A real-time soc detection and diagnosis method and system for power battery

Abstract

The application discloses a real-time SOC detection and diagnosis method and system for power batteries, relates to the technical field of pre-diagnosis, and comprises the following steps: generating a normalized image of a power battery pack, performing mask marking on the normalized image to obtain a region mask; determining the directions of line segments corresponding to the normalized image based on the region mask to obtain direction vectors of the line segments; performing anti-gravity bridge connection screening on all the line segments based on the direction vectors of the line segments to obtain a bridge connection line bundle set; generating a correction amount based on the bridge connection line bundle set, correcting a real-time SOC value corresponding to the power battery based on the correction amount to obtain a corrected SOC value. The application improves the reliability of a real-time SOC result in a structural abnormal scenario.

Description

Technical Field

[0001] This invention relates to the field of pre-diagnosis technology, and in particular to a method and system for real-time SOC detection and diagnosis of power batteries. Background Technology

[0002] The power battery pack is installed under the vehicle chassis. At the bottom of the battery pack casing is a bottom pressure relief valve for abnormal discharge, and a drainage labyrinth structure surrounding the valve. The drainage labyrinth structure consists of multi-level grooves, turning channels, and guide protrusions, designed to guide condensate, rainwater, or road splash water along a predetermined path and to block external liquids and impurities. In coastal salt spray environments, winter de-icing agent-treated road sections, or high humidity and temperature differences, salt spray deposition, condensation buildup, and periodic wet-drying cycles are prone to occur at the bottom of the power battery pack. The groove turning points of the drainage labyrinth structure and the low-lying area around the pressure relief valve are more susceptible to these issues. The accumulation of water sources and a moist film, coupled with vehicle vibrations and airflow disturbances, causes salt deposits to migrate and crystallize on the structural surface, forming elongated linear crystals around the pressure relief valve and on the surface of the drainage labyrinth structure. When these linear crystals extend from the water source area along the structural surface, across the drainage labyrinth, and approach conductive components such as metal casings or bolts, they exhibit an extension trend opposite to the direction of gravity, forming anti-gravity salt climbing whisker bridges. This creates undesigned potential conductive paths in the bottom structure of the power battery pack and affects the reliability of battery status monitoring.

[0003] In existing technologies, the real-time state of charge (SOC) of a power battery is usually obtained based on the cumulative calculation of the operating current over time, and SOC monitoring and diagnosis are performed on this basis. When there are anti-gravity salt creep whisker bridges in the pressure relief valve and drainage labyrinth structure area at the bottom of the power battery pack, weak and continuous hidden conductive paths will appear, causing some changes in charge not to come from the normal charging and discharging circuit but from the hidden discharge. This will cause the SOC results obtained based on current accumulation to deviate, and further affect the accuracy of fault prediction and pre-diagnosis of SOC deviation. Existing methods usually rely on abnormal electrical signals or post-fault manifestations to make judgments when dealing with such problems. It is difficult to identify abnormal bridging paths related to the pressure relief valve and drainage labyrinth structure from the appearance of the bottom structure of the power battery pack in a timely manner, and it is also difficult to convert the structural appearance information into quantitative basis that can be used for SOC correction and deviation pre-diagnosis. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies that make it difficult to identify abnormal bridging paths related to pressure relief valves and drainage labyrinth structures from the appearance of the bottom structure of the power battery pack in a timely manner. Therefore, this invention proposes a method and system for real-time SOC detection and diagnosis of power batteries.

[0005] To address the problems existing in the prior art, the present invention adopts the following technical solution:

[0006] A method for real-time SOC detection and diagnosis of power batteries includes the following steps: generating a normalized image of the power battery pack, and obtaining a region mask after mask marking;

[0007] The orientation of line segments corresponding to the normalized image is determined based on the region mask to obtain the orientation vector of the line segment; all line segments are then filtered by anti-gravity bridging based on the orientation vector of the line segment to obtain a set of bridging line bundles.

[0008] The correction amount is generated based on the bridge harness set, and the real-time SOC value corresponding to the power battery pack is corrected according to the correction amount to generate the corrected SOC value.

[0009] A pre-diagnosis report is obtained by pre-diagnosing the SOC deviation of the power battery pack by correcting the SOC value.

[0010] Preferably, the process of generating a normalized image is as follows:

[0011] Position the drainage labyrinth structure area around the pressure relief valve at the bottom of the power battery pack;

[0012] The original images of the power battery pack are acquired; the original images cover the pressure relief valve and the drainage labyrinth structure area; the original images are normalized to obtain normalized images.

[0013] Preferably, the process of obtaining the region mask is as follows:

[0014] Obtain structural design data for the power battery pack;

[0015] Based on the structural design data, the normalized image is divided into regions to obtain several different structural regions; among them, the several different structural regions include at least: a first structural region, a second structural region, and a third structural region;

[0016] The normalized image is masked based on several different structural regions to obtain the region masks corresponding to different structural regions.

[0017] Preferably, the process of obtaining the direction vector of the line segment is as follows:

[0018] Edge detection and morphological thinning are performed on the normalized image to obtain a binary image with a linear structure.

[0019] The binary image with a linear structure is skeletonized to obtain multiple line segments;

[0020] Determine the first and second endpoints of each line segment;

[0021] The first endpoint is redefined based on the region mask corresponding to the first structural region to obtain the endpoint of the first structural region; the second endpoint is redefined based on the region mask corresponding to the third structural region to obtain the endpoint of the third structural region.

[0022] Determine the direction vector of the line segment based on the endpoints of the first and third structural regions.

[0023] Preferably, the process of obtaining the bridge harness set is as follows:

[0024] Obtain the image coordinate system corresponding to the normalized image;

[0025] Determine the direction vector of gravity in the image coordinate system;

[0026] Perform a dot product operation between the direction vector of gravity in the image coordinate system and the direction vector of the line segment to obtain the dot product result;

[0027] The line segment is determined to pass through the second structural region based on the region mask corresponding to the second structural region, and the determination result is obtained.

[0028] A line segment that simultaneously has endpoints of the first structural region and the third structural region, and whose judgment result is that it passes through the second structural region and whose dot product result is less than 0, is defined as an anti-gravity line segment.

[0029] All anti-gravity line segments are gathered together to form a bridging bundle.

[0030] Preferably, the process of generating the correction amount is as follows:

[0031] Determine the length of each anti-gravity segment in the bridging harness set;

[0032] The total bridging length is obtained by summing the lengths of all anti-gravity line segments.

[0033] The structural risk coefficient is obtained by dividing the total bridge length by the sum of the total bridge length and 1.

[0034] Determine the proportional coefficient for structural risk coefficient;

[0035] The correction amount is obtained by multiplying the structural risk coefficient by the proportional coefficient of the structural risk coefficient.

[0036] Preferably, the process for generating the corrected SOC value is as follows:

[0037] The coulomb integral of the operating current of the power battery pack over time is performed to obtain the real-time SOC value; the correction value is obtained by subtracting the correction amount from the real-time SOC value.

[0038] Preferably, the process for obtaining a pre-diagnosis report is as follows:

[0039] Within a preset time period, a modified SOC sequence is generated based on the modified SOC value;

[0040] Within a preset time period, the change in the operating current of the power battery pack over time is calculated using coulomb integration based on the nominal capacity of the power battery pack to obtain the theoretical change in SOC.

[0041] The actual change in SOC is determined based on the corrected SOC sequence;

[0042] Subtract the theoretical SOC change from the actual SOC change to obtain the SOC offset.

[0043] Generate a pre-diagnostic report on the SOC offset status based on the SOC offset.

[0044] To address the aforementioned problems, the present invention also provides a real-time SOC detection and diagnostic system for power batteries, the system comprising:

[0045] The image masking module is used to generate a normalized image of the power battery pack, and after masking, a region mask is obtained.

[0046] The line segment orientation module is used to determine the orientation of line segments corresponding to the normalized image based on the region mask, and obtain the orientation vector of the line segment.

[0047] The anti-gravity bridging module is used to perform anti-gravity bridging filtering on all line segments based on the direction vector of the line segments, and obtain a set of bridging line bundles.

[0048] The SOC correction module is used to generate a correction amount based on the bridge harness set, and to correct the real-time SOC value of the power battery pack according to the correction amount, thereby generating a corrected SOC value.

[0049] The SOC diagnostic module is used to pre-diagnose the SOC deviation of the power battery pack by correcting the SOC value and obtain a pre-diagnostic report.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] 1. This invention uses the drainage labyrinth structure area around the pressure relief valve at the bottom of the power battery pack as the imaging object. By generating a normalized image and performing mask marking to obtain a region mask, a clear correspondence is established between the pixel positions in the image and the first, second, and third structural regions. Based on this, the direction of line segments in the normalized image is determined and anti-gravity bridging is screened. The bridging path that may correspond to the anti-gravity salt climbing whisker bridge can be directly located from the appearance image. This avoids the recognition lag caused by relying solely on abnormal electrical signals or post-fault manifestations, thus solving the problem of difficulty in timely identification of abnormal bridging paths from the appearance of the bottom structure of the power battery pack.

[0052] 2. This invention addresses the common problem that antigravity salt-climbing whisker bridges often appear as multiple adjacent, separate line segments in images, with a single line segment failing to reflect a complete bridging path. It aggregates line segments that meet the antigravity criteria into a bridging bundle set, enabling the reconstruction of continuous abnormal bridging patterns at the image level. Furthermore, it sums the antigravity line segment lengths to obtain the total bridging length, thereby calculating the structural risk coefficient and generating a correction amount. This transforms structural appearance information, which was previously difficult to use for electrical quantity estimation, into a quantitative basis that can participate in real-time state of charge correction. This reduces the impact of hidden conduction caused by antigravity salt-climbing whisker bridges on the Coulomb integral results, improving the reliability of real-time SOC results in structurally abnormal scenarios. Attached Figure Description

[0053] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0054] Figure 1 This is a flowchart illustrating a method for real-time SOC detection and diagnosis of power batteries according to an embodiment of the present invention. Detailed Implementation

[0055] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0056] Example: This example provides a method for real-time SOC detection and diagnosis of power batteries. See [link to example]. Figure 1 Specifically, including:

[0057] S1. Generate a normalized image of the power battery pack, and obtain a region mask after mask marking;

[0058] In an embodiment of the present invention, the process of generating a normalized image is as follows:

[0059] Position the drainage labyrinth structure area around the pressure relief valve at the bottom of the power battery pack;

[0060] Specifically, a power battery pack refers to an integrated unit for storing and outputting electrical energy installed on a vehicle. It includes a housing for accommodating battery cells or modules, as well as electrical connection structures, thermal management structures, and safety protection structures located inside and outside the housing. As a component of the vehicle's power system, the power battery pack has a defined external structure and installation location. The bottom pressure relief valve is a safety component located at the bottom of the power battery pack housing. It is used to release pressure in a directional manner when the internal pressure of the power battery pack rises abnormally, in order to avoid damage to the housing structure or safety risks. It has a fixed structural position at the bottom of the power battery pack and is connected to the external environment. The drainage labyrinth structure area refers to a physical area consisting of multi-level grooves, turning channels, and flow guiding structures surrounding the bottom pressure relief valve or located at the bottom of the power battery pack. It is used to guide liquid out of the power battery pack along a predetermined path, while simultaneously blocking external liquids and impurities. It has a continuous physical form and a defined spatial range at the bottom of the power battery pack.

[0061] In detail, the positioning process requires first placing the vehicle stably in a level testing environment to ensure the bottom structure of the power battery pack remains in a natural installation posture unaffected by vehicle tilt. Then, by consulting the vehicle's overall design drawings and the power battery pack assembly documents, the installation and fixing range of the power battery pack on the vehicle chassis and the outline parameters of the bottom casing are determined. Next, the vehicle is lifted to a safe height for easy observation and operation using a vehicle lifting device. Dust, dirt, salt, and other impurities adhering to the surface of the bottom casing of the power battery pack are cleaned, exposing all structural components to the outside. Then, based on the design information of the bottom structure of the power battery pack casing, a bottom pressure relief valve with directional pressure release function and communication with the external environment is identified. This bottom pressure relief valve has a fixed installation interface and a unique structural shape, and its position is consistent with the coordinate parameters marked in the design documents. This process is then completed. Centered on the bottom pressure relief valve, observe and trace the continuous physical structure in the surrounding area. This structure must exhibit multi-level grooved channels and flow-guiding protrusions, and form a complete path from liquid collection to drainage. By observing the extension direction and physical boundaries of this structure, distinguish its spatial relationship with other components such as the electrical connection interface, thermal management pipeline fixing bracket, etc. at the bottom of the power battery pack. Determine that the liquid collection end of this structure is the area near the bottom recess of the shell or the component joint, and the liquid discharge end is the area that connects to the external drain port of the power battery pack. Finally, define the continuous physical area containing all multi-level grooved channels and flow-guiding structures and located between the collection end and the discharge end as the drainage labyrinth structure area. Ensure that this area only covers the structural parts used for liquid guidance and blocking functions, forming a clear and non-overlapping boundary with other non-drainage functional structures.

[0062] Acquire raw images of the power battery pack; the raw images cover the pressure relief valve and the drainage labyrinth structure area;

[0063] Specifically, acquiring the original image of the power battery pack refers to using an image acquisition device to image the bottom structure of the power battery pack. The resulting image accurately reflects the appearance, surface condition, and spatial distribution of the pressure relief valve and drainage labyrinth structure area. The original image covering the pressure relief valve and drainage labyrinth structure area means that the image's field of view completely includes the aforementioned structure area, so that the pixels in the image can correspond one-to-one with the actual physical location of the bottom of the power battery pack.

[0064] In detail, the image acquisition device is fixed in a position that covers the pressure relief valve and the drainage labyrinth structure area using an adjustable bracket. The height of the bracket, the shooting angle, and the focal length of the device are slowly adjusted so that the field of view completely encompasses the entire structural outline of the pressure relief valve and the overall physical range of the drainage labyrinth structure area. This ensures that the shape features of the pressure relief valve, the installation position, and the groove distribution and turning shape of the drainage labyrinth in the image are presented in a reasonable proportion. The image acquisition device is started to perform single or multiple imaging operations. The image that is clear, without blurring or distortion is selected as the original image. The original image must accurately reflect the surface condition of the pressure relief valve and the drainage labyrinth structure area, such as the surface contamination and salt crystal adhesion. At the same time, it is ensured that each pixel in the image can accurately correspond to the actual physical position of the bottom of the power battery pack.

[0065] The original image is normalized to obtain a normalized image.

[0066] Specifically, normalization of the original image refers to performing a uniform scale transformation on the grayscale or brightness of each pixel in the original image, so that the brightness distribution of images obtained under different acquisition conditions is comparable, thereby obtaining a normalized image for subsequent processing. The normalized image maintains a one-to-one spatial relationship with the bottom structure of the power battery pack.

[0067] In detail, pixel information is extracted from the original image to obtain the grayscale or brightness values ​​of all pixels in the image. Data statistical methods are used to calculate the distribution range, maximum, minimum, mean, and variance of these grayscale or brightness values. Based on general image processing standards, a unified target distribution range for grayscale or brightness is determined. This target distribution range needs to be compatible with the requirements of subsequent edge detection and linear structure extraction. A linear transformation algorithm is used to adjust the grayscale or brightness values ​​of each pixel in the original image one by one. During the adjustment process, the mapping relationship between pixel values ​​and the target distribution range is strictly followed to eliminate the influence of factors such as differences in light intensity and fluctuations in acquisition device parameters under different acquisition environments on the image brightness distribution. This ensures that images obtained at different times and under different acquisition conditions have consistent and comparable brightness characteristics. After adjustment, a normalized image is obtained. This image must completely retain the appearance, surface state, and spatial distribution information of the pressure relief valve and drainage labyrinth structure area in the original image.

[0068] In an embodiment of the present invention, the process of obtaining the region mask is as follows:

[0069] Obtain structural design data for the power battery pack;

[0070] Specifically, the structural design data of a power battery pack refers to the set of engineering data formed during the design and manufacturing stages of the power battery pack, which describes the physical structure of the power battery pack. The structural design data includes the geometric dimensions of the power battery pack shell, the relative positional relationships of each functional component, and the layout information of structures such as the bottom pressure relief valve and the drainage labyrinth in the battery pack. The structural design data can accurately reflect the structural distribution characteristics of the power battery pack in its actual physical form and is used to determine the specific location and boundary range of different structures in space.

[0071] In detail, the formally reviewed and released engineering design documents are extracted, including 3D model files, 2D engineering drawings, bill of materials, and assembly process specifications. Then, using professional drawing analysis tools, the geometric information in the 3D model files is read to extract the overall outline dimensions, wall thickness, and internal cavity structural parameters of the power battery pack casing. The planar coordinate reference of the bottom of the casing is determined. Simultaneously, the annotation information in the 2D engineering drawings is analyzed to obtain the installation position parameters, outline dimensions, and relative distance to the edge of the casing for the bottom pressure relief valve. The groove dimensions, turning angles, extension paths, and inlet / outlet positions of the drainage labyrinth structure are clarified. Based on the bill of materials and assembly process specifications, the assembly relationships and spatial constraints of the bottom pressure relief valve, drainage labyrinth, and other components such as the cell module electrical connection structure and thermal management structure are analyzed. All extracted data is then organized and categorized, and the organized data is compared one by one with the actual power battery pack to verify its accuracy. Finally, a complete and standardized structural design data set that perfectly matches the actual product is integrated, comprehensively covering the geometric dimensions of the casing, the relative positions of various functional components, and the layout information of the bottom pressure relief valve and drainage labyrinth.

[0072] Based on the structural design data, the normalized image is divided into regions to obtain several different structural regions; among them, the several different structural regions include at least: a first structural region, a second structural region, and a third structural region;

[0073] Specifically, dividing the normalized image into regions based on the structural design data means using the spatial location information contained in the structural design data to divide the different physical structures corresponding to the bottom of the power battery pack in the normalized image into multiple distinct structural regions, thereby establishing a one-to-one correspondence between each structural region and the actual power battery pack structure in the image.

[0074] The first, second, and third structural regions are three independent and non-overlapping physical regions defined by the actual spatial distribution of different physical structures at the bottom of the power battery pack. The first structural region is located at the bottom of the power battery pack, near the pressure relief valve, and is characterized by easy liquid accumulation. In the actual structure of the power battery pack, this region typically corresponds to the pressure relief valve body and its surrounding recessed or low-lying structure. The second structural region is the transition area between the first structural region and the main body of the power battery pack shell. In the actual structure, this region is composed of drainage channels, guide channels, or labyrinthine structures, used to connect the first structural region with external or other structural regions, forming a continuous physical path at the bottom of the power battery pack. The third structural region is located on the outer side or edge of the bottom of the power battery pack. This region typically corresponds to the surface of the metal shell of the power battery pack or the area of ​​components fixedly connected to the shell, and has a stable structural form and fixed spatial position within the power battery pack.

[0075] In detail, a spatial correspondence is established between the structural design data of the power battery pack and the normalized image. First, based on the geometric contours, relative positional relationships, and boundary information of each structure at the bottom of the power battery pack recorded in the structural design data, the spatial coordinates in the structural design data are mapped to the pixel coordinate system of the normalized image, thereby determining the pixel range in the normalized image that corresponds one-to-one with the actual structure at the bottom of the power battery pack. Then, based on the distribution of different physical structures at the bottom of the power battery pack as reflected in the structural design data, the corresponding pixel ranges in the normalized image are grouped and divided. The pixel area corresponding to the pressure relief valve and its surrounding low-level structures is divided into the first structural area. The pixel area located between the first structural area and the external structure, corresponding to the drainage channel or labyrinth-type flow guide structure, is divided into the second structural area. The pixel area corresponding to the main body of the bottom shell of the power battery pack or the parts fixedly connected to the shell is divided into the third structural area.

[0076] The normalized image is masked based on several different structural regions to obtain the region masks corresponding to different structural regions.

[0077] Specifically, masking a normalized image based on several different structural regions refers to distinguishing and marking the pixel positions belonging to different structural regions in the normalized image, thereby forming a region mask that corresponds one-to-one with each structural region. The region mask is used to characterize the specific structural region at the bottom of the power battery pack corresponding to each pixel in the normalized image, so that the image data can reflect the spatial distribution of the actual bottom structure of the power battery pack.

[0078] In detail, under the pixel coordinate system of the normalized image, each pixel position in the normalized image is traversed one by one. According to the structural region belonging relationship corresponding to the pixel position in the region division step, the pixel position is marked as a structural region. When the pixel position corresponds to the first structural region, the pixel is marked as a valid pixel in the mask corresponding to the first structural region and as an invalid pixel in the mask corresponding to other structural regions. When the pixel position corresponds to the second structural region, the pixel is marked as a valid pixel in the mask corresponding to the second structural region and as an invalid pixel in the mask corresponding to other structural regions. When the pixel position corresponds to the third structural region, the pixel is marked as a valid pixel in the mask corresponding to the third structural region and as an invalid pixel in the mask corresponding to other structural regions. Thus, under the same pixel coordinate system of the normalized image, region masks corresponding to the first, second, and third structural regions are formed respectively, so that different structural regions can be clearly distinguished in the image and have a spatial correspondence consistent with the actual structure at the bottom of the power battery pack.

[0079] S2. Determine the direction of the line segments corresponding to the normalized image based on the region mask to obtain the direction vector of the line segments;

[0080] In an embodiment of the present invention, the process of obtaining the direction vector of the line segment is as follows:

[0081] Edge detection and morphological thinning are performed on the normalized image to obtain a binary image with a linear structure.

[0082] Specifically, edge detection of the normalized image refers to identifying locations with significant brightness changes between the surfaces of various physical structures in the normalized image reflecting the appearance of the bottom structure of the power battery pack, thereby extracting image features corresponding to the structural contours, sediment boundaries, or linear attachment edges. Morphological thinning refers to shrinking the extracted edge structures while maintaining the overall connectivity and geometric orientation of the linear structures, making them appear as narrow and continuous lines in the image. The resulting binary image of the linear structure is used to characterize the slender attachment structures or contour structures present on the bottom surface of the power battery pack.

[0083] In detail, the grayscale changes of each pixel in the normalized image are analyzed point by point. By comparing the grayscale differences between adjacent pixels, regions with significant grayscale changes are identified, thereby determining the edge positions in the image corresponding to the actual structural outline of the bottom of the power battery pack, the boundary of surface attachments, or slender depositional structures. After the edge positions are identified, they are uniformly represented as a set of linear pixels, making it clear that non-edge regions and edge regions in the image are clearly distinguished. Subsequently, the obtained edge pixel set is subjected to morphological thinning processing. The thinning process gradually removes redundant pixels in the edge structure that do not affect the overall connectivity, so that the edge region with a certain width gradually shrinks into a continuous linear structure with a single pixel width. This forms a linear structure binary map in the image that reflects the actual physical structure. Only pixel information related to the linear structure is retained in this binary map, and all other pixels are excluded.

[0084] The binary image with a linear structure is skeletonized to obtain multiple line segments;

[0085] Specifically, skeletonization of a binary image of a linear structure refers to further removing redundant width information from the linear structure and retaining only the single-pixel width path that reflects its central direction, thereby decomposing the continuous linear structure into multiple independent line segments.

[0086] In detail, using the binary image of the linear structure as input, each connected linear pixel is analyzed. By maintaining the overall topological relationship of the linear structure, redundant pixels that may remain in the linear structure are eliminated, so that each linear structure is represented as a single-pixel-width center path. After skeletonization, the pixel connectivity in the skeletonization result is traversed. Based on the connection between pixels, the continuously connected skeleton pixels are divided into several independent line segments. Each line segment consists of a group of pixels connected end to end, and the line segments are separated from each other at the pixel level, thereby obtaining multiple line segments that can respectively correspond to different linear attachment structures or deposition structures at the bottom of the power battery pack.

[0087] Determine the first and second endpoints of each line segment;

[0088] Specifically, the first and second endpoints of each line segment refer to the pixel positions located at both ends of the single-pixel-width line segment formed after the linear structure has been skeletonized, and which are connected to the other pixels of the line segment only through a single direction. In the normalized image, these two endpoints correspond to the starting and ending positions of the linear structure on the bottom surface of the power battery pack, respectively. They can accurately characterize the extension range and spatial boundary of the linear structure on the actual structural surface, and provide a clear geometric benchmark for subsequent judgment of the line segment direction, connection relationship and region attribution.

[0089] In detail, all pixels constituting a line segment are traversed one by one, and the connectivity between each pixel and its neighboring pixels is analyzed. The position of a pixel in the spatial structure is distinguished by the number of neighboring connected pixels it has in the line segment. When a pixel maintains a connectivity relationship with only one neighboring pixel in the line segment, the pixel is identified as the endpoint pixel of the line segment. After completing the connectivity relationship judgment of all pixels in the line segment, two endpoint pixel positions that satisfy only a single connectivity relationship can be obtained. Then, based on the spatial position of the two endpoint pixels in the normalized image, one endpoint is determined as the first endpoint of the line segment, and the other endpoint is determined as the second endpoint of the line segment, so that each line segment corresponds to a pair of clear endpoint positions. These endpoint positions are used to characterize the actual start and end positions of the line segment on the bottom structural surface of the power battery pack.

[0090] The first endpoint is redefined based on the region mask corresponding to the first structural region to obtain the endpoint of the first structural region;

[0091] Specifically, the region mask corresponding to the first structural region is a set of data used to identify the pixel positions in the normalized image corresponding to the first structural region at the bottom of the power battery pack. The pixels it covers correspond to the physical area around the pressure relief valve where liquid is prone to accumulate in the actual structure. Redefining the first endpoint of the line segment means determining whether the endpoint position of the line segment falls within the pixel range corresponding to the structural region, and determining the endpoint that meets the condition as the endpoint of the first structural region. In the actual structure, this endpoint corresponds to the starting position of the linear attachment structure or deposition structure within the first structural region.

[0092] In detail, the pixel position corresponding to the first endpoint is obtained in the pixel coordinate system of the normalized image. Then, the pixel position is compared point by point with the region mask corresponding to the first structural region. By determining whether the pixel position is marked as belonging to the first structural region in the region mask, it is determined whether the first endpoint is located within the spatial range corresponding to the first structural region at the bottom of the power battery pack. When the pixel position of the first endpoint is consistent with the set of pixels marked by the region mask of the first structural region, the endpoint is redefined as the endpoint of the first structural region. In the actual structure, the endpoint corresponds to the starting position of the linear attachment structure or deposition structure in the first structural region, thereby establishing a clear and consistent correspondence between the endpoint of the line segment and the specific structural region at the bottom of the power battery pack.

[0093] The second endpoint is redefined based on the region mask corresponding to the third structural region to obtain the endpoint of the third structural region;

[0094] Specifically, the region mask corresponding to the third structural region is a set of data used to identify the pixel positions in the normalized image corresponding to the area where the bottom shell of the power battery pack or its fixed connection parts are located. Redefining the second endpoint of the line segment based on the region mask means determining the endpoint of the line segment located in this region as the endpoint of the third structural region. This endpoint corresponds to the termination position of the linear structure in the shell region in the actual structure.

[0095] In detail, the pixel position corresponding to the second endpoint is obtained in the pixel coordinate system of the normalized image, and the pixel position is compared one by one with the region mask corresponding to the third structural region. By determining whether the pixel position falls within the pixel range marked by the region mask of the third structural region, it is determined whether the second endpoint is located at the physical structure position corresponding to the third structural region at the bottom of the power battery pack. When the pixel position of the second endpoint is consistent with the pixel marked in the region mask of the third structural region, the second endpoint is redefined as the endpoint of the third structural region. In the actual structure, the endpoint corresponds to the termination position of the linear structure in the bottom shell of the power battery pack or its fixed connection component area.

[0096] Determine the direction vector of the line segment based on the endpoints of the first and third structural regions.

[0097] Specifically, determining the direction vector of the line segment based on the endpoints of the first and third structural regions means using the spatial relationship between the two endpoints in the image to determine the extension direction of the linear structure on the bottom surface of the power battery pack. This direction vector is used to characterize the overall direction of the linear structure from the first structural region to the third structural region.

[0098] In detail, under the pixel coordinate system of the normalized image, the pixel coordinate positions of the endpoints of the first structural region and the third structural region corresponding to each line segment are obtained respectively. The pixel coordinates of the endpoints of the first structural region are determined as the starting reference position of the line segment direction, and the pixel coordinates of the endpoints of the third structural region are determined as the target reference position of the line segment direction. Then, based on the starting reference position, the position change of the target reference position in the pixel coordinate system is compared to determine the displacement change of the line segment in the horizontal and vertical directions of the image plane, thereby obtaining the pixel displacement result from the endpoint of the first structural region to the endpoint of the third structural region. The overall direction of the pixel displacement result in the image plane is determined. By comparing the changes in displacement in the horizontal and vertical directions, the spatial extension trend of the line segment from the endpoint of the first structural region to the endpoint of the third structural region in the image plane is determined, thereby clarifying the extension direction of the line segment. After clarifying the extension direction, the pixel displacement result is directly used as the direction expression carrier to construct a direction vector that can represent the extension direction. The direction vector is made consistent with the direction of the line segment on the actual structural surface at the bottom of the power battery pack, thereby completing the determination of the line segment direction vector.

[0099] S3. Based on the direction vector of the line segment, perform anti-gravity bridging screening on all line segments to obtain a set of bridging line bundles;

[0100] In an embodiment of the present invention, the process of obtaining the bridging harness set is as follows:

[0101] Obtain the image coordinate system corresponding to the normalized image;

[0102] Specifically, the image coordinate system corresponding to the normalized image refers to a two-dimensional coordinate system established with the upper left corner pixel position of the normalized image as the reference origin and along the horizontal and vertical directions of the image. Each coordinate position in this coordinate system corresponds one-to-one with a pixel position in the normalized image, and is used to describe the spatial distribution of the bottom structure of the power battery pack in the image plane.

[0103] In detail, the arrangement of pixels in the normalized image within the image plane is used as the basis for coordinate definition. By establishing a two-dimensional spatial description method with the upper left corner pixel position of the normalized image as the starting reference point, each pixel position in the image can be uniquely identified by its relative arrangement order in the horizontal and vertical directions, thus forming an image coordinate system that is completely consistent with the size and pixel distribution of the normalized image.

[0104] Determine the direction vector of gravity in the image coordinate system;

[0105] Specifically, the direction vector of gravity in the image coordinate system refers to the directional expression formed by converting and mapping the direction relationship of gravity on the power battery pack in actual space to the image coordinate system. This direction vector reflects the projection direction of gravity in the image plane and is used to describe the direction of gravity's effect on various structures in the image.

[0106] In detail, based on the spatial orientation information of the power battery pack under actual use, the direction of gravity relative to the bottom structure of the power battery pack in actual space is determined, and this direction is mapped to the previously established image coordinate system. By analyzing the vertical relationship and spatial orientation of the bottom structure of the power battery pack in the image, the overall pointing trend of gravity in the image plane is determined, thereby forming a direction vector in the image coordinate system that can reflect the direction of gravity. This direction vector is used to describe the spatial pointing relationship of gravity relative to each line segment and structural region in the normalized image.

[0107] Perform a dot product operation between the direction vector of gravity in the image coordinate system and the direction vector of the line segment to obtain the dot product result;

[0108] Specifically, the direction vector of a line segment refers to the directional expression constructed based on the extension relationship of the line segment from the endpoint of the first structural region to the endpoint of the third structural region in the normalized image. It is used to characterize the overall orientation of the line segment on the bottom structural surface of the power battery pack. The dot product result is used to represent the spatial consistency or relative deviation between the orientation of the line segment and the direction of gravity, thereby providing a quantitative basis for judging whether the line segment exhibits anti-gravity extension characteristics.

[0109] In detail, under a unified image coordinate system, the determined gravity direction vector and the direction vector of the corresponding line segment are obtained respectively, and both are represented in the direction expression form in the image plane. Then, the components of the gravity direction vector and the line segment direction vector in the image plane are matched accordingly. By multiplying each component and accumulating the product results, a dot product result that can reflect the spatial pointing relationship between the two is obtained.

[0110] The line segment is determined to pass through the second structural region based on the region mask corresponding to the second structural region, and the determination result is obtained.

[0111] Specifically, the region mask corresponding to the second structural region is a set of data used to identify the pixel positions in the normalized image corresponding to transition structures such as the drainage maze or flow channel at the bottom of the power battery pack. Determining whether a line segment passes through the second structural region based on the region mask means comparing the pixel positions contained in the line segment with the pixel range identified by the region mask to determine whether the line segment crosses the transition structural region in the image, thereby reflecting whether the linear structure extends along the drainage or flow path at the bottom of the actual power battery pack.

[0112] In detail, each identified line segment is used as the analysis object. Under the pixel coordinate system of the normalized image, all pixel positions constituting the line segment are obtained. These pixel positions are then compared one by one with the pixel range marked by the region mask corresponding to the second structural region. By determining whether there are any pixel positions in the line segment that fall within the coverage area of ​​the second structural region mask, it is determined whether the line segment passes through the second structural region in the image. When at least some pixels in the line segment are consistent with the pixel range marked by the second structural region mask, the line segment is determined to pass through the second structural region, thus obtaining the corresponding judgment result. This judgment result is used to characterize whether the line segment extends along the drainage labyrinth or transition structure region in the bottom structure of the power battery pack.

[0113] A line segment that simultaneously has endpoints of the first structural region and the third structural region, and whose judgment result is that it passes through the second structural region and whose dot product result is less than 0, is defined as an anti-gravity line segment.

[0114] Specifically, line segments that simultaneously possess endpoints of both the first and third structural regions, are determined to pass through the second structural region, and have a dot product result less than zero are defined as anti-gravity line segments. This is because such line segments simultaneously meet the key conditions for the formation of abnormal bridging structures in terms of spatial location and orientation. The presence of endpoints of both the first and third structural regions indicates that the line segment connects a liquid-prone area with the shell or outer structural region in the bottom structure of the power battery pack. Passing through the second structural region indicates that the line segment crosses an intermediate transition structure along a drainage maze or flow path. A dot product result less than zero indicates that the overall extension direction of the line segment is opposite to the direction of gravity in the image plane. This combination of features indicates that the actual structural morphology corresponding to the line segment violates the natural flow trend of liquid under gravity and has the potential to form abnormal bridging pathways. Therefore, defining line segments that meet the above conditions as anti-gravity line segments can effectively distinguish linear structures with potential structural risks.

[0115] Specifically, abnormal bridging structures refer to undesigned connection patterns formed in the bottom structure of a power battery pack. These connection patterns consist of linear attachments, deposits, or continuous wetted pathways, spatially spanning different structural regions that were originally isolated from each other. This allows liquids, moisture, or conductive media to form continuous pathways along these connection patterns. Abnormal bridging structures do not conform to the original structural design intent of the power battery pack, and their direction violates the natural flow path of liquids under gravity. This introduces hidden conductivity risks or environmental intrusion risks into the bottom structure of the power battery pack, and adversely affects the safety and accuracy of the battery pack's condition assessment.

[0116] All anti-gravity line segments are gathered together to form a bridging bundle.

[0117] Specifically, the anti-gravity line segment refers to a linear structure that simultaneously has endpoints of the first and third structural regions, passes through the second structural region in the image, and exhibits a trend opposite to the direction of gravity. This linear structure corresponds to an abnormal extension pattern at the bottom of the actual power battery pack that may violate the natural downward trend of liquids. The bridging bundle set refers to the data set formed by collecting all line segments that meet the above anti-gravity characteristics, used to comprehensively characterize the distribution of abnormal bridging structures that may exist at the bottom of the power battery pack.

[0118] In general, because antigravity salt-climbing whisker bridges do not appear as a single, continuous linear structure at the bottom of the power battery pack, but rather extend gradually in segments, grow, and branch along the surfaces of complex structures such as drainage labyrinths and flow guides, they often appear in images as multiple line segments that are in the same direction, adjacent in position, but separate from each other. If judgment is based solely on a single line segment, it is easy to separate structures that should belong to the same abnormal bridging path, failing to reflect the overall extension path from the water source area across the intermediate structure and finally approaching the metal shell. By aggregating these line segments that meet the same start and end area characteristics, pass through the same structural area, and have consistent antigravity direction characteristics, forming a bridging bundle set, the actual continuous whisker bridging morphology can be reconstructed at the image level, thus more realistically representing the complete length and distribution range of potential conductive paths, providing a reliable basis for subsequent latent leakage risk assessment and SOC offset prediagnosis.

[0119] Anti-gravity salt-climbing whisker bridges refer to the elongated, linear whisker structures formed by the crystallization of salt substances on the surface areas of pressure relief valves, drainage labyrinths, or flow guiding structures at the bottom of power battery packs under long-term environmental conditions such as salt spray, condensation, humidity, and vibration. These whisker structures originate from structural surfaces that are prone to water accumulation or liquid buildup, extend along drainage or flow guiding structures, and exhibit an upward trend opposite to the direction of gravity. They can cross the labyrinth structures originally used for isolation and flow guidance, and eventually approach or connect to the metal casing, bolts, or other conductive structural surfaces, thereby forming a non-designed potential conductive path at the bottom of the power battery pack.

[0120] S4. Generate a correction amount based on the bridge harness set, and correct the real-time SOC value corresponding to the power battery pack according to the correction amount to generate a corrected SOC value.

[0121] In an embodiment of the present invention, the process of generating the correction amount is as follows:

[0122] Determine the length of each anti-gravity segment in the bridging harness set;

[0123] Specifically, the length of the anti-gravity line segment refers to the spatial distance along the actual extension path of each linear structure identified as an anti-gravity line segment in the normalized image, which is used to characterize the actual extension scale of the linear structure on the bottom surface of the power battery pack.

[0124] In detail, the positions of all pixels constituting the anti-gravity line segment are obtained in the pixel coordinate system of the normalized image, and arranged according to the spatial connectivity order of the pixels in the line segment to form a pixel sequence that reflects the actual extension path of the line segment. Then, the spatial distance between adjacent pixels in the image plane is calculated one by one along the pixel sequence, and the spatial distance between all adjacent pixels in the anti-gravity line segment is continuously accumulated to obtain the overall path length of the anti-gravity line segment in the normalized image. This path length is used to characterize the actual extension scale of the corresponding anti-gravity line segment on the bottom structural surface of the power battery pack, thereby completing the determination of the length of each anti-gravity line segment in the bridging harness set.

[0125] The total bridging length is obtained by summing the lengths of all anti-gravity line segments.

[0126] Specifically, the total bridging length refers to the overall length obtained by summing the lengths of all anti-gravity segments in the bridging harness set. This overall length is used to reflect the overall scale of the continuous abnormal bridging pathways formed by the anti-gravity salt-climbing whisker bridge in the bottom structure of the power battery pack.

[0127] The structural risk coefficient is obtained by dividing the total bridge length by the sum of the total bridge length and 1.

[0128] Specifically, the antigravity salt-climbing whisker bridge, as the core carrier of structure-induced latent leakage, has its length directly related to the probability of forming a conductive path and its conductivity. It follows the physical law that the longer the conductive path, the more complete the charge migration path and the higher the leakage risk. The cumulative value of the total bridging length can intuitively reflect the development degree and spatial coverage of this type of whisker structure. Dividing the total bridging length by the sum of the total bridging length and 1 is a quantitative method designed based on the progressive physical characteristics of conductive path formation. This ensures that the risk coefficient increases monotonically with the increase of the total bridging length, and also keeps the risk coefficient in a reasonable range of 0 to 1 through the constraint of the denominator. This conforms to the physical evolution process of risk from zero to existence and from weak to strong. When the total bridging length is 0, the risk coefficient is 0, corresponding to the physical state of no whisker bridge structure and no related leakage risk. As the total bridging length gradually increases, the risk coefficient gradually approaches 1, corresponding to the physical state of continuous improvement of the whisker bridge structure, continuous enhancement of conductivity, and a higher level of leakage risk.

[0129] Determine the proportional coefficient for structural risk coefficient;

[0130] Specifically, the structural risk coefficient refers to the quantitative result obtained after normalization of the total bridging length, which is used to characterize the proportion and potential impact of abnormal bridging structures in the bottom structure of the power battery pack; the proportional coefficient of the structural risk coefficient refers to the proportional parameter used to establish a mapping relationship between the structural risk coefficient and the power battery state assessment quantity, which reflects the relative weight of the impact of abnormal bridging structures on the battery state.

[0131] In detail, the specific implementation steps for synchronously acquiring the battery state change obtained by integrating the operating current and the battery state change calculated based on the state before correction within the same time interval during the operation of the power battery pack include: firstly, determining a continuous battery operation time interval, and recording the state reference value of the power battery pack at the start and end times of this time interval; subsequently, continuously acquiring the operating current data of the power battery pack within this time interval, and performing integration calculations on the acquired operating current data in chronological order to obtain the change in charge caused by the current flow within this time interval, and converting the change in charge into the corresponding battery state change based on the nominal capacity of the power battery pack, which is used as the battery state change obtained by integrating the operating current; simultaneously, within the same time interval, reading the correction values ​​corresponding to the start and end times of this time interval respectively. The system obtains the battery state value before correction and calculates the difference between the two state values ​​to obtain the battery state change calculated based on the state before correction within the time interval, thereby achieving synchronous acquisition of the two battery state changes within the same time interval; and obtains the corresponding structural risk coefficient within this time interval; then, it calculates the difference between the battery state changes to obtain the state offset introduced by structural anomalies within this time interval, and performs a correspondence analysis between the state offset and the corresponding structural risk coefficient; after completing the above analysis, based on the relationship between the state offset and the structural risk coefficient, it determines the proportional relationship used to characterize the degree of influence of the unit structural risk coefficient on the battery state change, and determines this proportional relationship as the proportional coefficient of the structural risk coefficient, so that the proportional coefficient can convert the structural risk coefficient into a quantitative result with actual corrective significance for the battery state.

[0132] The correction amount is obtained by multiplying the structural risk coefficient by the proportional coefficient of the structural risk coefficient.

[0133] Specifically, the correction amount refers to the result obtained by combining the structural risk coefficient with its corresponding proportional coefficient. It is used to characterize the degree of correction effect introduced by the abnormal bridging structure on the state parameters of the power battery pack, and serves as the basis for subsequent correction and diagnosis of the battery state.

[0134] Specifically, the correction amount is obtained by multiplying the structural risk coefficient by its proportionality factor. This is because the structural risk coefficient characterizes the relative severity of the abnormal bridging structure at the bottom of the power battery pack, while the proportionality factor characterizes the weighted relationship of the impact of unit structural risk on battery state parameters. By multiplying the two, the structural risk coefficient, which only has relative scale significance, can be converted into a quantitative correction amount that has a real impact on battery state. The correction amount can change proportionally with the increase of the degree of abnormal bridging structure, thereby reflecting the cumulative impact of structural anomalies on battery state assessment results. This approach conforms to the objective law that the impact of continuous structural risks on battery state gradually accumulates with the degree of risk, enabling the correction amount to be stably and interpretably used for subsequent correction and diagnostic analysis of battery state parameters.

[0135] In an embodiment of the present invention, the process of generating the corrected SOC value is as follows:

[0136] The real-time SOC value is obtained by performing coulomb integration on the change of the operating current of the power battery pack over time.

[0137] Specifically, the operating current of the power battery pack refers to the magnitude of the current output or received by the power battery pack during vehicle operation. Its change over time reflects the charging and discharging state of the power battery pack under different operating conditions. The coulomb integral calculation of the change of operating current over time refers to the cumulative calculation of the operating current over a continuous period of time to obtain the change of the amount of electricity passing through the power battery pack within that time interval, thereby estimating the change ratio of the remaining amount of electricity in the power battery relative to the nominal capacity. The real-time SOC value is a state parameter obtained based on the above current accumulation results, used to characterize the proportion of the remaining amount of electricity in the power battery pack to its available capacity at the current moment. This parameter reflects the instantaneous energy state of the power battery pack during operation.

[0138] In detail, a starting moment is determined when the power battery pack begins operation, and the SOC state corresponding to that starting moment is obtained as the initial SOC value. This initial SOC value is used to represent the remaining charge state of the power battery pack at the starting moment. Subsequently, during the operation of the power battery pack, the operating current of the power battery pack is continuously acquired in chronological order at each moment, and the time interval between adjacent moments is recorded. For each time interval, the operating current magnitude within that time interval is multiplied by the corresponding time length to obtain the change in charge passing through the power battery pack within that time interval. The changes in charge obtained in each time interval since the starting moment are then calculated according to... The process involves multiple accumulations to obtain the cumulative change in power charge through the battery pack from the initial moment to the current moment. After obtaining the cumulative change in power charge, the cumulative change in power charge is divided by the nominal capacity of the battery pack to obtain the ratio of the cumulative change in power charge to the nominal capacity of the battery pack. This ratio represents the proportion of power consumption of the battery pack relative to its nominal capacity since the initial moment. Finally, based on the initial SOC value, the initial SOC value is subtracted according to the power consumption ratio to obtain the real-time SOC value of the battery pack at the current moment, so that the real-time SOC value can continuously reflect the remaining power status of the battery pack during operation.

[0139] Subtract the correction amount from the real-time SOC value to obtain the corrected SOC value.

[0140] Specifically, the corrected SOC value refers to the battery state result obtained by subtracting the correction amount from the real-time SOC value. It is used to more accurately reflect the actual remaining charge state of the power battery pack, taking into account the impact of potential structural risks.

[0141] Specifically, because the real-time SOC value is the result of coulomb integration based on the change of operating current over time, this result assumes that the change in the power battery pack's charge is only caused by normal charging and discharging behavior, without considering the impact of the hidden conduction path formed by the abnormal bridging structure at the bottom of the power battery pack on the charge statistics. When there are abnormal structures such as anti-gravity salt climbing whisker bridges, some current may be slowly discharged through undesigned paths, resulting in the real-time SOC value obtained based on coulomb integration relatively overestimating the actual remaining charge of the power battery pack. Therefore, by introducing a correction amount calculated from the structural risk coefficient and its proportional coefficient, and subtracting this correction amount from the real-time SOC value, the hidden charge deviation introduced by the abnormal structure can be compensated, making the corrected SOC value closer to the true remaining charge level of the power battery pack under actual operating conditions, thereby improving the reliability and diagnostic reference value of the SOC detection results under the condition of structural risk.

[0142] S5. By correcting the SOC value, a pre-diagnosis of the SOC deviation of the power battery pack is performed, and a pre-diagnosis report is obtained.

[0143] In an embodiment of the present invention, the process of obtaining a pre-diagnosis report is as follows:

[0144] Within a preset time period, a modified SOC sequence is generated based on the modified SOC value;

[0145] Specifically, the preset time period refers to a continuous time range selected during the operation of the power battery pack, used for unified analysis of changes in battery state; the corrected SOC sequence refers to a set of multiple corrected SOC values ​​formed continuously in chronological order within the preset time period, used to describe the change process of battery state over time.

[0146] In detail, a continuous time range for analyzing SOC changes during the operation of the power battery pack is determined, and the start and end times of this time range are used as the time boundaries for sequence generation. Subsequently, within this time range, the corrected SOC values ​​corresponding to each moment are continuously acquired in chronological order, where each corrected SOC value is obtained by subtracting the correction amount from the real-time SOC value. During the acquisition process, the corrected SOC values ​​at each moment are associated with their corresponding time sequence and stored sequentially to form an ordered data set, so that the data set can fully reflect the process of the corrected SOC value changing over time. Through the above method, the corrected SOC values ​​obtained continuously within the preset time period are arranged in chronological order to form a corrected SOC sequence.

[0147] Within a preset time period, the change in the operating current of the power battery pack over time is calculated using coulomb integration based on the nominal capacity of the power battery pack to obtain the theoretical change in SOC.

[0148] Specifically, the nominal capacity of a power battery pack refers to the capacity parameter used to characterize the amount of electricity that the power battery pack can store under design conditions; the theoretical SOC change refers to the battery state change result obtained by coulomb integration calculation based on the change of the operating current of the power battery pack over time and in combination with the nominal capacity within the preset time period. This result reflects the SOC change caused only by normal charging and discharging behavior under ideal conditions.

[0149] In detail, a continuous operating time period for analysis is determined, and the start and end times of this time period are used as the time boundaries for integration calculation. Subsequently, within the time period, the operating current of the power battery pack at each moment is continuously acquired in chronological order, ensuring that the operating current data fully covers the entire time range from the start to the end time. Based on this, the changes in operating current between adjacent moments within the time period are cumulatively calculated with the corresponding time intervals. By continuously accumulating the product of current and time within each time interval, the cumulative change in charge passing through the power battery pack within the time period is obtained. After obtaining the cumulative change in charge, the cumulative change in charge is divided by the nominal capacity of the power battery pack to obtain the ratio of the change in charge to the nominal capacity within the time period. This ratio is used to characterize the degree of SOC change caused only by normal charging and discharging behavior under ideal conditions, and this ratio is determined as the theoretical SOC change.

[0150] It should be noted that both the calculation of theoretical SOC change and the calculation of real-time SOC value use Coulomb integration. The calculation of real-time SOC value uses the SOC state of the power battery pack at a certain initial moment as a benchmark. The operating current of the power battery pack since that initial moment is continuously integrated during operation, and the cumulative change in charge is divided by the nominal capacity to update the current remaining charge state of the battery in real time. This calculation result evolves continuously over time and is used to describe the instantaneous energy level of the battery at any given moment. The calculation of theoretical SOC change is limited to a specific analysis period. Only the operating current within this period is integrated and divided by the nominal capacity to obtain the SOC change caused by normal charging and discharging behavior within this period. It is not used to directly represent the current state of the battery, but rather as a reference quantity for the SOC change that should occur under ideal conditions. This is then compared with the actual SOC change obtained based on the corrected SOC sequence to identify and quantify SOC deviation.

[0151] The actual change in SOC is determined based on the corrected SOC sequence;

[0152] In detail, the modified SOC sequence specifies the modified SOC value corresponding to the start time of the preset time period and the modified SOC value corresponding to the end time of the preset time period. The starting modified SOC value is used to represent the modified state of the power battery pack at the beginning of the time period, and the ending modified SOC value is used to represent the modified state of the power battery pack at the end of the time period. Then, using the starting modified SOC value as a benchmark, the ending modified SOC value is differentially calculated. By comparing the ending modified SOC value with the starting modified SOC value, the change range of the modified SOC value of the power battery pack within the preset time period is obtained, and this change range is determined as the actual SOC change. This actual SOC change is used to characterize the actual state change of the power battery pack within the time period after considering the impact of structural risk modification.

[0153] Subtract the theoretical SOC change from the actual SOC change to obtain the SOC offset.

[0154] Specifically, the actual SOC change refers to the battery state change result determined by the difference between the corrected SOC values ​​at the start and end of the preset time period in the corrected SOC sequence, which is used to characterize the actual state change after considering the impact of structural risks; the SOC offset refers to the difference obtained by comparing the actual SOC change with the theoretical SOC change, which is used to reflect the degree of deviation in the battery state change that cannot be explained by normal charging and discharging behavior.

[0155] Generate a pre-diagnostic report on the SOC offset status based on the SOC offset.

[0156] Specifically, the pre-diagnostic report on SOC offset refers to the status assessment results generated based on the SOC offset, which is used to characterize whether the power battery pack has SOC offset risk and its changing trend during operation, thereby providing a basis for subsequent fault prediction and maintenance decisions.

[0157] In detail, the data collected includes the specific values ​​of SOC offset, key node data of the corrected SOC sequence within a preset time period, specific values ​​of theoretical and actual SOC changes, structural risk-related parameters including total bridge length, structural risk coefficient, and proportional coefficient, environmental conditions during image acquisition including temperature, humidity, and whether the image experienced rain, snow, salt spray, etc., basic information of the power battery pack including model, nominal capacity, cumulative usage time, and operating condition records, and characteristic data of the bridge harness set including the number of anti-gravity segments and the length distribution of individual segments. Then, the magnitude of the SOC offset is quantitatively analyzed to clarify its positive and negative attributes. The absolute value of the SOC offset is measured by comparing the SOC offset of each sub-period within a preset time period. The magnitude and pattern of the offset's increase or decrease are tracked to determine whether the offset is continuously increasing, decreasing, or fluctuating. The correlation between the SOC offset and structural risk parameters is then verified. Based on the extension of the total bridging length and the level of the structural risk coefficient, the causal relationship between the offset and the structurally induced latent leakage caused by anti-gravity salt creep whisker bridges is determined, ruling out the possibility that other factors such as cell aging and natural self-discharge dominate the offset. Finally, considering the absolute value and rate of change of the SOC offset, as well as the severity of the structural risk, the final result is determined. The potential impact of the operating conditions of the power battery pack on the offset is divided into three risk levels: low, medium, and high. Low risk corresponds to small and gradual offset with a low structural risk coefficient; medium risk corresponds to moderate offset or accelerated offset with a medium structural risk coefficient; and high risk corresponds to large and continuously increasing offset with a high structural risk coefficient. A clear diagnostic conclusion is then drawn, clearly stating whether the current power battery pack exhibits structure-induced SOC offset, the core cause of the offset being the hidden leakage path formed by anti-gravity salt creep whisker bridges on the structural surface, the degree of impact of the current offset on the power battery pack's range display and charge / discharge control, and the future impact if no further measures are taken. Intervention may exacerbate the deviation trend. Then, practical maintenance recommendations for this deviation are added, including whether it is necessary to clean the pressure relief valve and drainage labyrinth structure area at the bottom of the power battery pack in a timely manner, whether it is necessary to check the actual conductivity of the hidden leakage path, and whether it is necessary to shorten the subsequent structural condition monitoring cycle. Finally, in a logically clear report structure, the collected basic information, data statistics, analysis process, risk level judgment, diagnosis conclusion, and maintenance recommendations are systematically integrated to ensure that each item in the report is supported by corresponding data analysis, forming a complete, standardized, and traceable SOC deviation pre-diagnosis report.

[0158] An embodiment of the present invention also provides a real-time SOC detection and diagnosis system for power batteries.

[0159] In this embodiment, the functions of each module / unit are as follows:

[0160] The image masking module is used to generate a normalized image of the power battery pack, and after masking, a region mask is obtained.

[0161] The line segment orientation module is used to determine the orientation of line segments corresponding to the normalized image based on the region mask, and obtain the orientation vector of the line segment.

[0162] The anti-gravity bridging module is used to perform anti-gravity bridging filtering on all line segments based on the direction vector of the line segments, and obtain a set of bridging line bundles.

[0163] The SOC correction module is used to generate a correction amount based on the bridge harness set, and to correct the real-time SOC value of the power battery pack according to the correction amount, thereby generating a corrected SOC value.

[0164] The SOC diagnostic module is used to pre-diagnose the SOC deviation of the power battery pack by correcting the SOC value and obtain a pre-diagnostic report.

[0165] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for real-time SOC detection and diagnosis of power batteries, characterized in that, The process includes the following steps: generating a normalized image of the power battery pack, and obtaining a region mask after mask marking; The orientation of line segments corresponding to the normalized image is determined based on the region mask to obtain the orientation vector of the line segment; all line segments are then filtered by anti-gravity bridging based on the orientation vector of the line segment to obtain a set of bridging line bundles. The correction amount is generated based on the bridge harness set, and the real-time SOC value corresponding to the power battery pack is corrected according to the correction amount to generate the corrected SOC value. A pre-diagnosis report is obtained by pre-diagnosing the SOC deviation of the power battery pack by correcting the SOC value.

2. The method for real-time SOC detection and diagnosis of power batteries according to claim 1, characterized in that, The process of generating a normalized image is as follows: Position the drainage labyrinth structure area around the pressure relief valve at the bottom of the power battery pack; The original images of the power battery pack are acquired; the original images cover the pressure relief valve and the drainage labyrinth structure area; the original images are normalized to obtain normalized images.

3. The method for real-time SOC detection and diagnosis of power batteries according to claim 1, characterized in that, The process of obtaining the region mask is as follows: Obtain structural design data for the power battery pack; Based on the structural design data, the normalized image is divided into regions to obtain several different structural regions; among them, the several different structural regions include at least: a first structural region, a second structural region, and a third structural region; The normalized image is masked based on several different structural regions to obtain the region masks corresponding to different structural regions.

4. The method for real-time SOC detection and diagnosis of power batteries according to claim 3, characterized in that, The process of obtaining the direction vector of a line segment is as follows: Edge detection and morphological thinning are performed on the normalized image to obtain a binary image with a linear structure. The binary image with a linear structure is skeletonized to obtain multiple line segments; Determine the first and second endpoints of each line segment; The first endpoint is redefined based on the region mask corresponding to the first structural region to obtain the endpoint of the first structural region; The second endpoint is redefined based on the region mask corresponding to the third structural region to obtain the endpoint of the third structural region; Determine the direction vector of the line segment based on the endpoints of the first and third structural regions.

5. A method for real-time SOC detection and diagnosis of power batteries according to claim 4, characterized in that, The process of obtaining the bridge harness set is as follows: Obtain the image coordinate system corresponding to the normalized image; Determine the direction vector of gravity in the image coordinate system; Perform a dot product operation between the direction vector of gravity in the image coordinate system and the direction vector of the line segment to obtain the dot product result; The line segment is determined to pass through the second structural region based on the region mask corresponding to the second structural region, and the determination result is obtained. A line segment that simultaneously has endpoints of the first structural region and the third structural region, and whose judgment result is that it passes through the second structural region and whose dot product result is less than 0, is defined as an anti-gravity line segment. All anti-gravity line segments are gathered together to form a bridging bundle.

6. The method for real-time SOC detection and diagnosis of power batteries according to claim 5, characterized in that, The process of generating the correction amount is as follows: Determine the length of each anti-gravity segment in the bridging harness set; The total bridging length is obtained by summing the lengths of all anti-gravity line segments. The structural risk coefficient is obtained by dividing the total bridge length by the sum of the total bridge length and 1. Determine the proportional coefficient for structural risk coefficient; The correction amount is obtained by multiplying the structural risk coefficient by the proportional coefficient of the structural risk coefficient.

7. The method for real-time SOC detection and diagnosis of power batteries according to claim 1, characterized in that, The process of generating the corrected SOC value is as follows: The coulomb integral of the operating current of the power battery pack over time is performed to obtain the real-time SOC value; the correction value is obtained by subtracting the correction amount from the real-time SOC value.

8. The method for real-time SOC detection and diagnosis of power batteries according to claim 1, characterized in that, The process of obtaining a pre-diagnosis report is as follows: Within a preset time period, a modified SOC sequence is generated based on the modified SOC value; Within a preset time period, the change in the operating current of the power battery pack over time is calculated using coulomb integration based on the nominal capacity of the power battery pack to obtain the theoretical change in SOC. The actual change in SOC is determined based on the corrected SOC sequence; Subtract the theoretical SOC change from the actual SOC change to obtain the SOC offset. Generate a pre-diagnostic report on the SOC offset status based on the SOC offset.

9. A real-time SOC detection and diagnostic system for power batteries, characterized in that, The system includes: an image mask module, used to generate a normalized image of the power battery pack, and obtain a region mask after mask marking; The line segment orientation module is used to determine the orientation of line segments corresponding to the normalized image based on the region mask, and obtain the orientation vector of the line segment. The anti-gravity bridging module is used to perform anti-gravity bridging filtering on all line segments based on the direction vector of the line segments, and obtain a set of bridging line bundles. The SOC correction module is used to generate a correction amount based on the bridge harness set, and to correct the real-time SOC value of the power battery pack according to the correction amount, thereby generating a corrected SOC value. The SOC diagnostic module is used to pre-diagnose the SOC deviation of the power battery pack by correcting the SOC value and obtain a pre-diagnostic report.

Images