Method and apparatus for evaluating the welding quality of a weld between end segments of two conductor elements

Abstract

The present invention relates to a method (40) and an apparatus (200) for evaluating the welding quality of a weld (14) between end segments of at least two elongate conductor elements (12), which conductor elements are welded to each other by irradiating end faces (13) of the end segments.

Description

[0001] The present invention relates to a method and a device for evaluating the welding quality of a weld seam, wherein end segments of at least two elongate conductor elements are welded to one another by irradiating end faces of the end segments through the weld seam.

[0002] Methods and devices for producing stators are known in the prior art. Stators are used in electric machines or electromechanical converters, for example in electric motors.

[0003] In the production of electromechanical transducers, in particular electric motors, for traction drives in particular, individual winding elements (insert coils, so-called "hairpins") are produced, which are further processed into a stator winding during further processes.

[0004] In order to achieve higher efficiency of an electric machine through a higher filling degree, a change from a circular conductor cross-section to a rectangular conductor cross-section has been made in hairpin technology. Within the scope of the present invention, any conductor cross-section can be used, i.e. for example a circular cross-section or a rectangular cross-section.

[0005] In the context of producing winding elements, corresponding conductor segments are separated from a continuous material, for example by cutting. The conductor segments are then stripped of insulation at the parts to be welded at a later step. The conductor segments can then be formed into suitable winding elements, for example the conductor segments can be bent into corresponding hairpins. A hairpin typically has two legs which are connected to one another by a connecting part (similar to the closed end of a U-shaped hairpin). On the side facing away from this connecting part, the legs each have a free end. The legs are typically parallel to one another and each have a stripped part in the region of their ends, where the hairpins are welded later.

[0006] A plurality of hairpins can be joined together to form a basket (stator basket), and optionally, paper insulation can be inserted into the stator. The hairpins or the basket are then inserted into the stator, and if necessary, their ends are twisted against one another ("twisting"), which allows the hairpins to be held in a positive fit in the stator core in a certain way. The winding elements or hairpins can be welded to one another at their ends after being positioned on the stator. In order to make this welding safe and reproducible, the hairpins to be welded together should be positioned relative to one another in precisely defined positions to ensure their desired mutual contact.

[0007] The welding itself is usually accomplished using laser radiation. For this purpose, a corresponding laser beam source and a so-called scanner optics with a camera are typically used. The scanner optics enables the camera to detect the position of the ends of the conductor elements before welding, and the scanner optics also uses two movable mirrors to enable the laser beam to move highly dynamically on the end faces of the two conductor elements to be welded. The number of conductor elements to be welded in each stator is typically in the three-digit range. A defective weld can cause a fault in the stator. Since the quality of the weld seam depends on various factors, quality control is necessary. The quality criteria for the weld seam can be determined individually, but they generally have in common a certain connection cross-section necessary to ensure the desired current flow.

[0008] Various methods for quality control of welded hairpins are known. This includes, for example, the cross-section through the weld seam. The disadvantage is that this type of quality control is a destructive test and is expensive and time-consuming. For example, it is also known to use optical coherence tomography or a laser sensor for profilometry to measure the weld beads produced in the weld seam. The disadvantage is that additional technical equipment is required for this, and this measurement is slow and thus expensive. In JP2019120650A, a lateral image of the welded hairpin is also proposed, i.e., when viewed in the plane of the end face. However, this also requires additional and expensive equipment. In addition, only one hairpin weld can be inspected at any one time. Finally, it has been proposed to measure the process light during the welding process. However, it is difficult to establish an exact correlation between the measured optical signal and the welding result.

[0009] In summary, the known quality control methods are disadvantageous in many respects; in particular, they are complex and expensive and error-prone.

[0010] The object on which the present invention is based is to overcome these and other disadvantages of the prior art and to provide the following method and device: with the method and the device, quality control of the weld seam between the end segments of at least two elongated conductor elements can be performed in a simple, cost-effective, rapid, and precise manner.

[0011] This object is achieved by the method according to claim 1 and the device according to claim 14.

[0012] Therefore, the method is a method for evaluating the welding quality of a weld seam between the end segments of at least two elongated conductor elements, which are welded to each other by irradiating (in particular by means of a laser) the end faces of the end segments.

[0013] The method comprises at least the following steps:

[0014] (a) Capture an image of the end faces of the conductor elements welded to each other via the weld seam. Here, the image is recorded from the end faces of the conductor elements, and the area captured by the image includes at least the entire surface of the weld seam on the end faces of the conductor elements. Furthermore, for each pixel of the image, the image is captured at different intensity levels. The intensity level is an indicator of the amount of light locally reflected at the recording position when the image is taken. The indicator or the amount of light is a measure of the local orientation of the surface of the weld seam. In particular, the weld seam is illuminated with light parallel to the direction in which the image is recorded.

[0015] (b) Divide the image into a first subset of pixels and a second subset of pixels, where the intensity levels of the pixels in the first subset are greater than a predetermined intensity level threshold, and the intensity levels of the pixels in the second subset are less than or equal to the intensity level threshold.

[0016] (c) Evaluate the welding quality of the weld seam by comparing the number and / or position of the pixels in the first subset and / or the second subset of pixels with at least one pixel target value and / or a pixel target range of the number and / or position of the pixels in the first subset and / or the second subset of pixels corresponding to a desired quality of the weld seam.

[0017] Thus, the present invention proposes a method for quality control of the welded end sections or ends of two or more conductor elements, by means of which the connection cross-section of the weld seam is evaluated. Quality control is performed by comparing the recorded pixels or the number of the recorded pixels in each subset and / or the position of the recorded pixels in each subset with target values (pixel target values or pixel target ranges). The method is non-destructive, simple, cost-effective, fast, and precise. For example, quality control may require a certain number of pixels in a subset or a certain spatial distribution of pixels in a subset. This will be discussed in more detail below.

[0018] The following explanations equally apply to the method and the device according to the invention.

[0019] According to the invention, an image of the welded end face of the conductor element is captured, in particular by an image capture device, in particular a camera. The image is taken from the end face of the conductor element and thus shows the weld seam when viewed from above or when the end face of the conductor element or conductor section is viewed vertically.

[0020] Viewing from above here means being positioned opposite to the end face. The image capture includes at least the entire weld seam or an area of the desired size of the weld seam at the expected position. Parts of the surface or end face of the conductor element and, if applicable, their surroundings can also be part of the captured image.

[0021] The weld seam is here usually formed in the shape of a (substantially hemispherical) bead. The bead is formed after a spherical or globular molten pool has been generated by irradiating (with a laser) the end section of the conductor element. The solidified bead thus also has a spherical or hemispherical shape.

[0022] The end section of the conductor element extends in the direction of the image-capturing device, and the end face is perpendicular to this direction. Thus, the viewing direction of the image-capturing device is perpendicular to the end face. Usually, the weld seam and its surroundings are illuminated from the direction of the image-capturing device or camera (i.e., in a direction parallel to its beam path), and for this purpose, a lighting device can be provided on the corresponding device or the lighting device can be used as part of the method.

[0023] The illumination light incident on the weld seam or bead is reflected by it and allows the bead or its spherical shape or the shape of the weld seam to be identified in the recorded image. When the illumination light is reflected, for example if the shape of the weld seam is predominantly spherical or hemispherical, a different part of this light is reflected depending on the local position on the surface of the weld seam. Thus, the hemispherical bead appears bright in the center of the captured image and darker towards the edge of the captured image. The light directed perpendicularly onto the end face of the conductor element is reflected to a greater extent in the direction of the image-capturing device for capturing the image in the region of the weld seam that extends parallel to the end face, compared to the light falling on the adjacent angled regions of the weld seam.

[0024] Thus, the weld seam can be seen on the captured image by the different intensity levels of the pixels of the image. Advantageously, the intensity levels are the gray levels of a gray-scale image. The image can be recorded directly in gray scale or the image can be converted from the recorded color image to a gray-scale image by image processing. As explained, the different intensity levels (especially gray levels) of the individual pixels in the captured image are caused by different amounts of light reflected from the region where the weld seam is located and which the image depicts. Depending on the shape of the surface of the weld seam (especially the bead), a smaller or larger amount of light is reflected back into the beam path of the image-capturing device at a specific point on the surface, which results in different exposures of the pixels of the image. Thus, the intensity levels (or gray-level values) serve not only as an indicator of the amount of reflected light but also as a measure of the local orientation of the surface of the weld seam via the amount of reflected light. The local orientation of the surface of the weld seam (especially the bead) refers to the geometry or shape of the weld seam, i.e., its extension in different spatial directions.

[0025] Furthermore, according to the present invention, the image is divided into a first subset of pixels and a second subset of pixels based on the intensity level of the pixels and comparison with a previously defined intensity level threshold, in particular based on a reference measurement performed on a weld seam considered to be good (i.e., a weld seam having sufficient or satisfactory quality according to the desired standard). The image can also be divided into other subsets of pixels by specifying multiple intensity level thresholds. Dividing the image based on the defined intensity level threshold now enables only two (or more, but rarely) different types or amounts of pixels to be evaluated for assessing the welding quality. This allows for simplified pixel data processing to simply, quickly, and precisely assess the welding quality of the weld seam in subsequent steps of the method.

[0026] If the pixels are divided into multiple subsets, quality control can be performed by comparing the number and / or position of the pixels in the different subsets with a pixel target value or pixel target range. This can be done using the first subset and the second subset as described herein.

[0027] Preferably, the gray level is used as the intensity level, and the gray level threshold is used as the intensity level threshold. For example, if the gray level is used as the intensity level for a gray level image with an 8-bit color depth or 256 gradations, each pixel of the gray level image can have up to 256 different gray levels, each of which is due to a different amount of light reflected onto the camera. By dividing the image into only a first subset of pixels (pixels having an intensity level above the threshold; the intensity level can be the gray level) and a second subset of pixels (pixels having an intensity level equal to or below the threshold; the intensity level can be the gray level), each pixel is assigned only one of two (the number of subsets) values instead of 256 possible gray level values (where 0 is the minimum of black or no exposure and 255 is the maximum of white, maximum exposure), where this value is the assignment to the first subset or the second subset. The gray level is an exemplary but preferred choice for the intensity level. However, hue can also be selected as the intensity level. Subsets with their own thresholds can also be created for each color channel.

[0028] When forming the subsets, the captured image can be binarized. This results in a binary image that visualizes the division of the image into the first subset of pixels and the second subset of pixels. The intensity level threshold can be adjusted individually for different processing parameters.

[0029] The binary image thus generated or the partitioning of the image into pixels of the first subset and pixels of the second subset - which can, for example, be stored in a table or be stored in a table - now shows the local orientation of the surface of the weld seam (in particular a weld bead) produced by welding based on the number and position of the pixels of the first subset and / or the pixels of the second subset, and allows the welding quality to be evaluated by evaluating the number and / or position (in particular a combination of number and position) of the pixels of the first subset and / or the pixels of the second subset. The position of a pixel here refers to the spatial position in the image or the position of the pixel in the image. The position of a pixel can be specified, for example, by specifying the two-dimensional coordinates of the pixel in the image plane spanned by the orthogonal axes or by specifying the quadrant in a grid superimposed on the image. For the evaluation, an optimal pixel range or a target pixel range for the number and / or position of the pixels (which is in particular given for a target weld seam or a weld seam with high welding quality) is used for comparison with the number and / or position given by the pixels of the first subset and / or the pixels of the second subset in the binary image or in the table.

[0030] The pixel target value can be specified, for example, by the target number of pixels of the first subset or the second subset, above or below which a weld seam is assumed to be properly formed. Thus, a single pixel target value, such as the target number of pixels of a certain subset, defines an easily verifiable quality criterion. The target quantity can be a minimum quantity or a maximum quantity or an interval bounded by a minimum quantity and a maximum quantity.

[0031] The pixel target range can be a specific specification regarding the position of the pixels of a subset in the image. For example, a criterion can be that more or fewer pixels of a subset must be present in a sub-region of an image.

[0032] The given weld seam is compared with these pixel targets based on the pixels of its first subset and / or second subset. If there is a large enough consistency as defined by the pixel target range, it can be evaluated that the welding quality is sufficient. Otherwise, the welding quality can be judged to be insufficient or faulty.

[0033] For example, the number of pixels in the first subset can be compared with the target value. It is also possible to compare the number of pixels of the first subset in a certain region of the image with the target value. Specifically, it can be set that the pixels in a certain specified region of the image are not considered in the comparison, and for example, only the number of pixels of a relevant subset outside this region is considered.

[0034] By evaluating the welding quality, it can be determined which weld seams are acceptable and which are not.

[0035] However, in addition, the proposed method also enables a more detailed specification of the quality based on the recognition of other features, in particular the shape of the weld bead. For example, the shape of the weld bead and thus, for example, its symmetry or asymmetry can be determined based on the position and number of the pixels of the first subset and / or the second subset in the image.

[0036] Generally, an assessment of the welding quality of conductor elements that are mutually welded to form the stator of an electric machine, in particular an electric motor, is performed as part of the method according to the invention.

[0037] In particular, the conductor elements are hairpins of the stator of an electric machine, in particular an electric motor. Given the large number of welds resulting from the large number of hairpins on the stator, the method according to the invention enables a particularly cost-effective, fast and reliable quality control of all welds.

[0038] It is possible that, within the scope of the invention, the image is a sub-region of a higher-level image that includes a plurality of welds. The assessment can then be performed for each image (a part of the higher-level image). In this way, a plurality of welds can be evaluated with one image at a time.

[0039] The step of cropping the image to a region that includes the weld can be performed, and the assessment can be carried out only in this sub-region. Generally, this sub-region is an image of the end face of the conductor element and, if applicable, an image of the part of the welds that protrude beyond them. If there are a plurality of welds in the image or the higher-level image, the respective regions of each weld can be cropped. This reduces the amount of data to be processed and also allows the method to be carried out only in the relevant regions of the welds, thus avoiding errors in the assessment.

[0040] As already explained, it can be provided that a specific region of the image is not considered in the evaluation. For example, this can be a central region around the center of the weld seam (in the case of a spherical weld seam on two conductor elements). Here, it can generally be assumed that if the weld seam is formed as intended, almost all pixels in this sub-region can be assigned to the first subset (strong reflection, bright). For example, including these pixels when considering the set of pixels of the first subset can reduce its relative accuracy. For example, the region excluded from the evaluation can be defined as the region in which the surface of the weld seam formed as intended is substantially "flat", i.e., orthogonal to the viewing direction of the image capture device. Simply put, this is advantageous because due to the reflection of the illumination light, the center of the bead appears bright in the image, and the bead mainly contains pixels of the first subset or the second subset assigned to the bright region. If for various reasons the individual pixels in this region are not assigned to the first subset (bright), this usually does not matter for the surface inclination. If these pixels are considered in the analysis, this may make the result less accurate. In other words, by excluding the region in the weld seam formed as intended that is almost completely assigned to one subset from consideration, the sensitivity of the control in the critical region can be increased.

[0041] The central region is preferably a circular or elliptical region with a predetermined radius. For example, when the weld seam is properly formed, a predetermined minimum number of pixels of the first subset (bright) fall into the circular or elliptical region, or there is a high degree of reflection in the circular or elliptical region. A properly formed weld seam should be understood as a weld seam with a specified welding quality, especially the lowest quality. In a weld seam connecting three or more conductor elements, the central region not considered in the evaluation is usually elliptical because the weld seam is not circular but elongated.

[0042] For example, a properly formed weld seam can have a certain number of pixels in the first subset or a certain number of pixels in the first subset outside the region not considered in the evaluation.

[0043] A properly formed weld seam can also have a certain symmetry, for example, in the distribution of pixels of the first type and the second type (pixel target range).

[0044] The maximum number of pixels of the first subset or the second subset can be specified, for example, absolutely or relatively with respect to the total number of pixels of the first subset or the second subset. This maximum number can be determined based on a sample of properly formed welds or welds with sufficient welding quality. For example, in the case of a circular area, as is particularly useful in the case of two conductor elements being welded to each other, the radius for the central area can be specified such that, based on the average value or mean of the sample, a minimum amount (e.g., 90%) of all pixels of the first subset or all pixels of the second subset lies within the cutout central area. Alternatively, if more than two conductor elements are welded to each other and an image of them is taken, the shape of the cutout area can be elliptical.

[0045] Preferably, the pixel target value of the number and / or the pixel target range of the position is determined by selecting the minimum connection cross-section as the quality characteristic of the welding quality. This can be done by a separate determination step in the method. This step is based on the recognition that the number and / or position of the pixels of the first subset and / or the second subset is associated with the connection cross-section, and thus the minimum connection cross-section can be determined via them.

[0046] Therefore, the data through which the number and / or position of the pixels of the first subset and / or the second subset is associated with the connection cross-section of the weld can advantageously be used to determine the pixel target value of the number and / or the pixel target range of the position. For example, the connection cross-sections of different welds can be determined, for example, by cross-sectioning, and set with respect to the number and / or position of the pixels of the first subset and / or the second subset to obtain the following database, which can be used to determine the pixel target value of the number and / or the pixel target range of the position based on the desired minimum connection cross-section.

[0047] In order to determine the welding quality, in principle, only the number of pixels of the first subset and / or the second subset can be evaluated, only the position of the pixels of the first subset and / or the second subset can be evaluated, or a combination of the number and the position can be evaluated. In the latter case, the number of pixels of the first subset and / or the second subset is considered based on the position of the pixels of the first subset and / or the second subset on the surface of the weld.

[0048] Advantageously, the welding symmetry of the weld seam can be determined based on the number and position of the pixels of the first subset and / or the second subset. Welding symmetry means the symmetry of the weld seam, in particular the bead, with respect to one or more imaginary symmetry lines passing through the weld seam, in particular the bead. Due to the highly non-uniform distribution of the pixels of the first subset or the second subset on each side of the symmetry line, an asymmetric bead shape can be detected. Excessive asymmetry indicates poor welding quality.

[0049] It can be provided that the weld seam is re-welded based on the number and position of the pixels of the first subset and / or the second subset. For example, if an asymmetric bead is detected due to non-uniform energy input during welding, targeted re-welding can be performed to produce a symmetric bead. Based on the number and position of the pixels of the first subset or the second subset, the geometry of the bead can be identified such that this geometry can be used for re-welding to enable the subsequent creation of the desired spherical shape of the bead. On the other hand, if the number of pixels of the first subset or the second subset is too high in the entire edge region of the bead, a too flat bead can be detected, such that in this case, re-welding can be performed in the entire edge region to rework the spherical shape of the bead and improve the welding quality. The edge region of the bead can be understood as the surface of the end faces of the conductor elements welded to each other minus the previously mentioned central region, in particular the cutout. This method enables not only quality control but also simple, rapid, and precise post-processing of the component to avoid rejection.

[0050] Here, the welding power for re-welding can be selected based on the number and / or position of the pixels of the first subset and / or the second subset. This enables not only the control of the position but also the control of the intensity of the energy input of the laser beam during re-welding, thus enabling the precise production of a quality-optimized weld seam.

[0051] It is also advantageous to evaluate the welding quality of the weld seams on the same component (in particular a stator) with multiple conductor elements and the welding of another different weld seam simultaneously. Parallel welding and evaluation enable further time savings in component production. Alternatively, it is also possible to alternate the evaluation and welding.

[0052] The object mentioned at the beginning is also achieved by a device for evaluating the welding quality of a weld seam between the end segments of at least two elongated conductor elements, which are welded to each other by irradiating the end faces of the end segments. The device includes: an image capture device having a beam path. The image capture device is designed to perform step (a) of the method according to the invention.

[0053] Furthermore, the device has at least one data processing system, which is set up to carry out steps (b) and (c) of the method according to the invention.

[0054] The device includes a lighting device, which is configured to illuminate the end face of the conductor element from the direction along the beam path of the image capture device.

[0055] The device also includes a laser beam source, which has a laser beam path for carrying out a welding process for connecting the conductor element on the end face of the conductor element. In particular, the beam path of the image capture device and the beam path of the laser beam are coupled to each other, in particular by means of a beam splitter. Therefore, the beam paths are the same. Thus, the image capture device enables an evaluation from the same "viewpoint", from which the laser also carries out its processing.

[0056] When carrying out the method, the image is preferably captured along the beam path of the laser used for the welding process. The illumination of the weld seam preferably takes place parallel to or along this beam path.

[0057] The device according to the invention hereby has the same advantages as have been explained in detail with reference to the method according to the invention. Furthermore, the features described herein with respect to the method can be equivalently applied to the device. In particular, the device can be configured to carry out the method described herein.

[0058] The invention is described in more detail below with reference to the drawings, where, if necessary, identical or functionally equivalent elements are designated only once with reference numerals.

[0059] In the drawings:

[0060] Figure 1 A schematic view of an exemplary embodiment of a device according to the invention is shown;

[0061] Figure 2a 、 Figure 2b A schematic side view and a plan view of two conductor units of a stator before welding are shown; Figure 1 from

[0062] Figure 3a 、 3b A schematic side view and a plan view of two conductor units after welding are shown; Figure 2a 、 Figure 2b from

[0063] Figure 4 A schematic view of a device according to the invention for evaluating from Figure 1Schematic representation of the steps of an exemplary embodiment of a method for the welding quality of welds on two conductor elements of a stator;

[0064] Figure 5 Shows an exemplary embodiment of a method for evaluating the welding quality of another weld of two conductor elements of a stator according to the present invention; Figure 1 Schematic representation of the steps of an exemplary embodiment of a method for the welding quality of another weld of two conductor elements of a stator;

[0065] Figure 6 Shows a binary image of a weld welded with different laser powers and a schematic representation of the weld beads of a weld of two conductor units of a stator; and Figure 1 Schematic representation of the weld beads of a weld of two conductor units of a stator; and

[0066] Figure 7 Shows for Figure 1 Schematic representation of the number of white pixels of each weld bead for multiple weld beads of multiple pairs of conductor elements of a stator.

[0067] Figure 1 Shows an exemplary embodiment of an apparatus 200 for evaluating the welding quality of weld 14 (see Figure 3a ). In the present case, the apparatus 200 is also designed to produce the weld 14 by laser beam welding with the aid of a laser beam source 102 and scanner optics 100.

[0068] In addition to the beam guide (beam path 114) of the laser beam source 102, the scanner optics 100 also includes an image capture device 116, which is also referred to hereinafter as the camera 116. The beam path 118 of the camera 116 is used to detect the positions of two (alternatively more) conductor elements 12 of the component 10 in the present case, which are welded before the actual laser beam welding and thus before the precise laser beam welding of the free end segments of the two conductor elements 12 of the component 10 by the beam path 114 of the laser beam source 102.

[0069] In the present case, the component 10 is a stator 10, which has hairpins (I-shaped clips or other plug-in coils are also preferred conductor elements within the scope of the present invention) as conductor elements 12 of an electric motor. For high electrical conductivity, the conductor elements 12 can be made of copper or at least contain copper, for example.

[0070] The beam paths 114, 118 are superposed coaxially with each other or guided in a common beam path 120 at the beam splitter 106. Here, the beam path 114 of the laser 102 is collimated by the collimating lens 104 in front of the beam splitter 106. One or more movable deflection mirrors 108 of the scanner optics 100 enable the coaxially combined beam path 120 to be displaced in a plane on the surface of the end segments of the conductor element 12. In the beam direction behind the deflection mirror 108, a focusing lens 110 and an additional illumination device 112 of the device 200 are also arranged. The focusing lens can also be designed as a lens package for focusing the beam path 120 onto the two end segments of the conductor element 12. The additional illumination device 112 is used to illuminate the end segments of the conductor element 12 for the camera 116.

[0071] The camera 116 of the scanner optics 100 is not only used for the process control of the laser beam welding; on the contrary, according to the method 40 described below (see Figure 4 ), it is also used for image recording in the context of evaluating the welding quality of the weld 14. Alternatively, an additional camera (not shown) can also be used. In addition to the advantages obtained with respect to the method 40 described in more detail below, this also has the following advantage: checking the welding quality does not require additional and expensive equipment; rather, the devices already provided for welding can be used. One or more data processing systems 202 of the device 200 - also referred to as computers 202 - can already exist or can be added. The functions of the data processing systems will be explained in more detail later.

[0072] Figure 2a Using an as-yet-unwelded pair of elongated conductor elements 12 or hairpins shows how the light 118a of the beam path 118 of the camera 116 or the illumination device 112 impinges on the surface of the end face 13 of the unwelded end segment of the conductor element 12. This light 118a is reflected back vertically as light 118b at the end face 13 of the unwelded end segment because the end face 13 of the end segment of the conductor element 12 is perpendicular to the direction of the light 118a. Figure 2b A schematic image 20 taken by the camera 116 is shown, and the illuminated end face 13 of the end segment of the conductor element 12 is shown.

[0073] On the other hand, Figure 3a Using a pair of welded end segments of the elongated conductor element 12 or hairpins shows how the light 118a of the beam path 118 of the camera 116 impinges on the end face 13 of the welded end segment of the conductor element 12. As compared with Figure 2a and Figure 2bIn contrast, the end face 13 is no longer visible separately here, or has been interconnected by welding (however, the end face 13 is still mentioned herein). Here, the light 118a irradiates the bead 16 of the weld 14, which is produced by laser welding using a laser beam source 102, and the weld interconnects two end segments of the conductor element 12 at the end face 13. The shape of the bead 16 is spherical. In this case, since the spherical shape of the bead 16 is uniform, a high welding quality can be presented, and thus a large connection cross-section between the conductor unit 12 or its ends can be presented. Different from that in Figure 2a , here the light 118a is not reflected at the center. Instead, the reflected light 118b is mainly reflected back to the camera 116 in the middle or central region of the bead 16, and otherwise mainly scattered outward. The closer the shape of the bead 16 is to spherical, the more inclined the surface of the bead 16 is towards the incident direction of the light 118a, so that less light 118b is reflected back to the camera 116 and captured by the camera when taking the image 20. However, the flatter the bead 16 is, the weaker the bead is and thus the more unstable it is. In the case of a flat bead 16 - which indicates a poor welding quality due to a small connection cross-section of the weld 14, more light 118b is reflected onto the camera 116 and captured when taking the image 20.

[0074] By means of different ratios or amounts of light reflection, conclusions can be drawn about the shape of the bead 16, and thus about the connection cross-section of the produced weld 14 and thus about the welding quality determined by the connection cross-section. The method 40 described below utilizes this correlation between the shape of the bead 16 and the associated welding quality of the weld 14 and the light in the image 20 recorded by the camera 116. Figure 3b Figure 20 shows an image of the end face 13 of the welded end segment of the conductor element 12 taken by the camera 116, where the bead 16 is clearly visible here, and the pixel density of the image 20 decreases towards the outside or edge region of the bead 16.

[0075] Figure 4 Now, the sequence of an exemplary method 40 for evaluating the welding quality of a pair of conductor elements 12 welded at the end segments of the pair of conductor elements 12 according to the present invention is shown. The method 40 described by way of example herein includes method steps 42, 44, 46, 48, 50, 52, and 54 (see Figure 7 ), where some of the listed method steps described previously herein are only advantageous and not absolutely necessary.

[0076] In a first method step 42, an image 20 of the weld seam 14 is taken by means of a camera 116, which is positioned such that its beam path 118 is above the bead 16 of the weld seam 14 or its beads 16. Although only one bead 16 or one weld seam 14 is mentioned here and hereinafter, it is also possible to take an image 20 of a larger area of the component 10, which includes a plurality of beads 16 or weld seams 14, such that the plurality of beads can be evaluated at least temporarily with respect to a plurality of beads 16, and the method 40 can be carried out particularly quickly.

[0077] The recorded image 20 includes an area larger than the weld seam 14, such that in a second method step 44 the image 20 is cropped to obtain an image 22 cropped to the area of the weld seam 14. From above (i.e., in the viewing direction of the camera 116 when vertically observing the end face 13), the individual conductor elements 12 and their end faces 13 are no longer visible, which end faces are only indicated here; instead, the bead 16 can be seen from above.

[0078] In a third method step 46, an intensity level threshold is determined for the cropped image 22 in the form of gray level values (in the case of a color image, color values are also possible); the image 20 is recorded as a gray image in method step 40. In the present example, this intensity level threshold determination is used for subsequent binarization according to a threshold method in a fourth method step 48.

[0079] Using the determined threshold, the pixels of the cropped image 22 are divided into two subsets. The pixels of the first subset 30 are those pixels whose intensity level is greater than the determined intensity level threshold. In a gray level image, these pixels are bright pixels, or pixels where a large amount of light is reflected back towards the camera. The pixels of the second subset 32 are those pixels whose intensity level is less than or equal to the intensity level threshold.

[0080] In the threshold image 24, for this purpose, only the pixels or regions 18 of the cropped image 22 that are above the determined intensity level threshold (pixels of the first subset 30) are marked. The remaining pixels in the threshold image 24 are below or at the threshold and belong to its second subset 32.

[0081] In a fourth method step 48, a binary image 26 is generated with white pixels of the first subset 30 and black pixels of the second subset 32. The intensity level threshold, which is a gray level threshold in this case, can be adjusted individually for different lighting devices 112.

[0082] All pixels above the threshold are grouped as white pixels, while all pixels at or below the threshold are grouped as black pixels. Thereby the "color depth" is reduced. In addition, the colors can be inverted to improve the visual recognition for the user monitoring the recognition process.

[0083] It is also possible to evaluate the image without binarization and directly after threshold analysis.

[0084] In the fifth method step 50, the central region 34 is cut out from the binary image 26 thus obtained, such that a re-clipped binary image 28 is obtained. The central region 34 can also simply be excluded from the subsequent evaluation and does not actually have to be cut out. This is because the central region 34 appears bright due to reflection and in any case mainly contains white pixels of the first subset 30, such that it generally does not provide any information about the welding quality of the weld 14. The cutout of the central region 34 can here be circular, thus following the spherical shape of the bead 16. The radius of the circular cutout depends on the size of the conductor element 12 or its end section and the welding strategy. When connecting more than 2 conductor elements, an elliptical central region 34 is generally masked out instead of a circular central region. The radius (for the circular central region 34) can be defined in different ways. One example is to use an image 20 of the weld 14 with sufficient welding quality and consider how large the radius has to be such that a certain amount (e.g., 90%) of all white pixels of the first subset 30 lies within the circular cutout. Thereby, an average value can be formed, which can then be defined as the radius of the circular central region 34. Advantageously, the amount of data to be processed by the method 40 can be reduced in this way.

[0085] Then, in the sixth method step 52, the number A of white pixels or the pixels of the first subset 30 in the clipped binary image 28 can be counted. In Figure 4 In this example of the weld 14 in the clipped binary image 28, the number A of white pixels of the first subset 30 = 508.

[0086] Finally, in the seventh method step 54, as Figure 7 shown in, the welding quality of the weld 14 is determined based on a comparison with a pixel target value or a pixel target range. For example, Figure 7 the pixel target value in is given particularly simply by a reference value R in the form of the number of white pixels of the first subset 30, which here is R = 1500 white pixels, i.e., the pixels of the first subset 30. Thus, for the number A = 508 white pixels of the first subset 30, it can be concluded that the bead 16 is sufficiently spherical to ensure a sufficiently high welding quality, because in the edge region of the bead 16, fewer than 1500 white pixels of the first subset 30 occur, i.e., are reflected back to the camera 116 in the form of light.

[0087] On the other hand, a number of more than 1500 white pixels in the first subset 30 will indicate that the solder bead 16 is flat rather than spherical because too much light is reflected back to the camera 116 in the edge region of the weld 14, which is captured as white pixels of the first subset 30 by binarization. Possible causes for this can be, for example, incorrect positioning of the beam path 114 of the laser beam source 102, too low power of the laser beam source 102, too short welding time, and too large a gap between the end segments of the conductor element 12. Thus, poor welding quality can be determined.

[0088] When determining the weld quality, in addition to the quantitative determination of the white pixels of the first subset 30 based on high-quality or properly formed welds 14 and improperly formed or poor-quality welds 14, other qualitative determinations, especially position-related determinations, are also possible, as shown, for example, by Figure 5 the embodiment of method 40 in

[0089] In Figure 5 it can be seen that the solder bead 16 is formed asymmetrically such that it is tilted towards one of the two conductor elements 12. This can be seen in addition to the very high number A = 2550 of white pixels in the first subset 30 in the cropped binary image 28 because the white pixels of the first subset 30 are visible or very predominantly visible only on one side of the edge region of the solder bead 16 in the cropped binary image 28. From this, it can be concluded that there is an asymmetric solder bead 16 here, as also shown in Figure 5 For example, this can be detected by a target range test. For example, it can be checked whether a certain number of pixels of the first subset are present in each of the individual quadrants of the binary image 28. It can also be checked for a certain form of distribution of the pixels of the first subset.

[0090] The evaluation method 40 can be controlled by at least one computer 202, which is part of the device 200, and as shown, the at least one computer can be connected to the camera 116 to perform the recording of the image 20 and other method steps of the method 40 in each case.

[0091] At least one computer 202 can also be connected to a control device (not shown) of the laser beam source 102 to transmit thereto information about the bead 16 obtained according to method 40 and / or to control the laser beam source 102. Thus, the laser beam source 102 can be given information about, for example, an asymmetric bead 16 such that the device 200 can then perform re-welding in areas of insufficient welding, which areas of insufficient welding are identified based on the missing white pixels of the first subset 30, such that the bead 16 becomes symmetric and spherical due to the re-welding. In other words, the number A and the position of the white pixels of the first subset 30 determine the position of the energy input during re-welding by means of the laser beam source 102 and preferably also determine the level of that energy.

[0092] Figure 6 Shown are binary images 28 and associated beads 16 cropped at exemplary, different laser powers P (indicated as % of the maximum power) of the laser beam source 102 with the same welding time. It can be seen that the selected laser power P as well as the duration have a considerable influence on the formation of the bead 16 and thus on the welding quality of the weld seam 14 during laser beam welding. As the laser power P decreases, the bead 16 becomes increasingly flat and the number A of white pixels of the first subset 30 in the edge region of the weld seam 14 increases. The connecting cross-section of the weld seam 14 decreases accordingly and the weld seam 14 becomes unstable and may not be able to withstand the loads during operation of the stator 10 in an electric motor.

[0093] If the number of white pixels of the first subset 30 above the reference value R or target value is distributed fairly evenly around the central region 34, for example due to too low a laser power or too short a welding time, as can be seen in Figure 6 this can not only be detected by means of method 40, but the entire pair of conductor elements 12 can also be re-welded. This can be done, for example, via pixel target area analysis or a separate analysis step to determine any necessary post-processing of the weld seam. Here, the local distribution of the individual pixels of the individual subsets can be analyzed in detail to draw more detailed conclusions about the shape and position of the existing weld seam and to further determine the optimal position and optimal level of additional energy input. The number of remaining white pixels of the first subset 30 can determine the energy input during re-welding.

[0094] already partly explained Figure 7 Shown are multiple pairs of conductor elements 12 already welded on the stator 10. Two outliers of the welded conductor elements 12 are shown, which outliers have a significant difference Δ in the number A of white pixels of the first subset 30 compared to the reference value R. The associated beads 16 are also formed asymmetrically.

Claims

1. A method (40) for evaluating the welding quality of a weld (14), wherein end segments of at least two elongate conductor elements (12) are welded to each other through the weld by irradiating end faces (13) of the end segments, wherein, The method (40) comprises the following steps: (a) Capturing an image (20) of the end faces (13) of the conductor elements (12) welded to each other via the weld seam (14), wherein the image (20) is recorded from the end faces (13) of the conductor elements (12), and the environment recorded by the image (20) at least comprises the complete surface of the weld seam (14), and wherein the image (20) is recorded with different intensity levels in the pixels of the image (20), wherein the intensity level is an indicator of the amount of light locally reflected from the recording position when recording the image (20), and thus the indicator is a measure of the local orientation of the surface of the weld seam (14), (b) Dividing the image (20) into at least one first subset (30) of pixels and at least one second subset (32) of pixels, the intensity levels of the pixels of the first subset being greater than a specified intensity level threshold, and the intensity levels of the pixels of the second subset being less than or equal to the intensity level threshold, and (c) Evaluating the welding quality of the weld seam (14) based on a comparison of the number and / or position of the pixels of the first subset (30) and / or the second subset (32) with at least one pixel target value and / or a pixel target range of the number and / or position of the pixels of the first subset (30) and / or the second subset (32) of the weld seam corresponding to a desired quality.

2. The method (40) according to claim 1, further comprising: In step (a), the weld seam is illuminated with light parallel to the direction in which the image is recorded.

3. The method (40) according to claim 1, wherein The intensity level is a gray level, and the intensity level threshold is a gray level threshold.

4. The method (40) according to any one of claims 1 to 3, wherein, The conductor element (12) is an inserted coil of the stator (10) of an electric machine.

5. The method (40) according to claim 4, wherein The conductor element (12) is a hairpin and / or an I-shaped clip.

6. The method (40) according to any one of claims 1 to 3, wherein Capturing a higher-level image (20), the higher-level image (20) comprising in each case a plurality of weld seams (14) of at least two end segments of the conductor element (12), and performing a welding quality assessment for each of the weld seams (14).

7. The method (40) according to any one of claims 1 to 3, wherein, Evaluating the welding quality of the weld seam (14), wherein the end segments of at least three conductor elements (12) are welded to each other at the end face of the end.

8. The method (40) according to any one of claims 1 to 3, wherein In the evaluation according to step (c), the central region (34) of the image (20), which corresponds to the weld seam, is not taken into account.

9. The method (40) according to any one of claims 1 to 3, wherein, In the evaluation according to step (c), the central region (34) of the image (20), which corresponds to the center of the expected position of the weld bead, is not taken into account.

10. The method (40) according to claim 8, wherein, The central region (34) is a circular or elliptical region having a predetermined size, and when the weld seam (14) is properly formed, a predetermined minimum number of pixels of the first subset (30) or the second subset (32) fall into the circular or elliptical region.

11. The method (40) according to any one of claims 1 to 3, wherein, The pixel target value of the number and / or the pixel target range of the position are determined by selecting the minimum connecting cross-section as a quality characteristic of the welding quality.

12. The method (40) according to claim 11, wherein, Data related to a plurality of welds considered to be good is used to determine the pixel target value of the number and / or the pixel target range of the position, so as to determine the pixel target value or the pixel target range by associating the position and / or the number of each intensity level of the pixels with the connection cross-section of the weld (14).

13. The method (40) according to any one of claims 1 to 3, wherein, Evaluating the welding quality of the weld (14) includes: evaluating the symmetry of the weld (14) based on the number and position of the pixels of the first subset (30) and / or the pixels of the second subset (32).

14. The method (40) according to any one of claims 1 to 3, wherein, Reweld the weld (14) according to the number and position of the pixels of the first subset (30) and / or the pixels of the second subset (32).

15. The method (40) according to claim 14, wherein, Select the welding power for the rewelding according to the number and / or position of the pixels of the first subset (30) and / or the pixels of the second subset (32).

16. The method (40) according to any one of claims 1 to 3, wherein, Simultaneously evaluate the welding quality of the weld (14) on the same component having a plurality of conductor elements (12) and the welding quality of another different weld (14).

17. An apparatus (200) for evaluating the welding quality of a weld (14) between end segments of at least two elongated conductor elements (12), the conductor elements being welded to each other by irradiating end faces (13) of the end segments, the apparatus (200) comprising: - An image capture device (116) having a beam path (118), the image capture device being designed to perform step (a) of the method according to the preceding claims - At least one data processing system (202), the data processing system being configured to perform step (b) and step (c).

18. The apparatus according to claim 17, wherein, The apparatus includes an illumination device (112), the illumination device being configured to illuminate the end face (13) of the conductor element (12) from a direction along the beam path (118) of the image capture device (116).

19. The device according to claim 17 or 18, wherein The apparatus further includes a laser beam source (102), the laser beam source having a beam path (114) of a laser beam for performing a welding process to connect the conductor elements (12) at the end face (13) of the conductor element.

20. The apparatus according to claim 19, wherein, The beam path (118) of the image capture device (116) and the beam path (114) of the laser beam are coupled to each other by a beam splitter (106).

Images