployed, with a temperature resolution of 0.05°C and a field of view of 45°×34°. It can completely cover the welding molten pool and the surrounding heat-affected area of about 100mm×80mm to obtain an accurate temperature field distribution.

[0027] Four high-sensitivity microphone arrays from Knowles, model SPK0830LR5H-B, are arranged around the welding station to omnidirectionally collect the sounds during the welding process. The frequency response range is 20Hz - 20kHz, effectively identifying various abnormal sounds.

[0028] Data preprocessing system unit:

[0029] A signal filtering module based on a field-programmable gate array (FPGA) is adopted. For electrical signals and sound signals, the IIR (Infinite Impulse Response) filtering algorithm is used to effectively filter out high-frequency noise. For image data, an Nvidia GPU acceleration card is used, and the adaptive histogram equalization algorithm is adopted for image enhancement, significantly improving the image quality.

[0030] Through a customized data normalization module, the current and voltage data are normalized to the [0,1] interval, and the temperature data are normalized to the [-1,1] interval to ensure the consistency of the data in subsequent analysis.

[0031] Data transmission system unit:

[0032] A gigabit Ethernet wired network with Huawei CloudEngine12800 series switches as the core is built inside the workshop. Each detection device is connected to the switch through a Category 6 network cable to ensure stable and high-speed data transmission. At the same time, Cisco's wireless access point AIR-CAP3802 I is deployed to provide wireless network access in the 5GHz band for mobile detection devices and temporary workstations, with a transmission rate of up to 1.3Gbps.

[0033] Data analysis system unit:

[0034] Equipped with a high-performance server, carrying machine learning and deep learning models based on the TensorFlow framework. The real-time analysis module uses a convolutional neural network (CNN) model to perform real-time analysis on the welding process data and judge the welding state. The historical data analysis module uses the pandas and scikit-learn libraries in Python to mine and analyze the stored massive historical welding data.

[0035] Quality assessment system unit:

[0036] The defect identification module uses self-developed image processing algorithms and combines with the OpenCV library to identify defects in the appearance images of welded joints, ultrasonic detection images, and X-ray detection images. The quality rating module determines the weights of each quality index using the analytic hierarchy process (AHP) according to the welding quality standards in the automotive industry, and realizes the quantitative assessment of the quality of welded joints.

[0037] Storage system unit:

[0038] Locally, a Dell PowerVault ME4084 storage array is adopted, configured with 8 solid-state drives of 4TB each, for real-time storage of raw data, preprocessed data, analysis results, and quality assessment reports. At the same time, through the object storage service (OSS) interface of Alibaba Cloud, important data is synchronized to the cloud to achieve remote backup and sharing of data.

[0039] Control and display system unit:

[0040] Each detection device is equipped with a 15-inch industrial-grade touch display screen. Operators can set detection parameters and start or stop the detection process through an intuitive graphical interface. The result display module displays the welding process data, detection images, and quality assessment results in various forms such as charts, images, and texts in real time.

[0041] Such as Figure 2As shown in the figure, a welding process and post-weld joint quality detection system includes a real-time monitoring module for the welding process, and is characterized in that: the real-time monitoring module for the welding process is connected to a signal preprocessing and feature extraction module, the signal preprocessing and feature extraction module is connected to a data fusion and analysis module, the data fusion and analysis module is connected to a post-weld joint quality detection module, the post-weld joint quality detection module is connected to a quality evaluation and report generation module. The real-time monitoring module for the welding process includes an electrical signal acquisition unit, an arc light and temperature field perception unit, and a sound monitoring unit. The data fusion and analysis module includes a data fusion system, a real-time analysis subsystem, and a historical data storage and mining subsystem. The post-weld joint quality detection module includes an ultrasonic phased array detection unit and an X-ray data program detection unit. The quality evaluation and report generation module includes a quality evaluation system and a report generation subsystem;

[0042] The signal preprocessing and feature extraction module includes data caching and prefetching, error control and data recovery, a signal preprocessing subsystem, and a feature extraction unit. In error control and data recovery, in error control, let the data to be sent be D(x), the generating polynomial be G(x), and its highest power be r. Shift D(x) to the left by r bits to get D(x)×x^r, and then divide D(x)×x^r by G(x) to get the remainder R(x). This remainder is the CRC check code. The data sent by the sending end is T(x) = D(x)×x^r + R(x). After the receiving end receives the data, divide T(x) by G(x). If the remainder is 0, it is considered that the data transmission is correct, otherwise it is considered that the data transmission is incorrect;

[0043] Specifically, the application of the detection system in the present invention to the actual situation in the welding workshop of a certain automobile manufacturing enterprise is as follows:

[0044] Real-time monitoring of the welding process:

[0045] At a certain automobile frame welding station, when the welding robot is performing welding operations, the electrical signal acquisition unit real-time collects welding current and voltage data. For example, during a certain welding period, the current is stable at about 200A, the voltage is maintained at 25V, the sampling frequency is 1000Hz, and a set of data is obtained every millisecond. The arc light and temperature field perception unit works synchronously. The high-speed arc light camera continuously captures arc light images at a frame rate of 90fps, and the infrared thermal imager generates 25 temperature field images per second, accurately monitoring that the highest temperature of the molten pool is about 1800°C, the average temperature is 1600°C, and the temperature gradient is about 50°C / mm. The sound monitoring unit synchronously collects welding sounds. After spectrum analysis, no abnormal sound characteristics are detected. These data are real-time transmitted to the signal preprocessing and feature extraction module.

[0046] Signal preprocessing and feature extraction:

[0047] The data cache and prefetch unit takes effect. Based on the analysis of historical data and real-time tasks, it caches in advance the welding process parameters that may be used in the next stage of welding, the historical detection data of similar welds, etc. from local storage or the cloud to the high-speed cache of the edge node. During the data transmission process, the error control and data recovery unit is activated. For example, for a block of current data containing 1024 bytes, let the data to be sent D(x) be the content of this data block, and the generating polynomial G(x) be x^16 + x^12 + x^5 + 1 (the common CRC-16 generating polynomial), and its highest power r = 16. Shift D(x) left by 16 bits to get D(x)×x^16, divide it by G(x) to get the remainder R(x) as the CRC check code, and send it together with the data. The receiving end verifies the data accuracy through the same calculation. If the verification fails, it will automatically retransmit. The signal preprocessing subsystem filters the collected electrical signals and sound signals, and enhances the image data. The feature extraction unit extracts key features from the processed data, such as the mean and standard deviation of the current, the area and mean brightness of the arc image, etc., and transmits the processed data and features to the data fusion and analysis module.

[0048] Data fusion and analysis:

[0049] The data fusion system uses the D-S evidence theory to fuse multi-source data such as electrical signals, arc light, temperature field, and sound. The real-time analysis subsystem, based on the fused data, uses the trained CNN model to judge the stability of the welding process in real time. After analysis, the current welding process is in a stable state and there is no risk of potential welding defects. The historical data storage and mining subsystem stores all the data of this welding process and regularly mines the historical data. For example, through association rule mining, it is found that when the welding current is between 190 - 210A, the voltage is between 23 - 27V, and the average temperature of the molten pool is between 1550 - 1650°C, the excellent rate of the welding joint quality is over 95%, providing a basis for the subsequent optimization of the welding process.

[0050] Welded joint quality inspection:

[0051] After welding, the ultrasonic phased array detection unit uses Olympus' OmniScan MX2 phased array flaw detector, equipped with a 5L64-W20 probe, to scan the welded joint. With a scanning step of 1mm, two-dimensional and three-dimensional images of the weld interior are generated, and no obvious defects are detected. The X-ray digital imaging detection unit uses GE's YXLON FF355 microfocus X-ray source and PerkinElmer's XRD1621 digital flat panel detector to radiograph the welded joint, sets the tube voltage at 150 kV, the tube current at 5 mA, and the exposure time at 5 s, generates high-quality X-ray images, and also finds no internal defects. The detection data is transmitted to the quality assessment and report generation module.

[0052] Quality Assessment and Report Generation:

[0053] The quality assessment system combines the real-time monitoring and analysis results of the welding process with the post-weld inspection data, and uses the Analytic Hierarchy Process (AHP) to quantitatively evaluate the quality of the welded joint. After calculation, the quality score of this welded joint reaches 90 points (out of 100), and it is judged as a high-quality grade. The report generation subsystem automatically generates a test report according to the welding quality inspection report template in the automotive industry, including welding process parameters, test data, quality assessment results and other information, supports report export and printing, and is convenient for archiving and traceability.

[0054] Finally, the implementation effects and beneficial effects of the present invention are verified as follows:

[0055] Significantly improve the real-time performance of the system:

[0056] Through the data caching and prefetching mechanism, in the actual application of the welding workshop of this automotive manufacturing enterprise, the data acquisition time is shortened by an average of 40%. For example, in a complex frame welding task involving multiple welding processes and a large amount of data calls, the original data waiting time was about 500 ms, and after implementing this system, it was shortened to within 300 ms, effectively avoiding the waiting time of the welding robot caused by data transmission delay, ensuring the efficient and continuous progress of the welding process, greatly improving the overall production efficiency, and meeting the strict requirements of the automotive manufacturing industry for the real-time performance of welding detection.

[0057] Ensure data accuracy and integrity:

[0058] The multi-layer error detection mechanism and automatic retransmission and recovery technology of the error control and data recovery unit have achieved remarkable results. After statistics, during the transmission of a large amount of data in this workshop, the data transmission error rate is reduced to less than 0.001%. During a high-intensity production cycle, about 10 GB of data was transmitted, and only 10 bytes of data errors occurred, and all were successfully corrected through automatic retransmission and data repair technology. It effectively avoids the misjudgment of welding quality caused by data errors, provides a highly reliable data basis for welding quality assessment, greatly improves the credibility and usability of the test results, reduces the scrap rate caused by quality misjudgment, and reduces the production cost.

[0059] Working principle: Real-time monitoring of the welding process: The electrical signal acquisition unit, arc and temperature field perception unit, and sound monitoring unit in the real-time monitoring module of the welding process work together to comprehensively collect various physical signals during the welding process. These signals are transmitted to the signal preprocessing and feature extraction module in real time.

[0060] Signal preprocessing and feature extraction: In the signal preprocessing and feature extraction module, the data cache and prefetch unit caches the data that may be used in advance to an appropriate location based on historical data and real-time task prediction, reducing data acquisition latency. Meanwhile, the error control and data recovery unit utilizes a multi-layer error detection mechanism. For example, in error control, CRC check codes are used (assuming the data to be transmitted is D(x), the generating polynomial is G(x), and its highest power is r. Shift D(x) to the left by r bits to get D(x)×x^r, then divide it by G(x) to obtain the remainder R(x) as the CRC check code. The transmitting end sends T(x) = D(x)×x^r + R(x), and the receiving end determines whether the data transmission is correct by dividing T(x) by G(x)) to ensure data accuracy. Once an error is detected, the data is recovered in a timely manner through mechanisms such as automatic retransmission. The signal preprocessing subsystem processes the collected signals, such as filtering and denoising. The feature extraction unit extracts the key feature parameters reflecting the welding process and quality from the processed data, and then transmits the processed data and features to the data fusion and analysis module.

[0061] Data fusion and analysis: The data fusion system in the data fusion and analysis module uses algorithms such as D-S evidence theory and Kalman filtering to fuse multi-source data and construct a comprehensive and accurate welding information model. The real-time analysis subsystem, based on the fused data, uses machine learning and deep learning models to judge the stability of the welding process in real time and predict welding defects. The historical data storage and mining subsystem stores all the data of the welding process and regularly uses data mining algorithms to mine potential patterns from the historical data, providing a basis for optimizing the welding process and detection model.

[0062] Post-weld joint quality inspection: The ultrasonic phased array detection unit and the X-ray digital imaging detection unit in the post-weld joint quality inspection module respectively use ultrasonic and X-ray technologies to detect the welded joint, obtain the internal structure images and data of the joint, and transmit them to the quality assessment and report generation module.

[0063] Quality assessment and report generation: The quality assessment system in the quality assessment and report generation module combines the real-time monitoring and analysis results of the welding process with the post-weld inspection data, and uses models such as fuzzy comprehensive evaluation and analytic hierarchy process to quantitatively evaluate the quality of the welded joint, judge the quality grade, analyze the causes of defects and give improvement suggestions. The report generation subsystem automatically generates a standardized inspection report according to the built-in industry standard report template based on the quality assessment results, supports report export and printing, and is convenient for archiving and traceability.

[0064] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. Welding process and post-weld joint quality inspection device, characterized in that: The detection device includes a data acquisition system unit, a data preprocessing system unit, a data transmission system unit, a data analysis system unit, a quality assessment system unit, a storage system unit, and a control and display system unit.

2. Welding process and post-weld joint quality detection system, including a real-time welding process monitoring module, characterized in that: The real-time welding process monitoring module is connected to a signal preprocessing and feature extraction module, the signal preprocessing and feature extraction module is connected to a data fusion and analysis module, the data fusion and analysis module is connected to a post-weld joint quality detection module, and the post-weld joint quality detection module is connected to a quality assessment and report generation module; The signal preprocessing and feature extraction module includes data caching and prefetching, error control and data recovery, a signal preprocessing subsystem, and a feature extraction unit. In error control of error control and data recovery, assume the data to be sent is D(x), the generating polynomial is G(x), and its highest power is r. Shift D(x) to the left by r bits to get D(x)×x^r, then divide D(x)×x^r by G(x) to get the remainder R(x). This remainder is the CRC check code. The data sent by the sending end is T(x) = D(x)×x^r + R(x). After the receiving end receives the data, divide T(x) by G(x). If the remainder is 0, it is considered that the data transmission is correct; otherwise, it is considered that the data transmission is incorrect.

3. The welding process and post-weld joint quality inspection system according to claim 2, characterized in that: The real-time welding process monitoring module includes an electrical signal acquisition unit, an arc light and temperature field perception unit, and a sound monitoring unit.

4. The welding process and post-weld joint quality inspection system according to claim 2, characterized in that: The data fusion and analysis module includes a data fusion system, a real-time analysis subsystem, and a historical data storage and mining subsystem.

5. The welding process and post-weld joint quality inspection system according to claim 2, wherein: The post-weld joint quality detection module includes an ultrasonic phased array detection unit and an X-ray data program detection unit.

6. The welding process and post-weld joint quality inspection system according to claim 2, characterized in that: The quality assessment and report generation module includes a quality assessment system and a report generation subsystem.