Battery shell quality judgment method based on battery steel before rolling release

Abstract

According to the battery shell quality judgment method based on the battery steel before rolling release, internal inclusion defect detection is established for each rolled strip steel before release, defect data are formed, and according to the defect data, whether the strip steel is blocked or not is judged in combination with simulation of user basketing and sampling inspection. According to the method for judging the quality of the battery shell before rolling release of the battery steel, a quantitative standard is established for blocking of the battery steel, and meanwhile, whether the quality of the battery shell meets the standard of a user side or not can be judged before release of the strip steel.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent steel manufacturing, and specifically relates to a method for determining the quality of a battery shell based on battery steel. Background Art

[0002] The internal inclusion defect (i.e., internal defect) of cold-rolled thin steel coil is one of the key qualities of the coil. The internal defect meter is an important device for detecting the internal inclusion quality of thin strip steel, and is generally installed in the cold rolling finishing unit. The data output by the internal defect meter detection equipment includes the defect position (i.e., position X in the length direction of the steel coil and position Y in the width direction) and the defect size (expressed by voltage value, the larger the voltage value, the larger the defect). In existing inspections, the detection of the internal defect meter is usually used to better guide the defect data analysis and better improve the production process and product quality services.

[0003] The invention application with application number: CN202211340624.9 discloses "a method and device for detecting defects in battery steel shells". This method inserts a columnar power source into the battery steel shell to allow light to pass through the battery steel shell, and then uses an imaging device to collect images, and performs binarization and recognition on the images to achieve fully automatic detection of sand hole defects in the battery steel shell, thereby preventing battery steel shells with defects such as sand holes and pores from flowing into the next process, thereby achieving automatic detection of battery steel shell defects and greatly reducing the quality risks brought by battery steel shell defects.

[0004] The invention application with application number: CN202110292168.4 discloses "a surface defect detection method and device". The surface defect detection method includes: obtaining an image of an object to be detected; inputting the image of the object to be detected into a pre-trained defect detection model to obtain the defect position of the object to be detected, wherein the network structure of the pre-trained defect detection model includes a feature extraction network and a prediction network, the feature extraction network includes multiple connected dense connection modules, the dense connection module includes an input layer, multiple intermediate layers, and an output layer, the output layer of the dense connection module is the input layer of the next dense connection module, the last intermediate layer of the dense connection module is superimposed with the input layer of the dense connection module and the input layer of the forward dense connection module to obtain the output layer of the dense connection module, and the prediction network is a multi-scale prediction module of YoloV3. Summary of the invention

[0005] The purpose of the present invention is, on the one hand, to quantify the standards for blocking or releasing battery steel, and on the other hand, to achieve the determination of whether the quality of the battery shell meets the user-side standards before the strip steel is released.

[0006] In order to achieve the above technical objectives, the present invention provides a method for determining the quality of a battery shell before rolling and releasing the battery steel, and the technical solution thereof is as follows:

[0007] A method for determining the quality of battery shells before battery steel rolling is released.

[0008] For each piece of rolled strip, internal inclusion defect detection is performed before release to generate defect data. Based on the defect data and the simulation of user basket loading and random inspection, the decision on whether the strip should be blocked is made.

[0009] Further,

[0010] The specific steps include:

[0011] S1: Establish internal inclusion defect detection for each rolled strip before release, form a defect data set with specific defect position and specific defect size and upload it to the data processing end;

[0012] S2: A data processing module is provided at the data processing end, a simulated steel strip is formed in the data processing module, and a simulated battery shell is simulated for the simulated steel strip according to the maximum yield rate. At the same time, the defect data set obtained based on step S1 is mapped to the simulated steel strip to form defect information of each simulated battery shell;

[0013] S3: Taking the blocking standard of the user end as the judgment basis, the unqualified rate is calculated and counted based on the simulated battery shell defect information. When the statistical result exceeds the blocking standard, a blocking instruction is issued for the roll of steel strip, otherwise it is released.

[0014] Further,

[0015] Step S3 specifically includes the following steps:

[0016] S31: The search section is set based on the number of random inspections at the user end. The simulated strip steel with battery shell simulation divisions is searched iteratively from the head position of the strip steel along the length direction of the strip steel in a full coverage search manner to the tail end of the strip steel according to the set search section, and the defect density of each search section under each iterative search is calculated according to the searched defect information;

[0017] S32: Determine the maximum defect density of each search segment corresponding to each iterative search number and the specific position of the search segment on the simulated steel strip, calculate the unqualified rate based on all defective battery shells in the position, and when the calculated result exceeds the blocking standard of the user end, issue a blocking instruction for the steel strip, otherwise release it.

[0018] Further,

[0019] The internal shortfall meter is used to establish defect detection for each piece of rolled strip before release to form specific defect position and specific defect size. The defect size is represented by a voltage value in the internal shortfall meter that is proportional to the defect size, and the defect position is represented by a two-dimensional number of points in the internal shortfall meter that carry the strip length and width information.

[0020] Further,

[0021] The “calculating the unqualified rate based on all defective battery shells in the position” specifically includes the following steps:

[0022] S321: dividing the detectable voltage section of the internal defect detector into several intervals from the lowest voltage to the highest voltage, each interval corresponding to a defect type, and each interval setting its own defect number limit value;

[0023] S322: for each defective battery shell, the defect type is determined based on the defect size and the number of defects under the corresponding type is counted. When the number of defects under any defect type exceeds the limit value of the number of defects under the defect type, the battery shell is determined to have abnormal quality, otherwise it is determined to have normal quality;

[0024] S333: Count the number of quality anomalies of all defective battery shells in the position, and calculate the failure rate based on the number of anomalies.

[0025] Further,

[0026] The defect number limit value of each interval in "setting a respective defect number limit value for each interval" in step S321 is determined according to the blocking standard of the user end, or according to experience, or according to a stamping test.

[0027] The battery shell quality determination method based on the battery steel rolling before release of the present invention, on the one hand, quantifies the standard for blocking or releasing the battery steel, and on the other hand, realizes the determination of whether the quality of the battery shell meets the user-end standard before the strip is released, thereby reducing quality objections and establishing a more friendly product interaction relationship. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the determination steps of the present invention;

[0029] Figure 2 Schematic diagram of full sample sampling in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] Below, a method for determining the quality of a battery shell before the battery steel is rolled and released according to the drawings and specific implementation methods of the present invention is further described in detail.

[0031] A method for determining the quality of battery shells before battery steel rolling is released.

[0032] For each piece of rolled steel strip, internal inclusion defect detection is established before release to generate defect data. Based on the defect data and combined with the simulation of user basket loading and random inspection, the determination of whether the steel strip is blocked or not is completed. Figure 1 As shown, the specific steps include:

[0033] S1: Establish internal inclusion defect detection for each rolled strip before release, form a defect data set with specific defect position and specific defect size and upload it to the data processing end;

[0034] S2: A data processing module is provided at the data processing end, a simulated steel strip is formed in the data processing module, and a simulated battery shell is simulated for the simulated steel strip according to the maximum yield rate. At the same time, the defect data set obtained based on step S1 is mapped to the simulated steel strip to form defect information of each simulated battery shell;

[0035] S3: Taking the blocking standard of the user end as the judgment basis, the unqualified rate is calculated and counted based on the simulated battery shell defect information. When the statistical result exceeds the blocking standard, a blocking instruction is issued for the roll of steel strip, otherwise it is released.

[0036] In order to facilitate a more comprehensive understanding of this technical solution, its working principle and process are described as follows:

[0037] When downstream users use steel coil products, they follow certain sampling rules for sampling inspection. In order to better control product quality, the present invention adopts a full inspection method for internal quality control, and outputs blocking suggestions for steel coils that exceed the sampling quality standards. The specific steps are: perform a carpet search of the defect distribution of the entire steel strip, and output the area with the largest density distribution; perform full sample sampling inspection on the area with the largest density, and calculate the unqualified rate; when the unqualified rate is higher than the set value, output a blocking suggestion. The specific steps are as follows:

[0038] (1) Downstream users conduct random inspections on a basket-by-basket basis. A large basket contains N battery shells. The diameter of the blanking disc for each battery shell is D, the strip width is B, and the blanking method is Figure 2 The staggered cutting method shown has the highest yield rate, and the circles are tangent to each other. The strip area corresponding to a basket of battery shells is S, and the corresponding strip length is L' (S = L' × B).

[0039] (2) Assuming the total length of the steel coil is L, a carpet-style full coverage search is performed along the length direction from the head to the tail of the steel strip based on the full width of the steel strip. The length of the steel strip covered by each search is S / B (i.e., L'), and the defect density dens1 in the area is calculated. After each search, the length is stepped forward stepwise, and the next search is started. The density dens2 of the next area is calculated, and so on, until the search ends at the tail of the steel strip. Since the circles are tangent to each other and a full coverage search is required, the recommended value of Step is D*sin60.

[0040] (3) Compare the sizes of dens1, dens2, dens3, dens4, ..., and find the maximum density dens max And their corresponding starting and ending positions L1 and L2.

[0041] (4) There are N battery shell discs within the length interval [L1, L2]. By mapping each defect point on the steel strip to each battery shell disc, the number of defects in different voltage intervals on each battery shell disc can be obtained. It is calculated whether each disc will cause quality objections, and the number of quality objections in these N battery shells is recorded. 质量异议 .

[0042] (5) The number of quality objections N 质量异议 Divided by the total number N, the failure rate is defined as N 质量异议 Indicates the number of battery shells that will cause quality problems in a basket.

[0043] (6) When the unqualified rate is greater than the set quality control standard, the volume is judged as unqualified and a blocking recommendation is output.

[0044] In step (4), the method adopted is as follows to determine how large the defects on each battery shell disc are and how many combinations of defects will cause quality issues.

[0045] Calculate the number of quality objections N that may occur after all N battery shell discs within the length interval [L1, L2] are made into battery shells 质量异议 .

[0046]

[0047] K i Indicates whether the i-th battery case will have quality disputes. If no quality disputes occur, K i =0, otherwise K i =1. K i The value is 0 or 1. The determination method is as follows:

[0048] (1) The voltage range that the internal short circuit instrument can detect is [1,12], which is divided into several small intervals [1,1.2), [1.2,1.5), [1.5,1.8), [1.8,2.0), [2.0,3), [3,4), [4,5), [5,12]. The limit values ​​of the number of defects that will cause quality disputes in each voltage interval on each battery shell are a, b, c, d, e, f, g, h (a, b, c, d, e, f, g, h are positive integers) as shown in Table 1. The above scheme for dividing small intervals is not the only one, but the general requirement is: it is generally recommended to divide it into 8 small intervals, and the intervals less than 2V can be divided into 4 intervals more closely, and the intervals greater than 2V can be divided into 4 intervals more widely. Too many interval divisions will cause a waste of computing resources and are not necessary.

[0049] Table 1 Different defect size ranges and corresponding quantities of a battery shell

[0050]

[0051] In the table:

[0052] a means that when the number of defect points on a battery shell with a voltage in the range of [1,1.2) is a or more, the battery shell will have quality issues.

[0053] b means that when the number of defect points on a battery shell with a voltage in the range of [1.2,1.5) is b or more, the battery shell will have quality issues.

[0054] c means that when the number of defect points on a battery shell with a voltage in the range of [1.5,1.8) is c or more, the battery shell will have quality issues.

[0055] And so on.

[0056] (2) If any voltage interval reaches the quality objection level, it is determined that the battery case will have a quality objection, that is, K i = 1. If multiple voltage intervals reach the quality objection level, it is considered a quality objection shell, that is, K i Still 1.

[0057] (3) The threshold values ​​a, b, c, d, e, f, g, and h for the occurrence of quality objections in each voltage range can be obtained through experiments, that is, through stamping experiments, to explore the critical range (i.e., threshold limit) for the occurrence of quality objections for different defect sizes and quantity combinations, but the cost of a large number of experiments is high. It can also be determined in combination with user feedback and experience. According to experience, the larger the defect, the smaller the threshold value for the occurrence of quality objections. That is, if there is a large defect on the battery shell, quality objections will inevitably occur (such as above 2V), and the threshold value is 1. For defects less than 2V, the threshold value for the occurrence of quality objections may be 1-15. The quality information fed back by users is used below to evaluate the reasonable values ​​of the quantities a, b, c, and d through optimization comparison.

[0058] Given the defect map distribution of n rolls and the ppm value fed back by users, calculate the quantity thresholds a, b, c, and d of quality objections occurring in each voltage interval less than 2.

[0059] (1) First, set the range of the four thresholds a, b, c, and d in the table to 1-15 (integers only). There are 15 different threshold combinations. 4 Combination of

[0060] (2) Secondly, for each combination, the battery shell is judged to be abnormal, and the number of abnormal battery shells among the N selected battery shells is counted. According to the formula Calculate ppm;

[0061] (3) Again, randomly draw 20 times, calculate the corresponding ppm value each time, and take the drawing plan with the largest ppm as the predicted ppm.

[0062] Contrast 15 4 Among the combinations, the group with the largest proportion of smaller errors between the predicted ppm and the user feedback ppm is selected, and the a, b, c, and d values ​​corresponding to this group are taken, which are the defect quantity thresholds corresponding to the four voltage intervals [1, 1.2), [1.2, 1.5), [1.5, 1.8), and [1.8, 2.0).

[0063] Example

[0064] like Figure 2 As shown, taking a coil of a certain specification of 0.25mm thick * 801mm wide * 2500m long as an example, it is used to produce No. 5 batteries, the diameter of the blanking disc is 51mm, and there are 17 discs in a row and 2 columns of staggered blanking. The downstream user has a basket of about 5000 batteries, and the strip length corresponding to a basket of battery shells is L'=5000 / 17*(51+51*sin60).

[0065] Based on the full width of the strip, a carpet search is performed from the head to the tail of the strip along the length direction. The strip length covered by each search is L'=5000 / 17*(51+51*sin60). The defect density dens1 in the area is calculated, and then the step length is increased by step (step=D*sin60) in the length direction. (It should be noted that: Figure 2 The step shown in the figure looks like a D value, which is just for the purpose of drawing a clear example. The specific step is set according to step = D*sin60, because this ensures that both the search and iteration are carried out under the premise of satisfying the full coverage constraint), and then starts the next search, calculates the density dens2 of the next area, and so on, until the search ends at the tail of the strip.

[0066] (3) Compare the sizes of dens1, dens2, dens3, dens4, ..., and find the maximum density dens max And its corresponding starting and ending positions L1=1800, L2=1800+5000 / 17*(51+51*sin60).

[0067] (4) There are about 5000 battery shell discs in the length interval [L1, L2]. Each defect point on the strip is mapped to each battery shell disc, that is, the defect size on each battery shell disc and the number of defects in different size intervals are obtained. It is calculated whether each disc will cause quality objections, and the number of battery shells that will cause quality objections in these 5000 discs is recorded. 质量异议 .

[0068] (5) The number of quality objections N 质量异议 Divided by the total number N, the failure rate is defined as N 质量异议 Indicates the number of battery shells that will cause quality problems in a basket.

[0069] (6) When the unqualified rate is greater than the set quality control standard, the volume is judged as unqualified and a blocking recommendation is output.

[0070] In step (4), the method adopted is as follows to determine how large the defects on each battery shell disc are and how many combinations of defects will cause quality issues.

[0071] Calculate the number of quality objections N that may occur after all N battery shell discs within the length interval [L1, L2] are made into battery shells 质量异议 .

[0072]

[0073] K iIndicates whether the i-th battery case will have quality disputes. If no quality disputes occur, K i =0, otherwise K i =1. K i The value is 0 or 1. The determination method is as follows:

[0074] (1) The voltage range that the internal short circuit detector can detect is [1,12], which is divided into several small ranges [1,1.2), [1.2,1.5), [1.5,1.8), [1.8,2.0), [2.0,3), [3,4), [4,5), [5,12]. The limit values ​​of the number of defects that may cause quality disputes in each voltage range on each battery shell are a, b, c, d, e, f, g, h (a, b, c, d, e, f, g, h are positive integers).

[0075] (2) If any voltage interval reaches the quality objection level, it is determined that the battery case will have a quality objection, that is, K i = 1. If multiple voltage intervals reach the quality objection level, it is considered a quality objection shell, that is, K i Still 1.

[0076] (3) The threshold values ​​a, b, c, d, e, f, g, and h for the occurrence of quality objections in each voltage range can be obtained through experiments, that is, through stamping experiments, to explore the critical range (i.e., threshold limit) for the occurrence of quality objections in different defect sizes and quantity combinations, but the cost of a large number of experiments is high. According to experience, the larger the defect, the smaller the threshold value for the occurrence of quality objections, that is, if there is a large defect on the battery shell, quality objections will inevitably occur (such as above 2V), and the threshold value is 1. For defects less than 2V, the threshold value for the occurrence of quality objections may be 1-15. The quality information fed back by users is used below to evaluate the reasonable values ​​of the quantities a, b, c, and d through optimization comparison.

[0077] Given the defect map distribution of n rolls and the ppm value fed back by users, calculate the quantity thresholds a, b, c, and d of quality objections in each voltage interval.

[0078] (4) First, set the range of the four thresholds a, b, c, and d in the table to 1-15 (integers only). There are 15 different threshold combinations. 4 =50625 combinations;

[0079] (5) Secondly, for each combination, the battery shell is judged to be abnormal, and the number of abnormal battery shells among the selected N battery shells is counted. According to the formula Calculate ppm;

[0080] (6) Again, randomly draw 20 times, calculate the corresponding ppm value each time, and take the drawing plan with the largest ppm as the predicted ppm.

[0081] (7) Comparison 15 4 = Among the 50625 combinations, the group with the smallest error between the predicted ppm and the user feedback ppm accounts for the largest proportion (Table 2 shows the error data between the two in some combinations). The a, b, c, and d values ​​corresponding to this group are [3, 2, 2, 1, 1, 1, 1] as the defect quantity threshold, that is, the defect quantity thresholds corresponding to the eight voltage intervals [1, 1.2), [1.2, 1.5), [1.5, 1.8), [1.8, 2.0), [2.0, 3), [3, 4), [4, 5), and [5, 12] are [3, 2, 2, 1, 1, 1, 1, 1].

[0082] Table 2 Absolute error ratio between predicted ppm and user feedback ppm (partial combination display)

[0083]

[0084] According to the above calculation method, the predicted ppm of this roll is 0.0002. If the blocking standard is greater than 0.0006ppm, the roll can be released.

Claims

1. A method for determining the quality of a battery shell before rolling and releasing the battery steel, characterized in that: For each piece of rolled strip, internal inclusion defect detection is performed before release to generate defect data. Based on the defect data, combined with the simulation of user basket loading and random inspection, the determination of whether the strip is blocked or not is completed.

2. A method for determining the quality of a battery shell before rolling and releasing battery steel according to claim 1, characterized in that: The specific steps include: S1: Establish internal inclusion defect detection for each rolled strip before release, form a defect data set with specific defect position and specific defect size and upload it to the data processing end; S2: A data processing module is provided at the data processing end, a simulated steel strip is formed in the data processing module, and a simulated battery shell is simulated for the simulated steel strip according to the maximum yield rate. At the same time, the defect data set obtained based on step S1 is mapped to the simulated steel strip to form defect information of each simulated battery shell; S3: Taking the blocking standard of the user end as the judgment basis, the unqualified rate is calculated and counted based on the simulated battery shell defect information. When the statistical result exceeds the blocking standard, a blocking instruction is issued for the roll of steel strip, otherwise it is released.

3. A method for determining the quality of a battery shell before rolling and releasing the battery steel according to claim 2, characterized in that: Step S3 specifically includes the following steps: S31: The search section is set based on the number of random inspections at the user end. The simulated strip steel with battery shell simulation divisions is searched iteratively from the head position of the strip steel along the length direction of the strip steel in a full coverage search manner to the tail end of the strip steel according to the set search section, and the defect density of each search section under each iterative search is calculated according to the searched defect information; S32: Determine the maximum defect density of each search segment corresponding to each iterative search number and the specific position of the search segment on the simulated steel strip, calculate the unqualified rate based on all defective battery shells in the position, and when the calculated result exceeds the blocking standard of the user end, issue a blocking instruction for the steel strip, otherwise release it.

4. A method for determining the quality of a battery shell before rolling and releasing battery steel according to claim 2, characterized in that: The internal shortfall meter is used to establish defect detection for each piece of rolled strip before release to form specific defect position and specific defect size. The defect size is represented by a voltage value in the internal shortfall meter that is proportional to the defect size, and the defect position is represented by a two-dimensional number of points in the internal shortfall meter that carry the strip length and width information.

5. A method for determining the quality of a battery shell before rolling and releasing the battery steel according to claim 3, characterized in that: The internal shortfall meter is used to establish defect detection for each piece of rolled strip before release to form specific defect position and specific defect size. The defect size is represented by a voltage value in the internal shortfall meter that is proportional to the defect size, and the defect position is represented by a two-dimensional number of points in the internal shortfall meter that carry the strip length and width information.

6. A method for determining the quality of a battery shell before rolling and releasing the battery steel according to claim 5, characterized in that: The "calculating the unqualified rate based on all defective battery shells in the position" specifically includes the following steps: S321: dividing the detectable voltage section of the internal defect detector into several intervals from the lowest voltage to the highest voltage, each interval corresponding to a defect type, and each interval setting its own defect number limit value; S322: for each defective battery shell, the defect type is determined based on the defect size and the number of defects under the corresponding type is counted. When the number of defects under any defect type exceeds the limit value of the number of defects under the defect type, the battery shell is determined to have abnormal quality, otherwise it is determined to have normal quality; S333: Count the number of quality anomalies of all defective battery shells in the position, and calculate the failure rate based on the number of anomalies.

7. A method for determining the quality of a battery shell before rolling and releasing the battery steel according to claim 6, characterized in that: The defect number limit value of each interval in "setting a respective defect number limit value for each interval" in step S321 is determined according to the blocking standard of the user end, or according to experience, or according to a stamping test.

Images