Method for determining the quality of a C-image acquired using ultrasound technology

Abstract

The present invention relates to a method for determining the quality or classifying a C-image (10) from ultrasonic imaging. The method comprises several steps. In a first step, the method involves receiving a C-image (10) of a weld area (11) to be examined on the workpiece. Subsequently, a first parameter is determined, which corresponds to the ratio between a welded area (13) and a potentially weldable area within a test circle (12) of the weld area (11) defined in the C-image (10). In a further process step, a second parameter is determined, which corresponds to a ratio between an area of ​​the welded area (13) within the test circle (12) and an area of ​​the welded area (13) of a main contour (16) of the weld nugget (15).Furthermore, a third parameter is determined, which corresponds to a measure of the roundness of the detected main contour (16) of the weld nugget (15). In a further process step, a fourth parameter is determined, which represents a measure of the maximum possible weld nugget diameter when using a welding cap and a cover plate inserted during the creation of the weld joint (11). Finally, a quality parameter is determined, which depends on the first, second, third, and fourth parameters and allows a statement about the quality of the C-scan.

Description

[0001] The present invention relates to a method for determining the quality or classifying a C-image from ultrasound imaging technology.

[0002] A widely used imaging and non-destructive testing technique today is the so-called single-transducer technique. Imaging ultrasonic testing is a further development of conventional single-transducer technology and enables detailed visualization of test objects. Three main methods are distinguished: phased array testing, phased array testing, and scanning single-transducer technology. All imaging ultrasonic techniques are based on the pulse-echo method, which uses a coupling medium for sound transmission.

[0003] In this process, sound pulses are coupled into the workpiece being inspected, for example, a welded body panel, using a probe, and their echo signal is evaluated. To accelerate the process, phased array transducers can be used, in which several individual transducers are housed in a single probe. These individual transducers can be activated sequentially or in groups. The evaluation and visualization of the numerous individual A-scans (amplitudes) is performed using the C-scan (top view), in which amplitudes or sound travel times are color-coded for imaging purposes. A variety of color palettes and display methods are available. The image represents a grid of the individual transducers. In the case of inspecting spot welds on body panels, the diameter of the weld nuggets can be determined using this grid.

[0004] The C-scan and the resulting weld nugget diameter can exhibit a relatively high measurement deviation compared to the actual values ​​when the weld surface is not optimal. This deviation can be significantly influenced by the inspector by ensuring the optimal contact of the probe with the weld under test. To improve quality and reduce testing effort, it is crucial to minimize the measurement deviation as much as possible.

[0005] A key concept in the inspection of welded joints is the weld nugget. A weld nugget refers to the molten and resolidified portion of a weld. It forms the load-bearing connection between the welded workpieces. The weld nugget is created by the heat generated during the flow of the electric current. As the welding time increases, the nugget grows in size and diameter. Under optimal parameters, it reaches its maximum size before the electrodes penetrate the workpiece. Ideally, the weld nugget is circular when viewed from above. The cross-sectional shape of the weld nugget is typically lenticular, although the exact geometry can vary depending on the material and welding parameters. In steel, a cylindrical heat-affected zone often forms, while in other materials such as aluminum, a more lenticular spread is more common.The size and characteristics of the weld nugget are crucial for the joint's strength. Maximum strength is generally achieved below the spatter limit. A high-quality weld nugget should be free of inclusions, pores, or cracks. Such irregularities, especially if they are off-center, can negatively affect the joint's mechanical properties. A heat-affected zone forms around the weld nugget, the size of which depends on the amount of heat applied. An excessively large heat-affected zone can impair the joint's strength.

[0006] Using multi-element ultrasonic matrix arrays, an ultrasonic device acquires data from the surface and internal structure of the weld pool. Special algorithms use this data to reconstruct an image of the weld pool and estimate its average diameter in real time. This estimation is performed within the framework of an optimization problem in the C-scan. The circle defined by the average diameter is referred to as the test circle in this application and defines a measure of the weld pool.

[0007] In the case of an ideal weld bead, the test circle corresponds in shape to the typically circular weld bead. It is positioned in the C-scan such that the maximum proportion of the welded area lies within the circle, while simultaneously minimizing the unwelded areas within the test circle. Depending on the quality of the imaging process and / or the quality of the weld, the area calculated (or graphically derived from the C-scan) thus represents a more or less precise substitute for the actual area of ​​the measured weld bead. Within the framework of the method described here, the marking of the assumed weld bead contained in the C-scan, which was previously determined by the software of the ultrasonic imaging device (i.e., the test circle), can be detected using image recognition.

[0008] Document KR 10 2020 0 041 647 A concerns a method for the automatic verification of weld quality. This is based on generated images of the weld point. Two different circles are created to characterize the weld joint, and their sizes are compared. The evaluation also considers the shape of the weld joint. Furthermore, the detection of inclusions or similar defects is factored into the assessment of the welding result.

[0009] Document US 2018 / 0259489A1 describes a system for testing a welded joint of a structure. This is also based on cross-sectional images of the weld. Document WO 2021 / 076358A1 also describes a method for evaluating a welded joint using C-scanning. This method determines the minimum geometric dimensions of the welded area by defining a circle or ellipse enclosed by a predefined threshold contour and making it as large as possible.

[0010] Document EP 4 296 662 A1 describes a method for the ultrasonic detection of defects in a welded joint. This involves defining a boundary around a potential defect and further examining the defect. Document US 2003 / 0 234 239 A1 also concerns the examination of weld quality using ultrasound-generated images. The images produced in this way are analyzed, particularly using AI, and their quality is evaluated. According to document EP 4 012 401 A2, C-scans are also created for the examination of welded joints. These are divided into pixels, and the quality of the joint is determined based on the pixel analysis. The teaching in document DE 10 2013 106 901 A1 relates to a method for the non-destructive testing of a test specimen for material defects using ultrasound.

[0011] Document US 9,733,219 B2 describes an automated system for the non-destructive testing of spot welds using ultrasound. The core component is a robot-mounted matrix phased array probe whose curved shape and flexible, fluid-filled tip adapt to the weld contour. Software analyzes the received signals and provides a clear pass / fail rating of the weld quality, enabling reliable, automated inspection.

[0012] The publication EP 3 539 683 A1 discloses a monitoring method for the production of resistance-welded steel pipes. Real-time images of the welding process are captured using a high-speed camera to detect defects caused by foreign materials such as scale. An image processing unit identifies these defects by detecting irregular arcs at or before the weld point. This enables the subsequent marking or tracking of the defective pipe sections.

[0013] In light of the aforementioned prior art, the object of the present invention can be seen as determining the quality of the data supplied by a measuring device of ultrasound imaging technology.

[0014] This problem is solved by means of the subject matter of the independent patent claim. Further preferred embodiments are found in the dependent patent claims.

[0015] According to the invention, a method for determining the quality of a C-image of a workpiece recorded using ultrasound technology is provided.

[0016] The procedure comprises several steps. The first step involves receiving a C-image of the weld area to be inspected on the workpiece. "Receiving" can refer to reading the C-image into software configured for executing this procedure. The term "C-image" used in this description should be understood according to its established meaning in the prior art and refers to an echo height image of a workpiece generated by ultrasonic testing.

[0017] In a further process step, a first parameter is determined, which corresponds to the ratio between a welded area and a potentially weldable area within a test circle of the weld defined in the C-scan. Whether a pixel within the C-scan represents a welded or unwelded area of ​​the workpiece depends on the distance between the two surfaces and / or the position of their interfaces relative to each other and is usually determined automatically by the software of the ultrasonic device used to perform the inspection.

[0018] In a further process step, a second parameter is determined, which corresponds to the ratio between the area of ​​the welded region within the test circle and the area of ​​the welded region of a main contour of the weld nugget. The main contour represents the largest contiguous welded area in the C-scan. The second parameter is thus a measure of what proportion of the welded area lies within the test circle.

[0019] Furthermore, in a subsequent process step, a third parameter is determined, corresponding to a measure of the roundness of the detected main contour of the weld nugget. The roundness is a measure of the measurement uncertainty of the weld nugget. A "non-round" C-scan has a higher measurement uncertainty than a round C-scan.

[0020] Preferably, a roundness index is determined based on the isoperimetric inequality. The isoperimetric inequality states that for every closed plane curve: 4πA≤U2 where A is the enclosed area and U is the circumference of the curve. The equals sign only holds true for a perfect circle. Based on this inequality, a roundness index R can be defined: R=4πAU2

[0021] This index has the following properties: R ≤ 1 for all forms R = 1 only for a perfect circle The closer R is to 1, the rounder the shape.

[0022] The method is independent of the object's size. It takes into account both local and global deviations from a circular shape. However, the index provides no information about the type of deviation (e.g., oval, angular). Very small local irregularities can significantly affect the perimeter without substantially changing the area. This mathematical method for estimating roundness offers a simple and effective way to quantify an object's overall shape, but it is less precise than specialized measurement techniques.

[0023] In a further step of the inventive process, a fourth parameter is determined, which represents a measure of the maximum possible weld nugget diameter when using a welding cap and a cover plate employed in the creation of the weld joint. The maximum possible weld nugget diameter is the diameter of an ideally formed, and therefore circular, weld nugget.

[0024] Subsequently, according to the inventive method, a quality parameter is determined which depends on the first, second, third, and fourth parameters and allows a statement about the quality of the C-scan. Preferably, all parameters, including the quality parameter, are normalized to a range, in particular between 0 and 1, so that a simple, especially weighted, calculation of the individual parameters leads to the quality parameter. Preferably, the quality parameter is determined by averaging the individual parameters.

[0025] In an advantageous embodiment, the determination of the first parameter is carried out taking into account inclusions in the weld nugget within the test circle as an unwelded area. Since inclusions in the weld nugget indicate that no material bond has occurred between the welding partners, these areas are to be considered unwelded.

[0026] A further preferred embodiment of the method according to the invention is one in which, when determining the fourth parameter, the measure for the maximum possible weld lenticular diameter is normalized to a range between 0 and 1 by the sigmoid function. The sigmoid function is a known mathematical function and has an S-shaped graph that ranges between 0 and 1. Experimental tests have shown that a slope of -10 for the sigmoid function is particularly advantageous for determining the fourth parameter.

[0027] The method according to the invention is explained in more detail below with reference to the accompanying figures. The figures show: Fig. Figure 1a shows an example C-image of a weld. Fig. Figure 1b shows a processing of the C-scan for the determination of the first parameter. Fig. Figure 1c shows a processing of the C-scan for the determination of the second parameter. Fig. Figure 1d shows a processing of the C-scan for the determination of the third parameter.

[0028] In Fig. Figure 1 shows an exemplary C-10 image of a weld 11. The C-image 10 also shows the computationally or graphically determined test circle 12 of the weld nugget 15, which was inserted into the C-image by the software of the associated ultrasonic imaging device. The computationally determined test circle 12 has a circular shape. Furthermore, it can be seen that a large portion of a welded area 13 lies within the determined test circle 12, and this area also corresponds topologically to a contiguous region. Outside the test circle 12, further smaller, distributed welded areas 13 can be seen. The welded area 13 that lies almost entirely within the test circle represents the main contour 16 of the welded area 13.

[0029] How well the diameter of the test circle 12 corresponds to the diameter of the actual weld nugget depends, among other things, on the quality of the C-image 10 and the quality of the actual weld nugget 11 on the workpiece. Generally, the better the quality of the C-image 10, the better the ultrasonic software can determine the size and position of the weld nugget 15 within the C-image. Similarly, only slight variations occur when the weld nugget 15 is repeatedly detected in the same C-image 10. Therefore, determining and classifying the quality of the C-image is of particular importance when inspecting the weld nuggets 11.

[0030] The in Fig.The C-scan shown in Figure 1a is received according to the first process step. Subsequently, a first parameter P1 is determined, which corresponds to the ratio between a welded area 13 and a potentially weldable area within the test circle 12 of the weld 11. Inclusions in the welded area are preferably taken into account. Mathematically, the first parameter P1 can therefore be determined as follows: P1=AvB,innen−AnvB,innen+AEinshclussARSW A where A vB,innen the area of ​​the welded region 13 within the test circle 12, A nvB,innen the area of ​​an unwelded region 14 within the test circle 12 and A RSWA corresponds to the area of ​​test circle 12.

[0031] Furthermore, a second parameter P2 is determined, which corresponds to a ratio between the area of ​​the welded region 13 within the test circle 12 and the area of ​​the welded region 13 of a main contour of the weld nugget 15. The main contour represents the largest contiguous welded area in C-Figure 10. The second parameter is thus a measure of what proportion of the welded area lies within the test circle.

[0032] The calculation of the second parameter P2 can therefore be determined as follows: P2=AvB,innerAHK where A vB,innen the area of ​​the welded region 13 within the test circle 12 and A HK corresponds to the area of ​​the main contour 16.

[0033] The determination of the third parameter P3, which corresponds to a measure of the roundness of the identified main contour of the weld nugget. This is preferably done, as already described, using the isoperimetric inequality. The third parameter P3 can thus be calculated as follows: P3=4*π*AHKUHK2 where A HK the area of ​​the main contour 16 and U HK corresponds to the extent of the main contour 16.

[0034] A fourth parameter, P4, is then determined, representing a measure of the maximum possible weld nugget diameter when using a welding cap and cover plate during weld creation. This can be calculated using a sigmoid function as follows: P4=11+e−a(d−(6.5*tcover plate−mdisplacement) where t DeckblechThe thickness of the cover plate at the weld joint is determined by the slope a. Experimental tests have shown that a slope of -10 is particularly advantageous, although other slope values ​​are also conceivable. The maximum possible displacement of the weld bead m is determined by the thickness of the cover plate at the weld joint. Verschiebung It preferably takes the value -0.5, as experiments have shown. However, other shifts are also conceivable. The variable d represents the possible lens diameter. (6.5*t deck plate) determines the vertex of the function. The formulation 6.5*tDeckblech−0.5 This results in an effective area of ​​6mm for a welding cap used in the creation of welding point 11.

[0035] Subsequently, in an advantageous embodiment, the individual parameters P1 to P4 are combined into a single quality parameter by averaging. The closer this parameter is to 1, the better the quality of the C-scan 10. REFERENCE MARK LIST 10 C-scan 11 Welding point 12 Test Circle 13 welded area 14 non-welded area 15 welding lens 16 Main contour A vB,innen Welded area, inside A nvB,innen Area not welded, inside A Prüf Area of ​​the test circle A HK Area of ​​the main contour U HK Scope of the main contour

Claims

[1] Method for determining the quality of a C-image (10) of a workpiece recorded using ultrasound technology, comprising (A) Receiving a C-image (10) of a weld area (11) to be examined on the workpiece; (B) Determining a first parameter which corresponds to the ratio between a welded area (13) and a potentially weldable area within a test circle (12) of the weld (11) defined in the C-image (10); (C) Determining a second parameter which corresponds to a ratio between an area of ​​the welded region (13) within the test circle (12) and an area of ​​the welded region (13) of a main contour (16) of a weld nugget (15); (D) Determining a third parameter, which corresponds to a measure of the roundness of the detected main contour (16) of the weld lens (15); (E) Determining a fourth parameter which represents a measure of a maximum possible weld nugget diameter using a welding cap and cover plate used in the production of the weld (11); (F) Determining the quality of the C-image (10) as a function of a quality parameter which depends on the first, the second, the third and the fourth parameter. [2] Method according to the preceding claim, wherein when determining the first parameter, inclusions in the weld nugget (15) within the test circle (12) are evaluated as an unwelded area (14). [3] Method according to one of the preceding claims, wherein to determine the third parameter a roundness index is determined which is based on the isoperimetric inequality. [4] Method according to one of the preceding claims, wherein when determining the fourth parameter the measure for the maximum possible weld lens diameter is normalized to a range between 0 and 1 by the sigmoid function. [5] Method according to the preceding claim, wherein the slope of the sigmoid function is set to -10. [6] Method according to one of the preceding claims, wherein the quality parameter is determined by averaging the first, second, third and fourth parameters.

Images