Square battery flatness detection system and method

Abstract

This invention discloses a planarity testing system and method for square batteries. The testing system includes two sets of laser scanning devices arranged longitudinally and longitudinally along the length of a linear slide rail, and a battery carrier plate that moves along the linear slide rail to transport the battery to be tested to the testing station. A toothed disc is rotatably connected to a slide block of the linear slide rail at the bottom of the battery carrier plate. Each set of laser scanning devices includes two line laser scanning devices symmetrically arranged along the central axis of the linear slide rail, used for vertically moving and scanning the battery. A rack is arranged between the two sets of laser scanning devices on one side of the linear slide rail. After the battery to be tested is scanned by the first set of laser scanning devices, when the battery carrier plate moves to the rack position, the toothed disc engages with the rack, causing the battery carrier plate and the battery it carries to rotate synchronously by 90 degrees, allowing the second set of laser scanning devices to perform a side scan. This invention improves testing efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the field of square battery testing technology, and in particular to a square battery flatness testing system and method. Background Technology

[0002] Square aluminum-cased batteries are the mainstream form of new energy power batteries. During the manufacturing process, the large surface and sides of the aluminum casing are prone to deformation. Several factors contribute to this deformation, including uneven stress release during the stamping and demolding process, stress differences between the large surface and sides during peripheral welding, pressure differences caused by vacuuming during helium testing, baking, electrolyte injection, and formation processes, and negative pressure generated by the consumption of free electrolyte within the casing during capacity testing and aging. Therefore, battery flatness directly determines the compatibility of the PACK module assembly. Excessive flatness can lead to module structural failure, abnormal thermal management, and compromised cell consistency, making it a core testing indicator that must be strictly controlled in mass production. Thus, square cell flatness testing is a necessary step in battery manufacturing. Traditional manual methods using feeler gauges and height gauges are prone to scratching and bumping the cells, resulting in subjective biases, low efficiency, and unsuitability for large-scale mass production. Some technologies employ automated 3D laser / vision inspection solutions, which generally require the battery to be paused and positioned on the logistics line. They lack motion compensation mechanisms, limiting mass production cycle time. Furthermore, the reference area often selects simple edges or overall fitting, failing to incorporate the battery's edge structure for precise positioning, resulting in poor reference stability. For example, CN110596143A discloses a 3D laser camera, combined with lifting, flipping, and rotating devices, to scan six sides of a square battery. The battery is placed vertically on a battery positioning module, with the largest face facing the largest face detection module. The battery positioning module rotates the battery 180°, so that the other largest face faces the largest face detection module, thus completing the large-face scan. The flipping and lifting device transfers the battery to a rotating component at the inspection station. Small face and edge detection modules, along with vertical edge detection modules, scan the battery. The rotating component drives the battery to rotate, achieving scanning of four small faces and 12 edges. This technology requires the battery to remain at the inspection station, making online detection of its motion impossible.

[0003] CN119533289A discloses a battery cell size measurement system and method, which uses a U-shaped conveyor belt in conjunction with a through-beam 3D laser to scan the large surface and sides of the battery cell to measure its size. Its structure is complex.

[0004] Some technologies lack targeted screening methods for abnormal point clouds, resulting in significant measurement interference. Furthermore, the measurement area is not finely divided, and the utilization rate of the laser detection limit is low, thus lowering detection accuracy. For example, CN121504898A discloses an image recognition-based method for detecting the flatness of power batteries. This method uses a 3D camera to acquire point cloud images of the battery surface, divides the point cloud into rectangular grids, and selects the points closest to the centroid within the grid as candidate points. An initial plane is constructed using three non-collinear candidate points. Through iterative projection and vertical deviation adjustment until convergence, a reference plane is fitted. The flatness of the measured surface is evaluated by calculating the vertical deviation of each vertex in the battery surface point cloud from the fitted reference plane. However, its flatness detection accuracy is not high. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings and defects of existing technologies and provide a flatness detection system and method for square batteries. Combining the structural characteristics of square battery casings and leveraging the advantages of non-contact measurement based on laser ranging principles, this invention enables efficient, non-destructive, and accurate detection of battery flatness, meeting the mass production testing needs of square aluminum-cased batteries.

[0006] According to one aspect of this application, a planar battery flatness detection system is provided, comprising two sets of laser scanning devices arranged back-to-back along the length of a horizontally arranged linear slide rail, and a battery carrier plate that moves along the linear slide rail to transport the battery to be tested to a detection station. The bottom of the battery carrier plate is provided with a toothed disc rotatably connected to a slide block of the linear slide rail. Each set of laser scanning devices includes two line laser scanning devices symmetrically arranged along the central axis of the linear slide rail, the line laser scanning devices being used for vertically moving and scanning the battery. Each set of laser scanning devices is arranged on a pair of vertical linear slide rails. Between the two sets of laser scanning devices, a rack is arranged on one side of the linear slide rail along its length. After the battery to be tested is scanned on a large surface by the first set of laser scanning devices at the first detection station, when the battery carrier plate moves forward to the rack position, the toothed disc engages with the rack and is driven to rotate. The battery carrier plate and the battery it carries rotate synchronously by 90 degrees, and then the battery is scanned on the side by the second set of laser scanning devices at the second detection station.

[0007] Preferably, a boss is coaxially provided at the top center of the gear disc, and the boss is fixed to the bottom surface of the battery carrier plate.

[0008] Preferably, the toothed disk is provided with a rotational position detector in the radial direction for detecting the position of the toothed disk.

[0009] Preferably, the rotational position detection includes at least a photoelectric detection component, wherein the emitting part and the receiving part of the photoelectric detection component are respectively arranged on the upper and lower parts of the toothed disk near the outer peripheral surface.

[0010] Preferably, the slide is provided with a limiting component, including at least a push rod that is perpendicular to the gear plate. The push rod cooperates with a limiting hole on the surface of the gear plate to limit the gear plate at a preset rotation position.

[0011] Preferably, the laser scanning device includes at least a laser scanning device.

[0012] Preferably, the laser scanning device is connected to a data processing module, which is used to detect and judge the flatness of the battery based on the scanning data of the large surface and side surface of the battery by the laser scanning device.

[0013] According to another aspect of this application, a method for detecting the flatness of a square aluminum-cased battery is provided, comprising the steps of detecting and judging the flatness of the battery using the square battery flatness detection system: At the first inspection station, the first set of laser scanning devices scans the large surface of the battery by vertical movement. Then, the battery is rotated 90 degrees during transport and transported to the second inspection station. At the second inspection station, the second set of laser scanning devices scans the side of the battery by vertical movement. Based on the movement speed of the battery carrier and the scanning frequency of the line laser scanning device, a compensation correction value in the vertical direction is determined. The compensation correction value is then used to compensate for each row of point cloud data obtained in the vertical direction to obtain the true coordinates of each row of point cloud. Based on the true coordinates of the point cloud, discrete points are removed by comparing the point cloud height value with the average height value to obtain a standard point cloud. A pre-selected area is selected based on the connection area between the battery's large / side surface and the shell's edge. The pre-selected area includes the connection area and the battery's large / side surface and the chamfered edge area on both sides of the connection area. A reference area is determined from the pre-selected area based on a standard point cloud using an iterative algorithm. The height of the measurement reference surface is determined based on the height of the reference area. A preset rectangular measurement area is selected on the large / side surface of the battery and uniformly gridded. Based on the height value of the point cloud in the rectangular grid area after gridding and the depth of field of the laser scanning device, the flatness of the rectangular grid area is detected by the bisection method. The height mean is calculated based on the height value of the point cloud in the rectangular grid area with the expected flatness. The height mean of all rectangular grid areas with the expected flatness constitutes a height matrix. The flatness of the battery's large / side surface is determined based on the difference between the height matrix and the height of the measurement reference plane.

[0014] Preferably, an iterative algorithm is used to determine a reference region from a pre-selected area based on a standard point cloud, and the height of the measurement reference surface is determined according to the height of the reference region, including: A region is selected in the Z0Y plane using a rectangular bounding box, with a section of the surface or side edge selected as the filtering area. An iterative algorithm is used to partition and compare the filtering area, eliminating discrete partitions until the partition width meets a threshold. The region whose partition width meets the threshold is then determined as the target region. The partitioning and elimination of discrete partitions in the filtering area includes: dividing the target region into multiple equal partitions along the Y-axis; calculating the absolute value of the height difference between adjacent partitions based on the average height of the point clouds within the multiple partitions; and deleting the right / left partition by comparing the left and right values ​​of the absolute height difference to obtain a new filtering area. Based on the average height of the point cloud in multiple partitions of the target area, calculate the absolute value of the difference in average height between adjacent partitions; calculate the absolute value of the difference in average height between the two partitions corresponding to the smallest difference in average height and their respective adjacent partitions, and determine a reference area for an edge based on the partition corresponding to the smaller calculated value; determine the measurement reference plane height of the large / side surface based on the height of the reference areas of two adjacent edges.

[0015] Preferably, the flatness of the rectangular grid area is detected using a bisection method based on the height values ​​of the point cloud within the rectangular grid area formed after meshing and the depth of field of the laser scanning device, including: Using a rectangular grid area as the target area, the X-axis height values ​​of the point cloud within the target area are sorted in ascending order to obtain a height sequence. The mean of the height values ​​of each preset proportion in the first and second segments of the height sequence is calculated, and the absolute value of the difference between the mean values ​​is calculated. The absolute value of the difference is compared with the preset proportion accuracy value of the depth of field of the laser scanning device. If the absolute value of the difference is less than the preset proportion accuracy value, the flatness of the target area is determined to meet the expectation. Otherwise, the length and width parameters of the target area are reduced proportionally to halve the area, forming a new target area for iterative judgment. The iteration ends when the absolute value of the difference is less than the preset proportion accuracy value.

[0016] The square battery flatness detection system and method of the present invention uses a rotatable battery carrier plate and achieves a single 90° change in battery posture through the meshing of a toothed disk and a rack to complete the detection of the large surface and side surface. After the detection is completed, the tangent position between the edge of the shell and the large surface / side surface is found as a reference plane. The measurement area of ​​the large surface / side surface is pixelated and the optimal measurement area is divided by a bisection method. The flatness of the measurement surface is reflected by the height difference between the optimal measurement area and the reference plane, thereby improving the accuracy of the detection results.

[0017] The prismatic battery flatness detection system and method of the present invention realizes continuous automated flatness detection of prismatic batteries in mass production lines. Through motion compensation, abnormal point cloud screening, reference area positioning and measurement area selection calculation methods, the flatness detection accuracy of batteries can be accurately measured and the detection efficiency can be greatly improved. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the overall structure of the square battery flatness detection system of the present invention.

[0019] Figure 2 This is an isometric schematic diagram of the engagement between the toothed disc and the rack below the battery carrier plate of the present invention.

[0020] Figure 3 This is a top view schematic diagram of the engagement between the toothed disc and the rack below the battery carrier plate of the present invention.

[0021] Figure 4 This is a front view schematic diagram of the engagement between the toothed disc and the rack below the battery carrier plate of the present invention.

[0022] Figure 5 This is a side view of the toothed disc and rack below the battery carrier plate of the present invention.

[0023] Figure 6 This is a schematic diagram of the laser scanning device and the electric motion direction of the square battery flatness detection system of the present invention.

[0024] Figure 7 This is a schematic diagram showing the cross-sectional position where the edge of the square aluminum-cased battery connects to the large / side surface.

[0025] Figure 8 This is a schematic diagram of the X-axis height at the beginning and end of each of the 10% points in the point cloud height sequence of 400 equally divided regions measured in an embodiment of the present invention.

[0026] Figure 9 This is a schematic diagram of the measured X-axis height of 400 equally divided regions in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0028] This invention achieves front-to-front scanning of the large surface and side surfaces during movement by automatically rotating the battery in posture in conjunction with symmetrical linear laser scanning. It uses a motion compensation formula to correct the coordinates of the linear laser point cloud, obtaining a distortion-free spatial point cloud of the large / side surfaces. Outliers are eliminated by filtering the height difference between the relative mean values ​​of the point clouds within a defined area using three times the standard deviation. Then, using the battery edge as a reference area, the flattest sub-region is determined through iterative division and gradient elimination, serving as the reference area. The height of the measurement reference surface is constructed using the average height of adjacent edge reference areas. Finally, the measurement area is pixelated and subdivided. Sub-regions with height fluctuations less than the laser accuracy tolerance are located through iterative binary division. The set of deviations of each sub-region relative to the reference surface is used as a flatness evaluation index.

[0029] See Figures 1 to 6As shown, the square battery flatness detection system of this application includes two sets of laser scanning devices arranged back-to-back along the length of a horizontally arranged linear slide rail 9, and a battery carrier plate 11 that moves along the linear slide rail to transport the battery 26 to be tested to the detection station. The bottom of the battery carrier plate 11 is provided with a toothed disk 13 that is rotatably connected to the slide seat 16 of the linear slide rail. Each set of laser scanning devices includes two line laser scanning devices symmetrically arranged along the length central axis of the linear slide rail. The line laser scanning devices are used to vertically move and scan the battery. Each set of laser scanning devices is arranged on a pair of vertical linear slide rails. Between the two sets of laser scanning devices, a rack 14 is arranged on one side of the linear slide rail along its length. After the battery 26 to be tested is scanned on a large surface by the first set of laser scanning devices at the first detection station, when the battery carrier plate moves forward to the rack position, the toothed disk 13 engages with the rack 14 and is driven to rotate. The battery carrier plate 11 and the battery it carries rotate synchronously by 90 degrees. Then, at the second detection station, the battery 26 is scanned on the side by the second set of laser scanning devices.

[0030] The first group of laser scanning devices consists of a first line scanning laser device 1 and a second line scanning laser device 2, and the second group of laser scanning devices consists of a third line scanning laser device 3 and a fourth line scanning laser device 4. The vertical linear slide rails include a first vertical linear slide rail 5, a second vertical linear slide rail 6, a third vertical linear slide rail 7, and a fourth vertical linear slide rail 8. Each group of laser scanning devices moves synchronously to scan the large surface or side surface.

[0031] In this application, the laser scanning device uses line-scanning lasers to accurately measure the concave and convex deformations of the battery's large surface and sides during the manufacturing process, which is suitable for automated flatness inspection of square aluminum-cased batteries in mass production lines. In this application, two sets of line-scanning laser devices are symmetrically arranged along a linear slide rail, respectively installed on the vertical slide rail and capable of synchronous lifting and lowering. During inspection, the battery 26 to be inspected is fed into the battery carrier plate 11 by the feeding conveyor belt 10 with its large surface parallel to the material flow line. A parallel positioning step 12 can be provided at the bottom of the battery carrier plate to position the battery. Subsequently, the battery carrier plate 11 passes through the first inspection station at a uniform speed along the linear slide rail. The two laser scanning devices of the first set of laser scanning devices move synchronously from top to bottom, scanning the large surface of the battery before resetting. After the inspection is completed, the battery carrier plate 11 continues to move along the linear slide rail towards the second inspection station. During the movement of the testing station, the bottom gear plate 12 of the battery carrier plate meshes with the rack 14, driving the battery carrier plate to rotate 90 degrees, so that the battery changes from facing the large side to facing the side of the second set of laser scanning devices. When the battery carrier plate continues to move to the side detection area, i.e. the second detection station, the second set of laser scanning devices moves synchronously from top to bottom at a constant speed, and resets after completing the side scan of the battery. After the detection is completed, the battery carrier plate 11 continues to move along the linear slide rail and is fed into the unloading material line to complete the full scanning of the four sides of the battery, including the large side and the side. The empty battery carrier plate is then reset to the loading area for the next cycle.

[0032] In some embodiments, after the second inspection station, a rack can be set up to cooperate with the battery carrier plate, so that the battery carrier plate and the battery on it can be rotated 90°. After the battery returns to the state where the large surface is parallel to the slide rail axis, the cylinder push rod will transfer the inspected battery to the unloading logistics line.

[0033] The relationship between the rack and the gear disk can be calculated and set according to the needs of rotation control, such as setting the rack length to be the same as the circumference of the gear disk. Preferably, a boss 27 is coaxially provided at the top center of the toothed disc, and the boss is fixed to the bottom surface of the battery carrier plate, thereby fixing the toothed disc to the battery carrier plate.

[0034] To achieve precise control of the rotation angle or position, a rotation position detector for detecting the position of the gear disk is provided in the radial direction of the gear disk. The rotation position detection includes at least a photoelectric detection component 18, such as a grating component. The emitting part 181 and the receiving part 182 of the photoelectric detection component are respectively arranged on the upper and lower parts of the gear disk near the outer peripheral surface. For example, the emitting part is provided on the rack 14, located above the plane of the gear disk, and the receiving part is provided on the linear slide rail platform 30.

[0035] After the battery carrier plate is rotated into position, to prevent movement during detection from affecting the detection effect, a limiting component is provided on the slide block to limit and fix the battery carrier plate after it has been rotated into position, preventing movement. The limiting component includes at least a push rod 17 perpendicular to the gear disk 13, which is disposed on the slide block 16. The push rod cooperates with the limiting hole on the surface of the gear disk to limit the gear disk at a preset rotation position. The push rod can be an electromagnetic push rod, used to pop out or push out from the positioning hole when the photoelectric detection component is blocked, and to spring back and lock into the next positioning hole after the battery carrier plate is rotated 90 degrees. In a preferred embodiment, there are four positioning holes, such as a first positioning hole 19, a second positioning hole 20, a third positioning hole 21, and a fourth positioning hole 22, arranged along the axial and vertical axes of the linear slide rail.

[0036] The slide block 16 and the gear plate 13 are rotatably connected by a bearing 15.

[0037] According to the photoelectric detection component and limiting component of this application, after the large-area inspection of the battery is completed at the first inspection station, during the movement of the battery carrier plate 11 to the second inspection station, when the gear plate 13 and the rack 14 just come into contact and mesh, the photoelectric detection component 18 is blocked, and the push rod 17 pops out from the first positioning hole 19 of the gear plate 13. The gear plate 13 can rotate. After the station completes a 90° rotation, the push rod 17 springs back and locks into the second positioning hole 20. The linear slide rail 9 of the battery carrier plate 11 moves at a constant speed to the side inspection area. The second set of linear lasers moves synchronously from top to bottom at a constant speed to complete the side scanning of the battery. After the scanning is completed, the laser moves up and resets, the battery carrier plate continues to move to the right, the bottom toothed disk 13 meshes with the tooth 23, the photoelectric detection component 18 is blocked, the electromagnetic push rod 17 pops out from the positioning hole 20, the battery carrier plate rotates 90° and the push rod 17 springs back and locks into the positioning hole 21, the battery returns to the state where the large surface is parallel to the slide rail axis, and the cylinder push rod 24 moves the battery to the unloading logistics line 25, completing the full scanning of the large surface and the side of the battery.

[0038] The system of this invention uses linear battery posture rotation control to achieve the conversion detection control from the large surface to the side surface. It does not require too many mechanical structures and movements. The posture is rotated 90 degrees by the meshing of the toothed disk under the battery carrier plate and the rack arranged on one side of the linear slide rail, so that the line scanning laser can smoothly switch from scanning the large surface to scanning the side surface. The battery posture is restored by rotating the posture again by 90 degrees before unloading, which is convenient for logistics line transfer.

[0039] According to an embodiment of this application, the laser scanning device includes at least a laser scanning unit connected to a data processing module. The data processing module is used to detect and determine the flatness of the battery based on the scanning data of the large and side surfaces of the battery from the laser scanning unit. Specifically, the data processing module is housed in an industrial control computer 28, which is connected to the laser scanning unit via a wire 29.

[0040] In this application, the linear scanning laser is specified to move along the Z-axis and to be directly obliquely incident. The laser emission direction is along the X-axis, facing the large / side surface of the battery. The linear slide rail moves along the Y-axis. Since the horizontal movement of the battery carrier plate and the vertical movement of the linear scanning laser are coupled, the original point cloud collected has spatial distortion on the Y-axis (corresponding to the length direction of the battery, i.e. the movement direction of the battery carrier plate). In order to restore the true appearance of the battery, Y-axis compensation is performed on each row of point cloud data to obtain accurate point cloud coordinates.

[0041] Specifically, the compensation formula for the movement is as follows, and the compensation value is calculated using the following formula: Where N represents the scanning line number of the linear laser from top to bottom, N=1,2,3,...; V2 represents the horizontal movement speed of the battery carrier, in m / s; a represents the laser scanning frequency, in Hz; X N This represents the compensation value required for the Nth row of the point cloud in the Y-axis direction, in meters (m).

[0042] Based on the compensation data calculated above, and based on the collected original spatial coordinates... The true spatial coordinates of the point cloud are calculated based on the following formula. : ; The coordinate correction and true shape reconstruction method involves setting the original spatial coordinates of any point P in the Nth row of the point cloud obtained from the original acquisition as its y-coordinate, i.e., the original reading of the direction of movement of the battery carrier (i.e., the Y-axis), and subtracting the corresponding compensation value X under the row number N. N The spatial coordinates of all point clouds on the large / side surface of the battery are obtained. By capturing and correcting the point clouds on the large / side surface, a spatial point cloud image consistent with the shape of the large / side surface is obtained.

[0043] Then, the height of the point cloud measured along the X-axis is calculated. and the average height of point clouds in the defined area The difference is used as the relative height of the measured point cloud. The relative height of all point clouds in the designated area Perform the mean This is used to eliminate discrete point clouds that do not meet the criteria. For example, suppose there are N point clouds within the defined measurement area, and the X-axis height of the i-th point cloud is... The mean point cloud height in this measurement area is: ; The relative height of the i-th point cloud is Sample mean and standard deviation They are as follows: ; ; Based on the mean of the valid point retention conditions Points outside the defined range are identified as discrete points and directly discarded, while valid points are retained for subsequent calculations. A three-standard-deviation filter is applied to the difference between the height of all point clouds within the defined region and the average height of the point clouds within that region, eliminating points outside this three-standard-deviation filter to improve the resolution of discrete point cloud removal.

[0044] Because the four edges 112 connecting the battery's large surface 100 and side surface 111 are curved chamfered structures, and the junction or connection area between the edges 112 and the large surface 100 / side surface 111 is a nearly horizontal cross-section 113, in order to achieve accurate battery flatness detection, this application creatively proposes to select the area where the cross-section 13 connecting the edges and the large surface / side surface is located as the reference plane of the large surface / side surface. This reference plane is parallel to the Z0Y plane, and the flatness of the large surface / side surface is detected by constructing a reference plane height along the X-axis using this reference plane. Figure 7 As shown.

[0045] The following example uses the cross-section where the battery edge connects to the large / side surface as the reference area.

[0046] After line-scan laser scanning and motion compensation are performed on the battery's large / side surface, the point cloud image exhibits different heights along the X-axis, while its Z0Y plane can represent the entire contour. In the Z0Y plane, a virtual wireframe is used to define a candidate region 114 at the edge connecting the battery's large / side surface. The selection rules are as follows: the defined area includes the cut surface 113; the width of the candidate region in the W direction must at least include the chamfered portion of the edge on both sides and extend to a portion of the large surface; the defined length cannot extend beyond the top and bottom surfaces to avoid the influence of the stamping steps on the battery's bottom surface and the welding excess height of the cover plate. The defined region serves as the candidate region for determining the final reference region, and its position and width are as follows... Figure 7 As shown. After selecting the target area, the wireframe area (i.e., the reference area) is divided into n equal parts along the Y-axis, named N from the outside in. 1-1 N 1-2 ...N 1-nFor each region N i Calculate the average height of all point clouds along the X-axis (perpendicular to the large surface direction). The absolute value of the height difference between adjacent areas is Let i = 1, 2, ..., n-1. Eliminate unqualified areas based on the absolute value of the height difference between adjacent areas; elimination rule: compare the difference at the beginning and end. Difference from the tail end Size, if Then remove the left-end region N1, if Then remove the right-end region N. n After removing the remaining n-1 regions, divide the remaining n-1 regions into n equal parts along the Y-axis and repeat the above difference comparison and removal process.

[0047] When the width of the divided region (i.e., the Y-axis span of the wireframe after equal division) Iteration stops when the following expression is satisfied:

[0048] in, The stopping iteration condition, defined as the laser point cloud contour interval between two adjacent laser scan lines in the Y-axis direction (by inherent laser parameters), ensures that the partitioned region width R... n It should contain at least 2-3 point clouds to avoid positioning errors caused by missing points in the point cloud.

[0049] After the iteration stops, the remaining n equally divided regions are denoted as N1´, N2´, …, N. n ´, calculate the average height of each region along the X-axis (j=1…n), take the absolute value of the mean height difference of all adjacent regions. : j=1,2,…,n-1; Find the region corresponding to the minimum absolute value of the difference between the height mean and the mean. Calculate by comparing the absolute value of the difference between the average height of this area and its adjacent areas. (p≥2) (p≤n-2), select the region corresponding to the smaller value as the reference region: if If the minimum value is selected, the corresponding region will be chosen. As the reference area; if If the minimum value is selected, the corresponding region will be chosen. If the baseline region is used, then the average height along the X-axis of the point cloud within each region is used. Or The height of the reference area corresponding to the left edge If the two are equal or in boundary cases, then the two regions are considered. and The average height along the X-axis is used as the reference area height corresponding to the position of the left edge. ; The average X-axis height of the reference area corresponding to the position of the adjacent edge is obtained using the same method described above. Then the height (X-axis) of the measurement reference plane for the entire battery surface or side is: .

[0050] Using the above method, the most gently changing sub-region can be located within the candidate region through iterative division and gradient elimination. This effectively identifies the region with the most gradual height change, which serves as the reference region. The height of the reference surface is then used to measure the flatness of the battery's side and large surface areas.

[0051] When the battery's large / side surface exhibits uneven deformation, various parts of the plane will display a certain curvature. If a large area is directly selected to calculate the average height, the curvature of the surface will cause the measurement result to deviate from the actual plane distance. In this application, we creatively propose to "pixelate" the measurement area and iteratively filter it, so that the area used for calculation tends to be horizontal, maximizing the accuracy of measuring the distance from the plane to the reference plane.

[0052] Specifically, when selecting the measurement area, on the large or side surface where Y-axis compensation has been completed and the cross-section has been positioned, select the rectangular area to be measured. The Y-axis (length direction) span of this rectangular area is... The Z-axis (height direction) span is Divide the rectangular region S evenly into... Each small region (i.e., pixel) is divided into several equal parts, where 'a' represents the number of equal parts along the Y-axis and 'b' represents the number of equal parts along the Z-axis. The initial values ​​of 'a' and 'b' are set according to the point cloud density. The size of each small region is (...). Each small area is denoted as ); Let i = 1…a, j = 1…b. For each small region… Extract the X-axis coordinates (i.e., the height values ​​perpendicular to the large surface / side surface) of all internal point clouds to form a height set. For any small region Perform the following levelness determination process: Sort the X-axis height values ​​of all point clouds within the region in ascending order to obtain an ordered sequence. Where M is the total number of point clouds within the small region; the mean is calculated by taking the first x% of the data (i.e., the smallest subset of height values). And calculate the mean from the last x% of the data (i.e., the largest portion of the height values). ; Calculate the absolute value of the difference between the means before and after. The linear accuracy of the linearly scanned laser sensor is known to be... ,in The depth of field of the laser scanning device (in mm). Precision percentage; Judgment criteria are: If this condition is met, it indicates that the point cloud height distribution within the small region is sufficiently concentrated, and the deviation caused by the region's tendency to be horizontal or curved is within the allowable range of sensor accuracy. In this case, the small region is accepted as a valid measurement point for subsequent height value calculation. If the condition is not met, it indicates that there is significant surface curvature or noise within the small region, resulting in an excessively wide height distribution. In this case, the area of ​​the small region is divided in half: the length and width of the current small region are proportionally reduced to the original size. (Area reduced to) That is, the new Y-axis span is the same as the original span. The new Z-axis span is 1 / 3 of the original span. Then, extract the point cloud from the new small region again, and repeat the above steps until the condition is met. The average height along the X-axis of all point clouds within the region corresponding to the effective measurement point. As the height of this area : ; in, To determine the final number of point clouds within a small region, a height matrix is ​​formed by the height values ​​of all valid measurement point regions. This matrix represents the set of height differences between each valid measurement point region and the measurement reference plane. This refers to the flatness of the entire measurement surface.

[0053] In this application, the battery moves along a linear slide rail on a battery carrier plate. After scanning the large surface and side surface by rotating its posture and cooperating with a linear laser scan, based on the structural characteristics of the battery edge, the battery edge and the cross-section of the large surface are selected as the reference area. The selected area is divided into multiple candidate areas. Through iterative division and gradient elimination, the candidate area with the most gradual height change is located and determined as the reference area. The height of the measurement reference surface is characterized by the average height of adjacent edge reference areas. Then, the measurement area on the large / side surface is selected and pixelated. Each sub-region is iteratively divided until the difference in the average height of the point cloud before and after x% is less than the laser accuracy tolerance. The flatness is characterized by the set of deviations of the average value of each sub-region relative to the reference surface. That is, the set of pixelated measurement areas is the measurement sub-region. The set of height differences of all measurement sub-regions relative to the reference surface reflects the flatness of the entire measurement surface, thus improving the flatness detection accuracy.

[0054] Taking the LP71173207 battery as an example, the battery length L1=173mm, width L2=71mm, height L3=207mm, linear laser movement speed V1=0.2m / s, linear laser scanning frequency a=3000Hz, and battery carrier horizontal movement speed V2=0.05m / s. The time required to complete the scanning of the entire large area / side surface is = L3*10. -3 / V1=207*10 -3 / 0.2=1.035s, number of scan lines in the Y-axis direction=aL3*10 -3 / V1=1.035*3000=3150 lines. The Y-axis compensation amount under different scan line numbers is as follows: Table 1

[0055] After correcting the y-axis coordinates of the 3150 rows of point cloud using the motion compensation formula, spatial point cloud coordinates consistent with the state of the battery scanning surface are obtained.

[0056] 3D laser point cloud contour interval The initial virtual wireframe has a total length of 5mm along the Y-axis, and is evenly divided into fractions n=10. After the 13th iteration, the region width... satisfy The iteration has stopped. Detailed data is shown in Table 2 below: Table 2

[0057] Taking the side of the battery as an example, the width of the area after the left and right edges of the large surface are divided into equal parts through 13 iterations meets the requirements, and the remaining total width is 1.271mm.

[0058] The width of each of the ten equally divided regions along the Y-axis is 0.1271 mm. The average height X along the X-axis of each equally divided region along the left and right edges is... 14-1 X 14-2 ...X 14-10 Perform pairwise subtraction and compare the absolute values ​​of the differences. Divide the region X by the left edge. 14-7 and X 14-8 The absolute value of the difference is the smallest, X 14-7 -X 14-6 The absolute value of the difference < X 14-9 -X 14-8 If the absolute value of the difference is N, then select the left edge. 14-7 As the reference region, X is evenly divided into regions by the right edge. 14-8 and X 14-9 The absolute value of the difference is the smallest, X 14-8 -X 14-7 The absolute value of the difference > X 14-9 -X 14-10The absolute value of the difference is selected from the right edge N. 14-9 As the reference area, the final reference height of the side reference area along the X-axis is obtained as (X... 14-7左 + X 14-9右 ) / 2, as shown in Table 3: Table 3

[0059] The final calculated reference height of the large surface reference area along the X-axis is 299.5978 mm.

[0060] The battery side width L2=71mm, height L3=207mm, and the width and height of the entire measurement area are set as follows: and The measurement area was pixelated and divided into 10*40 small equally divided regions (10 in the width direction and 40 in the height direction, for a total of 400 pixel regions). The point cloud heights along the X-axis of these 400 pixel regions were sorted in ascending order. The absolute value of the difference between the mean heights of the first 10% and the last 10% of the point cloud regions was <0.04mm (Z%*FS=0.02%*200, Z=0.02, FS=200mm). Figure 8 As shown.

[0061] The above division of the measurement area meets the requirements. The set of differences between the average point cloud height along the X-axis of all sub-regions and the height of the reference surface is taken as the flatness of the measurement surface, such as... Figure 9 As shown in the figure. Among the flatness measurements, a positive value indicates that the height of the evenly divided area is greater than the height of the reference surface, meaning that the area is lower than the reference surface and is defined as a concave surface; otherwise, it is a convex surface.

[0062] The battery flatness detection system and method of the present invention uses the edge of the battery as the reference area for selection. By iteratively dividing and gradient culling, the high point tangent position of the edge is accurately found. The measurement area is pixelated and iteratively filtered so that the small area used for calculation tends to be horizontal. This avoids the measurement area curvature being too large, which affects the accuracy of flatness, and realizes accurate measurement of the flatness of the large surface / side surface of the battery.

[0063] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. It will be apparent to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of the equivalents of the claims be included within the invention.

[0064] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A square battery flatness detection system, characterized in that, The system includes two sets of laser scanning devices arranged longitudinally and longitudinally along the length of a horizontally arranged linear slide rail, and a battery carrier plate that moves along the linear slide rail to transport the battery to be tested to the testing station. The bottom of the battery carrier plate is provided with a toothed disc rotatably connected to a slide block of the linear slide rail. Each set of laser scanning devices includes two line laser scanning devices symmetrically arranged along the central axis of the linear slide rail, used for vertically moving and scanning the battery. Each set of laser scanning devices is arranged on a pair of vertical linear slide rails. Between the two sets of laser scanning devices, a rack is arranged on one side of the linear slide rail along its length. After the battery to be tested is scanned at the first testing station by the first set of laser scanning devices, when the battery carrier plate moves forward to the rack position, the toothed disc engages with the rack and is driven to rotate. The battery carrier plate and the battery it carries rotate synchronously by 90 degrees, and then the battery is scanned laterally by the second set of laser scanning devices at the second testing station.

2. The square battery flatness detection system according to claim 1, characterized in that, A boss is coaxially provided at the top center of the gear disk, and the boss is fixed to the bottom surface of the battery carrier plate.

3. The square battery flatness detection system according to claim 1, characterized in that, The toothed disk is provided with a rotational position detector in the radial direction for detecting the position of the toothed disk.

4. The square battery flatness detection system according to claim 3, characterized in that, The rotational position detection includes at least a photoelectric detection component, wherein the emitting part and the receiving part of the photoelectric detection component are respectively arranged on the upper and lower parts of the toothed disk near the outer peripheral surface.

5. The square battery flatness detection system according to claim 4, characterized in that, The slide block is provided with a limiting component, including at least a push rod that can be perpendicular to the gear plate. The push rod cooperates with a limiting hole on the surface of the gear plate to limit the gear plate at a preset rotation position.

6. The square battery flatness detection system according to claim 1, characterized in that, The laser scanning device includes at least a laser scanning device.

7. The square battery flatness detection system according to claim 1, characterized in that, The laser scanning device is connected to the data processing module, which is used to detect and judge the flatness of the battery based on the scanning data of the large surface and side surface of the battery by the laser scanning device.

8. A method for detecting the flatness of square aluminum-cased batteries, characterized in that, The method for detecting and judging the flatness of a square battery according to any one of claims 1-7 includes the following steps: At the first inspection station, the first set of laser scanning devices scans the large surface of the battery by vertical movement. Then, the battery is rotated 90 degrees during transport and transported to the second inspection station. At the second inspection station, the second set of laser scanning devices scans the side of the battery by vertical movement. Based on the movement speed of the battery carrier and the scanning frequency of the line laser scanning device, a compensation correction value in the vertical direction is determined. The compensation correction value is then used to compensate for each row of point cloud data obtained in the vertical direction to obtain the true coordinates of each row of point cloud. Based on the true coordinates of the point cloud, discrete points are removed by comparing the point cloud height value with the average height value to obtain a standard point cloud. A pre-selected area is selected based on the connection area between the battery's large / side surface and the shell's edge. The pre-selected area includes the connection area and the battery's large / side surface and the chamfered edge area on both sides of the connection area. A reference area is determined from the pre-selected area based on a standard point cloud using an iterative algorithm. The height of the measurement reference surface is determined based on the height of the reference area. A preset rectangular measurement area is selected on the large / side surface of the battery and uniformly gridded. Based on the height value of the point cloud in the rectangular grid area after gridding and the depth of field of the laser scanning device, the flatness of the rectangular grid area is detected by the bisection method. The height mean is calculated based on the height value of the point cloud in the rectangular grid area with the expected flatness. The height mean of all rectangular grid areas with the expected flatness constitutes a height matrix. The flatness of the battery's large / side surface is determined based on the difference between the height matrix and the height of the measurement reference plane.

9. The method for detecting the flatness of a square aluminum-cased battery according to claim 8, characterized in that, Using an iterative algorithm, a reference region is determined from a pre-selected area based on a standard point cloud. The height of the measurement reference surface is then determined based on the height of the reference region, including: A region is selected in the Z0Y plane using a rectangular bounding box, with a section of the surface or side edge selected as the filtering area. An iterative algorithm is used to partition and compare the filtering area, eliminating discrete partitions until the partition width meets a threshold. The region whose partition width meets the threshold is then determined as the target region. The partitioning and elimination of discrete partitions in the filtering area includes: dividing the target region into multiple equal partitions along the Y-axis; calculating the absolute value of the height difference between adjacent partitions based on the average height of the point clouds within the multiple partitions; and deleting the right / left partition by comparing the left and right values ​​of the absolute height difference to obtain a new filtering area. Based on the average height of the point cloud in multiple partitions of the target area, calculate the absolute value of the difference in average height between adjacent partitions; calculate the absolute value of the difference in average height between the two partitions corresponding to the smallest difference in average height and their respective adjacent partitions, and determine a reference area for an edge based on the partition corresponding to the smaller calculated value; determine the measurement reference plane height of the large / side surface based on the height of the reference areas of two adjacent edges.

10. The method for detecting the flatness of a square aluminum-cased battery according to claim 8, characterized in that, Based on the height values ​​of the point cloud within the rectangular grid area formed after meshing and the depth of field of the laser scanning device, the flatness of the rectangular grid area is detected using a bisection method, including: Using a rectangular grid area as the target area, the X-axis height values ​​of the point cloud within the target area are sorted in ascending order to obtain a height sequence. The mean of the height values ​​of each preset proportion in the first and second segments of the height sequence is calculated, and the absolute value of the difference between the mean values ​​is calculated. The absolute value of the difference is compared with the preset proportion accuracy value of the depth of field of the laser scanning device. If the absolute value of the difference is less than the preset proportion accuracy value, the flatness of the target area is determined to meet the expectation. Otherwise, the length and width parameters of the target area are reduced proportionally to halve the area, forming a new target area for iterative judgment. The iteration ends when the absolute value of the difference is less than the preset proportion accuracy value.

Images