Resistance spot welding method

The resistance spot welding method employs a machine learning model to analyze workpiece expansion and contraction gradients, correcting for disturbances, thus enhancing the accuracy of welding quality assessment and nugget diameter estimation.

JP2024169832AActive Publication Date: 2024-12-06TOYOTA JIDOSHA KK

Patent Information

Application Number
JP2023086612
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2023-05-26
Publication Date
2024-12-06
Estimated Expiration
2043-05-26

AI Technical Summary

Technical Problem

Existing resistance spot welding methods lack accuracy in quality judgment, particularly when disturbances occur during the welding process.

Method used

A resistance spot welding method utilizing a machine learning model that learns the relationship between feature values of the workpiece's expansion and contraction during welding, including gradients of these changes, to determine the quality of the welding state with high accuracy, and corrects for disturbances to enhance precision.

Benefits of technology

The method enables precise determination of welding quality by accurately estimating the nugget diameter and identifying disturbances, thereby improving the overall accuracy of resistance spot welding.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2024169832000001_ABST
    Figure 2024169832000001_ABST
Patent Text Reader

Abstract

To improve the accuracy of the quality determination using an expansion amount of a workpiece in resistance spot welding.SOLUTION: A resistance spot welding method includes: a preparation process of preparing a machine learning model obtained by learning a relation between a feature amount of test data in which a time series variation of an expansion amount of a workpiece in test welding is recorded and quality of a welding state of the test welding; and a determination process of determining quality of a welding state of regular welding by using regular welding data in which a time series variation of an expansion amount of the workpiece in the regular welding is recorded and the machine learning model. The feature amount includes: a first feature amount related to an inclination of a variation of an expansion amount of the workpiece in an expansion period in which the workpiece is expanded through electric conduction; and a second feature amount related to an inclination of a variation of an expansion amount of the workpiece in a contraction period in which the workpiece contracts after the expansion period.SELECTED DRAWING: Figure 5
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present disclosure relates to a resistance spot welding method. [Background technology]

[0002] Regarding resistance spot welding, Patent Document 1 discloses a method for measuring the degree of thermal expansion of a workpiece while current is flowing and the degree of contraction of the workpiece at the end of welding, and determining whether or not an adequate nugget has been formed on the workpiece based on these degrees. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2001-300738 A Summary of the Invention [Problem to be solved by the invention]

[0004] As in Patent Document 1, the quality of welding can be determined by utilizing the amount of expansion of a workpiece during resistance spot welding. It is desirable to further improve the accuracy of such quality determination. [Means for solving the problem]

[0005] The present disclosure can be realized in the following forms.

[0006] (1) According to one aspect of the present disclosure, a resistance spot welding method is provided. The resistance spot welding method includes a preparation step of preparing a machine learning model that learns the relationship between feature values ​​of test data recording a time series change in an expansion amount of a workpiece during test welding and a quality of a welding state of the test welding, and a determination step of determining the quality of a welding state of the main welding using main welding data recording a time series change in an expansion amount of the workpiece during main welding and the machine learning model. The feature values ​​include a first feature value related to a slope of a change in the expansion amount of the workpiece during an expansion period in which the workpiece expands due to current flow, and a second feature value related to a slope of a change in the expansion amount of the workpiece during a contraction period in which the workpiece contracts after the expansion period. In this embodiment, the quality of the welding condition of the actual welding can be determined using a machine learning model that has learned the relationship between the feature amounts of the test data and the quality of the welding condition of the test welding. Since these feature amounts include the first feature amount and the second feature amount, which are feature amounts relating to the slope of the change in the expansion amount, the quality of the welding condition of the actual welding can be determined with high accuracy. (2) In the above embodiment, the preparation step may prepare the machine learning model that has learned the relationship between the feature amount and the type of disturbance in the test welding, and the determination step may include a first step of performing a first determination to determine whether the welding state of the main welding is poor using the main welding data and the machine learning model, and determining the type of disturbance in the main welding in which the welding state is not determined to be poor in the first determination, and a second step of performing a second determination to determine whether the welding state of the main welding in which the welding state is not determined to be poor in the first determination, using the determined type of disturbance and the main welding data. With this embodiment, in the second step, the quality of the welding state of the main welding in which the welding state is not determined to be poor in the first determination can be determined using the type of disturbance determined in the first step and the main welding data. Therefore, it is possible to suppress a decrease in the accuracy of determining the quality of the welding state of the main welding due to disturbance. (3) In the above embodiment, in the second determination, the main welding data may be corrected according to the type of disturbance, the nugget diameter of the nugget in the main welding may be estimated using the corrected main welding data, and the quality of the welding state of the main welding may be determined based on the estimated nugget diameter. In this embodiment, the nugget diameter is estimated using the main welding data corrected according to the type of disturbance, so that the nugget diameter can be estimated with high accuracy even when the main welding includes a disturbance. Then, the quality of the main welding may be determined with high accuracy based on the nugget diameter estimated in this manner. (4) In the above embodiment, the machine learning model may be generated by machine learning using the test data associated with a first label representing the test welding in a good welding state and not including the disturbance, the test data associated with a second label representing the type of the disturbance in the test welding in a good welding state and including the disturbance, and the test data associated with a third label representing the test welding in a poor welding state. In this embodiment, the first step can be easily performed by inputting the welding data into the machine learning model. (5) In the above embodiment, the first feature amount may include a feature amount related to a slope of a change in the expansion amount of the workpiece during a first period, and a feature amount related to a slope of a change in the expansion amount of the workpiece during a second period after the first period. With this embodiment, it is more likely that the quality of the welding state of the main welding can be determined with higher accuracy.

[0007] In addition to the form as the resistance spot welding method described above, the present disclosure can be realized in the form of, for example, a method for generating a machine learning model, a resistance spot welding apparatus, a control method for a resistance spot welding apparatus, a computer program for realizing the control method, a non-transitory recording medium on which the computer program is recorded, etc. [Brief description of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram showing a schematic configuration of a resistance spot welding device. [Diagram 2]FIG. 11 is an explanatory diagram of test data. [Diagram 3] FIG. 4 is an explanatory diagram showing examples of types of disturbances. [Figure 4] FIG. 11 is an explanatory diagram showing an example of an estimation result of a nugget diameter. [Diagram 5] FIG. 1 is a process diagram of a resistance spot welding method. [Figure 6] 13 is a flowchart of a generation process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] A. First embodiment: 1 is a diagram showing a schematic configuration of a resistance spot welding apparatus 100. The resistance spot welding apparatus 100 melts and joins a workpiece W, which is made up of multiple overlapping metal plates. More specifically, the workpiece W has a first metal plate W1 and a second metal plate W2 overlapping each other.

[0010] The resistance spot welding device 100 includes a welding gun 10, a resistance measuring device 16, and a control device 20. The welding gun 10 includes a gun body 11, a pair of electrodes, a lower electrode 12 and an upper electrode 13, and an electrode lifting device 14. The gun body 11 has an upper arm 11T and a lower arm 11B. The lower electrode 12 is fixed to the lower arm 11B. The upper electrode 13 is fixed to the upper arm 11T via the electrode lifting device 14. The electrode lifting device 14 is configured as an electric lifting device having a servo motor 15. The electrode lifting device 14 holds the upper electrode 13 and raises and lowers the upper electrode 13 by the rotational driving force of the servo motor 15. In other words, the upper electrode 13 can be moved by the electrode lifting device 14 in a direction along the opposing direction of the upper electrode 13 and the lower electrode 12.

[0011] When welding the workpiece W, the resistance spot welding device 100 supplies a welding current to the workpiece W via both electrodes while the workpiece W is sandwiched and pressed between the upper electrode 13 and the lower electrode 12. In this manner, the workpiece W melts due to Joule heat generated by the passage of current. The workpiece W is then cooled and solidified, joining the first metal sheet W1 and the second metal sheet W2. The workpiece W expands as it melts due to the passage of current, and contracts as it cools after the passage of current is completed. A nugget is formed at the joint interface between the welded metal sheets.

[0012] The resistance measuring device 16 measures the electrical resistance between the lower electrode 12 and the upper electrode 13. The electrical resistance measured by the resistance measuring device 16 is transmitted to the control device 20.

[0013] The resistance spot welding apparatus 100 further includes an encoder 30 and a strain gauge 40. The encoder 30 detects the amount of rotation of the servo motor 15 at predetermined time intervals and transmits a signal indicating the detected amount of rotation to the control device 20. The strain gauge 40 detects the amount of displacement of the lower electrode 12 due to an external force at predetermined time intervals and transmits a signal indicating the amount of displacement to the control device 20. The encoder 30 and the strain gauge 40 are used to detect the distance between the upper electrode 13 and the lower electrode 12.

[0014] The control device 20 is configured by a computer including a CPU 21 and a storage device 22. The storage device 22 stores a program 221 for controlling the operation of the resistance spot welding device 100, a machine learning model 222 described later, test data 223, actual welding data 224, estimation equation data 225, and correction equation data 226. The CPU 21 executes the program 221 stored in the storage device 22 to realize various functions such as the functions of an operation control unit 71, an expansion amount calculation unit 72, a time-series data generation unit 73, a correction unit 74, a nugget diameter estimation unit 75, a learning model generation unit 76, and a determination unit 77. The control device 20 may be configured by a circuit.

[0015] The operation control unit 71 controls the operation of the resistance spot welding apparatus 100 .

[0016] The expansion amount calculation unit 72 calculates the expansion amount of the workpiece W by calculating the distance between the upper electrode 13 and the lower electrode 12 at predetermined time intervals from the detection values ​​of the encoder 30 and the strain gauge 40. In this specification, the expansion amount is the amount of change in the distance between the upper electrode 13 and the lower electrode 12 from the state in which the workpiece W is clamped between the upper electrode 13 and the lower electrode 12 before a welding current is applied to the workpiece W. The expansion amount correlates with the size of the nugget (e.g., the nugget diameter, which is the diameter of the nugget), which is an index of the quality of the welding condition. Therefore, the quality of the welding condition can be determined based on the expansion amount.

[0017] The time-series data generating unit 73 generates time-series data recording the time-series change in the expansion amount of the workpiece W based on the expansion amount calculated by the expansion amount calculating unit 72. More specifically, the time-series data generating unit 73 generates test data 223 and main welding data 224 as the time-series data. The main welding data 224 is data recording the time-series change in the expansion amount in main welding for actually manufacturing a product. The test data 223 is data recording the time-series change in the expansion amount in test welding performed prior to the main welding.

[0018] FIG. 2 is an explanatory diagram of the test data 223. More specifically, FIG. 2 is a graph in which the horizontal axis represents the elapsed time and the vertical axis represents the expansion amount of the workpiece W. FIG. 2 shows test data TD1, test data TD2, and test data TD3 as examples of the test data 223. The test data TD1 is data related to test welding in which the welding state is good and does not include a disturbance, which will be described later. The test data TD2 is data related to test welding in which the welding state is good and includes a disturbance. More specifically, the test data TD2 is data related to test welding in which a "gap" is included as a disturbance, which will be described later. The test data TD3 is data related to test welding in which the welding state is poor. Although not shown in the figure, the main welding data 224 is recorded as data substantially similar to the test data 223 shown in FIG. 2.

[0019] The test data TD1 to TD3 are each associated with a label related to the result of the test welding for each test data 223. The test data TD1 is associated with a first label. The first label is a label representing a test welding in which the welding condition is good and does not include a disturbance. The test data TD2 is associated with a second label. The second label is a label representing a type of disturbance in a test welding in which the welding condition is good and includes a disturbance. For example, the second label associated with the test data TD2 represents "gap" as a type of disturbance. The test data TD3 is associated with a third label. The third label is a label representing a test welding in which the welding condition is poor.

[0020] In this embodiment, when a test welding includes multiple types of disturbances, a second label representing the type of disturbance that is more dominant in the test welding is associated with the test data 223 of the test welding. In other embodiments, when a test welding includes multiple types of disturbances, the test data 223 of the test welding may be associated with a second label representing, for example, that the test welding includes each type of disturbance.

[0021] As shown in FIG. 2, in each test data 223, the time series change in the expansion amount is recorded at least over the expansion period PE and the contraction period PC. The expansion period PE refers to a period in which the workpiece W expands due to energization. More specifically, in the expansion period PE, the workpiece W expands due to energization through the upper electrode 13 and the lower electrode 12 in a state in which it is pressurized by the upper electrode 13 and the lower electrode 12. The contraction period PC refers to a period after the expansion period PE, in which the workpiece W contracts. More specifically, in the contraction period PC, the workpiece W is compressed by the upper electrode 13 and the lower electrode 12 in a state in which energization is stopped, and thus the workpiece W is cooled and contracts. Each test data shown in FIG. 2 further includes the time series change in the expansion amount in the pre-pressure period PP. In the pre-pressure period PP, the workpiece W is pressurized by the upper electrode 13 and the lower electrode 12 in a state in which no current is applied.

[0022] The learning model generation unit 76 generates the machine learning model 222 by executing machine learning using the feature amount of the test data 223. The machine learning model 222 is a trained model that has learned the relationship between the feature amount of the test data 223 and the quality of the welding state of the test welding. As the machine learning model 222, various machine learning models such as a random forest, a support vector machine (SVM), and a neural network can be used. A method for generating the machine learning model 222 in this embodiment will be described in detail later.

[0023] The feature amount of the test data 223 used to generate the machine learning model 222 includes a first feature amount and a second feature amount. The first feature amount is a feature amount related to the gradient of the change in the expansion amount in the expansion period PE shown in FIG. 2. The second feature amount is a feature amount related to the gradient of the change in the expansion amount in the contraction period PC. Furthermore, the first feature amount in this embodiment includes a third feature amount and a fourth feature amount. The third feature amount is a feature amount related to the gradient of the change in the expansion amount in the first period PE1. The fourth feature amount is a feature amount related to the gradient of the change in the expansion amount in the second period PE2 after the first period PE1. More specifically, the first period PE1 in this embodiment is a period from the expansion start timing t1 at which the expansion period PE starts to the current value change timing tc. Moreover, the second period PE2 is a period from the current value change timing tc to the expansion end timing t2 at which the expansion period PE ends. The current value change timing tc represents the timing at which the magnitude of the welding current is changed during the expansion period PE. In this embodiment, the welding current supplied to the workpiece W in the second period PE2 is greater than the welding current supplied to the workpiece W in the first period PE1. The expansion start timing t1 can also be said to be the timing at which the pre-pressure period PP ends. The expansion end timing t2 can also be said to be the contraction start timing t3 at which the contraction period PC starts.

[0024] Fig. 2 shows various feature amounts of the test data TD1 as examples of the feature amounts of the test data 223. More specifically, Fig. 2 shows the slope s1, slope s2, slope s3, slope s4, maximum expansion amount EM, minimum expansion amount Em, and difference value Dv as the feature amounts of the test data TD1.

[0025] The slope s1 represents the slope of the change in the expansion amount from the expansion start timing t1 to the expansion end timing t2. More specifically, the slope s1 is calculated as a value obtained by dividing the difference between the expansion amount at the expansion end timing t2 and the expansion amount at the expansion start timing t1 by the elapsed time between the two timings. The slope s2 represents the slope of the change in the expansion amount from the contraction start timing t3 to the contraction start timing t4 at which the contraction period PC ends. More specifically, the slope s2 is calculated as a value obtained by dividing the difference between the expansion amount at the contraction start timing t4 and the expansion amount at the contraction start timing t3 by the elapsed time between the two timings. The slope s3 represents the slope of the change in the expansion amount from the expansion start timing t1 to the current value change timing tc. More specifically, the slope s3 is calculated as a value obtained by dividing the difference between the expansion amount at the current value change timing tc and the expansion amount at the expansion start timing t1 by the elapsed time between the two timings. The slope s4 represents the slope of the change in the expansion amount from the current value change timing tc to the expansion end timing t2. More specifically, the slope s4 is calculated as the difference between the expansion amount at the expansion end timing t2 and the expansion amount at the current value change timing tc divided by the elapsed time between the two timings. The maximum expansion amount EM represents the maximum expansion amount during the expansion period PE. The minimum expansion amount Em represents the minimum expansion amount during the contraction period PC. The difference value Dv represents the difference between the maximum expansion amount EM and the minimum expansion amount Em, i.e., the amount of change in the expansion amount during the contraction period PC.

[0026] Among the various feature amounts of the test data TD1 shown in FIG. 2, the slopes s1, s3, and s4 correspond to the first feature amount. The slope s2 corresponds to the second feature amount. The slope s3 corresponds to the third feature amount. The slope s4 corresponds to the fourth feature amount. Note that the first feature amount may be a feature amount related to the slope of the change in the expansion amount during the expansion period PE, and in other embodiments, it is not necessary for the first feature amount to be a value that represents the slope of the change in the expansion amount itself. For example, the first feature amount may be a value for calculating the slope of the change in the expansion amount during the expansion period PE. The same is true for the second feature amount, the third feature amount, and the fourth feature amount.

[0027] Furthermore, the machine learning model 222 in this embodiment has already learned the relationship between the feature amount of the test data 223 and the type of disturbance in the test welding.

[0028] Fig. 3 is an explanatory diagram showing examples of types of disturbances. In Fig. 3, examples of the types of disturbances are "edge", "gap", "push down", "push up", "axis tilt", "axis misalignment", "sealer", "tip wear", "thickness reduction", and "vertical wall". As described later, each disturbance shown in Fig. 3 acts to reduce the amount of expansion relative to the actual measured value of the nugget diameter in resistance spot welding including that disturbance, compared to spot welding not including the disturbance.

[0029] "Edge" refers to the central axis TX of the upper electrode 13 and the central axis BX of the lower electrode 12 being shifted from the desired position HP by a reference distance or more toward the end side in the horizontal direction of the workpiece W. "Gap" refers to the size of the gap gp between the metal plates to be welded being equal to or greater than a reference level at the start of welding. "Pushing down" refers to the position Ed of the upper end of the workpiece W being shifted downward from the reference position SEd by a reference level or more at the start of welding. "Pushing up" refers to the position Ed at the start of welding being shifted upward from the reference position SEd by a reference level or more. "Axis inclination" refers to the central axis TX or central axis BX being tilted by a reference level or more from an angle perpendicular to the plate surface of the workpiece W. "Axis misalignment" refers to the horizontal positional deviation between the central axis TX and the central axis BX being equal to or greater than a reference level. "Sealer" refers to welding being performed in a state where an electrically conductive rust inhibitor (sealer) Sr is applied to the part of the workpiece W to be welded. "Tip wear" refers to the tip of the upper electrode 13 or the lower electrode 12 being worn down to a predetermined standard level or more during welding. "Thickness reduction" refers to the plate thickness d1 of the workpiece W used in welding being smaller than a predetermined standard plate thickness ds by a standard level or more. "Vertical wall" refers to welding being performed near a vertical wall Vw of a workpiece W having a vertical wall Vw. The vertical wall Vw refers to a portion of the workpiece W that is erected so as to intersect with the plate surface of the workpiece W. Note that "no disturbance" shown in FIG. 3 refers to welding that does not include any of the above disturbances.

[0030] The machine learning model 222 in this embodiment has already learned the relationship between the "edge" and "gap" of the disturbances shown in FIG.

[0031] 1 judges whether the welding state of the main welding is good or bad by using the main welding data 224 and the machine learning model 222. Details of the quality judgment for judging whether the welding state of the main welding is good or bad will be described later.

[0032] The nugget diameter estimation unit 75 estimates the nugget diameter to be formed when the resistance spot welding apparatus 100 welds the workpiece W. In this embodiment, the nugget diameter estimation unit 75 estimates the nugget diameter using an estimation equation represented by the following equation (1). This estimation equation is included in the estimation equation data 225. φ=C1×E+C2×S+C3×R+C4 (1) φ represents an estimated value of the nugget diameter [mm]. E represents the expansion amount [mm] in the expansion period PE. More specifically, in this embodiment, the integral value of the expansion amount from the expansion start timing t1 to the expansion end timing t2 is used as the expansion amount E. S represents the contraction amount [mm] in the contraction period PC. More specifically, in this embodiment, the difference between the contraction amount at the contraction start timing t4 and the expansion amount at the contraction start timing t3 is used as the contraction amount S. Therefore, the contraction amount S takes a negative value. R represents an electrical resistance value [Ω]. More specifically, in this embodiment, the electrical resistance value R is used as the electrical resistance value in the second period PE2. C1, C2, C3, and C4 each represent a constant. The above estimation formula is defined using a multiple regression method based on, for example, experimental results of resistance spot welding without disturbances and simulation results of resistance spot welding without disturbances.

[0033] The inventors of the present application have found that when test welding or main welding includes a disturbance as shown in Fig. 3, the correspondence relationship between the expansion amount and the actual measurement value related to the size of the nugget (for example, the actual measurement value of the nugget diameter) changes compared to when test welding or main welding does not include a disturbance. More specifically, when test welding or main welding includes each disturbance shown in Fig. 3, the expansion amount becomes smaller relative to the actual measurement value of the nugget diameter compared to when test welding or main welding does not include a disturbance. Therefore, when determining whether the weld condition is good or bad based on the expansion amount in main welding including a disturbance, if the weld condition is determined in the same way as when no disturbance is included, the accuracy of determining whether the weld condition is good or bad may decrease.

[0034] FIG. 4 is an explanatory diagram showing an example of the estimation result of the nugget diameter. More specifically, FIG. 4 is a graph in which the horizontal axis indicates the estimated value of the nugget diameter and the vertical axis indicates the actual value of the nugget diameter. The estimated value of the nugget diameter in FIG. 4 represents the nugget diameter calculated using the above-mentioned formula (1). In FIG. 4, the result of welding without disturbance is indicated by a black marker, and the result of welding with disturbance is indicated by a white marker. As shown in FIG. 4, in welding with disturbance, the estimated value of the nugget diameter is relatively smaller than the actual value of the nugget diameter compared to welding without disturbance. This is because, in welding with disturbance, the expansion amount is relatively smaller than the actual nugget diameter, as described above.

[0035] When a disturbance is included in the main welding, the correction unit 74 shown in FIG. 1 corrects the main welding data 224 related to the main welding according to the type of the disturbance. More specifically, the correction unit 74 corrects the main welding data 224 by each correction formula included in the correction formula data 226. Each correction formula included in the correction formula data 226 is defined for each type of disturbance. The correction formula for a certain type of disturbance is determined as a correction formula for removing the influence of the disturbance from the main welding data 224 related to the main welding including the disturbance of that type. For example, the correction formula for a certain type of disturbance is defined based on the results of experiments and simulations as a correction formula for correcting the expansion amount in the test data 223 related to the test welding including the disturbance of that type to the same expansion amount as the expansion amount in the test data 223 related to the test welding not including the disturbance. When estimating the nugget diameter of the main welding including the disturbance, the above-mentioned nugget diameter estimation unit 75 uses the integrated value of the expansion amount based on the main welding data 224 corrected by the correction unit 74 as the expansion amount E.

[0036] 5 is a process diagram of the resistance spot welding method in this embodiment. In S100, the machine learning model 222 is prepared. A process of preparing the machine learning model 222, such as S100, is also called a preparation process. In S100 in this embodiment, a generation process described later is executed to generate the machine learning model 222. In the preparation process in other embodiments, for example, the machine learning model 222 may be prepared by the control device 20 acquiring the machine learning model 222 from another computer, a recording medium, or the like.

[0037] FIG. 6 is a flowchart of the generation process. In S110, the learning model generation unit 76 acquires the test data 223. In S120 and S130, the learning model generation unit 76 generates the machine learning model 222 by performing supervised learning using the test data 223 as teacher data. More specifically, the learning model generation unit 76 first extracts the feature amount of the test data 223 in S120. Next, the learning model generation unit 76 learns the relationship between the feature amount extracted from the test data 223 and the label associated with the test data 223 in S130. In S140, the learning model generation unit 76 determines whether the machine learning is completed. In S140, for example, when learning is completed for all the test data 223 acquired in S110, the learning model generation unit 76 determines that the machine learning is completed. When it is determined that the machine learning is not completed in S140, the learning model generation unit 76 returns the process to S120.

[0038] 5, main welding is performed in S200. In S200, the main welding is performed by the operation control unit 71, the expansion amount is calculated by the expansion amount calculation unit 72, and the main welding data 224 is generated by the time-series data generation unit 73. Next, a determination process is performed in S210 to S500. The determination process refers to a process of determining whether the welding state of the main welding is good or bad using the main welding data 224 and the machine learning model 222.

[0039] In S210, a first step is executed. The first step refers to a step of executing a first determination to determine whether or not the welding state of the main welding is poor using the main welding data 224 and the machine learning model 222, and determining the type of disturbance in the main welding in which the welding state is not determined to be poor in the first determination. In S210 in this embodiment, the determination unit 77 inputs the main welding data 224 to the machine learning model 222, thereby executing the first step.

[0040] In S300 to S500, a second step is executed. The second step refers to a step of executing a second determination to determine whether the welding state of the main welding determined to be not defective in the first step is good or bad, using the type of disturbance determined in the first step and the main welding data 224. Note that, if the welding state of the main welding is determined to be defective in S200, S300 to S500 are omitted.

[0041] In S300, the main welding data 224 is corrected according to the type of disturbance determined in S200. In S300 in this embodiment, the correction unit 74 refers to the correction equation data 226 based on the type of disturbance determined in S200, and corrects the main welding data 224 using the correction equation according to the type of disturbance.

[0042] In S400, the nugget diameter is estimated using the main welding data 224 corrected in S300. In S400 in this embodiment, nugget diameter estimation unit 75 estimates the nugget diameter using the main welding data 224 corrected in S300 and the estimation equation included in the estimation equation data 225.

[0043] In S500, the quality of the welding condition of the main welding is determined based on the nugget diameter estimated in S400. In S500 of this embodiment, the determination unit 77 determines whether the nugget diameter estimated in S400 is equal to or greater than a predetermined threshold value, thereby determining whether the welding condition of the main welding is good or bad. More specifically, when the nugget diameter is equal to or greater than the threshold value, the determination unit 77 determines that the welding condition of the main welding is good. Conversely, when the nugget diameter is less than the threshold value, the determination unit 77 determines that the welding condition of the main welding is poor.

[0044] In addition, in S200, main welding may be performed multiple times, in which case main welding data 224 corresponding to each main welding is generated. In this case, for example, S210 to S500 may be performed for each main welding after a series of main welding is completed, or S210 to S500 may be repeatedly performed for each main welding.

[0045] The resistance spot welding method according to the present embodiment described above includes a preparation step of preparing a machine learning model 222 that has learned the relationship between the feature amount of the test data 223 and the quality of the test welding, and a determination step of determining the quality of the main welding using the main welding data 224 and the machine learning model 222. The feature amount of the test data 223 includes a first feature amount related to the gradient of the change in the expansion amount in the expansion period PE and a second feature amount related to the gradient of the change in the expansion amount in the contraction period PC. In this embodiment, the quality of the welding state of the main welding can be determined using the machine learning model 222. Furthermore, since the feature amount of the test data 223 includes the first feature amount and the second feature amount each representing the gradient of the change in the expansion amount, the quality of the welding state of the main welding can be determined with high accuracy. More specifically, for example, even if the lengths of the expansion period PE and the contraction period PC in the test data 223 are different from the lengths of the expansion period and the contraction period in the main welding data 224, the quality of the welding state of the main welding can be determined with high accuracy using the machine learning model 222.

[0046] In the present embodiment, the determination step includes a first step and a second step. In the first step, a first determination is performed to determine whether the welding state of the main welding is poor or not using the machine learning model 222 that has learned the relationship between the feature amount of the test data 223 and the type of disturbance in the test welding, and the main welding data 224, and the type of disturbance in the main welding in which the welding state is not determined to be poor in the first determination is determined. In the second step, the quality of the welding state of the main welding in which the welding state is not determined to be poor in the first determination is determined using the determined type of disturbance and the main welding data 224. Therefore, it is possible to suppress a decrease in the accuracy of the determination of the quality of the welding state of the main welding due to disturbance. In addition, for example, the slope of the expansion amount in the expansion period PE of the test data 223 including a "gap" as a disturbance, such as the test data TD2 in FIG. 2, is smaller than the slope of the expansion amount in the expansion period PE of the test data 223 not including a disturbance, such as the test data TD1. Although not shown, for example, the absolute value of the slope of the expansion amount in the contraction period PC of the test data 223 including an "edge" as a disturbance is larger than the absolute value of the slope of the expansion amount in the contraction period PC of the test data 223 not including the disturbance. In this way, the characteristics of each disturbance are reflected in the slope of the expansion amount change in the expansion period PE and the contraction period PC. Therefore, as described above, the feature amount used to generate the machine learning model 222 includes the first feature amount and the second feature amount, so that the type of disturbance can be determined more accurately in the first step.

[0047] In the second step of this embodiment, the main welding data 224 is corrected according to the type of disturbance, the nugget diameter in the main welding is estimated using the corrected main welding data 224, and the quality of the welding state of the main welding is determined based on the estimated nugget diameter. In this manner, the nugget diameter is estimated using the main welding data 224 corrected according to the type of disturbance, so that the nugget diameter can be estimated with high accuracy even when the main welding includes a disturbance. Then, the quality of the main welding can be determined with high accuracy based on the nugget diameter estimated in this manner.

[0048] In the present embodiment, the machine learning model 222 is generated by machine learning using the test data 223 associated with a first label representing a test welding in which the welding state is good and does not include a disturbance, the test data 223 associated with a second label representing a type of disturbance in the test welding in which the welding state is good and includes a disturbance, and the test data 223 associated with a third label representing a test welding in which the welding state is poor. Therefore, the first process can be easily executed by inputting the main welding data 224 to the machine learning model 222.

[0049] In the present embodiment, the first feature amount includes a third feature amount related to the gradient of the change in the expansion amount in the first period PE1 and a fourth feature amount related to the gradient of the change in the expansion amount in the second period PE2. Therefore, in the determination step, the possibility of determining the quality of the welding state of the main welding with higher accuracy is increased. Particularly, in the present embodiment, the first period PE1 is a timing before the current value change timing tc, and the second period PE2 is a timing after the current value change timing tc. Therefore, even if the magnitude of the welding current is changed during the expansion period PE in the main welding, the quality of the welding state of the main welding can be determined with higher accuracy.

[0050] B. Other embodiments: (B1) In the above embodiment, the feature amount of the test data 223 used to generate the machine learning model 222 includes, in addition to the first feature amount and the second feature amount, the maximum value of the expansion amount in the expansion period PE, the minimum value of the expansion amount in the contraction period PC, and the change amount of the expansion amount in the contraction period PC. In contrast, the feature amount of the test data 223 only needs to include the first feature amount and the second feature amount, and may not include the maximum value of the expansion amount in the expansion period PE, the minimum value of the expansion amount in the contraction period PC, or the change amount of the expansion amount in the contraction period PC. In addition, the feature amount of the test data 223 may include feature amounts other than those described above.

[0051] (B2) In the above embodiment, the first judgment and the second judgment are executed in the judgment process. In contrast, the first judgment and the second judgment do not have to be executed in the judgment process. For example, in the judgment process, the quality of the welding state of the main welding may simply be judged by inputting the main welding data 224 to the machine learning model 222. In this case, the machine learning model 222 does not have to have learned the relationship between the feature amount of the test data 223 and the type of disturbance in the test welding.

[0052] (B3) In the above embodiment, in the second judgment, the main welding data 224 is corrected according to the type of disturbance, and the nugget diameter in the main welding is estimated using the corrected main welding data 224. In contrast, the nugget diameter does not have to be estimated using the corrected main welding data 224, and for example, the nugget diameter may be estimated according to the type of disturbance by using an estimation formula that differs according to the type of disturbance.

[0053] (B4) In the above embodiment, the quality of the welding condition of the main welding, which is not determined to be poor in the first determination, is determined based on the nugget diameter estimated by nugget diameter estimation unit 75. In contrast to this, the quality of the welding condition of the main welding does not have to be determined based on the nugget diameter, and for example, the quality of the welding condition of the main welding, which is not determined to be poor in the first determination, may be determined by inputting main welding data 224 corrected according to the determined type of disturbance into machine learning model 222.

[0054] (B5) In the above embodiment, the machine learning model 222 is generated by machine learning using the test data 223 associated with the first label, the test data 223 associated with the second label, and the test data 223 associated with the third label. In contrast, the machine learning model 222 does not have to be generated in this manner, and may be generated by machine learning using, for example, the test data 223 associated with the second label and the test data 223 associated with the third label. In addition, when the first step and the second step are not executed in the determination step, the machine learning model 222 may be generated by machine learning using, for example, the test data 223 associated with the first label and the test data 223 associated with the third label.

[0055] (B6) In the above embodiment, the first feature amount includes the third feature amount and the fourth feature amount. In contrast, the first feature amount does not have to include the third feature amount or the fourth feature amount.

[0056] (B7) In the above embodiment, the machine learning model 222 may have already learned the relationship between the feature amount of the test data 223 and a type of disturbance other than the “gap” and the “edge”.

[0057] The present disclosure is not limited to the above-mentioned embodiment, and can be realized in various configurations without departing from the spirit of the present disclosure. For example, the technical features in the embodiments corresponding to the technical features in each aspect described in the Summary of the Invention column can be appropriately replaced or combined to solve some or all of the above-mentioned problems or to achieve some or all of the above-mentioned effects. Furthermore, if the technical feature is not described as essential in this specification, it can be appropriately deleted. [Explanation of symbols]

[0058] 10...welding gun, 11...gun body, 11B...lower arm, 11T...upper arm, 12...lower electrode, 13...upper electrode, 14...electrode lifting device, 15...servo motor, 16...resistance measuring device, 20...control device, 21...CPU, 22...storage device, 30...encoder, 40...strain gauge, 71...operation control unit, 72...expansion amount calculation unit, 73...time series data generation unit, 74...correction unit, 75...nugget diameter estimation unit, 76...learning model generation unit, 77...determination unit, 100...resistance spot welding device, 221...program, 222...machine learning model, 223...test data, 224...actual welding data, 225...estimated formula data, 226...corrected formula data

Claims

1. 1. A resistance spot welding method, comprising: A preparation process of preparing a machine learning model that learns the relationship between a feature amount of test data that records a time series change in the expansion amount of the workpiece during test welding and the quality of the welding state of the test welding; A determination process for determining whether the welding state of the main welding is good or bad using main welding data that records a time-series change in the expansion amount of the workpiece during the main welding and the machine learning model, A resistance spot welding method, wherein the feature quantity includes a first feature quantity related to a slope of a change in an amount of expansion of the workpiece during an expansion period in which the workpiece expands due to current flow, and a second feature quantity related to a slope of a change in an amount of expansion of the workpiece during a contraction period in which the workpiece contracts after the expansion period.

2. 2. The resistance spot welding method according to claim 1, In the preparation step, the machine learning model that learns the relationship between the feature amount and the type of disturbance in the test welding is prepared, The determination step includes: A first step of performing a first determination of whether or not a welding state of the main welding is poor using the main welding data and the machine learning model, and determining a type of the disturbance in the main welding in which the welding state is not determined to be poor in the first determination; and a second step of performing a second determination of the quality of the welding state of the main welding, the welding state of which is not determined to be poor in the first determination, using the determined type of the disturbance and the main welding data. Resistance spot welding method.

3. 3. The resistance spot welding method according to claim 2, a nugget diameter of a nugget in the main welding using the corrected main welding data; and a quality of a welding state of the main welding based on the estimated nugget diameter.

4. 3. The resistance spot welding method according to claim 2, a first label representing the test welding in a good welding state and not including the disturbance, a second label representing a type of the disturbance in the test welding in a good welding state and including the disturbance, and a third label representing the test welding in a poor welding state.

5. A resistance spot welding method according to any one of claims 1 to 4, a feature amount relating to a slope of a change in an expansion amount of the workpiece during a first period and a feature amount relating to a slope of a change in an expansion amount of the workpiece during a second period after the first period.

Citation Information

Patent Citations

  • Resistance welding machine

    JP1993329661A

  • Method and device for controlling spot welding

    JP2000005882A

  • Resistance welding equipment and resistance welding quality monitoring device

    JP2002239745A

  • Welding quality decision method and device therefor

    JP2003181649A

  • Spot welding quality diagnosis system

    JP2016203246A

Cited By

  • Method for detecting surface cracks in spot welds

    JP2026168682A