Welded part characteristic prediction method, prediction model generation method, and prediction model

Abstract

This method for predicting characteristics of a welded section (5) is a method for predicting characteristics of a welded section (5) obtained by resistance spot welding in which a current flows between two electrodes (3, 4) sandwiched between a plate group in which two or more metal plates are stacked, the method comprising: a step for acquiring electrical information output during resistance spot welding; the electrical information comprises at least one of a current flowing between the two electrodes (3, 4), a voltage between the two electrodes (3, 4), a resistance between the two electrodes (3, 4), or a calorific value between the two electrodes (3, 4) in resistance spot welding; and a step of predicting characteristics of the welded portion (5) on the basis of the electrical information.

Description

Technical Field

[0001] The present invention relates to a method for predicting characteristics of a weld obtained by resistance spot welding, a method for generating a prediction model for characteristics of a weld, and a prediction model for characteristics of a weld. Background Art

[0002] Resistance spot welding is performed to assemble the car bodies of automobiles, etc. Resistance spot welding is performed at thousands of locations on a car body. Resistance spot welding is a welding method in which two or more steel plates are overlapped, clamped by a pair of upper and lower welding electrodes, and current is passed between the welding electrodes while applying pressure, thereby forming a nugget of a specified size between the steel plates to obtain a weld joint. The joint strength of the welded portion of the resistance spot welding is determined by the diameter of the nugget. In welding that requires high joint strength, such as welding of automobile parts, it is important to form a nugget of a specified diameter or larger between the steel plates. A method for manufacturing a spot welding joint having a nugget diameter of a desired size is known (for example, see Patent Document 1).

[0003] Patent Document 1: Japanese Patent Application Laid-Open No. 2005-262259

[0004] In order to select resistance spot welding conditions that can produce spot welds with a desired nugget diameter, experiments are conducted on steel plates welded under various conditions, and the cross-sections of the welded joints obtained are observed to measure the nugget diameter. As the number of experimental conditions increases, the number of welded joints for which the nugget diameter is measured increases. Because measuring the nugget diameter of each welded joint takes a considerable amount of time, the efficiency of selecting resistance spot welding conditions decreases. There is a need to improve the efficiency of selecting resistance spot welding conditions. In particular, in stacks of three or more metal plates, where the total thickness of the stack is divided by the thickness of the thinnest of the plates, i.e., a thickness ratio of 5 or more, it is difficult to form a nugget between the thinnest and thickest plates. Therefore, Patent Document 1 discloses a welding method that stabilizes nugget formation by applying multiple levels of power and pressure. However, as welding parameters increase, the efficiency of selecting welding conditions further decreases. Therefore, in a plate set formed by stacking three or more metal plates with a thickness ratio of 5 or more, there is a demand for further improvement in the efficiency of the operation of selecting conditions for resistance spot welding. Summary of the Invention

[0005] The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a weld property prediction method, a prediction model generation method, and a prediction model that can efficiently select resistance spot welding conditions.

[0006] The method for determining welding conditions for resistance spot welding according to the present invention to achieve the above-mentioned object is as follows.

[0007] [1] A method for predicting the characteristics of a weld portion of a stack of two or more metal plates, obtained by resistance spot welding with a current flowing between two electrodes sandwiching the stack of two or more metal plates, comprising the following steps:

[0008] The step of obtaining electrical information output during resistance spot welding, wherein the electrical information includes at least one of a current flowing between the two electrodes, a voltage between the two electrodes, a resistance between the two electrodes, or a heat generation between the two electrodes during resistance spot welding; and

[0009] A step of predicting characteristics of a welded portion of a plate assembly formed by stacking the two or more metal plates based on the electrical information.

[0010] [2] The method for predicting weld characteristics according to [1], wherein:

[0011] In the step of predicting the characteristics of the weld, the electrical information is input into a prediction model for the characteristics of the weld, and a prediction result of the characteristics of the weld outputted from the prediction model is obtained.

[0012] [3] The method for predicting weld characteristics according to [2] above, wherein:

[0013] The method further includes the step of performing learning using the electrical information and the analysis results of the characteristics of the welded portion as learning data to generate the prediction model.

[0014] [4] The method for predicting weld characteristics according to any one of [1] to [3] above, wherein:

[0015] In the step of obtaining the predicted results of the characteristics of the weld, at least one predicted result of the weld nugget diameter, tensile shear strength, cross tensile strength, or fracture strength based on a mechanical peel test of the weld is obtained as the characteristic of the weld.

[0016] [5] The method for predicting weld characteristics according to any one of [1] to [4] above, wherein:

[0017] In the step of acquiring the electrical information, a total calorific value from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

[0018] [6] The method for predicting weld characteristics according to [5], wherein:

[0019] In the step of obtaining the electrical information, the electrical resistance between the two electrodes when the current flow is completed is obtained as the electrical information.

[0020] [7] The method for predicting weld characteristics according to [5] or [6] above, wherein:

[0021] In the step of acquiring the electrical information, a heat value during a portion of a period from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

[0022] [8] The method for predicting weld characteristics according to any one of [1] to [4] above, wherein:

[0023] The resistance spot welding is performed on a plate group formed by stacking at least three metal plates including a first metal plate, a second metal plate, and a third metal plate.

[0024] The thickness ratio, which is a value obtained by dividing the total thickness of the plate group by the thickness of the thinnest metal plate in the plate group, is 5 or more.

[0025] [9] The method for predicting weld characteristics according to [8], wherein:

[0026] In the step of acquiring the electrical information, a total calorific value from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

[0027]

[10] The method for predicting weld characteristics according to [9], wherein:

[0028] In the step of obtaining the electrical information, the electrical resistance between the two electrodes when the current flow is completed is obtained as the electrical information.

[0029]

[11] The method for predicting weld characteristics according to [9] or

[10] , wherein:

[0030] The resistance spot welding includes a first stage in which a first current flows and a second stage in which a second current flows.

[0031] In the step of acquiring the electrical information, a heat value during the first stage of the period from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

[0032]

[12] A prediction model generation method is a method for generating a model for predicting the characteristics of a weld portion of a stack of two or more metal plates obtained by resistance spot welding in which a current flows between two electrodes sandwiching the stack of two or more metal plates, wherein the prediction model generation method comprises the following steps:

[0033] The step of obtaining electrical information output during resistance spot welding, wherein the electrical information includes at least one of a current flowing between the two electrodes during resistance spot welding, a voltage between the two electrodes, a resistance between the two electrodes, or a heat generation between the two electrodes;

[0034] a step of obtaining analysis results of characteristics of the welded portion when resistance spot welding is performed on a plate group formed by overlapping the two or more metal plates; and

[0035] A step of performing machine learning based on learning data that establishes a correlation between the above-mentioned electrical information and the analysis results of the characteristics of the above-mentioned welded portion when resistance spot welding is performed on the plate group formed by overlapping the above-mentioned two or more metal plates, thereby generating a model for predicting the characteristics of the welded portion of the plate group formed by overlapping the above-mentioned two or more metal plates.

[0036]

[13] A prediction model, which is generated by the prediction model generation method described in

[12] above and predicts the characteristics of a weld obtained by resistance spot welding.

[0037] According to the present invention, a weld property prediction method, a prediction model generation method, and a prediction model are provided, which are capable of efficiently selecting resistance spot welding conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 This is a schematic diagram explaining the outline of resistance spot welding.

[0039] Figure 2 This is a block diagram showing a configuration example of a prediction system according to the present invention.

[0040] Figure 3A This is a graph showing an example of temporal changes in the resistance value between electrodes in resistance spot welding.

[0041] Figure 3B This is a graph showing an example of temporal changes in the cumulative amount of heat generated by resistance spot welding.

[0042] Figure 4 This is a flowchart showing an example of the procedure of the prediction model generation method.

[0043] Figure 5 This is a flowchart showing an example of the procedure of a weld zone characteristic prediction method.

[0044] Figure 6 This is a cross-sectional view showing an example of a welded portion of a plate assembly formed by stacking three metal plates.

[0045] Figure 7 This is a graph showing an example of a pattern in which the current flowing between electrodes is divided into two stages.

[0046] Figure 8A This is a cross-sectional view showing an example of a welded portion when a first-stage current flows through a plate group formed by stacking three metal plates.

[0047] Figure 8B This is a cross-sectional view showing an example of a welded portion when a second-stage current flows through a plate group formed by stacking three metal plates.

[0048] Figure 9 This is a graph showing an example of the correlation between the total calorific value and the nugget diameter.

[0049] Figure 10 This is a graph showing an example of the correlation between the initial calorific value and the nugget diameter. DETAILED DESCRIPTION

[0050] The following describes embodiments of the weld property prediction method, prediction model generation method, and prediction model according to the present invention, based on the accompanying drawings. The drawings are schematic and may differ from actual examples. The following embodiments illustrate devices and methods for embodying the technical concepts of the present invention and do not limit the structure to the following. In other words, the technical concepts of the present invention can be modified in various ways within the technical scope of the claims.

[0051] By executing the weld property prediction method of the present invention, the properties of the weld obtained by resistance spot welding are predicted. Figure 1 As shown in FIG. , resistance spot welding, performed to produce the welded portion predicted by the method of the present invention, involves sandwiching a pair of overlapping metal plates 1 and 2 between electrodes 3 and 4, and joining the plates by passing a current between electrodes 3 and 4 while applying pressure. The current flowing between electrodes 3 and 4 causes the temperatures of the metal plates 1 and 2 to rise due to heat generated by the electrical resistance of each plate and the contact resistance of the plates. When current is applied between electrodes 3 and 4, the high-temperature portions of the metal plates 1 and 2 melt and solidify after the current flow ends, forming a welded portion 5 joining the metal plates 1 and 2. The welded portion 5 can be disk-shaped, with the axis of rotation being the line connecting the electrodes 3 and 4. The welded portion 5 is also called a "nugget."

[0052] The joint strength of the weld 5 obtained by performing resistance spot welding is determined by the area of the weld 5 joining the metal plates 1 and 2. In other words, the joint strength of the weld 5 is determined by the size of the weld 5 when the metal plates 1 and 2 are viewed from above. The size of the weld 5 when the metal plates 1 and 2 are viewed from above can be expressed as the diameter d of the disc-shaped weld 5. The diameter d of the disc-shaped weld 5 is also called the nugget diameter. In cases where high joint strength is required, such as in automotive parts, it is important to form the weld 5 so that the nugget diameter is greater than a specified diameter.

[0053] Welding section 5 is located between metal plates 1 and 2. To confirm that the nugget diameter of weld section 5 is greater than a specified diameter, it is necessary to conduct a test by actually welding metal plates 1 and 2. Samples are collected from the weld joint obtained through the test and the nugget diameter is measured through cross-sectional observation. To determine the welding conditions that ensure that the nugget diameter of weld section 5 is greater than the specified diameter, it is necessary to repeat the test and cross-sectional observation while changing the welding conditions. However, performing cross-sectional observation requires a lot of time and cost.

[0054] The joint strength of the weld 5 is expressed as the tensile shear strength or cross-tensile strength of the weld 5. In cases where high joint strength is required for automotive parts, etc., it is important to form the weld 5 in a manner that improves the tensile shear strength or cross-tensile strength of the weld 5. In order to confirm the tensile shear strength or cross-tensile strength of the weld 5, it is necessary to conduct a test in which the metal plates 1 and 2 are actually welded together, and then conduct a fracture test in which a load is applied to the weld joint obtained through the test. In order to determine the welding conditions so that the joint strength of the weld 5 reaches the required strength, it is necessary to repeat the test and fracture test while changing the welding conditions. However, the fracture test consumes a lot of man-hours and costs.

[0055] Therefore, in order to reduce the time and cost required to determine welding conditions and improve the efficiency of welding condition selection, it is necessary to predict the nugget diameter of the weld 5 without cross-sectional observation or to predict the strength of the weld 5 without fracture testing. The weld property prediction method of the present invention makes it possible to predict the properties of the weld 5, including the nugget diameter and strength of the weld 5, without cross-sectional observation of the weld joint obtained through testing or fracture testing. As a result, the selection of welding conditions is improved.

[0056] (Configuration Example of Prediction System 100)

[0057] like Figure 2 As shown, a prediction system 100 according to an embodiment of the present invention includes a prediction model generation device 10 , a prediction device 20 , a welding device 30 , and an analysis device 40 .

[0058] The welding device 30 performs resistance spot welding by using two electrodes 3 and 4 (refer to Figure 1 ) with overlapping, for example Figure 1 The metal plates are joined together by applying pressure and passing electricity between the two electrodes 3 and 4. The welding device 30 is configured to control the pressure and the welding current flowing between the electrodes during welding. The welding device 30 may be in various forms such as a stationary or robotic welding gun. The power supply of the welding device may be either a DC power supply or an AC power supply. When the power supply is an AC power supply, calculations at the point where V=0 is ignored to prevent the resistance from becoming infinite, or a moving average is obtained to prevent the resistance from becoming 0. The shapes of the electrodes 3 and 4 of the welding device 30 may be Figure 1 The shape of the tip having a curved surface is illustrated, but is not limited thereto and may have various other shapes. The welding device 30 outputs electrical information during resistance spot welding. The electrical information output from the welding device 30 during resistance spot welding includes at least one of the current flowing between the two electrodes during resistance spot welding, the voltage between the two electrodes during resistance spot welding, the resistance between the two electrodes during resistance spot welding, or the amount of heat generated between the two electrodes during resistance spot welding. Hereinafter, the electrical information output during resistance spot welding is also referred to as electrical information during resistance spot welding.

[0059] A plate assembly can be formed by stacking two or more steel plates. If one or more of the two or more steel plates constituting the plate assembly is a high-tensile steel plate of 590 MPa or higher, the analysis and confirmation of the weld status is relatively complex. Therefore, the method of the present invention selects welding conditions for a plate assembly containing high-tensile steel plates, thereby improving the efficiency of the welding condition selection process. Furthermore, if the strength difference between the thinnest metal plate and the thickest metal plate of the two or more steel plates constituting the plate assembly is 590 MPa or higher, the analysis and confirmation of the weld status is even more complex. Therefore, the method of the present invention selects welding conditions for a plate assembly containing metal plates with a strength difference of 590 MPa or higher, thereby improving the efficiency of the welding condition selection process.

[0060] The number of metal plates constituting the plate set to be joined by resistance spot welding is not limited to two and may be three or more. The two or more metal plates to be joined by resistance spot welding may be steel plates. At least one of the two or more metal plates to be joined by resistance spot welding may be a plated steel plate. Alternatively, two or more metal plates may be plated steel plates. The plated steel plates may be plated on only one side or on both sides. For example, the plated steel plates may be galvanized steel plates.

[0061] The analysis device 40 analyzes the properties of the welded portion of the plate assembly after resistance spot welding performed by the welding device 30 and outputs the analysis results of the welded portion properties. The analysis device 40 may include a cross-sectional observation device or a non-destructive inspection device to measure the diameter of the weld nugget. The analysis device 40 may also include a fracture testing device to measure the joint strength of the weld. The welded portion properties include at least one of the weld nugget diameter, tensile shear strength, cross-tension strength, or fracture strength based on a mechanical peel test.

[0062] The prediction model generation device 10 includes a data acquisition unit 12, a database 14, and a model generation unit 16. The data acquisition unit 12 acquires electrical information from a welding device 30 during a test in which resistance spot welding is actually performed on a plate assembly consisting of two or more stacked metal plates. The data acquisition unit 12 acquires analysis results of weld characteristics of a sample welded joint obtained through the welding test from an analysis device 40. The data acquisition unit 12 stores data in the database 14 that associates the electrical information from the resistance spot welding during the welding test with the analysis results of the weld characteristics of the sample welded joint obtained through the welding test. The model generation unit 16 performs learning using the data stored in the database 14 as learning data. Through learning, the model generation unit 16 generates a model for predicting weld characteristics. The model for predicting weld characteristics is configured to, when input with electrical information from resistance spot welding, output a prediction result of the weld characteristics when resistance spot welding is performed using an operation determined by the electrical information. The model generation unit 16 outputs the generated model to the prediction device 20.

[0063] Prediction device 20 uses the model obtained from prediction model generation device 10 as prediction model 22. Prediction device 20 obtains electrical information during resistance spot welding during a welding test performed by welding device 30 and inputs the information into prediction model 22. Prediction device 20 obtains the prediction results of weld characteristics output from prediction model 22 and outputs them externally.

[0064] The prediction model generation device 10, prediction device 20, welding device 30, or analysis device 40 are communicatively connected to one another. The prediction model generation device 10, prediction device 20, welding device 30, or analysis device 40 may include a communication interface based on a wired or wireless communication standard. For example, wireless communication standards may include cellular phone communication standards such as 3G, 4G, or 5G. Other examples of wireless communication standards include IEEE 802.11 or Bluetooth (registered trademark). The communication interface may support one or more of these communication standards. The communication interface is not limited to the examples above and may communicate with other devices or input and output data based on various standards. The prediction model generation device 10, prediction device 20, welding device 30, or analysis device 40 may be connected via a network or directly (e.g., peer-to-peer (P2P)).

[0065] The prediction device 20 can also be communicatively connected to an external device to transmit the predicted results of the weld characteristics to the external device. The prediction device 20 can include an output device for outputting the predicted results of the weld characteristics. The prediction device 20 can also include a display device for displaying the predicted results of the weld characteristics as an output device. The display device can be configured to include, for example, an LCD (Liquid Crystal Display), an organic EL (Electro-Luminescence) display, an inorganic EL display, or a PDP (Plasma Display Panel). The display device is not limited to the above-mentioned displays and can also include various other types of displays.

[0066] The prediction model generation device 10 or the prediction device 20 can be configured to include at least one processor, such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). The prediction model generation device 10 or the prediction device 20 can be configured with a single processor or multiple processors. The processor constituting the prediction model generation device 10 or the prediction device 20 can implement the functions of the prediction model generation device 10 or the prediction device 20 by reading and executing a program stored in a storage unit described below.

[0067] The prediction model generation device 10 or the prediction device 20 may include a storage unit. The storage unit stores various information or data. For example, the storage unit may store a program executed in the prediction model generation device 10 or the prediction device 20, or data used for processing executed in the prediction model generation device 10 or the prediction device 20 or the results of processing. In addition, the storage unit may function as a working memory of the prediction model generation device 10 or the prediction device 20. The storage unit may be configured to include, for example, a semiconductor memory, etc., but is not limited thereto. For example, the storage unit may be configured as an internal memory used as a processor of the prediction model generation device 10 or the prediction device 20, or may be configured as a hard disk drive (HDD) that can be accessed from the prediction model generation device 10 or the prediction device 20. The storage unit may also be configured as a non-temporary readable medium. The storage unit may be configured integrally with the prediction model generation device 10 or the prediction device 20, or may be configured separately from the prediction model generation device 10 or the prediction device 20.

[0068] Prediction system 100 may not include prediction model generation device 10. If prediction system 100 does not include prediction model generation device 10, prediction device 20 obtains prediction model 22 from an external device. Prediction system 100 may not include welding device 30. If prediction system 100 does not include welding device 30, prediction model generation device 10 or prediction device 20 obtains electrical information from the external device performing welding during resistance spot welding in a welding experiment. Prediction system 100 may not include analysis device 40. If prediction system 100 does not include analysis device 40, prediction model generation device 10 obtains analysis results of weld characteristics of a sample welded joint obtained through welding experiments from an external device performing analysis.

[0069] (Operation Example of Prediction System 100)

[0070] Hereinafter, an example of the operation of the prediction system 100 will be described in detail.

[0071] Data acquisition unit 12 of prediction model generation device 10 acquires electrical information from welding device 30 during resistance spot welding performed in a welding experiment. Data acquisition unit 12 acquires analysis results of weld characteristics of welded joint samples obtained in the welding experiment from analysis device 40.

[0072] The data acquisition unit 12 selects one or more performance data from the electrical information during resistance spot welding and stores the selected data in the database 14. The performance data of the electrical information during resistance spot welding includes at least one of the current flowing between the two electrodes or the voltage applied between the two electrodes during resistance spot welding, the resistance between the two electrodes during resistance spot welding, or the amount of heat generated between the two electrodes during resistance spot welding.

[0073] The data acquisition unit 12 selects one or more performance data from the analysis results of the weld characteristics and stores them in the database 14. The performance data of the weld characteristics include the weld 5 (refer to Figure 1 The analysis results of the weld characteristics are information obtained through offline testing in the analysis device 40.

[0074] The data acquisition unit 12 stores the performance data of electrical information during resistance spot welding and the performance data of weld characteristics in a database 14 in a manner that allows for association, for example, along with supplementary information such as a welding management number. The performance data of electrical information during resistance spot welding and the performance data of weld characteristics are associated with each other using supplementary information such as the welding management number and stored as a set of data sets in the database 14. One data set corresponding to one welding condition in a welding experiment is stored in the database 14. The number of data sets used to generate the prediction model 22 is preferably 100 or more, more preferably 200 or more, and even more preferably 300 or more.

[0075] The model generation unit 16 uses one or more performance data selected from the electrical information in the resistance spot welding stored in the data set of the database 14 as input performance data, and generates a prediction model 22 obtained by performing machine learning using the input performance data. The machine learning method performed by the model generation unit 16 can be a well-known learning method. The machine learning method can be any method as long as it can obtain a practically sufficient prediction accuracy of the weld characteristics in the learned prediction model 22. For example, a well-known machine learning method based on a neural network including deep learning, convolutional neural network (CNN), or recurrent neural network (RNN) can be used. As other methods, regression tree learning, support vector regression, or Gaussian process can also be used. In addition, an integrated model composed of a combination of multiple models can also be used. In addition, the prediction model 22 can also be appropriately updated by relearning using the latest learning data after it is generated by performing machine learning.

[0076] <Performance Data on Electrical Information in Resistance Spot Welding>

[0077] As described above, in the prediction system 100 according to this embodiment, one or more performance data are selected from the electrical information in the resistance spot welding to be used for the prediction of the weld characteristics. The one or more performance data selected from the electrical information in the resistance spot welding include, for example, the welding current value I, the inter-electrode voltage V, the inter-electrode resistance R, or the calorific value Q. The welding current value I is the value of the current flowing between the two electrodes 3 and 4 (see FIG. 1 ). Figure 1). The inter-electrode voltage V is the voltage applied between the two electrodes 3 and 4 during resistance spot welding. The inter-electrode resistance R is the resistance between the two electrodes 3 and 4 and is calculated using the inter-electrode voltage V and the welding current value I as R = V / I. The calorific value Q is the calorific value between the two electrodes 3 and 4 and is calculated using the welding current value I, the inter-electrode resistance R, the resistivity r of the metal plates constituting the plate group, and the plate thickness t as Q = (I·R)^2 / (r·t^2). When the metal plates are steel plates, the resistivity r is the inherent resistance value of pure iron at 1200°C.

[0078] During resistance spot welding, the metal sheets melt and the resistance of the metal sheets changes. Figure 3A As shown in the example, during the execution of resistance spot welding, the inter-electrode resistance R changes with time. Figure 3B As shown in the example, during the execution of resistance spot welding, the cumulative value of the amount of heat increases over time. Figure 3A and Figure 3B In the graph of , the horizontal axis represents the welding time of resistance spot welding. Figure 3A The vertical axis of the graph represents the inter-electrode resistance R. Figure 3B The vertical axis of the graph represents the cumulative value of heat generated between the electrodes, that is, the cumulative calorific value.

[0079] Figure 3A and Figure 3B The reference numerals indicating specific time on the horizontal axis of are defined as follows.

[0080] Tm: The time when the inter-electrode resistance R reaches its maximum value (Rmax)

[0081] Ts: The time after a specified time has passed since the start of power supply for resistance spot welding

[0082] Tf: The time when the power supply of resistance spot welding ends

[0083] Tml: The time between the time (Tm) when the inter-electrode resistance R reaches its maximum value (Rmax) and the time (Tf) when the energization ends. This time is calculated as Tm + (Tf - Tm) / 2.

[0084] Figure 3A The reference numerals indicating specific resistance values on the vertical axis are defined as follows.

[0085] Rfirst: The value of the inter-electrode resistance R after a specified time (Ts) has passed since the start of energization for resistance spot welding.

[0086] Rmax: Maximum value of inter-electrode resistance R

[0087] Rfm: The value of the inter-electrode resistance R at half the time (Tm) when the inter-electrode resistance R reaches its maximum value (Rmax)

[0088] Rlast: The value of the resistance R between electrodes at the end of the power supply of resistance spot welding

[0089] Rml: The value of the inter-electrode resistance R at the time (Tml) between the time (Tm) when the inter-electrode resistance R reaches its maximum value (Rmax) and the time (Tf) when the energization ends

[0090] Figure 3B The reference symbols indicating specific calorific values on the vertical axis are defined as follows.

[0091] Qm: Cumulative heat generation until the time (Tm) when the inter-electrode resistance R reaches its maximum value (Rmax)

[0092] Qi: The cumulative heat generated from the start of resistance spot welding until a specified time (Ts) has passed

[0093] Qt: The cumulative heat generated from the start to the end of the resistance spot welding

[0094] Qi is also called the initial calorific value. Qt is also called the total calorific value.

[0095] The performance data of electrical information during resistance spot welding is determined based on the values indicated by the respective reference numerals. In other words, the performance data of electrical information during resistance spot welding includes the values indicated by the respective reference numerals. It has been found that the total calorific value Qt in the performance data of electrical information during resistance spot welding significantly affects weld properties. For example, in the examples described below, it was confirmed that the total calorific value Qt is correlated with the nugget diameter. Therefore, the total calorific value Qt can be used as the performance data used to generate the prediction model 22, and as the performance data input into the prediction model 22 to predict weld properties, from the performance data of electrical information during resistance spot welding.

[0096] Furthermore, it was discovered that the initial calorific value Qi significantly impacts the nugget growth process during the initial phase of resistance spot welding, during the initial phase of energization. In other words, it was discovered that the initial calorific value Qi in the performance data of electrical information in resistance spot welding significantly impacts weld characteristics. The initial calorific value Qi is the calorific value for a portion of the period from the start of energization to the end of energization in resistance spot welding. The prescribed time Ts defining the initial calorific value Qi can be set to, for example, 200 ms (milliseconds). The initial calorific value Qi can also be calculated by excluding the calorific value during a very short period of time, such as 10 ms, at the beginning of welding, after the start of energization in resistance spot welding. For example, in the examples described below, it was confirmed that the initial calorific value Qi is correlated with the nugget diameter. Therefore, the initial calorific value Qi can be used as the performance data in the performance data of electrical information in resistance spot welding for generating the prediction model 22, as well as the performance data input into the prediction model 22 to predict weld characteristics. By using the initial calorific value Qi, prediction accuracy is improved.

[0097] In addition, it has been found that the various values of the inter-electrode resistance R in the performance data of electrical information in resistance spot welding (Rfirst, Rfm, Rmax, Rml, or Rlast, or the average value of the inter-electrode resistance R over the entire period from the start to the end of welding, i.e., Rmean) have a significant impact on the characteristics of the weld. The value of the inter-electrode resistance R is greatly affected by the energized diameter. The energized diameter is affected by the nugget diameter. For example, the value of the inter-electrode resistance R at the end of welding (Rlast) is correlated with the final nugget diameter. Therefore, the value of the inter-electrode resistance R at the end of welding (Rlast) can be used to predict the nugget diameter of the weld.

[0098] <Welding properties>

[0099] As a characteristic of the weld portion, the nugget diameter d of the weld portion 5 can be cited (see Figure 1 ), cross tensile strength, shear tensile strength, or fracture strength based on a mechanical peel test, etc. The weld properties are measured by collecting a test piece from a weld joint obtained by performing resistance spot welding on a plate group consisting of two or more overlapping metal plates. The nugget diameter in the weld properties can be measured by the method described in JIS Z3140, or by non-destructive testing such as ultrasonic testing or eddy current testing. The cross tensile strength can be measured by the method described in JIS Z3137. The shear tensile strength can be measured by the method described in JIS Z3136. The fracture strength based on a mechanical peel test can be measured by the method described in ISO 14270.

[0100] <Welding conditions>

[0101] The welding conditions for resistance spot welding, which is the subject of weld characteristics prediction in the prediction system 100 according to this embodiment, are not particularly limited. For example, the welding conditions can be set so that a constant current flows from the start to the end of resistance spot welding. Alternatively, the welding conditions can be set so that the period from the start to the end of resistance spot welding is divided into a first stage and a second stage, with different welding conditions applied during the first stage and the second stage, respectively. In particular, when welding a stack of three or more metal plates, where the total thickness of the stack is divided by the thickness of the thinnest of the metal plates, i.e., a thickness ratio of 5 or more, it is preferable to perform two-stage welding under the following conditions.

[0102] In the first stage of welding, the pressing force P1 (kN) is preferably set to satisfy the following equation (1) in relation to the thickness tm of the thinnest metal plate among the plurality of metal plates. Figure 6 The middle corresponds to the plate thickness (mm) of the upper steel plate 6.

[0103] 0.8tm≤P1≤5tm………(1)

[0104] When the pressure P1 in the first stage of welding is 5tm (kN) or more, the pressure becomes too high, the heat generated by the contact resistance becomes small, and no nugget is formed between the upper steel plate 6 and the middle steel plate 7. On the other hand, when the pressure P1 is 0.8tm or less, the contact resistance between the electrode contact and the upper steel plate 6 becomes large, sparks are easily generated, and spatter is also easily generated from between the upper steel plate 6 and the middle steel plate 7.

[0105] In the first stage of welding, the current-carrying time T1 (ms) is preferably set so as to satisfy the following formula (2).

[0106] 30≤T1≤120………(2)

[0107] When the energizing time T1 is 30 ms or less, the energizing time is too short, so a nugget of desired size is not formed between the upper steel plate 6 and the middle steel plate 7. On the other hand, when the energizing time T1 is longer than 120 ms, spattering occurs.

[0108] In the first stage of welding, the welding current I1 (kA) is preferably set so as to satisfy the following equation (3) in relation to the thickness tm of the thinnest metal plate among the plurality of metal plates. Figure 6 The middle corresponds to the plate thickness (mm) of the upper steel plate 6.

[0109] 3tm+5≤I1………(3)

[0110] When the welding current I1 in the first stage of welding is (3tm+5) or less, heat generation by contact resistance cannot be effectively utilized, and a nugget is not formed between the upper steel plate 6 and the middle steel plate 7 .

[0111] Therefore, in the first stage of welding, in order to form a nugget of desired size between the upper steel plate 6 and the middle steel plate 7, it is preferable to set the pressure P1, welding current I1, or power-on time T1 so as to satisfy the above-mentioned equations (1), (2), and (3).

[0112] In the second stage welding, the pressure P2 (kN) is preferably set so as to satisfy the following formula (4) in relation to the pressure P1 in the first stage welding.

[0113] P1≤P2≤10P1………(4)

[0114] Furthermore, it is preferable that the welding current I2 (kA) in the second stage is set so as to satisfy the following formula (5) in relation to the welding current I1 in the first stage.

[0115] 0.5≤I2≤I1………(5)

[0116] Furthermore, the second-stage energization time T2 (ms) is preferably set so as to satisfy the following formula (6) in relation to the first-stage energization time T1 of welding.

[0117] T1≤T2≤10T1………(6)

[0118] If the second-stage welding conditions deviate from the above ranges, it will be difficult to prevent spattering or achieve the specified nugget diameter. Furthermore, excessively increasing the pressure can lead to larger heat marks or lift. When performing two-stage welding, the specified time (Ts) in the electrical information performance data during resistance spot welding is preferably set to the time at which the first stage ends and the second stage begins.

[0119] When selecting welding conditions for performing welding with two-stage energization as described above, there are many combinations of conditions. Therefore, by applying the prediction system 100 and prediction method according to this embodiment to selecting welding conditions for performing welding with two-stage energization, the efficiency of the welding condition selection operation is improved.

[0120] Furthermore, the calorific value during the first stage of resistance spot welding can be obtained as electrical information during resistance spot welding. Using the calorific value during the first stage of resistance spot welding as electrical information during resistance spot welding improves the accuracy of prediction of weld characteristics.

[0121] The prediction system 100 and prediction method according to this embodiment are applicable not only to selecting welding conditions for the welding methods described above, but also to selecting welding conditions for various other welding methods. For example, the prediction system 100 and prediction method according to this embodiment can be applied to selecting welding conditions for methods such as single-stage energization that maintains a constant current value and pressure, multi-stage energization that changes the current value or pressure setting midway, or adaptive control that controls the current value based on a reference heat amount. Furthermore, the prediction system 100 and prediction method according to this embodiment can also be applied to selecting welding conditions for methods that include post-energization after the main energization to form the weld.

[0122] <Flowchart Example>

[0123] In the prediction system 100 according to this embodiment, the Figure 4 The prediction model 22 is generated according to the illustrated sequential process.

[0124] The welding device 30 performs a welding test on a plate group for determining welding conditions for resistance spot welding (step S1). The welding conditions for the welding test can be set as appropriate.

[0125] The data acquisition unit 12 of the prediction model generation device 10 acquires electrical information from resistance spot welding performed by the welding device 30 during the welding test (step S2). Furthermore, the data acquisition unit 12 of the prediction model generation device 10 acquires analysis results of weld characteristics for samples of welded joints obtained during the welding test (step S3). The analysis results of weld characteristics are obtained by offline analysis of the samples of welded joints obtained during the welding test using the analysis device 40. The data acquisition unit 12 stores data in the database 14, in which the electrical information from the resistance spot welding and the analysis results of the weld characteristics are associated.

[0126] The model generation unit 16 of the prediction model generation device 10 performs learning using the data stored in the database 14 as learning data, and generates a prediction model 22 of the weld characteristics (step S4). The prediction system 100 ends after the procedure of step S4 is completed. Figure 4 Execution of the process.

[0127] In the prediction system 100 according to this embodiment, the following steps can be performed to determine welding conditions through welding tests: Figure 5 The flow of the sequence is illustrated.

[0128] The test conditions for performing a welding test on the plate set for determining the welding conditions of the resistance spot welding are set (step S11). The test conditions may be set by a person who performs the welding test or the like.

[0129] Prediction device 20 obtains electrical information from welding device 30 regarding resistance spot welding performed by welding device 30 during the welding test (step S12). Based on the electrical information regarding resistance spot welding, prediction device 20 predicts the weld characteristics of the welded joint obtained from the welding test (step S13). Prediction device 20 can input the electrical information regarding resistance spot welding into prediction model 22, thereby obtaining a prediction result of the weld characteristics outputted by prediction model 22.

[0130] A determination is then made as to whether the weld characteristics prediction result, performed by the prediction device 20, is satisfactory (step S14). This determination of the weld characteristics prediction result can be made by the person conducting the welding test, for example, or by the prediction device 20 or an external device that has obtained the prediction result from the prediction device 20. If the weld characteristics prediction result is not satisfactory (step S14: No), the procedure returns to step S11, and the test is repeated.

[0131] If the predicted weld characteristics are favorable (step S14: Yes), the welded joint obtained through the test is actually analyzed to determine whether the analysis results are favorable (step S15). The determination of the weld characteristics analysis results can be performed by the person conducting the welding test, or by a device such as the analysis device 40. If the weld characteristics analysis results are unfavorable (step S15: No), the process returns to the test condition setting sequence of step S11 and the test is repeated.

[0132] If the analysis result of the weld characteristics is good (step S15: Yes), the test conditions when the good weld characteristics were obtained are determined as welding conditions (step S16). The prediction system 100 ends after the procedure of step S16 is completed. Figure 5 Execution of the process.

[0133] (summary)

[0134] As described above, the prediction system 100 according to the present invention predicts weld properties of welded joints based on actual electrical information from resistance spot welding during welding tests on a set of plates for determining welding conditions, without actually analyzing the welded joints obtained through the welding tests. The ability to predict weld properties without actual analysis makes it easier to increase the number of welding tests with varying welding conditions. As a result, the efficiency of selecting welding conditions suitable for achieving a weld with a predetermined nugget diameter or strength when welding the set of plates is improved.

[0135] <Example>

[0136] The following describes a specific embodiment. In this embodiment, a single-phase AC resistance spot welder with a servo motor pressurization method is used as the welding device 30. The spot welder has a DR type electrode made of chromium copper with a tip diameter of 6 mm and a curvature radius of 40 mm. In this embodiment, the determination Figure 6 The illustrated welding conditions for a plate set consisting of three stacked steel plates. The plate set for which the welding conditions are determined includes an upper steel plate 6, a middle steel plate 7, and a lower steel plate 8. The upper steel plate 6 is also referred to as the first metal plate. The middle steel plate 7 is also referred to as the second metal plate. The lower steel plate 8 is also referred to as the third metal plate. In this embodiment, the upper steel plate 6 and the middle steel plate 7 are in contact with each other, and the middle steel plate 7 and the lower steel plate 8 are in contact with each other. The upper steel plate 6 uses a steel plate with a tensile strength of 270 MPa and a thickness of 0.7 mm. The middle steel plate 7 and the lower steel plate 8 use steel plates with a tensile strength of 1470 MPa and a thickness of 1.6 mm. In other words, in this embodiment, the plate set for which the welding conditions are determined is configured such that the first metal plate is thinner than the second and third metal plates, and the second and third metal plates have the same thickness.

[0137] The welding test of the plate group was carried out by the welding device 30. Figure 7 As illustrated, the welding conditions for the welding test in this embodiment are set to two-stage energization with different welding currents flowing between the electrodes during the first stage represented by time T1 and the second stage represented by time T2. Figure 7 The horizontal axis represents the current-carrying time of resistance spot welding. The vertical axis represents the current flowing between the electrodes. The welding current flowing between the electrodes during the first stage is represented by I1. The welding current flowing between the electrodes during the second stage is represented by I2.

[0138] like Figure 8A As shown in the example, by the conduction of electricity during the first stage, the contact surface between the upper steel plate 6 and the middle steel plate 7, which are thinner than the middle steel plate 7 and the lower steel plate 8, rises in temperature earlier than the contact surface between the middle steel plate 7 and the lower steel plate 8, causing the steel plates to melt, thereby forming the first weld 51. Figure 8B As shown in the example, by the conduction of electricity during the second stage, the temperature of the contact surface between the middle steel plate 7 and the lower steel plate 8 also rises and the steel plates melt, thereby forming the second weld 52. As a result, Figure 6 As shown in the example, a weld 5 is formed by joining three steel plates. Figure 6As shown, the nugget diameter of the weld 5 is measured separately as the nugget diameter d1 at the contact surface between the upper steel plate 6 and the middle steel plate 7, and the nugget diameter d2 at the contact surface between the middle steel plate 7 and the lower steel plate 8. The nugget diameters d1 and d2 are the diameters of the weld 5 when viewed from above from the individual steel plates comprising the plate pack. When the weld 5 is formed into a disk shape with the normal to the individual steel plates comprising the plate pack as the rotation axis, the nugget diameters d1 and d2 are measured on a cross-section of the plate pack cut so as to pass through the rotation axis.

[0139] Specifically, during the first stage, the welding conditions were set to an electrode force of 3 to 5 kN (kilonewtons), a welding current I1 of 5 to 13 kA (kiloamperes), and a current-on time T1 of 60 ms (milliseconds). Furthermore, during the second stage, the welding conditions were set to an electrode force of 3 to 5 kN, a welding current I2 of 5 to 10 kA, and a current-on time T2 of 360 ms.

[0140] By varying the electrode pressure and welding current within the aforementioned ranges, 180 combinations of welding conditions were set. Resistance spot welding was performed on the plate assembly for which welding conditions were determined in this example, applying each of the 180 welding conditions to produce 180 welded joints.

[0141] In this example, the nugget diameter was analyzed as a characteristic of the welded portion of 180 welded joints. Specifically, the nugget diameter was measured by cutting the welded portion of the welded joint obtained by welding, etching the cross section, and observing it with an optical microscope. Figure 6 The nugget diameter d1 between the upper steel plate 6 and the middle steel plate 7 and the nugget diameter d2 between the middle steel plate 7 and the lower steel plate 8 are exemplified.

[0142] The data acquisition unit 12 of the prediction model generation device 10 acquires performance data on electrical information during resistance spot welding for each of 180 welding conditions, and performance data on nugget diameters for 180 weld joints. The data acquisition unit 12 associates the performance data on electrical information during resistance spot welding for each welding condition with the performance data on nugget diameters, and stores them in a database 14 of the prediction model generation device 10.

[0143] The model generation unit 16 of the prediction model generation device 10 obtains data that associates the performance data of electrical information in resistance spot welding with the performance data of the nugget diameter from the database 14. The model generation unit 16 performs machine learning using the obtained data as learning data, and generates a prediction model 22 that outputs a prediction result of the nugget diameter when the performance data of electrical information in resistance spot welding is input. As a machine learning model, ridge regression, which is one of the regression models, is used. As described above, it is found that the total calorific value Qt in the performance data of electrical information in resistance spot welding has a significant impact on the characteristics of the weld. For example, Figure 9 As shown, the total heat generation Qt is related to the weld nugget diameters d1 and d2. Figure 9 The horizontal axis of the graph represents the total calorific value Qt. The vertical axis represents the nugget diameter. Based on the correlation between total calorific value Qt and nugget diameter, in this example, total calorific value Qt was selected as the performance data for electrical information during resistance spot welding. Qm was also selected. Furthermore, Rfirst, Rfm, Rmax, Rml, Rlast, and Rmean, among the inter-electrode resistance values, were also selected as performance data for electrical information during resistance spot welding.

[0144] The prediction device 20 can use the prediction model 22 generated by the prediction model generation device 10 as described above to predict the nugget diameter of the weld portion of a weld joint obtained from a welding test performed under new welding conditions, without observing the cross section of the weld joint. In this embodiment, 50 new welding conditions were set and 50 weld joints were produced. The results of the prediction of the nugget diameter of the weld portion of the 50 weld joints using the prediction model 22 were compared with the results of the actual measurement of the nugget diameter of the weld portion of the 50 weld joints. The nugget diameter d1 between the upper steel plate 6 and the middle steel plate 7 was predicted with an error of ±9%. In addition, the nugget diameter d2 between the middle steel plate 7 and the lower steel plate 8 was predicted with an error of ±2%.

[0145] As described above, it has been found that the initial calorific value Qi in the performance data of electrical information in resistance spot welding has a significant impact on the characteristics of the weld. Figure 10 As shown, the initial heat generation Qi is related to the nugget diameter d1. Figure 10The horizontal axis of the graph represents the initial calorific value Qi. The vertical axis represents the nugget diameter. Based on the correlation between the initial calorific value Qi and the nugget diameter, as another embodiment, a prediction model 22 was generated by performing learning using the initial calorific value Qi as actual performance data for electrical information during resistance spot welding and further adding it to the learning data. The initial calorific value Qi is the cumulative calorific value from the start of power application until 60 ms (milliseconds). Using the prediction model 22 generated with the initial calorific value Qi added, the nugget diameter d1 between the upper steel plate 6 and the middle steel plate 7 was predicted with an error of ±7%. Furthermore, the nugget diameter d2 between the middle steel plate 7 and the lower steel plate 8 was predicted with an error of ±2%. By using the initial calorific value Qi as the actual performance data input to the prediction model 22, the prediction accuracy of the nugget diameter d1 between the upper steel plate 6 and the middle steel plate 7 was improved. In other words, by using the initial calorific value Qi as the actual performance data input to the prediction model 22, the prediction accuracy of the nugget diameter at the contact surface between thin and thick steel plates was improved.

[0146] As described as an embodiment, the prediction system 100 according to the present invention accurately predicts weld characteristics such as nugget diameter based on actual performance data of electrical information during resistance spot welding. This highly accurate prediction of the nugget diameter eliminates the need to cut a cross section of a weld joint produced during a welding test to measure the nugget diameter. This reduces the time and cost associated with this process. Consequently, the number of welding conditions that can be set during a welding test is reduced, allowing for more efficient selection of welding conditions.

[0147] By executing the prediction system 100 and prediction method of the present invention on a stack of two or more metal plates, the efficiency of selecting welding conditions for the stack of two or more metal plates can be improved. Furthermore, executing the prediction system 100 and prediction method of the present invention on a stack of three or more metal plates can also improve the efficiency of selecting welding conditions for the stack of three or more metal plates. Selecting welding conditions for a stack of three or more metal plates is more difficult than selecting welding conditions for a stack of two metal plates. Therefore, executing the prediction system 100 and prediction method of the present invention on a stack of three or more metal plates further enhances the efficiency of selecting welding conditions.

[0148] In the above-mentioned embodiment, the thickness of the upper steel plate 6 is 0.7 mm, and the thickness of the middle steel plate 7 is 1.6 mm. That is, the value obtained by dividing the total thickness of the plate group by the thickness of the thinnest metal plate in the plate group, that is, the thickness ratio, is about 5.6. It is known that the larger the thickness ratio of the plate group to be resistance spot welded, the more difficult it is to form a weld with the required nugget diameter on the contact surface between the thin steel plate and the thick steel plate. In particular, in resistance spot welding of plate groups with a thickness ratio exceeding 5, it is difficult to determine the welding conditions for forming a weld with the required nugget diameter. The prediction system 100 involved in the present invention can also predict the weld characteristics of plate groups formed by overlapping steel plates with a combination with a further increased thickness ratio, or especially a combination with a thickness ratio exceeding 5 with high accuracy, and can improve the efficiency of the operation of determining welding conditions.

[0149] The embodiments of the present invention are described based on the accompanying drawings and embodiments, but it should be noted that those skilled in the art can make various deformations or changes based on the present invention. Therefore, it should be noted that these deformations or changes are included in the scope of the present invention. For example, the functions included in each structural part or each step can be reconfigured in a logically non-contradictory manner, and multiple structural parts or steps can be combined into one or divided. The embodiments of the present invention can also be implemented as a program executed by a processor possessed by the device or a storage medium having a program recorded thereon. It should be understood that these are also included in the scope of the present invention.

[0150] Description of Reference Numerals

[0151] 100…prediction system; 1, 2…metal plates; 3, 4…electrodes; 5…welding portion (51…first welding portion, 52…second welding portion); 6…upper steel plate; 7…middle steel plate; 8…lower steel plate; 10…prediction model generating device (12…data acquisition portion, 14…database, 16…model generating portion); 20…prediction device (22…prediction model); 30…welding device; 40…analysis device.

Claims

1. A method for predicting properties of a weld portion of a stack of two or more metal plates, the method comprising: The following steps are included: a step of obtaining electrical information output during resistance spot welding, the electrical information including at least one of a current flowing between the two electrodes during resistance spot welding, a voltage between the two electrodes, a resistance between the two electrodes, or a heat generation between the two electrodes; and A step of predicting characteristics of a welded portion of a plate assembly formed by stacking the two or more metal plates based on the electrical information.

2. The weld property prediction method according to claim 1, wherein: In the step of predicting the characteristics of the weld, the electrical information is input into a prediction model for the characteristics of the weld, and a prediction result of the characteristics of the weld outputted from the prediction model is acquired.

3. The weld property prediction method according to claim 2, wherein: The method further includes the step of performing learning using the electrical information and analysis results of the characteristics of the welded portion as learning data to generate the prediction model.

4. The method for predicting weld characteristics according to any one of claims 1 to 3, wherein: In the step of obtaining the predicted results of the characteristics of the weld, a predicted result of at least one of the nugget diameter, tensile shear strength, cross tensile strength, or fracture strength based on a mechanical peel test of the weld is obtained as the characteristic of the weld.

5. The method for predicting weld characteristics according to any one of claims 1 to 4, wherein: In the step of acquiring the electrical information, a total calorific value from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

6. The method for predicting weld characteristics according to claim 5, wherein: In the step of acquiring the electrical information, resistance between the two electrodes when energization is completed is acquired as the electrical information.

7. The weld property prediction method according to claim 5 or 6, wherein: In the step of acquiring the electrical information, a heat value during a portion of a period from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

8. The weld property prediction method according to any one of claims 1 to 4, wherein: The resistance spot welding is performed on a plate group formed by overlapping three or more metal plates including at least a first metal plate, a second metal plate, and a third metal plate. The thickness ratio, which is a value obtained by dividing the total thickness of the plate group by the thickness of the thinnest metal plate in the plate group, is 5 or more.

9. The method for predicting weld characteristics according to claim 8, wherein: In the step of acquiring the electrical information, a total calorific value from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

10. The weld property prediction method according to claim 9, wherein: In the step of acquiring the electrical information, resistance between the two electrodes when energization is completed is acquired as the electrical information.

11. The method for predicting weld characteristics according to claim 9 or 10, wherein: The resistance spot welding includes a first stage in which a first current flows and a second stage in which a second current flows. In the step of acquiring the electrical information, a heat value during the first stage of the period from the start of energization to the end of energization of the resistance spot welding is acquired as the electrical information.

12. A method for generating a prediction model for predicting characteristics of a weld portion of a stack of two or more metal plates obtained by resistance spot welding in which a current flows between two electrodes sandwiching the stack of two or more metal plates, wherein: The prediction model generation method comprises the following steps: a step of obtaining electrical information output during resistance spot welding, the electrical information including at least one of a current flowing between the two electrodes during resistance spot welding, a voltage between the two electrodes, a resistance between the two electrodes, or a heat generation between the two electrodes; a step of obtaining analysis results of characteristics of the welded portion when resistance spot welding is performed on a plate group formed by overlapping the two or more metal plates; as well as A step of performing machine learning based on learning data in which the electrical information is associated with an analysis result of the characteristics of the weld portion when resistance spot welding is performed on a plate group formed by overlapping the two or more metal plates, thereby generating a model for predicting the characteristics of the weld portion of the plate group formed by overlapping the two or more metal plates.

13. A prediction model, wherein: The prediction model is generated by the prediction model generation method according to claim 12, and predicts the characteristics of a weld obtained by resistance spot welding.

Images