Method for determining the stacking accuracy of a plurality of electrode sheets in a stack

Abstract

The invention relates to a method for determining the stacking accuracy of a plurality of electrode sheets (1, 2, 3), wherein the electrode sheets (1, 2, 3) extend in mutually parallel planes (4) and are arranged stacked on one another and form a stack (5); wherein the stacking accuracy describes the positions (6, 7, 8) of the edges (9) of all electrode sheets (1, 2, 3) relative to one another in the stack (5); wherein the method is carried out by means of a measuring device (10) having a two-dimensionally resolving X-ray device (11) with at least one radiation source (12) for X-ray radiation and a detector (13).

Description

Technical Field

[0001] This invention relates to a method for determining the stacking accuracy of multiple electrode sheets in a stack. The electrode sheets extend in mutually parallel planes and are arranged and stacked to form a stack. The stacking accuracy describes the position of the edges of all electrode sheets relative to each other in the stack. Background Technology

[0002] Batteries, especially lithium-ion batteries, are increasingly used in the propulsion of motor vehicles. Batteries are typically composed of individual cells, each of which has an anode plate, a cathode plate, and, if necessary, separator plates or stacks of separating materials. These anode plates, cathode plates, and, if necessary, separator plates are referred to below as electrode plates.

[0003] Electrode sheets are typically manufactured by punching or cutting, such as laser cutting.

[0004] The stacking accuracy of individual electrode pieces in a lithium-ion battery stack has a significant impact on the safety-related quality standards of a single cell and its performance. Stacking accuracy, i.e., the deviation of the relative positions of the individual electrode pieces, must therefore be kept within a very small range. One known method for determining stacking accuracy is computed tomography (CT). This involves generating a three-dimensional image of the stack through lengthy measurements. This method is very expensive and therefore only has limited applicability in production lines.

[0005] Compared to computed tomography (CT), there are also measurement devices that use X-ray radiation to achieve two-dimensional imaging of stacked or inspected objects. These devices are significantly cheaper and faster than CT. However, this method cannot directly measure the relative positions of the electrode pads.

[0006] A conveying device for a single battery cell is known from document KR 10 2020 0088 222 A. Here, the conveying device includes an X-ray source through which the edges of the individual layers of the single battery cell are scanned. The X-ray source is oriented such that the X-rays are parallel to the edges of the layers to be inspected.

[0007] As known from document WO 2016 / 114257 A1, the coating region of an electrode film is detected by X-rays. The X-ray source is oriented such that the X-rays are parallel to the edge of the layer to be detected.

[0008] A device for examining a single cell of a battery by X-ray radiography is known from DE 10 2020 112 814 A1. Here, the single cell of the battery can be moved relative to an X-ray source. The X-ray source is oriented such that the X-rays are parallel to the edge of the layer to be inspected. Summary of the Invention

[0009] Therefore, the technical problem to be solved by the present invention is to at least partially solve the above-mentioned problems. In particular, a method for determining the stacking accuracy of multiple electrode sheets in a stack is proposed. Specifically, it is possible to determine the position of each individual electrode sheet so as to enable the allocation of deviations to specific electrode sheets in the stack.

[0010] The above-mentioned problems are solved by a method having the features of claim 1. Advantageous improved designs are the subject of the dependent claims. Features described individually in the claims can be combined with each other in a technically suitable manner and can be supplemented by details from the description and / or drawings, which indicate further embodiments of the invention.

[0011] A method is proposed for determining the stacking accuracy of multiple electrode sheets. The electrode sheets extend in mutually parallel planes and are arranged and stacked to form a stack. Stacking accuracy describes the precision of the relative positions of the edges of all electrode sheets in the stack, or the accuracy of the overlap of the electrode sheet edges. The method is implemented by a measuring device having a two-dimensional resolution X-ray apparatus with at least one radiation source and detector for X-ray radiation. The method includes at least the following steps:

[0012] a) Provide a stack and arrange the stack in the measuring device between at least one radiation source and a detector;

[0013] b) Irradiate the stack from a first spatial coordinate with at least one radiation source, wherein the irradiation direction is at least perpendicular to the plane and extends toward the detector, wherein the radiation source detects the overlapping edges of the electrode sheets and projects a two-dimensional first profile of the edge of the stack onto the detector.

[0014] c) Irradiate the stack from at least one second spatial coordinate, which is different from the first spatial coordinate, using at least one radiation source, wherein the edges of the overlapping arrangement of the radiation detection electrode sheets are irradiated and a two-dimensional second profile of the edges of the stack is projected onto the detector.

[0015] d) Detect the first contour using the detector;

[0016] e) Detect at least one second profile using the detector;

[0017] f) Analyze different contours and determine the position of the electrode edge.

[0018] The above (non-closed) division of method steps a) to f) is primarily for differentiation only and is not a mandatory order and / or association. The frequency of method steps may, for example, change during the implementation of the method. It is also possible that the method steps overlap at least partially in time. It is particularly preferred that method step a) is performed in time before steps b) to f). In particular, steps b) and c) are performed in time staggered. In particular, step d) is performed in time concurrently with step b), and step e) is performed in time concurrently with step c). In particular, steps c) and e) may be performed multiple times, wherein the corresponding spatial coordinates are different from the previous spatial coordinates. In particular, steps a) to f) are performed in the following order: step a), step b), step d), step c), step e), step f).

[0019] The method is used particularly within the scope of methods for manufacturing single-cell batteries. Here, electrode sheets cut to appropriate shapes, namely anode sheets, cathode sheets, and, if necessary, separator sheets, are arranged in a predetermined order and stacked and oriented relative to each other. In the resulting stack, the individual electrode sheets should be arranged relative to each other with the most aligned (edge) positions possible.

[0020] The electrode sheets are arranged in a stack, extending in mutually parallel planes and stacked on top of each other. The stack includes at least two electrode sheets, specifically at least one anode sheet and one cathode sheet. A spacer material is arranged between the anode sheet and the cathode sheet. This can be designed, for example, as a coating on one of the anode and cathode sheets, or as a separate spacer sheet.

[0021] The electrode plates are typically rectangular in shape. Where necessary, discharge lugs extend from this rectangular shape. They are usually designed to be uncoated, i.e., without an active material coating, and serve as electrical contacts for the respective electrode plates, i.e., the anode or cathode plates.

[0022] Stacking accuracy specifically refers to the relative positions of the edges of all electrode sheets in the stack. The electrode sheets should be arranged relative to each other in a predetermined position. Because the dimensions of the anode and cathode sheets, as well as any necessary separators, can differ, stacking accuracy is determined particularly at the edges of the electrode sheets, which are aligned with each other along a first direction extending perpendicular to the plane.

[0023] In particular, the stacking precision of the electrode sheets is determined only at one edge of each electrode sheet.

[0024] The method is implemented, in particular, by a measuring device having a two-dimensional resolution X-ray apparatus with at least one radiation source and detector for X-ray radiation. Multiple fixed or even movable radiation sources can be configured. Multiple fixed or even movable detectors can also be configured.

[0025] A radiation source emits X-ray radiation along the direction of illumination. A detector records the X-ray radiation to display an X-ray image.

[0026] The detector, in particular, enables the display of two-dimensional images of X-ray radiation. The proposed method should enable the accurate determination of electrode stacking from these two-dimensional images of the detector.

[0027] The stack is provided according to step a), and the stack is arranged in the measuring device between at least one radiation source and a detector. It is feasible to form the stack separately and then arrange it as a whole in the measuring device. However, it is also feasible to form the stack itself (partially) in the measuring device and arrange it therein simultaneously.

[0028] According to step b), the stack is irradiated from a first spatial coordinate using at least one radiation source, wherein the irradiation direction is at least perpendicular to the plane and extends toward the detector. The rays from the radiation source or the overlapping edges of the radiation detection electrodes project a two-dimensional first profile of the stack's edges onto the detector. Specifically, the radiation source is positioned directly above the edges, i.e., without any lateral displacement relative to the edges.

[0029] For the arrangement of the stack, it is assumed that the edge of the electrode is in a predetermined target position. Within the scope of this method, the actual position of the edge deviating from this target position is determined.

[0030] According to step d), the first contour is detected using the detector.

[0031] According to step c), the stack is irradiated from at least one second spatial coordinate, which is different from the first spatial coordinate, by at least one radiation source, wherein the edges of the overlapping arrangement of the X-ray detection electrode sheets are projected onto the detector in a two-dimensional second profile of the edges of the stack.

[0032] According to step e), the detector is used to detect the at least one second profile.

[0033] According to step f), different contours are analyzed and the positions of the electrode edges are determined. This analysis can be achieved, in particular, through a system used for data processing.

[0034] In particular, the measuring device includes a system for data processing, the system having devices that are suitably equipped, configured, or programmed to implement the method, or the devices implementing the method. The devices, for example, include a processor and a memory storing commands to be executed by the processor, and the devices also include data lines or transmission means for transmitting commands, measurements, data, or the like between the aforementioned elements.

[0035] The contours detected by the detector comprise two-dimensional images, where the edges of the electrode patches can be identified based on the transitions between color intensities. Assigning the edges present in the contours to individual electrode patches is not easily achievable. The following proposes a feasible scheme for implementing such assignment.

[0036] Within the scope of this method, multiple contours of a stack are generated by radiation sources and detected by detectors. Due to the different arrangements of the radiation sources relative to the stack or edges, distinct contours are generated. These contours are particularly inferred from the position of the corresponding edge in the stack through linear equation analysis, i.e., from the known arrangement of the radiation sources and detectors and the linear direction of the rays generated by the radiation sources (i.e., straight lines). Specifically, multiple distinct contours are generated as needed to determine the position of the edge for each electrode sheet present in the stack.

[0037] The two different designs of the method described herein are feasible and can be combined with each other if necessary.

[0038] In the first design, the radiation source moves parallel to the plane to the second spatial coordinate system between steps b) and d). In the second design, the radiation source moves perpendicular to the plane to the second spatial coordinate system between steps b) and d). The resulting different contours allow for a one-to-one correspondence between the edges of each electrode sheet.

[0039] In particular, the first spatial coordinate differs from at least one second spatial coordinate in that it is different from the spacing of the stack. The spacing extends along a first direction perpendicular to the plane. Alternatively or additionally, the first spatial coordinate differs from at least one second spatial coordinate in that its distance from the edge is different from the edge, wherein this distance extends along a second direction parallel to the plane and perpendicular to the edge.

[0040] In particular, steps c) and e) are performed multiple times, in such a manner that, according to the first design scheme and / or the second design scheme, the second spatial coordinates of each step c) are different from the respective previous second spatial coordinates of the previous step c).

[0041] In particular, in step f), each edge in the corresponding contour is assigned to the corresponding spatial coordinates or, in other words, corresponds to the corresponding spatial coordinates according to the equation of the straight line.

[0042] The following example illustrates this correspondence. Here, the stack, radiation source, and detector are observed in a common plane. Therefore, the spatial coordinate z is the same for all components. The first direction extends perpendicularly to the plane along the y-axis. The second direction extends parallel to the plane along the x-axis.

[0043] The first spatial coordinate is, for example, (x) q1 |yq1 The second spatial coordinate is, for example, (x... q2 |y q2 The spatial coordinates to be determined for the edge of a defined electrode sheet are called (x...). e |y e In the first profile detected by the detector, the position of the edge of the electrode plate with respect to the first spatial coordinates of the radiation source is (x... d1 |y d1 In the second contour, the position of the edge of the electrode plate relative to the second spatial coordinates of the radiation source is (x... d2 |y d2 ).

[0044] The equation of a straight line applies to the first spatial coordinates of the radiation source:

[0045] (1)

[0046] (2) y = m1·(xx q1 )+y q1

[0047] (3) y=m1x+b1

[0048] The linear equations applicable to the second spatial coordinates of the radiation source are as follows:

[0049] (1)

[0050] (2) y = m2·(xx) q2 )+y q2

[0051] (3) y = m2x + b2

[0052] If these two equations are equal, then the following applies:

[0053] (4) m1·x+b1=m2·x+b2

[0054] (5)→(m1-m2)·x=b2-b1

[0055] (6)

[0056] (7)

[0057] In particular, the stacking accuracy is analyzed in step g). For the stack, a limit value for the maximum deviation of the target position of the contour and edge is predetermined. It is assumed that the maximum deviation is generated by the electrode piece closest to the detector.

[0058] The maximum deviation is the maximum permissible difference between the target position of an edge in a stack and the actual position of that edge. Measures are introduced when the maximum deviation is exceeded, such as marking the stack as abnormal or adjusting the stacking manufacturing process.

[0059] The maximum deviation is determined, especially at the first contour, preferably when using a straight line equation, i.e., when considering the target position of the first spatial coordinates relative to the edge.

[0060] The assumption for the maximum deviation is that it occurs on the electrode closest to the detector. This electrode is therefore positioned furthest from the radiation source. Considering the ray equation, the deviation between the identifiable edge position and the target position in the contour is most pronounced for the electrode closest to the radiation source, meaning the magnitude of this deviation is the largest. Conversely, the deviation between the identifiable edge position and the target position in the contour is least pronounced for the electrode furthest from the radiation source, meaning the magnitude of this deviation is the smallest.

[0061] The assumption that the maximum deviation is generated by the electrode plate furthest from the radiation source ensures that other electrode plates will not have a deviation from the target position exceeding the maximum deviation.

[0062] Specifically, for determining stacking accuracy, only steps a), b), d), f), and g) are performed first. Steps c) and e) are only performed if the limit value is exceeded in step g).

[0063] Within the scope of this method, for example, each stack is checked for stacking accuracy, but the position of the edges of all electrode sheets in the stack is determined only if the limit value is exceeded. Otherwise, for each stack, only whether the limit value is exceeded is checked.

[0064] In particular, if it is determined that the limit value is exceeded, steps c) and e) are performed exactly twice with different (second) spatial coordinates, and then steps f) and g) are repeated.

[0065] In particular, the first spatial coordinates are chosen such that the radiation source is positioned directly above the target location on the edge of the stack, i.e., aligned with the target location on the edge of the stack in the first direction. In particular, two second spatial coordinates are chosen such that the radiation source is displaced relative to the first spatial coordinates in the second direction, specifically by one displacement towards the stack so that the radiation source is aligned with the stack, and another displacement away from the stack so that the radiation source is positioned adjacent to the stack on the side.

[0066] In particular, if the determination in step g) exceeds the limit value, then steps c) and e) are performed as many times as are required to determine the position of all edges in a one-to-one correspondence.

[0067] This staged approach eliminates the need for comprehensive measurement of every stack (i.e., determining the position of all electrode edges) during the manufacturing of production stacks or individual cells. Maximum deviations within a stack can be detected or estimated based on a smaller profile, i.e., a limited number of probe records. If a limit is exceeded, the corresponding electrode position can be determined through additional measurements. If necessary, the position can be determined only for electrodes exceeding the limit, or only for selected electrodes, such as those with the maximum deviation.

[0068] Artificial intelligence is used, at least for step f). In particular, artificial intelligence can assist in determining the location of edges in the contour. Furthermore, artificial intelligence can, if necessary, reduce the number of linear equations required to determine the location of all electrode edges.

[0069] In particular, the analysis of the contours is achieved through a convolutional neural network (CNN). The CNN learns from a dataset of synthetic, i.e. artificially generated stacks of known positions of the electrode sheet edges, in order to determine the position of the edge of each electrode sheet from the contours detected in step d) of the stack.

[0070] In order to apply CNN and thereby determine the set of equations proposed here, it is especially necessary to perform at least the recording or generation and analysis of profiles according to steps b) and c) on a number that is the same as the number of electrode sheets to be measured.

[0071] To train a CNN, a synthetic dataset can be generated that maps random stacking layouts (number of electrode pieces in the stack, orientation in the plane along with known edge positions, etc.) to different contours, a first contour, and at least one second contour. The CNN then models the inverse mapping, i.e., the mapping of the contours to the stacking layout.

[0072] Analysis can also be performed using another machine-based or automated learning method without using a convolutional neural network. The following section introduces convolutional neural networks and the procedures and terminology used herein.

[0073] This type of CNN is used to analyze images or contours, i.e., images of the detector, which are essentially known. Here, it is suggested that CNN be used for quality analysis of the (cut) edges of electrode sheets, i.e., within the scope of manufacturing battery components.

[0074] Within the scope of CNN-based analysis, to automate and enable online edge analysis, a training dataset, i.e., a synthetic dataset, can be generated first. The direction of the edges can be manually labeled on each contour of the detector in this training dataset, or on each first image. This manual labeling, i.e., the direction of the edge markings, is then manually exported from the tool. The direction of the edges in the contours, encoded as a pixel matrix, reflects the stacked geometry or edge layout for the training dataset, i.e., the so-called ground truth.

[0075] These quantities of images contain only pixels with values ​​of zero ("0") or one ("1"), for the corresponding locations in the outline or first image where there is no edge (e.g., value 0) or an edge (e.g., value 1).

[0076] Next, a convolutional neural network is used to learn a first mathematical mapping from the edges shown in the first image to their corresponding geometric shapes. The CNN trained in this way can then identify geometric shapes from a second image of a detector or contour that has not been previously learned. Due to the low variance of the first and second images of essentially the same object—in this case, the edges of stacked electrode sheets with defined edge target geometry—and due to the statistical significance of the large data volume, this identification is more accurate than similarity methods, such as trend edge recognition.

[0077] A CNN is known to consist of a series of so-called convolutional layers, which discretely convolve a fixed number of filters with local areas of an image. For each filter, the layer computes a so-called feature map. This feature map indicates whether a pattern defined by the filter parameters is recognized at a corresponding location in a corresponding second image or contour. The size of these feature maps is reduced by so-called max-pooling or average-pooling layers to decrease computational complexity. Max-pooling or average-pooling layers move an n x n window across the feature map and, in particular, only transfer the maximum value from the local area to the next layer.

[0078] The order and number of convolutional layers and max-pooling or average-pooling layers, along with the corresponding window and filter sizes, are known as hyperparameters. Optimization of these hyperparameters is primarily accomplished using a validation dataset, which has no influence on the optimization of the model parameters.

[0079] In the final step, the values ​​of all feature maps are concatenated into a vector, a process called flattening, and this vector is used as the input to the feedforward neural network. This network is characterized by a variable number of hidden layers, and the number of neurons in each hidden layer is also variable. These numbers constitute additional hyperparameters.

[0080] As an alternative to flattening, the compressed feature map can first be converted back to its original size through transposed convolution, and then its number can be reduced to 1 through convolutional layers.

[0081] In its output layer, the network attempts to approximate the manually generated stacking geometry of the stacking truth by assigning each pixel a zero or a one (“1”).

[0082] At the start of training, the filter parameters and the parameters of the feedforward neural network (which together constitute the CNN) can be randomly initialized, which initially leads to inaccurate geometry predictions. During training, all model parameters are adapted using a method called gradient descent to minimize the number of misclassified pixels across all training instances.

[0083] After training, a CNN can be used to identify the orientation or location of at least one edge in a second image of a contour or detector for unknown stacks or newly created contours, for example, within the scope of steps f) and / or g).

[0084] In particular, in another step h), at least one process parameter used to manufacture the corresponding stack is determined and changed from the analysis of stacking accuracy according to step g) in order to improve the stacking accuracy for the other stack.

[0085] In particular, if, for example, it is determined that the limit value is exceeded and / or verified to exceed the limit value within the range of supplementary measurements, the mispositioned electrode piece and its deviation from the target position can be identified. Accordingly, the manufacturing process can be traced back from the understanding of the electrode piece, and the settable process parameters can be changed if necessary.

[0086] In particular, a system for data processing is proposed, the system having devices that are suitably equipped, configured or programmed to implement the method, or the devices implementing the method.

[0087] The device includes, for example, a processor and a memory, the memory storing commands to be executed by the processor, and the device also includes data lines or transmission means for transmitting commands, measurements, data or the like between the aforementioned components.

[0088] Furthermore, a computer program is proposed that includes commands that, when executed by a computer, cause the computer to perform the method or the steps of the method.

[0089] Furthermore, a computer-readable storage medium is proposed, comprising a command that, when executed by a computer, causes the computer to perform the method or the steps of the method.

[0090] The design scheme of the method can be transferred in particular to the system used for data processing and / or the computer-implemented method (i.e., computer program and computer-readable storage medium), and vice versa.

[0091] The use of the indefinite article "a" is not understood as a quantifier, especially in the claims and the description interpreting the claims. Accordingly, the relevant terms and components are understood to exist at least once, but in particular, they may exist multiple times.

[0092] It should be noted that the ordinal numbers used herein (“first,” “second,” etc.) are primarily (only) used to distinguish multiple objects, quantities, or processes of the same kind; that is, they do not, in particular, mandate the relationship and / or order between these objects, quantities, or processes. If a relationship and / or order is required, it will be explicitly stated herein or will be obvious to those skilled in the art when studying the specifically described design. As long as a component can appear multiple times (at least one), the description of one such component also applies to all or most of that component, but this is not necessarily the case. Attached Figure Description

[0093] The invention and its technical field are further described below with reference to the accompanying drawings. It should be noted that the invention is not limited to the embodiments described. In particular, unless otherwise explicitly stated, certain aspects can be extracted from the facts illustrated in the figures and combined with other components and knowledge derived from this description. It should be particularly noted that the drawings and the dimensional relationships shown are merely schematic. Wherein:

[0094] Figure 1 The first embodiment of the method;

[0095] Figure 2 The second embodiment of the method;

[0096] Figure 3 Step f) of the method according to the first implementation scheme;

[0097] Figure 4 Steps b), d), and g) of the method for the first stacking;

[0098] Figure 5 Steps b), d), and g) of the method for the second stacking;

[0099] Figure 6 Steps c) and e) of the method for the first stacking;

[0100] Figure 7 Steps c) and e) are repeated in the method for the first stacking.

[0101] Figure 8 Steps c) and e) of the method for the second stacking;

[0102] Figure 9 Steps c) and e) are repeated in the method for the second stacking. Detailed Implementation

[0103] Figure 1 A first embodiment of the method is shown. The method is used within the scope of methods for manufacturing single-cell batteries. Here, electrode sheets 1, 2, 3, cut to appropriate shapes—namely, anode sheets, cathode sheets, and, if necessary, separator sheets—are arranged in a predetermined order into a stack 5 and oriented relative to each other. In the stack 5 thus generated, the individual electrode sheets 1, 2, 3 should be arranged relative to each other in a position that is as aligned as possible.

[0104] Electrode sheets 1, 2, and 3 extend in mutually parallel planes 4 and are stacked on top of each other to form a stack 5. The stack 5 includes multiple electrode sheets 1, 2, and 3.

[0105] Electrode plates 1, 2, and 3 are each substantially rectangular in shape. Discharge tubes extend beyond the rectangular shape at the edges 9 of electrode plates 1, 2, and 3.

[0106] The stacking precision specifies the relative positions 6, 7, and 8 of the edges 9 of all electrode plates 1, 2, and 3 within stack 5. Electrode plates 1, 2, and 3 should be arranged in a predetermined relative position. Because the anode and cathode plates, as well as any necessary separators, can have different dimensions, the stacking precision is determined at the edges 9 of electrode plates 1, 2, and 3, which are aligned with each other along a first direction 21 extending perpendicular to plane 4. The stacking precision of electrode plates 1, 2, and 3 is determined only at one edge 9 of each electrode plate 1, 2, and 3.

[0107] The method is implemented using a measuring device 10 having a two-dimensional resolution X-ray apparatus 11, which includes a radiation source 12 and a detector 13 for X-ray radiation. The radiation source 12 emits X-ray radiation along an irradiation direction 15. The detector 13 records the X-ray radiation to display an X-ray image.

[0108] Detector 13 generates two-dimensional images (hereinafter referred to as contours 17, 19) displaying X-ray radiation. The proposed method enables the accurate determination of the stacking of electrode plates 1, 2, and 3 from these two-dimensional images of detector 13.

[0109] According to step a), a stack 5 is provided and arranged in the measuring device 10 between the radiation source 12 and the detector 13. According to step b), the stack 5 is irradiated by the radiation source 12 from a first spatial coordinate 14, wherein the irradiation direction 15 is perpendicular to the plane 4 and extends toward the detector 13 (substantially along the first direction 21). The rays 16 of the radiation source 12 detect the overlapping edges 9 of the electrode plates 1, 2, 3 and project a two-dimensional first profile 17 of the edges 9 of the stack 5 onto the detector 13. The radiation source 12 is positioned exactly above the edges 9, i.e., without any lateral displacement relative to the edges 9.

[0110] To arrange the stack 5, it is assumed that the edges 9 of electrode plates 1, 2, and 3 are in a predetermined target position. Within the scope of this method, the actual positions 6, 7, and 8 of the edges 9 deviating from this target position are determined.

[0111] According to step d), the first contour 17 is detected by the detector 13.

[0112] According to step c), the stack 5 is irradiated from at least one second spatial coordinate 18 different from the first spatial coordinate 14 by radiation source 12, wherein the ray 16 detects the edge 9 of the overlapping arrangement of electrode sheets 1, 2, 3 and projects the two-dimensional second contour 19 of the edge 9 of the stack 5 onto the detector 13.

[0113] According to step e), the second contour 19 is detected by the detector 13.

[0114] As shown in the figure, steps c) and e) are repeated, wherein the second spatial coordinate 18 of each step c) is different from the previous second spatial coordinate 18 of the previous step c). This additional spatial coordinate of the repetition of steps c) and e) is called the third spatial coordinate 25, and the contour detected here is called the third contour 26.

[0115] According to step f), the different contours 17, 19, 26 are analyzed and the positions 6, 7, 8 of the edges 9 of the electrode plates 1, 2, 3 are determined. This analysis is performed by system 27 for data processing.

[0116] The measuring device 10 includes a system 27 for data processing, the system having devices (Mittel) suitably equipped, configured, or programmed to implement the method, or the device implementing the method. The devices include, for example, a processor and a memory storing commands to be executed by the processor, and the devices also include data lines or transmission means for transmitting commands, measurements, data, or the like between the aforementioned components.

[0117] The contours 17, 19, and 26 detected by detector 13 comprise two-dimensional images, in which the edges 9 of electrode plates 1, 2, and 3 can be identified based on the transitions between color intensities. Assigning the edges 9 present in contours 17, 19, and 26 to the individual electrode plates 1, 2, and 3 is not easily achievable. The method proposes a feasible scheme to achieve this assignment.

[0118] Within the scope of this method, multiple contours 17, 19, and 26 of the stack 5 are generated by radiation source 12 and detected by detector 13. Due to the different arrangements of radiation source 12 relative to the stack 5 or edge 9, contours 17, 19, and 26 that are different from each other are generated. These contours 17, 19, and 26 can be analyzed by linear equations, i.e., from the known layout of radiation source 12 and detector 13 and the linear direction of ray 16 generated by radiation source 12, i.e., straight line direction, and the positions 6, 7, and 8 of the corresponding edges 9 in the stack 5 can be inferred from the positions of the edges 9 in the corresponding contours 17, 19, and 26.

[0119] The first spatial coordinate 14 differs from at least one second spatial coordinate 18, 25 (i.e., the second and third spatial coordinates) in that its distance 22 from the edge 9 is different from each other, wherein the distance 22 extends along a second direction 23 that is parallel to the plane 4 and perpendicular to the edge 9. The third direction 28 extends parallel to the edge 9 as measured by the measuring device 10.

[0120] Each contour 17, 19, and 26 corresponds to its spatial coordinates 14, 18, and 25 via arrows.

[0121] Figure 2 A second embodiment of the method is shown. (References to...) Figure 1 Explanation.

[0122] In the first embodiment, the radiation source 12 moves parallel to plane 4 to the second spatial coordinate 18 between steps b) and d). Unlike the first embodiment, in the second embodiment, the radiation source 12 moves perpendicular to plane 4 to the second spatial coordinate 18 or to the third spatial coordinate 25 between steps b) and d). The resulting different contours 17, 19, 26 also achieve a one-to-one correspondence between the edges 9 of each electrode sheet 1, 2, 3.

[0123] The first spatial coordinate 14 and at least one second spatial coordinate 18, 25 (i.e., the second and third spatial coordinates) are distinguished by their spacing 20 from the stack 5. The spacing 20 extends along a first direction 21 perpendicular to the plane 4.

[0124] Each contour 17, 19, and 26 corresponds to its spatial coordinates 14, 18, and 25 via arrows.

[0125] Figure 3 Step f) of the method according to the first embodiment is shown. References to... Figure 1 Explanation.

[0126] According to step f), the different contours 17, 19, 26 are analyzed and the positions 6, 7, 8 of the edges 9 of the electrode plates 1, 2, 3 are determined. This analysis is performed by system 27 for data processing.

[0127] The stack 5, radiation source 12, and detector 13 are observed in a common plane. Therefore, the spatial coordinate z is the same for all components 1, 2, 3, 5, 12, and 13. The first direction 21 is perpendicular to plane 4 along the y-axis (here, z is the direction of the first direction). Figure 3 The vertical axis in the upper diagram extends. The second direction 23 is parallel to plane 4 along the x-axis (here it is...). Figure 3 The horizontal axis in the upper diagram extends. The third direction 28 extends along the z-axis (here, it is...). Figure 3 The axis pointing backward in the upper figure extends.

[0128] The first spatial coordinate 14 is (x q1 |y q1 The second spatial coordinate 18 is (x q2 |y q2 The spatial coordinates to be determined for the edge 9 of the first electrode plate 1 are called (x...). e |y e In the first contour 17 detected by detector 13, the position of the edge 9 of the first electrode plate 1 with respect to the first spatial coordinates 14 of the radiation source 12 is (x... d1 |y d1 In the second contour 19, with respect to the second spatial coordinates 18 of the radiation source 12, the position of the edge 9 of the first electrode plate 1 is (x... d2 |y d2 ).

[0129] The equation of the straight line for the first spatial coordinate 14 of radiation source 12 applies:

[0130] (1)

[0131] (2) y = m1·(xx q1 )+y q1

[0132] (3) y=m1x+b1

[0133] The equation for the straight line at the second spatial coordinate 18 of radiation source 12 applies:

[0134] (1)

[0135] (2) y = m2·(xx) q2 )+y q2

[0136] (3) y = m2x + b2

[0137] If these two equations are equal, then the following applies:

[0138] (4) m1·x+b1=m2·x+b2

[0139] (5)→(m1-m2)·x=b2-b1

[0140] (6)

[0141] (7)

[0142] (x e |y e The spatial coordinates indicate the first position 6 of the edge 9 of the first electrode sheet 1.

[0143] Figure 4 Steps b), d), and g) of the method for the first stack 5 are shown. References to... Figure 1 Explanation.

[0144] In the first stack 5, the bottommost first electrode piece 1 has a maximum deviation 24 from the target position. Stacking accuracy is analyzed in step g). For stack 5, the maximum deviation 24 of the contours 17, 19, 26 (or the identifiable position of edge 9 within the contours) from the target position of edge 9 is predetermined. It is assumed that the maximum deviation 24 is generated by the electrode pieces 1, 2, 3 closest to detector 13.

[0145] The maximum deviation 24 is the maximum permissible difference between the target position of edge 9 in stack 5 and the actual positions 6, 7, and 8 of edge 9. The maximum deviation 24 is determined at the first profile 17, taking into account the target position of the first spatial coordinate 14 relative to edge 9.

[0146] To determine the stacking accuracy, only steps a), b), d), f), and g) are performed first. Only in step g) (see...) Figure 4 Steps c) and e) are only performed when the limit value is exceeded (see...). Figure 8 and Figure 9 ).

[0147] Within the scope of this method, for example, each stack 5 is checked for stacking accuracy, but the positions of the edges 9 of all electrode pieces 1, 2, 3 of stack 5 are determined only if the limit value is exceeded. Otherwise, for each stack 5, only whether the limit value is exceeded is checked.

[0148] If it is determined that the limit value is exceeded, then steps c) and e) are performed exactly twice with different (second) spatial coordinates 18 and 25, and then steps f) and g) are repeated (see Figure 8 and Figure 9 ).

[0149] This selection of the first spatial coordinate 14 means that the radiation source 12 is positioned exactly above the target location on the edge 9 of the stack 5, that is, aligned with the target location on the edge of the stack in the first direction 21 (see...). Figure 4 By selecting two second spatial coordinates 18 and 25 (i.e., the second and third spatial coordinates), the radiation source 12 is arranged in a displaced manner relative to the first spatial coordinate 14 in the second direction 23. It is first displaced towards the stack 5, aligning the radiation source 12 with the stack 5, and then displaced away from the stack 5, placing the radiation source 12 adjacent to the stack 5 on its side (see...). Figure 8 and Figure 9 ).

[0150] If the determination in step g) exceeds the limit value, then steps c) and e) are performed as many times as are needed to determine the positions 6, 7, and 8 of all edges 9 in a one-to-one correspondence.

[0151] Using this staged approach, it is not necessary to measure every stack in production stack 5 (i.e., determine the positions 6, 7, 8 of the edges 9 of all electrode pieces 1, 2, 3). The maximum deviation 24 in stack 5 can be detected or estimated based on a smaller set of contours 17, 19, 26, i.e., a small number of records from detector 13. If the limit is exceeded, the corresponding positions 6, 7, 8 of electrode pieces 1, 2, 3 can be determined through additional measurements.

[0152] Figure 5 Steps b), d), and g) of the method for the second stack 5 are shown. References to... Figure 4 Explanation.

[0153] In the second stack 5, the topmost first electrode piece 1 has a maximum deviation 24 from the target position. Stacking accuracy is analyzed in step g). For stack 5, a limit value for the maximum deviation 24 between the contours 17, 19, 26 (or the identifiable position of edge 9 within the contours) and the target position of edge 9 is predetermined. It is assumed that the maximum deviation 24 is generated by the electrodes 1, 2, 3 closest to detector 13. For the second stack 5 shown, the maximum deviation 24 identifiable in the first contour 17 is much smaller (if the first electrode piece 1 is arranged at the bottom in stack 5). In this case, an unacceptably large deviation 24 cannot be identified.

[0154] If it is determined that the limit value is exceeded, then steps c) and e) are performed exactly twice with different (second) spatial coordinates 18 and 25, and then steps f) and g) are repeated (see Figure 6 and Figure 7 ).

[0155] Choosing the first spatial coordinate 14 in this way ensures that the radiation source 12 is positioned exactly above the target location on the edge 9 of the stack 5, that is, aligned with the target location on the edge of the stack in the first direction 21 (see...). Figure 5 By selecting two second spatial coordinates 18 and 25 (i.e., the second and third spatial coordinates), the radiation source 12 is arranged in a displaced manner relative to the first spatial coordinate 14 in the second direction 23. It is first displaced towards the stack 5, aligning the radiation source 12 with the stack 5, and then displaced away from the stack 5, placing the radiation source 12 adjacent to the stack 5 on its side (see...). Figure 6 and Figure 7 ).

[0156] List of reference numerals in the attached diagram:

[0157] 1 First electrode plate

[0158] 2 Second electrode plate

[0159] 3 Third electrode plate

[0160] 4 planes

[0161] 5. Stacking

[0162] 6 First position

[0163] 7 Second position

[0164] 8 Third position

[0165] 9. Edge

[0166] 10 Measuring Device

[0167] 11 X-ray equipment

[0168] 12 radiation sources

[0169] 13 detectors

[0170] 14 First spatial coordinates

[0171] 15. Irradiation direction

[0172] 16 rays

[0173] 17 First Outline

[0174] 18 Second Space Coordinates

[0175] 19 Second Outline

[0176] 20 spacing

[0177] 21 First direction (y-axis)

[0178] 22 Distance

[0179] 23 Second direction (y-axis)

[0180] 24 Deviation

[0181] 25 Third Space Coordinates

[0182] 26 Third Outline

[0183] 27 System

[0184] 28. Third direction (x-axis)

Claims

1. A method for determining the stacking accuracy of multiple electrode sheets (1, 2, 3), wherein, Electrode sheets (1, 2, 3) extend and stack on top of each other in mutually parallel planes (4) to form a stack (5); wherein, stacking accuracy describes the relative positions (6, 7, 8) of the edges (9) of all electrode sheets (1, 2, 3) in the stack (5); wherein, the method is implemented by a measuring device (10) having a two-dimensional resolution X-ray device (11) with at least one radiation source (12) and a detector (13) for X-ray radiation, and the method includes at least the following steps: a) Provide a stack (5) and arrange the stack (5) in the measuring device (10) between at least one radiation source (12) and a detector (13); b) Irradiate the stack (5) from a first spatial coordinate (14) with the at least one radiation source (12), wherein the irradiation direction (15) is at least perpendicular to the plane (4) and extends toward the detector (13), wherein the rays (16) of the radiation source (12) detect the overlapping edges of the electrode plates (1, 2, 3) and project a two-dimensional first profile (17) of the edge (9) of the stack (5) onto the detector (13); c) Irradiate the stack (5) from at least one second spatial coordinate (18) different from the first spatial coordinate (14) using the at least one radiation source (12), wherein the ray (16) detects the overlapping edge (9) of the electrode sheet (1) and projects a two-dimensional second profile (19) of the edge (9) of the stack (5) onto the detector (13). d) Detect the first profile (17) using the detector (13); e) Detect at least one second profile (19) using the detector (13); f) Analyze the different contours (17, 19) and determine the positions (6, 7, 8) of the edges (9) of the electrode plates (1, 2, 3), wherein in step f), each edge (9) in the corresponding contour (17, 19) is mapped to the corresponding spatial coordinates (14, 18) according to the equation of the straight line. g) Analyze the stacking accuracy; wherein, for the stack (5), a limit value for the maximum deviation (24) of the target position (25) of the contour (17, 19) and the edge (9) is predetermined; wherein, for the maximum deviation (24), it is assumed that the electrode (1, 2, 3) closest to the detector (13) produces the maximum deviation (24). Specifically, to determine the stacking accuracy, only steps a), b), d), f), and g) are performed initially. Steps c) and e) are only performed if the limit value is exceeded in step g).

2. The method according to claim 1, wherein, The first spatial coordinate (14) differs from at least one second spatial coordinate (18) in that it is different from the spacing (20) of the stack (5), wherein the spacing (20) extends along a first direction (21) perpendicular to the plane (4), or it differs from the distance (22) of the edge (9), wherein the distance (22) extends along a second direction (23) parallel to the plane (4) and perpendicular to the edge (9).

3. The method according to claim 1, wherein, If it is determined that the limit value is exceeded, then steps c) and e) are performed exactly twice with different spatial coordinates (18, 25), and then steps f) and g) are repeated.

4. According to the method of claim 3, if the determination in step g) exceeds the limit value, then steps c) and e) are performed as many times as required to determine the positions (6, 7, 8) of all edges (9) in a one-to-one correspondence.

5. The method according to claim 1 or 2, wherein, Artificial intelligence should be used at least for step f).

6. The method according to claim 5, wherein, The analysis of the contours (17, 19) is achieved by a convolutional neural network; wherein the convolutional neural network learns from a dataset of a stack (5) of synthesized material having known positions (6, 7, 8) of the edges (9) of the electrode pieces (1, 2, 3) so as to subsequently determine the position (6, 7, 8) of the edge (9) of each electrode piece (1, 2, 3) from the contours (17, 19) detected according to step d).

7. The method according to claim 1 or 2, wherein, In another step h), at least one process parameter used to manufacture the corresponding stack (5) is determined and changed from the analysis of stacking accuracy according to step g) in order to improve the stacking accuracy for the other stack (5).

Images