Welding quality detection method and system

Abstract

The invention discloses a welding quality detection method and system, and relates to the technical field of welding quality detection.The method comprises the steps that a plurality of to-be-detected welding parts are determined, the temperature of a molten pool is continuously detected through a thermal imaging module in the welding process, and a molten pool temperature field distribution time sequence is generated. Based on the time series data, heat conduction analysis and cooling rate identification are performed to generate a heat conduction path and a cooling curve distribution. By analyzing the data, whether a heat accumulation or loss abnormal area exists or not is judged, and if the heat accumulation or loss abnormal area is abnormal, a quality detection signal is sent out. The technical problem that an existing welding quality detection method cannot accurately recognize welding defects caused by uneven heat conduction and abnormal cooling rate in the welding process in real time is solved, and the technical effect of improving the welding quality detection precision and timeliness is achieved.

Description

Technical Field

[0001] The present application relates to the technical field of welding quality detection, and in particular to a welding quality detection method and system. Background Art

[0002] Welding is an indispensable key process in the manufacturing industry, especially in the production of vehicle frames, where welding quality directly affects structural strength, safety and service life. Vehicle frames usually use metal welding technology. Any welding defects, such as lack of fusion, incomplete weld penetration or cracks, may lead to structural instability, thereby affecting safety. Traditional welding quality detection methods, such as visual inspection, ultrasonic testing and X-ray testing, although they can detect some obvious defects, often cannot monitor the temperature changes, heat distribution and cooling rate abnormalities during the welding process in real time. Moreover, most of the detection methods are carried out after welding is completed, and hidden defects in the welding process cannot be detected in time. In the production process of vehicle frames, due to the complex structure and changeable welding environment, welding defects are often difficult to detect and repair in time, especially in those difficult-to-reach welding parts. In order to improve the detection accuracy and efficiency of welding quality, an advanced technology that can monitor the temperature, heat distribution and cooling rate in real time during the welding process is urgently needed to overcome the limitations of existing methods and ensure that welding defects can be identified in a timely and accurate manner, thereby ensuring the quality and safety of the vehicle frame.

[0003] In the current related technologies, there is a technical problem that the welding quality detection method cannot accurately identify welding defects caused by uneven heat conduction and abnormal cooling rate during the welding process in real time. Summary of the invention

[0004] The present application solves the technical problem that the existing welding quality detection methods are unable to identify welding defects caused by uneven heat conduction and abnormal cooling rate during the welding process in real time and accurately by providing a welding quality detection method and system.

[0005] The present application provides a welding quality detection method, comprising: Determine multiple preset welding positions of the target to be detected; during the welding process of the first welding position among the multiple preset welding positions, continuously detect the molten pool temperature through the thermal imaging module to generate a molten pool temperature field distribution time series; perform heat conduction analysis based on the molten pool temperature field distribution time series to generate a heat conduction path; perform cooling rate identification based on the molten pool temperature field distribution time series to generate a cooling curve distribution; determine whether there is an abnormal heat accumulation area or an abnormal heat loss area based on the heat conduction path and the cooling curve distribution, and if so, issue a quality detection abnormality signal.

[0006] The present application also provides a welding quality detection system, comprising: A preset welding position determination module, the preset welding position determination module is used to determine multiple preset welding positions of the target to be detected; a temperature continuous detection module, the temperature continuous detection module is used to continuously detect the molten pool temperature through the thermal imaging module during the welding process of the first welding position among the multiple preset welding positions, and generate a molten pool temperature field distribution time series; a heat conduction analysis module, the heat conduction analysis module is used to perform heat conduction analysis based on the molten pool temperature field distribution time series, and generate a heat conduction path; a cooling rate identification module, the cooling rate identification module is used to identify the cooling rate based on the molten pool temperature field distribution time series, and generate a cooling curve distribution; a quality detection abnormal signal issuing module, the quality detection abnormal signal issuing module is used to determine whether there is an abnormal heat accumulation area or an abnormal heat loss area based on the heat conduction path and the cooling curve distribution, and if so, issue a quality detection abnormal signal.

[0007] The welding quality detection method and system proposed in this application first determine multiple welding parts to be detected, and continuously detect the molten pool temperature through the thermal imaging module during the welding process to generate a molten pool temperature field distribution time series. Based on the time series data, heat conduction analysis and cooling rate identification are performed to generate heat conduction paths and cooling curve distributions. By analyzing these data, it is determined whether there are abnormal areas of heat accumulation or loss. If there are abnormalities, a quality detection signal is issued. By real-time monitoring of the temperature distribution and cooling rate changes during the welding process, welding defects caused by uneven heat conduction or improper cooling during the welding process can be accurately identified, achieving the technical effect of improving the detection accuracy and timeliness of welding quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0009] Figure 1 A schematic flow chart of a welding quality detection method provided in an embodiment of the present application.

[0010] Figure 2 A structural schematic diagram of a welding quality detection system provided in an embodiment of the present application.

[0011] Explanation of the reference numerals: preset welding position determination module 10, temperature continuous detection module 20, heat conduction analysis module 30, cooling rate identification module 40, quality detection abnormal signal issuing module 50. DETAILED DESCRIPTION

[0012] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0013] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0014] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0015] The present application embodiment provides a welding quality detection method, such as Figure 1 As shown, the method includes: Step S100, determine multiple preset welding parts of the target to be detected. Specifically, when determining multiple preset welding parts of the target to be detected of the frame, first collect the design drawings, welding process specifications, quality standard documents and past similar project case materials of the frame for in-depth analysis. By analyzing the design drawings, decomposing the frame structure, and using mechanical analysis methods to determine the key areas where each part is subjected to greater force during use, especially the parts that are subjected to high loads or frequent force, such as frame connection points, support areas, connection joints, etc. The welding quality of these parts is crucial to the stability and safety of the entire vehicle. Then, combined with the welding process specifications, considering the heat input of special welding methods, molten pool characteristics, and thermal stress conditions of welded joints of different materials (such as steel, aluminum alloy, etc.), evaluate areas where welding quality problems may occur, especially in areas where temperature and thermal stress changes greatly during welding. Finally, a comprehensive screening of potential critical areas is carried out to remove some minor or repeated parts, and a number of representative and critical preset welding parts are determined, such as frame joints, support frame connections, drive system joints, etc., so that the welding process and quality of these parts can be monitored in the future to ensure that the structural strength and performance of the frame meet the design requirements and ensure the safety and reliability of the entire vehicle.

[0016] Step S200, during the welding process of the first welding position among the plurality of preset welding positions, the temperature of the molten pool is continuously detected by the thermal imaging module to generate a time series of the temperature field distribution of the molten pool. Specifically, before welding the first welding position among the plurality of preset welding positions, the thermal imaging module is first installed in a suitable position and fully calibrated, including calibrating the temperature with a standard black body source, and adjusting the image acquisition parameters to ensure that the thermal radiation signal of the molten pool can be clearly and accurately captured without interference from strong light and smoke. After the welding starts, the thermal imaging module performs real-time continuous thermal radiation signal acquisition of the molten pool at a high frequency. Based on the principle of thermal radiation of the molten pool, the detector receives infrared rays and converts them into electrical signals, which are processed and reconstructed to generate a sequence of thermal images of the molten pool, which can reflect the temperature change of the molten pool in real time. After that, a large number of molten pool thermal images arranged in chronological order are formed into a time series of the temperature field distribution of the molten pool. In the process, the data is preprocessed and optimized, such as denoising and interpolation to fill in missing points, to ensure that the data is complete and continuous, and to provide key thermal information for subsequent heat conduction analysis and cooling rate identification, so as to comprehensively evaluate the welding quality.

[0017] Step S300, based on the molten pool temperature field distribution time series, heat conduction analysis is performed to generate a heat conduction path. Specifically, the molten pool temperature field distribution time series data is preprocessed, noise is reduced using a filtering algorithm, missing data points are supplemented by an interpolation method, and the temperature value is mapped to a unified range using normalization, laying the foundation for subsequent analysis. Then, the molten pool temperature field images at multiple key time points are extracted from the preprocessed data in chronological order, the molten pool center, boundary and heat-affected zone range are determined using image processing technology, and the temperature gradient vector field is calculated by a differential algorithm, etc., which is used as the basis for the heat conduction direction and diffusion range. Then, a heat conduction path is constructed based on the temperature gradient information, starting from the high temperature area of ​​the molten pool at adjacent time points, and a heat diffusion path is drawn along the conduction direction combined with the diffusion range. After processing adjacent time points in sequence, the paths are connected in chronological order to form a complete heat conduction path, which provides key information for evaluating welding quality, etc.

[0018] In a possible implementation, heat conduction analysis is performed based on the molten pool temperature field distribution time series to generate a heat conduction path, and step S300 further includes step S310, extracting the first time series molten pool temperature field, the second time series molten pool temperature field, the third time series molten pool temperature field, and the Nth time series molten pool temperature field in the molten pool temperature field distribution time series. Specifically, from the generated molten pool temperature field distribution time series data set, the first time series molten pool temperature field, the second time series molten pool temperature field, the third time series molten pool temperature field, and so on are separated according to the time sequence, until the Nth time series molten pool temperature field. It is necessary to identify and extract the time tags of the time series data to ensure the integrity and independence of each time series molten pool temperature field, and provide accurate data samples for the subsequent adjacent time series heat diffusion path analysis. For example, in a molten pool temperature field distribution time series collected at 0.1 second time intervals, if you want to analyze the heat conduction in the first 10 seconds, then N is 100 (10 seconds divided by 0.1 seconds), so that the molten pool temperature field data corresponding to these 100 time points can be extracted in turn. Each data contains the spatial temperature distribution information of the molten pool at that moment. The information will serve as the basic data for subsequent steps, and its accuracy and completeness will directly affect the reliability of the entire heat conduction path analysis.

[0019] Step S320, the heat diffusion path of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field under adjacent time sequences is identified to establish a first diffusion path. Specifically, the first time sequence molten pool temperature field and the second time sequence molten pool temperature field are simultaneously loaded into the thermal analysis software or algorithm module. By comparing the temperature changes of the two time sequence molten pool temperature fields at various positions in space, the basic principle of heat conduction, that is, heat is transferred from the high temperature area to the low temperature area, is used to identify the diffusion direction and approximate range of heat in the molten pool and its surrounding areas from the first time sequence to the second time sequence. For example, by calculating the temperature difference between adjacent pixels or regions, the area with a large temperature gradient is determined, and the area is the main diffusion direction of heat. Combined with the geometric shape of the molten pool and the range of the heat affected zone, the diffusion path of heat from the high temperature area (such as the center of the molten pool) of the first time sequence molten pool temperature field to the relatively low temperature area in the second time sequence molten pool temperature field is drawn in a visual way in the software, which is the first diffusion path, which intuitively shows the propagation trajectory of heat between the two adjacent time sequences, and provides key local information for the subsequent analysis of the conduction of heat in the entire welding process.

[0020] Step S330, the heat diffusion paths of the second time sequence molten pool temperature field and the third time sequence molten pool temperature field are identified under adjacent time sequences, and the second diffusion path is established. This is repeated until the Nth time sequence molten pool temperature field is traversed and the N-1th diffusion path is established. Specifically, according to the same method and process as that of establishing the first diffusion path, the heat diffusion paths of the molten pool temperature fields of adjacent time sequences such as the second time sequence molten pool temperature field and the third time sequence molten pool temperature field, the third time sequence molten pool temperature field and the fourth time sequence molten pool temperature field are identified and drawn in sequence. For each pair of molten pool temperature fields of adjacent time sequences, the heat diffusion direction and range are determined by analyzing the temperature change, and the corresponding diffusion paths are established, which are named the second diffusion path, the third diffusion path, etc., respectively, and continue until the N-1th diffusion path. In this process, it is necessary to ensure that the identification and establishment of each diffusion path have high accuracy and consistency, because the local diffusion paths will eventually be combined into a complete heat conduction path, and the error in any link may affect the reliability of the overall result. At the same time, as the traversal proceeds, the heat diffusion information from the beginning of welding to different stages is gradually accumulated. The continuity and integrity of the information are crucial to a comprehensive understanding of the thermal behavior during the welding process.

[0021] Step S340, connect the first diffusion path, the second diffusion path and the N-1 diffusion path to generate the heat conduction path. Specifically, when all adjacent sequential diffusion paths (i.e., the first diffusion path to the N-1 diffusion path) have been established, use graphics processing or path connection algorithm to connect the diffusion paths in chronological order. For example, match and connect the end point of the first diffusion path with the starting point of the second diffusion path, and then connect the end point of the second diffusion path with the starting point of the third diffusion path, and so on, to finally form a continuous and complete heat conduction path. The heat conduction path shows the propagation trajectory and distribution of heat inside the welded part after the heat is generated from the molten pool over time during the entire welding process, which reflects the thermal behavior of the welding process from a macroscopic perspective, and provides a key basis for further analyzing the welding quality and judging whether there are welding defects (such as lack of fusion caused by heat accumulation, cracks caused by uneven cooling, etc.). Through in-depth research on the heat conduction path, the welding process parameters can be optimized in a targeted manner to improve the welding quality and product reliability.

[0022] In a possible implementation, the heat diffusion paths of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field are identified under adjacent time sequences to establish a first diffusion path. Step S320 further includes step S321, comparing the first time sequence molten pool temperature field and the second time sequence molten pool temperature field, identifying the temperature changes of the center, boundary and heat affected zone of the molten pool, and establishing a temperature gradient. Specifically, the data of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field are imported into thermal analysis software or a tool with image processing function. Using image recognition algorithm and temperature data analysis technology, the center position of the molten pool in the two time sequence molten pool temperature fields is accurately located, and it is determined by finding the area with the highest temperature and relatively concentrated. For the molten pool boundary, an edge detection algorithm is used to identify the transition area where the temperature drops sharply from a higher value inside the molten pool to a value close to the surrounding environment, thereby determining the boundary contour of the molten pool. The identification of the heat affected zone is based on a pre-set temperature threshold range. Outside the molten pool boundary, an area with a temperature higher than the ambient temperature but lower than the temperature inside the molten pool is selected as the heat affected zone. The temperature difference between the two time series is calculated for each pixel point or small area unit in the center, boundary and heat-affected zone of the molten pool. The temperature gradient field is established through the distribution of the temperature difference, in which the direction and area with a large temperature difference represent the direction and position where the temperature changes more drastically, providing an intuitive quantitative basis for the subsequent determination of the direction of heat conduction. For example, in a certain area, if the temperature drops by 50°C from the first time series to the second time series, and the temperature of the adjacent area only drops by 10°C, then there is a large temperature gradient between the two areas, indicating that there is an obvious trend of heat transfer between the two areas.

[0023] Step S322, based on the temperature gradient, identify the direction and diffusion range of heat conduction. Specifically, according to the established temperature gradient field, based on the basic physical principles of heat conduction, heat is always conducted from the high temperature area to the low temperature area. Therefore, in the temperature gradient field, the direction of temperature reduction is the direction of heat conduction. By analyzing the temperature gradient vector, the specific conduction path of heat in the molten pool and its surrounding areas can be determined. For example, if there is a large temperature gradient vector in the direction from the center of the molten pool to the boundary, then the heat diffuses from the center of the molten pool to the boundary. Combining the actual physical properties of the molten pool and the mathematical model of heat conduction, the diffusion range of heat is estimated according to the size and rate of change of the temperature gradient. Near the area with a large temperature gradient, the heat diffuses relatively quickly and over a wide range; while in the area where the temperature gradient gradually decreases, the heat diffusion rate slows down and the range gradually narrows. For example, near the molten pool boundary, although the temperature gradient is reduced from the center of the molten pool to the boundary, there is still a certain temperature change, which indicates that the heat is still diffusing to the heat-affected zone at the boundary, but the diffusion range is smaller than that inside the molten pool. Through analysis, the range contour of the heat diffusion from the center of the molten pool to the boundary and the heat-affected zone between these two time series can be roughly determined, providing key spatial information for generating the first diffusion path.

[0024] Step S323, generating the first diffusion path according to the heat conduction direction and diffusion range. Specifically, after determining the heat conduction direction and diffusion range, using a graphics drawing tool or a path generation algorithm, taking the high temperature area (usually the center of the molten pool) in the first time series molten pool temperature field as the starting point, along the heat conduction direction, according to the boundary of the diffusion range, draw a continuous curve or broken line, this line is the first diffusion path. In the drawing process, the actual physical process of heat conduction should be analyzed to ensure the rationality and accuracy of the path. For example, if the heat is diffused evenly in a certain direction, the path can be drawn as a relatively smooth curve; if there are obstacles or sudden changes in the thermal conductivity of the material in certain areas, resulting in a change in the heat conduction direction, the path should also turn or bend accordingly. At the same time, in order to more clearly indicate the diffusion of heat, the temperature values ​​or temperature change rates at different positions can also be marked on the path, so that the first diffusion path can not only intuitively display the propagation trajectory of heat from the first time-series molten pool temperature field to the second time-series molten pool temperature field, but also provide some quantitative information about the heat transfer process, which provides important basic data for subsequent heat conduction analysis and welding quality evaluation. Through the study of the first diffusion path, it is possible to preliminarily judge whether the thermal behavior in the welding process is normal and whether there is abnormal heat distribution that may lead to welding defects.

[0025] Step S400, based on the molten pool temperature field distribution time series, the cooling rate is identified and the cooling curve distribution is generated. Specifically, multiple point temperature values ​​at key positions of the molten pool center, boundary and heat-affected zone on the molten pool temperature field image at each time point are extracted from the molten pool temperature field distribution time series data set, and the data is quality checked and preprocessed, and outliers are removed and smoothed to ensure accuracy. Then, for each point, the temperature change rate over time, that is, the cooling rate, is calculated using numerical calculation methods such as the difference method, so as to generate a temperature rate change curve of each point over time, showing the change in cooling speed of each point. Afterwards, a clustering algorithm or other similarity analysis method is used to group points with similar cooling rate change trends according to characteristics such as the slope, curvature, and peak time of the curve, determine the key aggregation area, and simplify the cooling rate curves of many single points into a key aggregation area temperature change curve that can represent the typical cooling characteristics of different regions. Finally, based on the temperature change curve of the key polymerization area, it is plotted and displayed in the same coordinate system, with the horizontal axis being time and the vertical axis being temperature or cooling rate. The cooling curve distribution is generated to analyze key information such as temperature uniformity and cooling rate differences during the cooling process of the welded parts, providing a basis for evaluating welding quality and judging defects.

[0026] In one possible implementation, the cooling rate is identified based on the molten pool temperature field distribution time series, and a cooling curve distribution is generated. Step S400 further includes step S410, which calculates the rate of change curve of the temperature of each point over time based on the molten pool temperature field distribution time series, and generates the temperature change curve of each point. Specifically, a large number of representative points are accurately selected from the molten pool temperature field distribution time series data set. These points should be evenly distributed in key positions such as the center, boundary and heat-affected zone of the molten pool to fully reflect the temperature conditions of the welding area. For each selected point, in the time dimension, according to the temperature values ​​of the point at different time points recorded by the molten pool temperature field distribution time series, a numerical differential method (such as central difference method, forward difference method or backward difference method, etc.) is used to calculate the rate of change of its temperature over time. For example, if it is known that a point is at time The temperature is , at time The temperature is , the temperature change rate (i.e. cooling rate) of this point during this time period is calculated using the central difference method: . By calculating all time intervals of each point in the entire welding process, a data series of the rate of change of temperature at each point over time is obtained. The data series is plotted with time as the horizontal axis and temperature or cooling rate as the vertical axis to generate the temperature change curve of each point. The curve shows the temperature evolution of each point in the welding process in detail, providing basic data for subsequent analysis. They reflect the differences in heat transfer and cooling characteristics at different locations and are a key source of information for in-depth understanding of the welding thermal process.

[0027] Step S420, the temperature change curves of each point are consistently aggregated to establish the temperature change curves of each key aggregation area. Specifically, due to the complexity and unevenness of the welded parts, the temperature change curves of each point often present a variety of forms. In order to grasp the cooling characteristics of the welding area as a whole, a cluster analysis method is used to consistently aggregate these curves. Determine a suitable clustering algorithm, such as a distance-based K-Means clustering algorithm or a density-based DBSCAN clustering algorithm. The characteristic parameters of the curve are used as the basis for clustering. The characteristic parameters may include the slope change trend, curvature characteristics, peak time and size, temperature level of the stable segment, etc. of the curve. For example, for curves with similar slope change ranges and peak characteristics, they are classified into one category. Through cluster analysis, a large number of single point temperature change curves are divided into several categories, each category represents an area with similar cooling characteristics, and these areas are key aggregation areas. For each key aggregation area, by performing data fusion and feature extraction on the point temperature change curves contained therein, a temperature change curve that can represent the overall cooling characteristics of the area is established. This can be achieved by taking a weighted average of the temperature values ​​of the curves at each point in the region (the weight can be determined according to the importance or representativeness of the point), or by extracting the main features of the curve (such as the average slope, average peak temperature, etc.) and constructing a representative curve. The generated temperature change curve of the key polymerization area can summarize and simplify the complex point temperature information, highlight the commonalities and differences in cooling of different areas of the welded parts, and provide more macro and targeted information for further analysis of welding quality.

[0028] Step S430, generating the cooling curve distribution with the temperature change curve of each key polymerization area. Specifically, the temperature change curves of each key polymerization area are integrated in a unified coordinate system to construct the cooling curve distribution. In the coordinate system, the horizontal axis usually represents time, and the vertical axis represents temperature or cooling rate (selected according to the specific analysis purpose and needs). By plotting the temperature change curves of different key polymerization areas in the same figure, the temperature changes of different areas during the welding cooling process can be intuitively compared. For example, it can be clearly seen which areas cool faster, which areas cool relatively slowly, and the temperature difference between different areas changes over time. The cooling curve distribution shows the thermal behavior characteristics of the welded parts during the cooling stage, providing a key basis for evaluating the welding quality. By analyzing the cooling curve distribution, it can be determined whether there is a problem of uneven cooling during the welding process. For example, if the cooling curve of a key polymerization area deviates significantly from other areas, there is a premature or late temperature drop, abnormal cooling rate fluctuations, etc., indicating that there is a risk of welding defects in the area, such as cracks caused by uneven cooling. Based on such analysis results, the welding process can be optimized and adjusted in a targeted manner to improve welding quality and product reliability, ensuring that the welded components can meet performance requirements during subsequent use.

[0029] Step S500, judging whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution, if so, sending a quality detection abnormal signal. Specifically, collect and analyze the information of the first welding part, clarify its metal type, structural characteristics and weld continuity direction characteristics, etc., the characteristics will affect heat conduction and distribution. Then, with the characteristics as constraints, screen and collect from the qualified welding data in the past, randomly select the standard heat conduction path and cooling curve distribution, and analyze the deviation of the remaining samples from the standard, generate the corresponding first tolerance constraint and second tolerance constraint, and determine the acceptable fluctuation range during normal welding. Then compare the actual welding heat conduction path with the standard path, calculate the key position difference by image recognition and data matching technology, and mark the area beyond the first tolerance constraint range as the abnormal heat accumulation area; at the same time, compare the actual and standard cooling curve distributions, and determine the abnormal heat loss area by calculating the slope, temperature peak deviation and other indicators. Once the above abnormal area is found, the welding personnel are immediately reminded to stop the operation through the sound and light alarm device, and the detailed abnormal information is sent to the monitoring center by wireless communication, and the relevant data is integrated with the production management system to ensure timely handling of welding quality problems and improve production quality stability and reliability.

[0030] In a possible implementation, it is determined whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution. If so, a quality detection abnormality signal is issued. Step S500 further includes step S510, determining the first welding part feature of the first welding part, wherein the first welding part feature includes the metal type of the welding part, the structural feature of the welding part, and the continuity direction feature of the weld. Specifically, feature recognition is performed for the first welding part. The metal type of the welding part is accurately determined by detecting the material composition of the welding part and querying the relevant material standards, because different metals have significant differences in thermal conductivity, such as the thermal conductivity of copper is much higher than that of iron, which will directly affect the heat transfer speed and distribution during the welding process. Using measurement tools and three-dimensional modeling technology, the structural characteristics of the welding part are analyzed, including its geometric shape, size specifications, wall thickness changes, and whether there are complex internal structures (such as reinforcing ribs, holes, grooves, etc.). Structural elements will change the path and efficiency of heat conduction, forming local heat concentration or excessive heat dissipation areas. In addition, according to the welding process design drawings and the actual weld layout, the directional characteristics of the weld continuity are clarified, that is, whether the weld is straight, curved, or has a specific angle and direction, as well as whether it is longitudinally penetrating, horizontally connected, or obliquely staggered, etc. Different weld directions will cause the heat to diffuse along the weld direction and its surroundings during welding. Each has its own characteristics, which has an important impact on the overall heat distribution. By combining this information, the characteristics of the first welding part can be completely determined, providing a key basic basis for subsequent standard collection and abnormal judgment.

[0031] Step S520, taking the first welding part feature as a constraint, collecting the standard heat conduction path and the corresponding first tolerance constraint when the welding quality is qualified, as well as the standard cooling curve distribution and the corresponding second tolerance constraint. Specifically, based on the determined first welding part feature, screening is performed in the historical welding data. The historical data comes from past welding projects of the same type of metal, similar structural design and the same weld direction, and the welding quality of the project has been strictly tested and determined to be qualified. From the qualified samples, the heat conduction path and cooling curve distribution data obtained by thermal imaging analysis and temperature monitoring during the welding process are extracted. For the heat conduction path, image processing and path fitting algorithms are used to convert it into a standardized mathematical model representation, and a representative path is randomly selected as the standard heat conduction path. Then, the heat conduction paths of the remaining qualified samples are compared and analyzed point by point with the standard path, and the mean and standard deviation of the position deviation of each point are calculated by statistical methods, so as to determine a reasonable deviation range, that is, the first tolerance constraint, which represents the fluctuation range allowed for the heat conduction path under normal welding conditions. The cooling curve distribution is standardized through data fitting and feature extraction. A standard cooling curve distribution is randomly selected, and then the deviations between other sample curves and the standard curve in key indicators such as temperature value and cooling rate are analyzed. The second tolerance constraint is obtained through statistical analysis, which defines the acceptable range of variation of the curve during the normal cooling process. The determination of standards and constraints provides an accurate quantitative reference for subsequent abnormal judgment.

[0032] Step S530, compare the heat conduction path with the standard heat conduction path, locate the area that does not meet the first tolerance constraint, and generate the abnormal heat accumulation area. Specifically, in the current welding process, the heat conduction path is monitored and generated in real time by the thermal imaging module, and a detailed comparison operation is performed with the standard heat conduction path collected previously. Using image recognition algorithm and path matching technology, the actual path is compared point by point with the standard path at the same time node and spatial position, and the distance deviation, angle deviation and area change rate of key areas (such as near the molten pool, heat affected zone boundary, etc.) between the corresponding points are calculated. Once these parameters are found to exceed the preset first tolerance constraint range in a certain area or node, the area is marked as a potential abnormal heat accumulation area. Due to factors such as unstable welding current, uneven welding speed, impurities or gaps in the welding joint, heat is excessively accumulated in this area and cannot be diffused according to the normal conduction path, which may cause welding defects such as unfused, overheated tissue, etc. Positioning can timely discover abnormal conditions in the heat conduction process, providing key clues for subsequent quality assessment and problem solving.

[0033] Step S540, compare the cooling curve distribution with the standard cooling curve distribution, locate the area that does not meet the second tolerance constraint, and generate the abnormal heat loss area. Specifically, the cooling curve distribution obtained in the actual welding process is compared and analyzed with the standard cooling curve distribution. Using numerical calculation methods and curve fitting techniques, the temperature values, cooling rates and other key data of the two curves at each time point on the time axis are compared one by one, and the absolute value of the temperature difference, the rate of change of the cooling rate and other indicators are calculated. When the indicator exceeds the established second tolerance constraint range in a specific time period or local area, the area is determined to be an abnormal heat loss area. Due to insufficient flow of the cooling medium, uneven temperature, unreasonable cooling method (such as local excessive cooling or insufficient cooling) or abnormal fluctuations in the ambient temperature around the welded parts, the heat loss process does not conform to the normal cooling law, resulting in uneven distribution of thermal stress in the welding area, and prone to defects such as cracks and deformation. By comparison, the abnormal situation in the cooling process can be accurately identified, providing an important basis for ensuring welding quality.

[0034] Step S550, generating a quality detection abnormal signal for the abnormal heat accumulation area and the abnormal heat dissipation area. Specifically, it is determined that there is an abnormal heat accumulation area or an abnormal heat dissipation area, and the quality detection abnormal signal generation mechanism is immediately started. The signal is transmitted and expressed in a variety of ways. At the welding site, a strong sound and light alarm is issued through the sound and light alarm device connected to the control system to warn the welding operator to stop the current welding operation so as to check and adjust the welding parameters or process conditions in time. At the same time, the detailed abnormal information (including the position coordinates, area size, quantitative indicators of the abnormal degree and comparison data with the standard of the abnormal area) is sent to the remote welding quality monitoring center through the wireless communication module, and a striking alarm window pops up on the computer system of the monitoring center, and the abnormal information is recorded in a special quality detection database for subsequent analysis and decision-making by quality management personnel and technical experts. In addition, the abnormal signal can also be integrated with the automatic control system on the production line to automatically trigger the shutdown or adjustment program of the relevant equipment to prevent the continued production of unqualified products, and provide data support for subsequent quality tracing and process improvement. Through the abnormal signal processing mechanism, it is ensured that the welding quality problems can be handled in a timely and effective manner, and the stability and reliability of the entire welding production process are guaranteed.

[0035] In a possible implementation, the first welding part feature is used as a constraint to collect the standard heat conduction path and the corresponding first tolerance constraint when the welding quality is qualified, as well as the standard cooling curve distribution and the corresponding second tolerance constraint. Step S520 further includes step S521, taking the first welding part feature as a constraint to collect multiple heat conduction paths and multiple cooling curve distributions when the welding quality is qualified. Specifically, according to the determined first welding part feature, including the metal type of the welding part, the structural feature of the welding part and the weld continuity direction feature, accurate screening is performed in the welding process database of the enterprise or institution. The database stores detailed data of past welding projects, and finds welding case records with the same metal type (such as aluminum alloy, stainless steel, etc.), similar structural features (such as shape complexity, wall thickness distribution, rib arrangement, etc.) and the same weld continuity direction (such as longitudinal, transverse, curved, etc.) as the current first welding part. For the qualified welding cases screened out, the heat conduction path information and cooling curve distribution data of each case during the welding process are extracted through thermal imaging analysis technology and temperature sensor data records. The heat conduction path is obtained by analyzing the distribution time series of the molten pool temperature field, presenting the heat propagation trajectory from the molten pool to the welded parts over time in the form of images or data sequences; the cooling curve distribution is based on the monitoring of the temperature changes of key points in the welding process over time, and the temperature drop curves of different areas are drawn. Multiple heat conduction path and cooling curve distribution samples that match the characteristics of the current first welding part and have qualified welding quality have been collected, providing a rich data basis for the determination of subsequent standards and the generation of tolerance constraints.

[0036] Step S522, randomly select one of the multiple heat conduction paths as the standard heat conduction path, and analyze the concentration range of the heat conduction deviation between two adjacent points of the remaining standard heat conduction paths relative to the standard heat conduction path, and generate the first tolerance constraint. Specifically, from the multiple qualified heat conduction paths collected, one of the paths is randomly selected as the standard heat conduction path. The selection is random, but in the subsequent analysis, it will be corrected and verified by comparing with other paths. Its representativeness. For the remaining heat conduction paths, a mathematical calculation method is used to determine the heat conduction deviation between the two adjacent points on the path. For example, the actual distance deviation between the two points (reflecting the deviation of the conduction direction of heat in space), the rate of change of the temperature difference (reflecting the rate deviation of heat transfer), and the change of the heat flux density between the two points are calculated. By performing statistical analysis on the deviation parameters corresponding to the two adjacent points on all the remaining paths, its concentration trend and dispersion are determined. Statistics such as mean and standard deviation can be used to describe the concentration range of deviations. For example, the mean and standard deviation of the distance deviations between all two adjacent points are calculated, and the mean is used as the central trend of the deviation. The standard deviation is multiplied by an appropriate coefficient (such as 2 or 3, determined based on statistical experience and actual data distribution) as the acceptable range of deviations, thereby generating the first tolerance constraint. The constraint defines the range of heat conduction deviations allowed between two adjacent points on the heat conduction path under normal welding quality conditions, providing a quantitative reference standard for subsequent judgment of heat conduction anomalies in the actual welding process.

[0037] Step S523, randomly select one of the multiple cooling curve distributions as the standard cooling curve distribution, and analyze the cooling rate deviation concentration range between two adjacent points of the remaining cooling curve distribution relative to the standard cooling curve distribution to generate the second tolerance constraint. Specifically, randomly determine one of the multiple qualified cooling curve distributions as the standard cooling curve distribution, and the curve will be used as the benchmark for subsequent comparative analysis. For the remaining cooling curve distributions, analyze the cooling rate between two adjacent points on the curve. The cooling rate is calculated by taking the first-order derivative of the change in temperature over time, that is, the temperature drop value per unit time. Calculate the difference in cooling rate between two adjacent points of each remaining cooling curve and the standard cooling curve at the same time point, and then perform statistical analysis on the difference. Similar to the analysis method of heat conduction deviation, calculate the mean and standard deviation of the cooling rate difference to determine the concentration range of the deviation, and then generate the second tolerance constraint. For example, if the difference between the cooling rates of two adjacent points in a certain area of ​​the cooling curve at a certain time point is within the second tolerance constraint range compared with the standard curve, the cooling rate of the area is considered normal; if it is out of range, there may be abnormal cooling, such as too fast cooling leading to thermal stress concentration or too slow cooling affecting the microstructure and properties of the welded joint. The second tolerance constraint provides an accurate quantitative basis for judging the cooling uniformity and stability during the welding process, which helps to timely discover potential welding quality problems and ensure the quality and reliability of welded products.

[0038] In a possible implementation, a quality detection abnormality signal is generated for the abnormal heat accumulation area and the abnormal heat dissipation area, and step S550 further includes step S551, establishing a first local abnormality vector of the abnormal heat accumulation area relative to other areas in the heat conduction path. Specifically, after the abnormal heat accumulation area is determined, the entire heat conduction path is first divided into regions to clarify the boundaries and ranges of the abnormal area and other normal areas. A coordinate system is established with the geometric center of the heat conduction path as the origin, and for each key point in the abnormal heat accumulation area (such as the temperature peak point, the temperature gradient mutation point, etc.), its position vector relative to the origin and other representative points in the normal area is calculated. At the same time, the characteristic parameters of the abnormal area are extracted, such as the average temperature, temperature variance, heat flux, etc., and these parameters are combined with the position vector to form a multi-dimensional vector representation, namely the first local abnormality vector. For example, if the center position coordinates of the abnormal area are ( ), whose average temperature is , the temperature variance is , the relative position vector relative to a reference point in a normal area is ( ), then the first local outlier vector can be expressed as In this way, the unique characteristics and relative position relationship of the abnormal heat accumulation area in the entire heat conduction path are comprehensively described, providing key quantitative information for subsequent abnormal analysis and defect type judgment.

[0039] Step S552, establish a second local anomaly vector for the abnormal heat loss area relative to other areas in the cooling curve distribution. Specifically, for the abnormal heat loss area, an analysis is performed based on the time and temperature dimensions of the cooling curve distribution. In the cooling curve distribution, the time period and temperature range corresponding to the abnormal area are determined, and representative time points and temperature values ​​in the abnormal area and other normal areas are selected. The difference values ​​of the temperature change rate (cooling rate), temperature integral (indicating the difference in the total amount of heat loss) and other aspects of each point in the abnormal area relative to the reference point in the normal area are calculated, and the second local anomaly vector is constructed in combination with the position information of the abnormal area in the cooling curve (such as the start time, duration, etc.). For example, if the start time of the abnormal area in the cooling curve is , and the end time is The difference between the average cooling rate during this period and the normal area is , the difference in the total amount of heat loss is , the relative time position relative to a reference time point in the normal area is , then the second local outlier vector can be expressed as [ ], this vector describes in detail the abnormal characteristics and relative conditions of the abnormal heat loss area from the perspective of the cooling curve, which helps to further analyze the nature of cooling anomalies in the welding process.

[0040] Step S553, based on the first local abnormal vector and the second local abnormal vector, traverse and compare in the local abnormal vector database, screen the defect types whose similarity with the first local abnormal vector and the second local abnormal vector meets the preset similarity, and generate the quality detection abnormal signal. Specifically, the constructed first local abnormal vector and the second local abnormal vector are input into the local abnormal vector database. In the database, a similarity calculation algorithm (such as Euclidean distance algorithm, cosine similarity algorithm, etc.) is used to calculate the similarity between the input first local abnormal vector and the first local abnormal vector samples of each defect type in the database, and the similarity between the second local abnormal vector and the corresponding second local abnormal vector samples. When both similarities meet the preset similarity threshold (this threshold is determined by combining the statistical analysis of historical data with the actual welding quality requirements, for example, for key welding parts, the similarity threshold is set higher to ensure the accuracy of detection; for general welding parts, the threshold can be appropriately relaxed to balance the detection efficiency and accuracy), it can be determined that there are quality problems similar to the sample defect type in the current welding process. At this time, the system generates a corresponding quality detection abnormality signal according to the matched defect type. The signal may include detailed information of the defect type (such as name, possible severity, etc.), location information of the abnormal area (determined by the location parameters in the first local abnormality vector and the second local abnormality vector), and related processing suggestions (such as recommended welding process adjustment parameters, repair methods, etc.), so that welding operators and quality management personnel can understand welding quality problems in a timely and accurate manner, and take effective measures to deal with them, thereby ensuring the quality and reliability of welding products.

[0041] Step S554, wherein the local abnormal vector database contains first local abnormal vector samples and second local abnormal vector samples corresponding to a plurality of defect type samples. Specifically, the database is established by analyzing and extracting features from a large number of historical welding defect cases, wherein the first local abnormal vector samples and second local abnormal vector samples corresponding to a plurality of known defect types (such as lack of fusion, incomplete penetration of welds, cracks caused by uneven cooling, etc.) are stored.

[0042] The embodiment of the present application determines multiple welding parts to be detected, and continuously detects the molten pool temperature through the thermal imaging module during the welding process to generate a time series of the molten pool temperature field distribution. Based on the time series data, heat conduction analysis and cooling rate identification are performed to generate heat conduction paths and cooling curve distributions. By analyzing these data, it is determined whether there are abnormal areas of heat accumulation or loss. If there are abnormalities, a quality detection signal is issued, so that the temperature distribution and cooling rate changes during the welding process are monitored in real time, and the welding defects caused by uneven heat conduction or improper cooling during the welding process are accurately identified, achieving the technical effect of improving the detection accuracy and timeliness of welding quality.

[0043] In the above, refer to Figure 1 A welding quality detection method according to an embodiment of the present invention is described in detail. Figure 2 A welding quality inspection system according to an embodiment of the present invention is described.

[0044] A welding quality detection system according to an embodiment of the present invention is used to solve the technical problem that the existing welding quality detection method cannot accurately identify welding defects caused by uneven heat conduction and abnormal cooling rate during welding in real time. By real-time monitoring of the temperature distribution and cooling rate changes during welding, welding defects caused by uneven heat conduction or improper cooling during welding can be accurately identified, achieving the technical effect of improving the detection accuracy and timeliness of welding quality. A welding quality detection system includes: a preset welding position determination module 10, a temperature continuous detection module 20, a heat conduction analysis module 30, a cooling rate identification module 40, and a quality detection abnormal signal sending module 50.

[0045] The preset welding position determination module 10 is used to determine a plurality of preset welding positions of the target to be detected.

[0046] The temperature continuous detection module 20 is used to continuously detect the molten pool temperature through the thermal imaging module during the welding process of the first welding position among the multiple preset welding positions, and generate a molten pool temperature field distribution time series.

[0047] The heat conduction analysis module 30 is used to perform heat conduction analysis based on the molten pool temperature field distribution time series to generate a heat conduction path.

[0048] The cooling rate identification module 40 is used to identify the cooling rate based on the molten pool temperature field distribution time series and generate a cooling curve distribution.

[0049] The quality detection abnormal signal issuing module 50 is used to determine whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution, and if so, issue a quality detection abnormal signal.

[0050] Next, the specific configuration of the heat transfer analysis module 30 will be described in detail. As described above, heat conduction analysis is performed based on the molten pool temperature field distribution sequence to generate a heat conduction path. The heat conduction analysis module 30 further includes: a temperature field extraction unit, the temperature field extraction unit is used to extract the first time sequence molten pool temperature field, the second time sequence molten pool temperature field, the third time sequence molten pool temperature field, and the Nth time sequence molten pool temperature field in the molten pool temperature field distribution sequence; a heat diffusion path identification unit, the heat diffusion path identification unit is used to identify the heat diffusion paths of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field under adjacent time sequences to establish a first diffusion path; a second diffusion path establishment unit, the second diffusion path establishment unit is used to identify the heat diffusion paths of the second time sequence molten pool temperature field and the third time sequence molten pool temperature field under adjacent time sequences to establish a second diffusion path, and so on, until the Nth time sequence molten pool temperature field is traversed and the N-1th diffusion path is established; a heat conduction path generation unit, the heat conduction path generation unit is used to connect the first diffusion path, the second diffusion path, and the N-1th diffusion path to generate the heat conduction path.

[0051] Among them, the heat diffusion paths of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field are identified under adjacent time sequences to establish a first diffusion path, and the heat diffusion path identification unit further includes: a temperature gradient construction subunit, the temperature gradient construction subunit is used to compare the first time sequence molten pool temperature field and the second time sequence molten pool temperature field, identify the temperature changes in the center, boundary and heat affected zone of the molten pool, and establish a temperature gradient; a heat conduction identification subunit, the heat conduction identification subunit is used to identify the heat conduction direction and diffusion range based on the temperature gradient; a first diffusion path generation subunit, the first diffusion path generation subunit is used to generate the first diffusion path according to the heat conduction direction and diffusion range.

[0052] The specific configuration of the cooling rate identification module 40 will be described in detail below. As described above, the cooling rate identification is performed based on the molten pool temperature field distribution time series to generate a cooling curve distribution. The cooling rate identification module 40 further includes: a rate change curve calculation unit, the rate change curve calculation unit is used to calculate the rate change curve of the temperature of each point over time based on the molten pool temperature field distribution time series, and generate the temperature change curve of each point; a consistency aggregation unit, the consistency aggregation unit is used to perform consistency aggregation on the temperature change curve of each point, and establish the temperature change curve of each key aggregation area; a cooling curve distribution generation unit, the cooling curve distribution generation unit is used to generate the cooling curve distribution with the temperature change curve of each key aggregation area.

[0053] The specific configuration of the quality detection abnormal signal issuing module 50 will be described in detail below. As described above, it is determined whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution. If so, a quality detection abnormal signal is issued. The quality detection abnormal signal issuing module 50 further includes: a first welding part feature determination unit, the first welding part feature determination unit is used to determine the first welding part feature of the first welding part, wherein the first welding part feature includes the metal type of the welding part, the structural feature of the welding part and the continuity direction feature of the weld; a first tolerance constraint acquisition unit, the first tolerance constraint acquisition unit is used to collect the standard heat conduction path and the corresponding first tolerance when the welding quality is qualified based on the first welding part feature as a constraint. constraint, and a standard cooling curve distribution and a corresponding second tolerance constraint; a first tolerance constraint positioning unit, the first tolerance constraint positioning unit is used to compare the heat conduction path and the standard heat conduction path, locate the area that does not meet the first tolerance constraint, and generate the heat accumulation abnormal area; a second tolerance constraint positioning unit, the second tolerance constraint positioning unit is used to compare the cooling curve distribution and the standard cooling curve distribution, locate the area that does not meet the second tolerance constraint, and generate the heat loss abnormal area; a quality detection abnormal signal generating unit, the quality detection abnormal signal generating unit is used to generate a quality detection abnormal signal for the heat accumulation abnormal area and the heat loss abnormal area.

[0054] Among them, taking the first welding part feature as a constraint, collecting the standard heat conduction path and the corresponding first tolerance constraint when the welding quality is qualified, as well as the standard cooling curve distribution and the corresponding second tolerance constraint, the first tolerance constraint collection unit further includes: a data collection subunit, the data collection subunit is used to collect multiple heat conduction paths and multiple cooling curve distributions when the welding quality is qualified with the first welding part feature as a constraint; a first tolerance constraint generation subunit, the first tolerance constraint generation subunit is used to randomly select one of the multiple heat conduction paths as the standard heat conduction path, and analyze the heat conduction deviation concentration range between two adjacent points of the remaining standard heat conduction paths relative to the standard heat conduction path to generate the first tolerance constraint; a second tolerance constraint generation subunit, the second tolerance constraint generation subunit is used to randomly select one of the multiple cooling curve distributions as the standard cooling curve distribution, and analyze the cooling rate deviation concentration range between two adjacent points of the remaining cooling curve distribution relative to the standard cooling curve distribution to generate the second tolerance constraint.

[0055] Wherein, for the abnormal heat accumulation area and the abnormal heat dissipation area, a quality detection abnormal signal is generated, and the quality detection abnormal signal generating unit further includes: a first local abnormal vector constructing subunit, the first local abnormal vector constructing subunit is used to establish a first local abnormal vector of the abnormal heat accumulation area relative to other areas in the heat conduction path; a second local abnormal vector constructing subunit, the second local abnormal vector constructing subunit is used to establish a second local abnormal vector of the abnormal heat loss area relative to other areas in the cooling curve distribution; a quality detection abnormal signal generating subunit, the quality detection abnormal signal generating subunit is used to traverse and compare in a local abnormal vector database based on the first local abnormal vector and the second local abnormal vector, screen defect types whose similarity with the first local abnormal vector and the second local abnormal vector meets a preset similarity, and generate the quality detection abnormal signal; a local abnormal vector database composing subunit, the local abnormal vector database composing subunit is used, wherein the local abnormal vector database contains first local abnormal vector samples and second local abnormal vector samples corresponding to multiple defect type samples.

[0056] A welding quality detection system provided by an embodiment of the present invention can execute a welding quality detection method provided by any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0057] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0058] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A welding quality detection method, characterized in that: include: Determine a plurality of preset welding positions of the target to be inspected; During welding of a first welding position among the plurality of preset welding positions, continuously detecting the temperature of the molten pool by means of a thermal imaging module to generate a time series of the distribution of the temperature field of the molten pool; Performing heat conduction analysis based on the molten pool temperature field distribution time series to generate a heat conduction path; Identify the cooling rate based on the molten pool temperature field distribution time series and generate a cooling curve distribution; It is determined whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution. If so, a quality detection abnormality signal is issued.

2. A welding quality detection method as claimed in claim 1, characterized in that: Based on the molten pool temperature field distribution time series, heat conduction analysis is performed to generate a heat conduction path, including: Extracting the first time series molten pool temperature field, the second time series molten pool temperature field, the third time series molten pool temperature field, and up to the Nth time series molten pool temperature field in the molten pool temperature field distribution time series; Identify heat diffusion paths of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field in adjacent time sequences to establish a first diffusion path; Identify the heat diffusion paths of the second time sequence molten pool temperature field and the third time sequence molten pool temperature field under adjacent time sequences, establish a second diffusion path, and so on, until the Nth time sequence molten pool temperature field is traversed to establish an N-1th diffusion path; The first diffusion path, the second diffusion path, and the N-1th diffusion path are connected to generate the heat conduction path.

3. A welding quality detection method as claimed in claim 2, characterized in that: The heat diffusion paths of the first time sequence molten pool temperature field and the second time sequence molten pool temperature field under adjacent time sequences are identified to establish a first diffusion path, including: Comparing the first time-series molten pool temperature field with the second time-series molten pool temperature field, identifying the temperature changes at the center, boundary and heat-affected zone of the molten pool, and establishing a temperature gradient; Identifying the heat conduction direction and diffusion range based on the temperature gradient; The first diffusion path is generated according to the heat conduction direction and diffusion range.

4. A welding quality detection method as claimed in claim 1, characterized in that: Based on the molten pool temperature field distribution time series, cooling rate identification is performed to generate cooling curve distribution, including: Based on the molten pool temperature field distribution time series, calculating the rate change curve of the temperature of each point over time, and generating the temperature change curve of each point; Perform consistency aggregation on the temperature change curves of each point to establish the temperature change curves of each key aggregation area; The cooling curve distribution is generated by the temperature variation curve of each key polymerization zone.

5. A welding quality detection method as claimed in claim 1, characterized in that: Judging whether there is an abnormal heat accumulation area or an abnormal heat loss area according to the heat conduction path and the cooling curve distribution, and if so, issuing a quality detection abnormality signal, including: Determining first welding position characteristics of the first welding position, wherein the first welding position characteristics include welding component metal type, welding component structural characteristics and weld continuity direction characteristics; Taking the first welding part feature as a constraint, collecting a standard heat conduction path and a corresponding first tolerance constraint when the welding quality is qualified, as well as a standard cooling curve distribution and a corresponding second tolerance constraint; Comparing the heat conduction path with the standard heat conduction path, locating an area that does not meet the first tolerance constraint, and generating the abnormal heat accumulation area; Comparing the cooling curve distribution with the standard cooling curve distribution, locating the area that does not meet the second tolerance constraint, and generating the heat loss abnormal area; A quality detection abnormality signal is generated for the abnormal heat accumulation region and the abnormal heat dissipation region.

6. A welding quality detection method as claimed in claim 5, characterized in that: Taking the first welding part feature as a constraint, collecting the standard heat conduction path and the corresponding first tolerance constraint when the welding quality is qualified, as well as the standard cooling curve distribution and the corresponding second tolerance constraint, including: Taking the first welding part feature as a constraint, collecting multiple heat conduction paths and multiple cooling curve distributions when the welding quality is qualified; Randomly selecting one of the multiple heat conduction paths as the standard heat conduction path, and analyzing the concentration range of heat conduction deviations between two adjacent points of the remaining standard heat conduction paths relative to the standard heat conduction path to generate the first tolerance constraint; One of the plurality of cooling curve distributions is randomly selected as the standard cooling curve distribution, and a cooling rate deviation concentration range between two adjacent points of the remaining cooling curve distributions relative to the standard cooling curve distribution is analyzed to generate the second tolerance constraint.

7. A welding quality detection method as claimed in claim 6, characterized in that: For the abnormal heat accumulation area and the abnormal heat loss area, generating a quality detection abnormal signal includes: Establishing a first local abnormal vector of the abnormal heat accumulation region relative to other regions in the heat conduction path; Establishing a second local abnormal vector of the heat loss abnormal area relative to other areas in the cooling curve distribution; Based on the first local abnormality vector and the second local abnormality vector, a traversal comparison is performed in a local abnormality vector database, and a defect type whose similarity with the first local abnormality vector and the second local abnormality vector meets a preset similarity is screened, and the quality detection abnormality signal is generated; The local abnormal vector database includes first local abnormal vector samples and second local abnormal vector samples corresponding to various defect type samples.

8. A welding quality detection system, characterized in that: The system is used to implement a welding quality detection method according to any one of claims 1 to 7, and the system comprises: A preset welding position determination module, wherein the preset welding position determination module is used to determine a plurality of preset welding positions of a target to be detected; A temperature continuous detection module, wherein the temperature continuous detection module is used to continuously detect the temperature of the molten pool through a thermal imaging module during the welding process of the first welding position among the plurality of preset welding positions, and generate a time series of the temperature field distribution of the molten pool; A heat conduction analysis module, the heat conduction analysis module is used to perform heat conduction analysis based on the molten pool temperature field distribution time series to generate a heat conduction path; A cooling rate identification module, the cooling rate identification module is used to identify the cooling rate based on the molten pool temperature field distribution time series and generate a cooling curve distribution; The quality detection abnormal signal issuing module is used to determine whether there is an abnormal heat accumulation area or an abnormal heat loss area based on the heat conduction path and the cooling curve distribution, and if so, issue a quality detection abnormal signal.

Images