Apparatus for monitoring and operating method of the same

Abstract

A monitoring device according to one embodiment disclosed in this document may include: an image acquisition unit that acquires an image of at least one process processing device related to the manufacture of a battery cell; an analysis unit that inputs the acquired image into an artificial intelligence model for determining the welding state of the battery cell to generate a first determination result that determines whether the battery cell is defective; and a determination unit that determines whether the battery cell is defective based on the first determination result.

Description

Technology Field

[0001] The embodiments disclosed in this document relate to a monitoring device and a method of operating the same. Background Technology

[0002] Electric vehicles generate power by receiving electricity from an external source to charge battery cells, and then driving a motor using the voltage charged in the battery cells. Electric vehicle battery cells are manufactured by housing an electrode assembly in a battery case and injecting an electrolyte into the battery case.

[0003] Battery cells are classified into cylindrical, prismatic, and pouch types depending on the type of battery case, and cylindrical battery cells include an electrode assembly, a battery case in the form of a cylindrical metal can that accommodates the electrode assembly and electrolyte, and a cap assembly assembled on the top of the cylindrical can.

[0004] Here, the positive tab of the cylindrical battery cell is welded to the top cap of the cap assembly. However, if part or all of the weld detaches due to external impact or welding defects, battery cell degradation or explosion may occur. Therefore, it is important to inspect the battery cells to prevent defective welds from being released. The problem to be solved

[0005] One objective of the embodiments disclosed in this document is to provide a monitoring device capable of detecting defective battery cells by accurately analyzing the welding status of battery cells, and a method of operation thereof.

[0006] The technical problems of the embodiments disclosed in this document are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0007] A monitoring device according to one embodiment disclosed in this document may include: an image acquisition unit that acquires an image of at least one process processing device related to the manufacture of a battery cell; an analysis unit that inputs the acquired image into an artificial intelligence model for determining the welding status of the battery cell to generate a first determination result that determines whether the battery cell is defective; and a determination unit that determines whether the battery cell is defective based on the first determination result.

[0008] A method of operation of a monitoring device according to one embodiment disclosed in this document may include: a step of acquiring an image of at least one process processing device related to the manufacture of a battery cell; a step of inputting the acquired image into an artificial intelligence model for determining the welding status of the battery cell to generate a first determination result for determining whether the battery cell is defective; a step of generating a second determination result for determining whether the battery cell is defective based on gray level information of the acquired image; and a step of determining whether the battery cell is defective based on the first determination result and the second determination result. Effects of the invention

[0009] According to the monitoring device and the method of operation thereof according to one embodiment disclosed in this document, defective battery cells can be detected by accurately analyzing the welding status of battery cells. Brief explanation of the drawing

[0010] FIG. 1 is a drawing for explaining, in general, a battery process system according to one embodiment disclosed in this document. FIG. 2 is a block diagram showing the configuration of a monitoring device according to one embodiment disclosed in this document. FIG. 3 is a welding state classification table according to one embodiment disclosed in this document. Figure 4 is a graph showing the change in tensile strength according to the change in welding length according to one embodiment disclosed in this document. FIG. 5 is a diagram classifying battery units according to an embodiment disclosed in this document. FIG. 6 is a drawing illustrating the welding location of a battery cell according to one embodiment disclosed in this document. FIG. 7 is a drawing illustrating a welding area of ​​a battery cell according to one embodiment disclosed in this document. FIG. 8 is a drawing illustrating the welding form of a battery cell according to one embodiment disclosed in this document. FIG. 9 is a drawing illustrating a method of operation of a monitoring device according to an embodiment disclosed in this document. Specific details for implementing the invention

[0011] Some embodiments disclosed in this document are described in detail below with reference to exemplary drawings. It should be noted that in assigning reference numerals to the components of each drawing, the same components are given the same reference numeral whenever possible, even if they are shown in different drawings. Furthermore, in describing the embodiments disclosed in this document, detailed descriptions of related known configurations or functions are omitted if it is determined that such detailed descriptions would hinder understanding of the embodiments disclosed in this document.

[0012] In describing the components of the embodiments disclosed in this document, terms such as first, second, A, B, (a), (b), etc., may be used. These terms are intended only to distinguish the components from other components and do not limit the nature, order, or sequence of the components. Furthermore, unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments disclosed in this document belong. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this document.

[0013] According to various embodiments, the battery may include a battery cell, which is the basic unit of the battery capable of charging and discharging electrical energy. The battery cell may be a lithium-ion (Li-ion) battery, a lithium-ion polymer (Li-ion polymer) battery, a nickel-cadmium (Ni-Cd) battery, a nickel-hydrogen (Ni-MH) battery, etc., but is not limited thereto. The battery cell may supply power to a target device (not shown). To this end, the battery cell may be electrically connected to the target device. Here, the target device may include an electrical, electronic, or mechanical device that operates by receiving power from a battery pack (not shown) comprising a plurality of battery cells. For example, the target device may be a small product such as a digital camera, P-DVD, MP3P, mobile phone, PDA, portable game device, power tool, and E-bike, as well as a large product requiring high output such as an electric vehicle or hybrid vehicle, and a power storage device or backup power storage device that stores surplus generated power or new and renewable energy, but is not limited thereto.

[0014] A battery cell may consist of an electrode assembly, a battery case in which the electrode assembly is housed, and an electrolyte injected into the battery case to activate the electrode assembly. An electrode assembly is formed by interposing a separator between a positive plate, formed by coating a positive active material onto a positive current collector, and a negative plate, formed by coating a negative active material onto a negative current collector. Depending on the type of battery case, the electrode assembly may be manufactured in a jelly roll type, stack type, etc., and housed inside the battery case. The battery case serves as an exterior material that maintains the shape of the battery and protects it from external impacts; battery cells can be classified into cylindrical, prismatic, or pouch types depending on the type of battery case.

[0015] According to an embodiment, a battery cell can be manufactured through a series of manufacturing processes including an electrode manufacturing process, an assembly process, and a formation process. Here, the assembly process may include a process of assembling a positive electrode plate and a negative electrode plate made through the electrode manufacturing process and injecting an electrolyte, and may include a notching process, a winding process, an assembly process, and a packaging process.

[0016] The packaging process can be defined as the process of injecting the electrode assembly and electrolyte into the battery case and sealing it. In the case of a cylindrical battery cell, the electrode assembly is mounted in a cylindrical metal can, a negative tab extending from the negative electrode of the electrode assembly is welded to the bottom of the can, and a positive tab extending from the positive electrode assembly is welded to the top cap of the cap assembly while the electrode assembly and electrolyte are embedded.

[0017] However, if part or all of the positive tab of the electrode assembly and the top cap of the cap assembly detach due to external impact or welding defects—that is, if the positive tab and top cap are weakly welded—the battery cell may deteriorate or explode; therefore, it is important to inspect the battery cells to prevent defective welding from leaking out.

[0018] Below, the battery process system is explained using the assembly process system as an example.

[0019] FIG. 1 is a drawing for explaining, in general, a battery process system according to one embodiment disclosed in this document.

[0020] Referring to FIG. 1, the battery process system may include a monitoring device (100) and at least one process processing device (210, 220, 230).

[0021] The monitoring device (100) can collect and analyze data in real time from at least one process processing device (210, 220, 230) operating in a battery process system. The monitoring device (100) can collect and analyze data from at least one process processing device (210, 220, 230).

[0022] For example, the monitoring device (100) can collect and analyze data or graph data generated in the battery process system, such as the progress of the battery process system, whether an alarm occurs, temperature, pressure, quantity, etc.

[0023] The monitoring device (100) can acquire images of at least one process processing device (210, 220, 230) captured by a plurality of camera modules (not shown).

[0024] The monitoring device (100) can determine whether the battery cell is defective based on an image taken of at least one process processing device (210, 220, 230). According to one embodiment, the monitoring device (100) can determine whether the positive tab and top cap of the battery cell are weakly welded based on an image taken of at least one process processing device (210, 220, 230).

[0025] At least one process processing device (210, 220, 230) may include a first process processing device (210), a second process processing device (220), and a third process processing device (230). In FIG. 1, at least one process processing device (210, 220, 230) is shown as having three devices, but is not limited thereto. According to an embodiment, at least one process processing device (210, 220, 230) may include n devices (where n is a natural number greater than or equal to 1).

[0026] According to one embodiment, at least one process processing device (210, 220, 230) can weld the positive tab of the electrode assembly of the battery cell and the top cap of the cap assembly.

[0027] FIG. 2 is a block diagram showing the configuration of a monitoring device (100) according to one embodiment disclosed in this document.

[0028] In one embodiment, the monitoring device (100) can determine the state of the battery unit (105) based on an image of the battery unit (105) acquired by the image acquisition device (103). Here, the image acquisition device (103) may correspond to a plurality of camera modules (not shown) described in FIG. 1. Here, the battery unit (105) may be a battery cell.

[0029] In one embodiment, the monitoring device (100) can determine the state of the battery unit (105) based on the result of a rule-based inspection and / or the result of an artificial intelligence model-based inspection for the battery unit (105). In one embodiment, the rule-based inspection may include an inspection of at least one of the presence or absence of a can of the battery unit (105), the presence or absence of a top cap of the can assembly, the center distance between the can and the top cap, the position of the positive tab, or whether there is over-welding. In one embodiment, the artificial intelligence model-based inspection may include an inspection of at least one of the presence or absence of a top cap of the cap assembly, the welding position, the welding area, or the welding type.

[0030] In one embodiment, the monitoring device (100) may determine that the state of the battery unit (105) is good if both the rule-based inspection result and the artificial intelligence model-based inspection result for the battery unit (105) indicate normal. In one embodiment, the monitoring device (100) may determine that the state of the battery unit (105) is defective if at least one of the rule-based inspection result or the artificial intelligence model-based inspection result for the battery unit (105) indicates a defect.

[0031] Below, the components of the monitoring device (100) are described in a general manner, and then the specific operation method of the monitoring device (100) is described.

[0032] Referring to FIG. 2, the monitoring device (100) may include a communication circuit (110), a memory (120), and a processor (130). According to an embodiment, the monitoring device (100) shown in FIG. 2 may further include at least one component (e.g., a display, an input device, or an output device) other than the components shown in FIG. 2.

[0033] In one embodiment, the communication circuit (110) establishes a wired communication channel and / or a wireless communication channel between the monitoring device (100) and the image acquisition device (103), and can transmit and receive data with the image acquisition device (103) through the established communication channel.

[0034] In one embodiment, the memory (120) may include volatile memory and / or non-volatile memory.

[0035] In one embodiment, the memory (120) may store data used by at least one component of the monitoring device (100) (e.g., processor (130)). For example, the data may include a program (125) (or an instruction associated therewith), input data, or output data. In one embodiment, the instruction may cause the monitoring device (100) to perform operations defined by the instruction when executed by the processor (130).

[0036] In one embodiment, the memory (120) may include a program (125) (e.g., an artificial intelligence model learning unit (141), artificial intelligence models (145, 146, 147, 148), an image acquisition unit (150), a state determination unit (160), a location detection unit (170), a region detection unit (180), an analysis unit (190), and / or a determination unit (195)).

[0037] In one embodiment, the processor (130) may include a central processing unit, an application processor, a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor.

[0038] In one embodiment, the processor (130) can execute a program (125) (e.g., an artificial intelligence model learning unit (141), artificial intelligence models (145, 146, 147, 148), an image acquisition unit (150), a state determination unit (160), a location detection unit (170), a region detection unit (180), an analysis unit (190), and / or a determination unit (195)) to control at least one other component (e.g., a hardware or software component) of a monitoring device (100) connected to the processor (130) and can perform various data processing or operations.

[0039] In one embodiment, the artificial intelligence model learning unit (141) can train artificial intelligence models (145, 146, 147, 148) based on learning data. In one embodiment, the artificial intelligence models (145, 146, 147, 148) may be models trained to produce output results of different purposes based on different learning data.

[0040] In one embodiment, the image acquisition unit (150) can acquire an image of the battery unit (105) from the image acquisition device (103). In one embodiment, the image acquisition unit (150) can automatically log images of at least one process processing device (210, 220, 230) captured by the image acquisition device (103). For example, the image acquisition unit (150) can collect images of at least one process processing device (210, 220, 230) captured by setting the sampling period of the automatic image data logging to 0.1 sec. Here, the image acquisition device (103) can acquire operation data of at least one process processing device (210, 220, 230). The image acquisition device (103) can acquire images of at least one process processing device (210, 220, 230) captured in relation to the manufacture of the battery unit (105). According to one embodiment, the image acquisition device (103) can acquire an image of at least one process processing device (210, 220, 230) welding the positive tab of the electrode assembly of the battery unit (105) and the top cap of the cap assembly.

[0041] In one embodiment, the state determination unit (160) can determine the state of the can assembly based on an image of the battery unit (105). In one embodiment, the position detection unit (170) can determine the state of the welding position based on an image of the battery unit (105). In one embodiment, the area detection unit (180) can detect and / or crop the welding area in an image of the welding position of the battery unit (105). In one embodiment, the analysis unit (190) can analyze the welding state based on the image cropped from the welding position. In one embodiment, the judgment unit (195) can determine whether the battery unit (105) is a good product based on the judgment results of the state determination unit (160), the position detection unit (170), and / or the analysis unit (190).

[0042] Hereinafter, with reference to FIGS. 3 to 8, a method for a monitoring device (100) to determine the state of a battery unit (105) through an artificial intelligence model learning unit (141), artificial intelligence models (145, 146, 147, 148), an image acquisition unit (150), a state determination unit (160), a position detection unit (170), a region detection unit (180), an analysis unit (190), and / or a determination unit (195) is specifically described.

[0044] FIG. 3 is a table classifying welding conditions according to one embodiment disclosed in this document. FIG. 4 is a graph showing the change in tensile strength according to the change in welding length according to one embodiment disclosed in this document.

[0045] The battery unit (105) may be determined to be normal based on various criteria. For example, the battery unit (105) may be determined to be normal based on criteria related to the welding condition and / or criteria related to tensile strength. Criteria for determining the normality of the battery unit (105) may be set in advance. For example, the welding condition of the battery unit (105) may be determined to be normal if the welding length (or bead length) of at least one welding line is 75% or more of the welding section, and the tensile strength of the battery unit (105) may be determined to be normal if it is 0.4 kgf or more. Here, the welding bead may be the weld metal solidified on the base material when welding is performed along the welding line.

[0046] The welding condition of the battery unit (105) can be confirmed through an image of the battery unit (105), but the tensile strength of the battery unit (105) may be difficult to confirm through an image.

[0047] Referring to FIG. 3, in Type 1, the welding condition of the battery unit (105) can be determined as normal welding because the welding length of two welding lines is 100% of the welding section. In Type 2, the welding condition of the battery unit (105) can be determined as normal welding because the welding length of one welding line is 75% or more of the welding section. Also, in Type 1 and Type 2, the tensile strength of the battery unit (105) is 1.90 kgf and 1.61 kgf, respectively, so the battery unit (105) can be determined as normal.

[0048] However, in Type 3 and Type 4, since the welding length of the welding line is 50% or more or 50% or less of the welding section, the welding condition of the battery unit (105) can be determined to be weak welding. Nevertheless, in Type 3 and Type 4, since the tensile strength of the battery unit (105) is 1.10 kgf and 0.73 kgf, the battery unit (105) can be determined to be normal.

[0049] Referring to Type 1 and Type 2, Type 3 and Type 4, it can be seen that if the criteria for the welding condition are satisfied, the battery unit (105) is likely to also satisfy the criteria related to tensile strength.

[0050] Referring to FIG. 4, it can be seen that the weld length of the battery unit (105) and the tensile strength of the battery unit (105) exhibit a linear proportional relationship. For example, the Pearson correlation coefficient (PCC) between the tensile strength of the battery unit (105) and the weld length of the battery unit (105) can be calculated as 0.764. Here, the Pearson correlation coefficient is a numerical value that quantifies the linear correlation between two variables. That is, the tensile strength of the battery unit (105) can be calculated by multiplying the weld length (mm) of the battery unit (105) by the Pearson correlation coefficient of 0.764.

[0051] Considering the experimental results of FIGS. 3 and FIGS. 4, whether the criteria related to the tensile strength of the battery unit (105) are satisfied can be determined by considering only the criteria related to the welding condition of the battery unit (105). Therefore, it is important to determine whether the criteria for the welding condition of the battery unit (105) are satisfied through an image of the battery unit (105).

[0052] Artificial intelligence model training

[0053] FIG. 5 is a diagram classifying battery units according to an embodiment disclosed in this document.

[0054] In one embodiment, the artificial intelligence model learning unit (141) may input an image of a battery unit into the artificial intelligence model (145). In one embodiment, the artificial intelligence model (145) may be a model capable of multinomial classification (e.g., a model based on a convolutional neural network (CNN)). In one embodiment, the artificial intelligence model (145) may be trained to classify the presence or absence of a can (515, 535) and a top cap (511) based on images (510, 530, 550) of the battery unit.

[0055] In one embodiment, the data for training the artificial intelligence model (145) may include an image (510) of a battery unit in which a can (515) and a top cap (511) are normally present, an image (530) of a battery unit in which only a can (535) is present without a top cap, and an image (550) in which there is no can or top cap. In one embodiment, the images (510, 530, 550) used for training the artificial intelligence model (145) may be prepared in advance.

[0056] In one embodiment, the artificial intelligence model (145) may be trained to classify an input image into at least three states based on training data. Here, the at least three states may include a state in which a can and a top cap are normally present, a state in which only a can is present without a top cap, and a state in which neither a can nor a top cap is present.

[0057] FIG. 6 is a drawing illustrating the welding location of a battery unit according to an embodiment disclosed in this document. FIG. 6 includes images (610, 630) of a battery unit including a top cap (611, 631) and a positive tab (613, 633), and enlarged images (620, 640) of a portion of the images (610, 630).

[0058] In one embodiment, the artificial intelligence model learning unit (141) may input an image of the battery unit into the artificial intelligence model (146). In one embodiment, the artificial intelligence model (146) may be a model capable of multinomial classification. In one embodiment, the artificial intelligence model (146) may be trained to identify and / or crop welding locations (621, 641) based on images (610, 630) of the battery unit. Here, the welding location may include an area where welding was performed.

[0059] In one embodiment, the data for training the artificial intelligence model (146) may include an image (610) in which welding exists only in the area where the top cap (611) and the anode tab (613) overlap, and an image (630) in which welding exists at least partially in the area where the top cap (631) and the anode tab (633) do not overlap. In one embodiment, the images (610, 630) used for training the artificial intelligence model (146) may be prepared in advance.

[0060] In one embodiment, an artificial intelligence model (146) may be trained to classify an input image into at least two states based on training data. Here, the at least two states may include a state in which welding exists only in the area where the top cap and the anode tab overlap, and a state in which welding exists at least partially in the part where the top cap and the anode tab do not overlap.

[0061] In one embodiment, the artificial intelligence model (146) may be trained to crop a weld location in an input image based on training data. For example, the artificial intelligence model (146) may be trained to crop a weld location (621, 641) in images (610, 630). In one embodiment, the artificial intelligence model (146) may attach metadata indicating at least two states of the input image to the cropped weld location.

[0062] FIG. 7 is a drawing illustrating a welding area of ​​a battery cell according to one embodiment disclosed in this document.

[0063] In one embodiment, the artificial intelligence model learning unit (141) may input an image of the battery unit into the artificial intelligence model (147). In one embodiment, the artificial intelligence model (147) may be a model capable of multinomial classification. In one embodiment, the artificial intelligence model (147) may be trained to identify and / or crop a welding area (721, 741) based on images (710, 730) of the battery unit. Here, the welding area may refer to an area within a welding section within a welding location where actual welding was performed. Here, the welding location may be identified and / or cropped by the artificial intelligence model (146).

[0064] In one embodiment, the data for training the artificial intelligence model (147) may include images of welding positions having a welding length between 0 and 100% of the welding sections (711, 713, 741, 743). In one embodiment, the images (710, 730) used for training the artificial intelligence model (147) may be prepared in advance.

[0065] In one embodiment, the artificial intelligence model (147) may be trained to identify welding areas (721, 741) in an input image based on training data. In one embodiment, the artificial intelligence model (147) may be trained to crop welding locations in an input image based on training data. For example, the artificial intelligence model (147) may be trained to crop welding locations (721, 741) in images (710, 730). In one embodiment, the artificial intelligence model (147) may be trained to output images (720, 740) with the welding locations (721, 741) cropped.

[0066] FIG. 8 is a drawing illustrating the welding state of a battery cell according to one embodiment disclosed in this document.

[0067] In one embodiment, the artificial intelligence model learning unit (141) may input an image of a battery unit into the artificial intelligence model (148). In one embodiment, the artificial intelligence model learning unit (141) may train the artificial intelligence model (148) to classify the welding status of the battery unit based on an image cropped from the welding location. In one embodiment, the artificial intelligence model (148) may be a model capable of multinomial classification.

[0068] In one embodiment, the data for training the artificial intelligence model (148) may include an image (810) having a welding length within a reference length range (e.g., 75 to 100%) in a welding section (811, 813), an image (820) having a welding length within a reference length range (e.g., 0 to 75%) in a welding section (821, 823), and other welding images (830, 840, 850) having defects. Here, other defects may include an over-welding state (e.g., a welding length exceeding the reference length range), a top cap damage state (e.g., top cap burnt), or a tap damage state. In one embodiment, the images (810, 820, 830, 840, 850) used for training the artificial intelligence model (148) may be prepared in advance.

[0069] In one embodiment, the artificial intelligence model (148) may be trained to classify an input image into at least three states based on training data. Here, the at least three states may include a state having a weld length within a reference length range, a state having a weld length less than the reference length range, and a state having other defects.

[0070] State determination through artificial intelligence models

[0071] In one embodiment, the state determination unit (160) can determine the state of the can assembly based on an image of the battery unit (105). In one embodiment, the state determination unit (160) can determine the state of the can assembly by inputting an image of the battery unit (105) into an artificial intelligence model (145). In one embodiment, the state determination unit (160) can determine the state of the can assembly of the battery unit (105) based on the output of the artificial intelligence model (145). For example, the state determination unit (160) can determine the state of the can assembly of the battery unit (105) based on the state of the battery unit (105) classified by the artificial intelligence model (145) (e.g., a state where the can and top cap are normally present (e.g., image (510), a state where only the can is present without the top cap (e.g., image (530), and a state where the can and top cap are absent (e.g., image (550)).

[0072] In one embodiment, if the can and top cap are normally present in the battery unit (105), the state determination unit (160) may determine that the state of the can assembly is normal. In one embodiment, if only the can is present without the top cap, or if there is no can and top cap, the state determination unit (160) may determine that the state of the can assembly is defective.

[0073] In one embodiment, the state determination unit (160) may transmit to the position detection unit (170) an image of a battery unit (105) in which the state of the can assembly is determined to be normal among a plurality of images. In one embodiment, the state determination unit (160) may transmit to the determination unit (195) the result of determining the state of the can assembly of the battery unit (105).

[0074] In one embodiment, the position detection unit (170) can detect and / or crop the welding location in an image of the battery unit (105). In one embodiment, the state determination unit (160) can input an image of the battery unit (105) into an artificial intelligence model (146) to detect and / or crop the welding location (e.g., welding location (621, 641)).

[0075] In one embodiment, the position detection unit (170) can determine the state of the welding location based on an image of the battery unit (105). For example, the position detection unit (170) can determine the state of the welding location of the battery unit (105) based on the state of the battery unit (105) classified by the artificial intelligence model (146) (e.g., a state where the welding exists only in the area where the top cap and the positive tab overlap, and a state where at least some welding exists in the part where the top cap and the positive tab do not overlap).

[0076] In one embodiment, if welding exists only in the area where the top cap and the positive tab overlap on the battery unit (105), the position detection unit (170) can determine that the condition of the welding location is normal. In one embodiment, if welding exists at least partially in the area where the top cap and the positive tab do not overlap, the position detection unit (170) can determine that the condition of the welding location is poor.

[0077] In one embodiment, the position detection unit (170) may transmit to the area detection unit (180) an image of a battery unit (105) in which the state of the welding position among a plurality of images is determined to be normal. In one embodiment, the position detection unit (170) may transmit to the area detection unit (180) an image of a welding position (e.g., welding position (621)) of a battery unit (105) in which the state of the welding position among a plurality of images is determined to be normal. In one embodiment, the position detection unit (170) may transmit to the judgment unit (195) the result of determining the state of the welding position of the battery unit (105).

[0078] In one embodiment, the area detection unit (180) can detect and / or tag a weld area in an image of a weld location of a battery unit (105). In one embodiment, the area detection unit (180) can detect and / or tag a weld area (e.g., weld area (721, 741)) by inputting an image of a weld location of a battery unit (105) (e.g., 710, 730) into an artificial intelligence model (147).

[0079] In one embodiment, the area detection unit (180) can transmit an image (720, 740) tagged with a welding area (721, 741) to the analysis unit (190).

[0080] In one embodiment, the analysis unit (190) can analyze the welding condition based on an image tagged with a welding area. In one embodiment, the analysis unit (190) can analyze the welding condition by inputting an image tagged with a welding area into an artificial intelligence model (148). In one embodiment, the analysis unit (190) can determine the welding condition of the welding area based on the output of the artificial intelligence model (148). For example, the analysis unit (190) can analyze the welding condition based on the welding condition of the welding area classified by the artificial intelligence model (148) (e.g., a state having a welding length within a reference length range (e.g., image (810), a state having a welding length less than a reference length range (e.g., image (820)) and other defects (e.g., images (830, 840, 850)).

[0081] In one embodiment, if the battery unit (105) has a weld length within a reference length range, the analysis unit (190) may determine that the weld condition is normal. In one embodiment, if the battery unit (105) has a weld length less than the reference length range or has other defects, the analysis unit (190) may determine that the weld condition is defective.

[0082] In one embodiment, the analysis unit (190) can transmit the result of the determination of the welding state to the determination unit (195).

[0083] In one embodiment, the judgment unit (195) can determine whether the battery unit (105) is a good product based on the judgment results of the state judgment unit (160), the position detection unit (170), and / or the analysis unit (190). For example, the judgment unit (195) can determine that the battery unit (105) is a good product if the judgment results of the state judgment unit (160), the position detection unit (170), and the analysis unit (190) all indicate normal. For example, the judgment unit (195) can determine that the battery unit (105) is not a good product if at least one of the judgment results of the state judgment unit (160), the position detection unit (170), or the analysis unit (190) indicates a defect.

[0084] Additionally, the judgment unit (195) can determine whether the battery unit (105) is good based not only on the judgment result based on the artificial intelligence model but also on the judgment result based on the rule-based inspection. Here, the rule-based inspection may be an inspection that determines whether the battery unit (105) is defective based on the gray level information of the image of the battery unit (105).

[0085] For example, the judgment unit (195) may generate a judgment result by determining at least one of the presence or absence of a can of the battery unit (105), the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on the gray level information of the image of the battery unit (105). According to one embodiment, the judgment unit (195) may sequentially determine the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on the gray level information of the image of the battery unit (105). The judgment unit (195) may generate a judgment result determining whether the battery cell is over-welded, whether the top cap is sooted, or whether the tab is damaged by determining at least one of the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on the gray level information of the image of the battery unit (105). The judgment unit (195) can determine whether the battery unit (105) is defective based on a judgment result based on an artificial intelligence model and a judgment result based on a rule-based inspection. That is, the judgment unit (195) can finally determine whether the battery unit (105) is defective based on a judgment result determining whether the battery unit (105) is under-welded and a judgment result determining whether the battery unit (105) is defective or over-welded.

[0086] According to one embodiment, the judgment unit (195) can determine whether the battery unit (105) is defective based on a previously stored image file. Specifically, the judgment unit (195) can determine whether the battery unit (105) is defective based on a previously stored image file of a defective battery unit. Here, the image file of the defective battery unit may include an image of a battery unit that is under-welded, an image of a battery unit that is judged to be over-welded with the size of the soot being larger than a reference size, an image of a battery unit that is judged to have a degree of soot being larger than a reference degree, an image of a battery unit that is judged to have a lump in the weld, and an image of a battery unit that is judged to have an intrusion of the top cap. Accordingly, the judgment unit (195) can determine whether the battery unit is defective based on the image file of the defective battery unit.

[0087] As described above, according to the monitoring device (100) according to one embodiment disclosed in this document, the welding condition of the battery unit can be accurately analyzed to detect a defective battery unit.

[0088] The monitoring device (100) can improve the detection capability of weak welding defects in the battery unit by utilizing a deep learning algorithm in combination with the problem that it is difficult to clearly classify the welding area due to gray level noise in the image of the process processing device that manufactures the battery unit.

[0089] In addition, the monitoring device (100) can prevent leakage of the battery unit, thereby improving the stability and quality of the product equipped with the battery unit.

[0090] In addition, the monitoring device (100) can determine whether the battery unit is defective in real time without the need for a separate device or connection, thereby reducing the cost and time required for defect detection.

[0091] In FIG. 2, various program modules of the program (125) are shown as separate, but this is merely an example. In one embodiment, the state determination unit (160), the location detection unit (170), the area detection unit (180), and / or the analysis unit (190) may be integrated into a single program module. For example, the state determination unit (160), the location detection unit (170), the area detection unit (180), and / or the analysis unit (190) may be integrated into the analysis unit (190).

[0092] In FIG. 2, the artificial intelligence models of the program (125) are depicted as separate, but this is merely an example. In one embodiment, the artificial intelligence models (145, 146, 147, 148) may be implemented as a single artificial intelligence model.

[0093] FIG. 9 is a drawing illustrating a method of operation of a monitoring device (100) according to an embodiment disclosed in this document. FIG. 9 can be described with reference to FIG. 2 through FIG. 8.

[0094] Referring to FIG. 9, in operation 910, the monitoring device (100) can acquire an image. Here, the image may be an image of the battery unit (105) acquired by the image acquisition device (103).

[0095] In operation 920, the monitoring device (100) can perform rule-based inspection. In one embodiment, the monitoring device (100) can perform rule-based inspection based on gray level information of an image.

[0096] For example, the monitoring device (100) can generate a determination result by determining at least one of the presence or absence of a can of the battery unit (105), the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on gray level information of an image of the battery unit (105). According to one embodiment, the monitoring device (100) can sequentially determine the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on gray level information of an image of the battery unit (105). The monitoring device (100) can generate a determination result determining whether the battery cell is over-welded, whether the top cap is sooted, or whether the tab is damaged by determining at least one of the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the location of the positive tab, and whether the positive tab is over-welded, based on gray level information of an image of the battery unit (105).

[0097] In operation 930, the monitoring device (100) can perform an artificial intelligence model-based inspection. For example, the monitoring device (100) can perform an artificial intelligence model-based inspection on an image based on previously trained artificial intelligence models (145, 146, 147, 148).

[0098] In one embodiment, the monitoring device (100) can determine the state of the can assembly based on an image of the battery unit (105). In one embodiment, the monitoring device (100) can determine the state of the weld location based on an image of the battery unit (105). In one embodiment, the monitoring device (100) can analyze the weld status based on a cropped image of the weld location.

[0099] In operation 940, the monitoring device (100) can determine whether there is a defect based on the inspection result.

[0100] In one embodiment, the monitoring device (100) can determine that the battery unit (105) is a good product if both the result of the rule-based inspection and the result of the artificial intelligence model-based inspection indicate normal.

[0101] The above description is merely an illustrative explanation of the technical concept of the present disclosure, and those skilled in the art to which the present disclosure pertains may make various modifications and variations within the scope of the essential characteristics of the present disclosure without departing from its nature.

[0102] Accordingly, the embodiments disclosed in this disclosure are intended to illustrate, not limit, the technical concept of this disclosure, and the scope of the technical concept of this disclosure is not limited by such embodiments. The scope of protection of this disclosure shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of this disclosure.

Claims

Claim 1 A monitoring device comprising: an image acquisition unit for acquiring an image of at least one process processing device related to the manufacture of a battery cell; an analysis unit for inputting the acquired image into an artificial intelligence model for determining the welding state of the battery cell to generate a first determination result for determining whether the battery cell is defective; and a determination unit for determining whether the battery cell is defective based on the first determination result, wherein the determination unit generates a second determination result for determining whether the battery cell is defective based on gray level information of the acquired image, and determines whether the battery cell is defective based on the first determination result and the second determination result. Claim 2 A monitoring device according to claim 1, wherein the artificial intelligence model includes an artificial intelligence model for detecting a welding area of ​​the battery cell based on the image and an artificial intelligence model for analyzing the welding state of the welding area. Claim 3 A monitoring device according to paragraph 2, wherein the artificial intelligence model for analyzing the welding state of the welding area classifies the welding state into a state having a welding length within a reference length range, a state having a welding length less than the reference length range, and a state having other defects. Claim 4 A monitoring device according to claim 1, wherein the determination unit determines at least one of the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the position of the positive tab, and whether the positive tab is over-welded, based on the gray level information of the acquired image, and generates a second determination result. Claim 5 A monitoring device according to claim 1, wherein the artificial intelligence model comprises an artificial intelligence model for determining the state of the battery cell can assembly based on the image and an artificial intelligence model for detecting and / or cropping the welding location in the image when the state of the can assembly is normal. Claim 6 delete Claim 7 delete Claim 8 A method of operation of a monitoring device comprising: a step of acquiring an image of at least one process processing device related to the manufacture of a battery cell; a step of inputting the acquired image into an artificial intelligence model for determining the welding state of the battery cell to generate a first determination result for determining whether the battery cell is defective; a step of generating a second determination result for determining whether the battery cell is defective based on gray level information of the acquired image; and a step of determining whether the battery cell is defective based on the first determination result and the second determination result. Claim 9 A method of operation of a monitoring device according to claim 8, wherein the artificial intelligence model includes an artificial intelligence model for detecting a welding area of ​​the battery cell based on the image and an artificial intelligence model for analyzing the welding state of the welding area. Claim 10 A method of operation of a monitoring device according to claim 9, wherein the artificial intelligence model for analyzing the welding state of the welding area classifies the welding state into a state having a welding length within a reference length range, a state having a welding length less than the reference length range, and a state having other defects. Claim 11 A method of operation of a monitoring device according to claim 9, wherein the step of generating the second judgment result is characterized by generating the second judgment result by determining at least one of the presence or absence of a can of the battery cell, the presence or absence of a top cap, the center distance between the can and the top cap, the position of the positive tab, and whether the positive tab is over-welded, based on the gray level information of the acquired image. Claim 12 A method of operation of a monitoring device according to claim 8, wherein the artificial intelligence model comprises an artificial intelligence model for determining the state of the battery cell can assembly based on the image and an artificial intelligence model for detecting and / or cropping the welding location in the image when the state of the can assembly is normal.

Images