Welding support method, welding support system and program

Abstract

To provide a welding support method for supporting stabilization of a welding work.SOLUTION: A welding support method includes the steps of: acquiring a value of a welding parameter during welding for a welding parameter indicating a state of welding; evaluating the degree of abnormality of the welding parameter on the basis of an evaluation model for evaluating the degree of abnormality of the welding parameter, and a value of the welding parameter; specifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or more than a threshold, and calculating an improvement direction of the value of the parameter; and displaying information supporting welding relating to the welding parameter together with the evaluation result of the degree of abnormality.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a welding support method, a welding support system, and a program. [Background technology]

[0002] At welding work sites, there is a demand for stabilizing welding quality, reducing rework, and other work efficiency improvements. Patent Document 1 discloses a welding system that can set optimal welding conditions for an arc welding robot without requiring a skilled worker or specialized knowledge. The welding system of Patent Document 1 performs welding using a welding robot while adjusting the welding conditions and taking pictures of the welding part with a camera, and is equipped with a machine learning device that determines rewards for the results in terms of welding quality, amount of spatter generation, welding wire consumption, and the like, and acquires optimal welding conditions through reinforcement learning. The method of determining welding conditions described in Patent Document 1 is difficult to apply when it is difficult to take pictures of the welding part. In addition, since it is intended for welding robots, it is difficult to apply it to welding work by welders, and there is a possibility that it cannot respond to disturbances and fluctuations that occur during welding work, for example. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 6126174 Summary of the Invention [Problem to be solved by the invention]

[0004] There is a demand for a method to stabilize welding quality and reduce rework by supporting the welder during welding work.

[0005] The present disclosure provides a welding assistance method, a welding assistance system, and a program that can solve the above problems. [Means for solving the problem]

[0006] The welding assistance method disclosed herein includes the steps of acquiring values ​​of welding parameters during welding that indicate a welding state, evaluating the degree of abnormality of the welding parameters based on an evaluation model that evaluates the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters, identifying a parameter that should be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction of improvement for the value of the parameter, and displaying information to assist welding regarding the welding parameters together with the evaluation result of the degree of abnormality. The evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, and in the evaluating step, a density ratio between a probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal value is calculated as a degree of abnormality, and in the calculating step, a direction in which an average value of the probability density distribution of the evaluation target value approaches an average value of the probability density distribution of the normal value is calculated as the improvement direction. Alternatively, the evaluation model is a model that reproduces the normal value constructed so as to output the normal value when a normal value of each parameter included in the welding parameters is input, and in the evaluating step, a reproducibility indicating to what extent a value output by the evaluation model when a value of each parameter included in the welding parameters is input reproduces the input parameter value is calculated as a degree of abnormality, and in the calculating step, a direction in which the reproducibility is improved is calculated as the improvement direction.

[0007] The welding support system of the present disclosure includes a means for acquiring values ​​of welding parameters during welding that indicate a welding state, a means for evaluating the degree of abnormality of the welding parameters based on an evaluation model for evaluating the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters, a means for identifying a parameter that should be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value and calculating a direction of improvement for the value of the parameter, and a means for displaying information to assist welding regarding the welding parameters together with the evaluation result of the degree of abnormality. The evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, the evaluating means calculates a density ratio between the probability density distribution of an evaluation target value of each parameter included in the welding parameters and the probability density distribution of the normal values ​​as the degree of abnormality, and the calculating means calculates a direction in which an average value of the probability density distribution of the evaluation target value approaches an average value of the probability density distribution of the normal values ​​as the improvement direction. Alternatively, the evaluation model is a model that reproduces the normal values ​​constructed to output the normal values ​​when normal values ​​of each parameter included in the welding parameters are input, and the evaluating means calculates a reproducibility indicating to what extent a value output by the evaluation model when a value of each parameter included in the welding parameters is input reproduces the input parameter value as the degree of abnormality, and the calculating means calculates a direction in which the reproducibility is improved as the improvement direction.

[0008] The program disclosed herein causes a computer to execute the steps of acquiring values ​​of welding parameters during welding that indicate a welding state, evaluating the degree of abnormality of the welding parameters based on an evaluation model that evaluates the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters, identifying a parameter that should be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction in which to improve the value of the parameter, and displaying information to assist welding regarding the welding parameters together with the evaluation result of the degree of abnormality. The evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, and in the evaluating step, a density ratio between a probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal value is calculated as a degree of abnormality, and in the calculating step, a direction in which an average value of the probability density distribution of the evaluation target value approaches an average value of the probability density distribution of the normal value is calculated as the improvement direction. Alternatively, the evaluation model is a model that reproduces the normal value constructed so as to output the normal value when a normal value of each parameter included in the welding parameters is input, and in the evaluating step, a reproducibility indicating to what extent a value output by the evaluation model when a value of each parameter included in the welding parameters is input reproduces the input parameter value is calculated as a degree of abnormality, and in the calculating step, a direction in which the reproducibility is improved is calculated as the improvement direction. Effect of the Invention

[0009] According to the above-described welding support method, welding support system, and program, it is possible to stabilize welding quality and reduce rework. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of a welding support system according to an embodiment. [Figure 2A] 1 is a first diagram for explaining a process for analyzing a method for determining a welding state and a method for adjusting parameters for stabilizing the welding state according to an embodiment. FIG. [Figure 2B] FIG. 2 is a second diagram for explaining the process of analyzing the method for determining the welding state and the method for adjusting parameters for stabilizing the welding state according to the embodiment. [Figure 2C] FIG. 11 is a third diagram for explaining the process of analyzing the method for determining the welding state and the method for adjusting parameters for stabilizing the welding state according to the embodiment. [Diagram 3] FIG. 4 is a diagram showing an example of a transition of a welding parameter and an abnormality degree according to the embodiment. [Figure 4] FIG. 4 is a diagram showing an example of a display screen displayed on the display device according to the embodiment. [Diagram 5] 4 is a flowchart illustrating an example of an operation of the welding support system according to the embodiment. [Figure 6] FIG. 10 is a diagram showing an example of work result information of a welder according to the embodiment. [Figure 7] FIG. 1 is a diagram illustrating an example of a hardware configuration of a welding support system according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the welding assistance method of the present disclosure will be described with reference to FIGS. <Embodiment> (composition) FIG. 1 is a diagram illustrating an example of a welding support system according to an embodiment. The welding support system 100 includes a welding machine 10, a data logger 20, a welding support device 30, a display device 40, and a model construction device 50. The welding machine 10 includes an operation panel provided with levers, switches, and other instruments for adjusting welding conditions such as the welding current. A welder 60 performs welding work on an object to be welded 12 using a welding rod 11 connected to the welding machine 10 while adjusting the welding conditions using the operation panel of the welding machine 10. The data logger 20 is connected to the welding machine 10 and a sensor (not shown) installed near the welding site, and acquires measurement values ​​measured by the sensor equipped in the welding machine 10 and the sensor near the welding site in real time during welding. The measurement values ​​acquired by the data logger 20 and the feature values ​​obtained based on the measurement values ​​(for example, the product of the welding current and the welding voltage) are called welding parameters. The welding parameters are information necessary for determining the state of welding. The welding parameters include a plurality of parameters required for determining the quality of welding, such as the welding voltage, the welding current, the welding speed, and the product of the welding current and the welding voltage. The data logger 20 is connected to the welding support device 30 so as to be able to communicate with it. The measured values ​​of the welding parameters acquired by the data logger 20 are transmitted to the welding support device 30 in real time. When the welding support device 30 acquires the welding parameters, it evaluates the welding parameters using the evaluation model M and judges whether the welding state is stable or not. The welding state refers to the stability and normality of the welding work and welding quality estimated from the values ​​of the welding parameters. When the welding state becomes unstable, the welding support device 30 identifies the parameter that caused the welding state to become unstable from among the welding parameters and analyzes a method for improving the identified parameter. The welding support device 30 is connected to the display device 40 and outputs information such as the judgment result of the welding state, the latest parameter value, and a method for improving the parameter value to the display device 40. The welder 60 performs the welding work while referring to the information displayed on the display device 40. For example, when the welding state is stable, the current welding parameter value is confirmed, and when the welding state becomes unstable, the welding condition is adjusted by operating the operation panel of the welding machine 10 according to the method for improving the parameter value displayed on the display device 40. This enables high-quality and stable welding work.Furthermore, model construction device 50 acquires welding parameters stored in welding support device 30 and the like, and constructs an evaluation model M used to evaluate the welding parameters. Evaluation model M constructed by model construction device 50 is output to welding support device 30, and welding support device 30 evaluates the welding parameters based on this evaluation model M.

[0012] The welding support device 30 includes a data acquisition unit 31, a setting reception unit 32, a control unit 33, an output unit 34, and a memory unit 35. Data acquiring unit 31 acquires measurement values ​​of welding parameters from data logger 20 from time to time while welder 60 is performing welding work. The measurement values ​​of welding parameters include values ​​of multiple welding-related parameters and information on the time when those values ​​were measured. When the welding parameters include feature values ​​calculated from the measurement values, data acquiring unit 31 calculates the feature values ​​using the measurement values ​​acquired from data logger 20. The setting receiving unit 32 receives settings such as a threshold value for determining whether the welding state is stable or not, and an upper limit value for the number of parameters presented as parameters to be improved when the welding state becomes unstable.

[0013] The control unit 33 evaluates the welding parameters to determine the welding state, and when the welding state is determined to be unstable, controls the process of analyzing which parameters should be improved and how. The control unit 33 includes a determination unit 331 and an analysis unit 332.

[0014] The determination unit 331 determines whether the welding state is stable based on the welding parameters and the evaluation model M. The stable welding state means that the welding work is performed in a state where each parameter constituting the welding parameters is generally within a normal or ideal range, and the quality of the welding performed in such a state is good. For example, the determination unit 331 calculates the distance between the evaluation model M, which is constructed by learning the measured values ​​of the welding parameters measured when the welding state is stable and the feature amounts obtained based on the measured values, and the welding parameters to be evaluated, and if this distance is less than a threshold value, it determines that the welding state is stable, and if this distance is equal to or greater than the threshold value, it determines that the welding state is unstable. This distance is called the abnormality degree.

[0015] When the analysis unit 332 determines that the welding condition is unstable, it selects a predetermined number of parameters that have a large influence from among the parameters that caused the determination that the welding condition is unstable, and analyzes how to stabilize the welding condition by changing the values ​​of those parameters (for example, by increasing or decreasing them).

[0016] Here, an example of the processing of the determination unit 331 and the analysis unit 332 will be described in detail with reference to FIG. 2A. Point x shown in the upper part of FIG. 2A1 indicates the welding parameter to be evaluated, and point k indicates the normal value and ideal value of the welding parameter. For example, when the welding parameter is composed of three parameters, the welding parameter can be expressed as one point in a three-dimensional space as shown in FIG. 2A1. In reality, the welding parameter includes more types of parameters, but here, the welding parameter is expressed as a point in a three-dimensional space as shown in FIG. 2A1, assuming that the welding parameter is composed of three parameters. In addition, the three parameters are respectively defined as parameters 1, 2, and 3, and the parameters 1 to 3 of the welding parameter corresponding to point x are respectively represented by x1, x2, and x3, and the parameters 1 to 3 of the welding parameter corresponding to point k are respectively represented by k1, k2, and k3. The evaluation model M is an evaluation model that determines the range of normal values ​​for each of the parameters 1 to 3, which is created by learning the welding parameters (normal data) measured when the welding state is stable. Point k indicates a point closest to point x (it may be the closest point or the nth closest point) among the points included in evaluation model M. When judging the welding condition based on the welding parameter (point x) to be evaluated, judgment unit 331 identifies nearby point k included in judgment model M and calculates the distance (degree of abnormality) between point x and point k. The method of calculating the distance is shown in the "Distance (degree of abnormality)" row in the lower part of Fig. 2A2 in Fig. 2.

[0017] The determination unit 331 calculates the distance D(x) between the evaluation target data x and the neighboring point k by the formula (1a) in FIG. 2A2. As shown in the formula (1a), the distance D(x) is a single value obtained by adding up the differences between the parameters. If the distance D(x) is equal to or greater than a threshold, the determination unit 331 determines that the welding state is unstable, and if the distance D(x) is less than the threshold, the determination unit 331 determines that the welding state is stable. The determination unit 331 may also calculate the distance D(x) by weighting each of the parameters 1 to 3. For example, the abnormality determination unit 132 multiplies each of the parameters 1 to 3 by the weighting coefficients A1 to A3, and calculates the distance D(x) by the formula (1b) in FIG. 4. If it is known that the parameter 1 has a large effect on the welding state and the parameter 1 is to be emphasized, the user sets a large value for A1. The distance may be calculated by calculating the Mahalanobis distance or the Manhattan distance in addition to the Euclidean distance.

[0018] When the distance D(x) is equal to or greater than the threshold and the welding condition is determined to be unstable, the analysis unit 332 identifies the parameter that contributes most to the distance D(x) among the parameters 1 to 3 that constitute the points x and k, compares the identified parameters with each other, and analyzes how the values ​​of the identified parameters at the point x should be changed to stabilize the welding condition.

[0019] Specifically, the analysis unit 332 calculates (k1-x1) as shown in the "Factor" row of FIG. 2 , (k2-x2) 2 , (k3-x3) 2 In other words, the analysis unit 332 identifies parameters that contribute to and have a large influence on the distance D(x) as factors that destabilize the welding state. For example, if there is one parameter to be selected, (k1-x1) 2is the largest, the analysis unit 332 identifies parameter 1 as the destabilizing factor. The analysis unit 332 may identify the parameter causing the destabilizing factor by weighting each of parameters 1 to 3. For example, if parameter 1 is a dominant factor for stabilizing the welding state, the user can set a large value for A1 so that parameter 1 is more likely to be selected as the destabilizing factor.

[0020] Furthermore, as shown in the "Direction" row in Fig. 2A2, the analysis unit 332 calculates the direction of improvement of the parameter value for the parameter identified as the destabilizing factor by a formula for calculating the difference before squaring. For example, when parameter 1 is identified as the destabilizing factor, the analysis unit 332 calculates (k1-x1) and calculates the direction of bringing (k1-x1) closer to 0. Specifically, if (k1-x1) is a positive value, the analysis unit 332 evaluates that the direction of increasing the value of parameter 1 (x1) is the direction of improving (stabilizing) the welding condition, and if (k1-x1) is a negative value, the analysis unit 332 evaluates that the direction of decreasing the value of parameter 1 (x1) is the direction of improving the welding condition.

[0021] The evaluation model M and the calculation process of the degree of abnormality and the direction of improvement of the parameter value described in FIG. 2A are only examples, and are not limited thereto. As examples of other evaluation methods, FIG. 2B shows an example in which a probability density distribution is used as an evaluation model, and FIG. 2C shows an example of an evaluation model that performs abnormality judgment based on the reproducibility. The upper left diagram of FIG. 2B shows the probability density distribution of parameter 1 among multiple parameters included in normal data. The probability density distribution calculated from the normal data exemplified in the upper left diagram is the evaluation model. The lower left diagram of FIG. 2B shows the probability density distribution of parameter 1 among the welding parameters to be evaluated. The vertical axis of each graph shows the probability density, and the horizontal axis shows the value of parameter 1. The judgment unit 331 calculates the probability density distribution using a data set of welding parameters measured during a predetermined period (for example, t1 to t2, which are the period immediately before welding), and calculates a density ratio with respect to the probability density distribution of normal data (=evaluation model) calculated in advance. More specifically, the determination unit 331 calculates the density ratio (ratio of probability density functions representing the curve of the probability density distribution) of the probability density distribution of normal data and the data to be evaluated in the range of v1 to v2 in which the value of the parameter 1 to be evaluated exists as the degree of abnormality, and determines that the welding state is unstable when the density ratio deviates from the allowable range centered on "1". An example of the calculated density ratio is shown in the right diagram of FIG. 2B. In the example of the diagram, since the density ratio deviates from "1", the parameter 1 is determined to be abnormal. The determination unit 331 calculates the density ratio (degree of abnormality) for each of the multiple parameters to be evaluated, and determines that the welding state indicated by the parameters to be evaluated measured during a predetermined period is unstable if there is at least one parameter whose density ratio deviates from the allowable range centered on "1". Alternatively, the determination unit 331 may be configured to determine that the welding state is unstable when there are a predetermined number or more parameters whose density ratio deviates from the allowable range.

[0022] Furthermore, the analysis unit 332 may perform logarithmic transformation of the density ratios (degrees of abnormality) calculated for each parameter, and identify a predetermined number of parameters as destabilizing factors in ascending order of the density ratio, or may identify a predetermined number of parameters as destabilizing factors in descending order of the variance of the probability density distribution, or may identify a predetermined number of parameters as destabilizing factors in descending order of the area of ​​the range deviating from the allowable range centered on the density ratio "1", or may calculate the distance L between the average value of the probability density distribution of normal data and the average value of the probability density distribution of the parameter to be evaluated, and identify a predetermined number of parameters as destabilizing factors in descending order of the distance L. Furthermore, the analysis unit 332 may identify destabilizing factors by combining these. For example, the analysis unit 332 may identify a parameter identified as a destabilizing factor by a plurality of methods as a final destabilizing factor, or may identify all of the parameters identified as a destabilizing factor by any of a plurality of methods as a final destabilizing factor. In addition, for each parameter identified as an instability factor, the analysis unit 332 evaluates that the direction of the value with the highest probability density of normal data (average value of the probability density distribution) (the direction of the thick arrow in the lower left diagram of Figure 2B) is the direction for improving the welding condition.

[0023] Next, a process of judging the welding state by an evaluation model based on the reproducibility will be described with reference to FIG. 2C. (Step 1) The model construction device 50 constructs an evaluation model (for example, an autoencoder) that learns to output any normal data inputted by using normal data including parameters 1 to 3 as learning data and reproduces the normal data. (Step 2) The judgment unit 331 inputs parameters 1 to 3 (Q1 to Q3) of the evaluation target data to the constructed evaluation model. If the values ​​of parameters 1 to 3 outputted by the evaluation model are Q1' to Q3', the abnormality judgment unit 132 calculates |Q1-Q1'|, |Q2-Q2'|, and |Q3-Q3'|, and regards these values ​​as the reproducibility (abnormality). The smaller the reproducibility (abnormality), the better, and the judgment unit 331 judges the evaluation target data as abnormal if the reproducibility of at least one parameter exceeds a threshold value. Alternatively, the judgment unit 331 may be configured to judge the evaluation target data as abnormal if there are a predetermined number or more parameters whose reproducibility exceeds a threshold value. Furthermore, the analysis unit 332 identifies a predetermined number of destabilizing factors in descending order of reproducibility (degree of abnormality). For example, when |Q1-Q1'|=0, |Q2-Q2'|=2, and |Q3-Q3'|=5 and only one destabilizing factor is identified, the analysis unit 332 identifies parameter 3 as the destabilizing factor. Furthermore, the analysis unit 332 evaluates the direction in which reproducibility (degree of abnormality)=0 as a direction in which the welding condition is improved. For example, the analysis unit 332 evaluates the direction in which |Q3-Q3'|=5 becomes |Q3-Q3'|=0 for parameter 3 as a direction in which the welding condition is improved.

[0024] Alternatively, the judgment model M may be constructed by learning the correlations between a plurality of parameters when the welding state is stable. In this case, for example, if the difference between the correlation coefficients between the parameters constituting the welding parameters to be evaluated and the correlation coefficients between the same parameters determined in the judgment model M is equal to or greater than a threshold value, the welding state is judged to be unstable, and these parameters are identified as factors that cause the welding state to be unstable. Also, the direction in which the difference between the correlation coefficients between the parameters to be evaluated and the correlation coefficients between the parameters of the judgment model M approaches 0 is calculated as the direction in which the welding state is improved.

[0025] In addition, when welding parameters for various states, such as when the welding state is stable, when the welding state is unstable, when the unstable state can be restored to a stabilized state, and when the unstable state cannot be stabilized, can be obtained, these data can be analyzed, and a trained model can be constructed using machine learning or deep learning, etc., which, when the welding parameters obtained during welding are input, outputs information such as the degree of abnormality, the parameters that are destabilizing the welding state, and the direction in which those parameters are improved, and this can be used as the evaluation model M.

[0026] When the control unit 33 judges the welding state during welding and analyzes the direction of improvement of the parameters, etc., the control unit 33 records this information in the memory unit 35 in association with the welding parameters related to the analysis acquired from the data logger 20, the acquisition time of the welding parameters, and the identification information of the welder 60 in charge of welding. This record can be used for the welder 60 to review his / her own work or for analyzing the work tendencies of each of the multiple welders 60 (for example, Person A tends to use a high welding current, Person B tends to have an unstable welding speed, etc.).

[0027] The output unit 34 outputs to the display device 40 the results of the welding condition determination made by the determination unit 331, the values ​​of the welding parameters obtained from the data logger 20, the analysis results made by the analysis unit 332, and the like. The storage unit 35 stores various information such as welding parameters, judgment models, and threshold settings.

[0028] Model construction device 50 acquires welding parameters stored in welding support device 30 or the like, and constructs evaluation model M by a predetermined method. The construction method of evaluation model M is not particularly limited. For example, model construction device 50 extracts data acquired when the welding state is stable from among the welding parameters, learns the extracted normal data, and constructs evaluation model M in which normal ranges of each parameter are defined. Model construction device 50 outputs the constructed evaluation model M to welding support device 30.

[0029] (Operation during welding) FIG. 3 shows an example of the transition of the welding parameters and the abnormality degree. The upper part of FIG. 3A shows the transition of the welding parameters during welding, and the lower part of FIG. 3B shows the transition of the abnormality degree. The vertical axis of FIG. 3A is the magnitude of the measured value of each parameter, and the horizontal axis is time. The vertical axis of FIG. 3B is the magnitude of the abnormality degree, and the horizontal axis is time. The same position on the horizontal axis of FIG. 3A and FIG. 3B shows the same time. Among the welding parameters, graph P1 of FIG. 3A shows the transition of the welding current, graph P2 shows the transition of the welding voltage, and graph P3 shows the transition of the welding speed. The welding support device 30 acquires the welding parameters every moment and calculates the abnormality degree. Graph P4 of FIG. 3B shows the transition of the abnormality degree calculated in real time by the determination unit 331. For example, when the welding speed increases at time t1, the abnormality degree calculated by the formula (1a) increases. Similarly, when the welding current increases at time t2, the abnormality degree increases, and when the welding current decreases at time t3, the abnormality degree increases. Then, when the abnormality degree becomes equal to or greater than the threshold, the determination unit 331 determines that the welding state has become unstable, and instructs the analysis unit 332 to analyze a method for stabilizing the welding state. If the abnormality degree becomes equal to or greater than the threshold at all of times t1 to t3, the analysis unit 332 identifies, by the process described with reference to FIG. 2A and the like, that the increase in the welding speed is the cause of the increase in the abnormality degree at time t1, and analyzes that slowing down the welding speed is a method for reducing the abnormality degree below the threshold. Then, the output unit 34 displays a guide for supporting welding, such as "Please reduce the welding speed," on the display device 40 based on the analysis result. The welder 60 sees this guide, reduces the welding speed, and continues the welding work. Then, the abnormality degree decreases below the threshold. Similarly, analysis unit 332 identifies parameters (both welding currents) that are factors that destabilize the welding state (factors that increase the degree of abnormality above the threshold) for times t2 and t3, and analyzes the direction of improvement of the identified parameter values ​​(a decrease at time t2, an increase at time t3). In response to this, output unit 34 displays a guide such as "Please decrease the welding current" on display device 40 at time t2, and a guide such as "Please increase the welding current" on display device 40 at time t3. Welder 60 sees these guides and adjusts the welding current during welding based on his own judgment to stabilize the welding state.

[0030] FIG. 4 shows an example of a screen displayed on the display device 40. In FIG. The upper, interrupted, and lower rows of FIG. 4 respectively show an example of an image (screen 41) displayed on display device 40 at times T1-T2, T2-T3, and T3-T4 during welding, and a transition in the degree of abnormality. At time T1, the welding state is stable. In this case, output unit 34 displays the welding parameters on screen 41. Output unit 34 also displays that the determination result by determination unit 331 is stable (normal) by, for example, displaying the background color of screen 41 in a color (e.g., blue) indicating that the welding state is stable. Output unit 34 continues to display the same until the degree of abnormality becomes equal to or greater than the threshold value, while reflecting the latest values ​​acquired from data logger 20 in the values ​​of each parameter of welding current, welding voltage, and welding speed on screen 41.

[0031] When the degree of abnormality becomes equal to or greater than the threshold at time T2, the output unit 34 switches to a display mode as exemplified by a screen 41 shown in the middle part of FIG. 4. The output unit 34 displays the destabilizing factor parameters calculated by the processing described with reference to FIGS. 2A to 2C and 3, and a guide on how to improve them. At this time, the output unit 34 displays the parameters in descending order of contribution to the degree of abnormality ((k1-x1) in formula (1a)). 2In the example shown in the middle of FIG. 4, the calculation result shows that the distance value calculated for the welding current is the largest, followed by the welding current, and the welding speed is the third largest in the calculation of the degree of abnormality using formula (1a). It also shows that the number of parameters to be improved and displayed on the screen 41 (the maximum number of parameters) is set to three. The output unit 34 displays the background of the screen 41 in a color (e.g., red) indicating that the welding state is unstable, or displays a warning mark, to indicate that the judgment result by the judgment unit 331 is unstable (abnormal). The output unit 34 continues to display the same information until the degree of abnormality falls below the threshold. The welder 60 continues the welding work while referring to this display and reducing the welding current, increasing the welding voltage, reducing the welding speed, etc. (only some of them may be performed). If the welding current is reduced too much during these adjustments, for example, the analysis by the analysis unit 332 displays "Increase the welding current" on the screen 41.

[0032] When the degree of abnormality falls below the threshold at time T3 due to adjustment of the welding parameters by welder 60, output unit 34 switches the display content on screen 41 to a display mode in which the background is, for example, blue, as in times T1 to T2, and the latest parameter values ​​are displayed in the foreground. Output unit 34 continues to display in the same manner as long as the state in which the degree of abnormality remains below the threshold continues (until time T4).

[0033] By displaying information to support the welding work in this way, the welder 60 can proceed with the work while appropriately managing the welding conditions. For example, by checking the values ​​of the welding parameters, the welder 60 can work while understanding the actual welding conditions, and can independently respond to fluctuations in each parameter during welding. In addition, by monitoring the degree of abnormality, if the value of the welding parameter falls outside the ideal range, a warning is immediately displayed and a method of improvement is suggested. This allows the welder 60 to recognize instability in the welding state that would not be noticed if the welder 60 were working by himself, and the welding state can be quickly stabilized by the suggestion of a method of improvement.

[0034] (operation) Next, an example of the operation of welding support system 100 will be described taking as an example a case where the determination method described with reference to FIG. 2A is applied. FIG. 5 is a flowchart showing an example of the operation of the welding support system according to the embodiment. First, model construction device 50 constructs evaluation model M (step S10). For example, model construction device 50 acquires time-series data of multiple parameters, such as the welding current, welding voltage, welding speed, and the product of the welding current and welding voltage, collected during a time period when ideal welding was performed, and constructs evaluation model M that defines a normal value range for each parameter. Model construction device 50 outputs evaluation model M to welding support device 30, and welding support device 30 receives evaluation model M and stores it in memory unit 35.

[0035] In addition, before starting welding, threshold values ​​and the like are set (step S11). For example, the user inputs to welding support device 30 a threshold value used to determine the degree of abnormality, a weighting coefficient such as A1 when formula (1b) is used, a setting of the number of parameters to be displayed on screen 41 when the welding state becomes unstable (e.g., three), and the like. Setting reception unit 32 receives these settings and records the set information in storage unit 35.

[0036] Next, welder 60 starts welding (step S12). Welder 60 activates welding machine 10, data logger 20, and sensors provided at the welding site. This causes the welding parameters to be transmitted to welding support device 30. Data acquisition unit 31 acquires the welding parameters (step S13). Specifically, data acquisition unit 31 acquires the measured values ​​of the welding parameters by the sensors, and calculates the welding parameters that need to be calculated as necessary. As an example, the welding parameter to be evaluated is assumed to be x, and x is composed of three parameters 1 to 3. Data acquisition unit 31 outputs the acquired welding parameters to be evaluated to control unit 33 and records them in memory unit 35.

[0037] Next, the determination unit 331 calculates the degree of abnormality (step S14). For example, the determination unit 331 calculates the distance between the evaluation model M and each welding parameter to be evaluated, and calculates the sum of the distances as the degree of abnormality. The degree of abnormality is the distance D(x) described with reference to FIG. 2. The determination unit 331 calculates the nth (n is any natural number) neighboring point k in the evaluation model M, and calculates the degree of abnormality using formula (1a) or formula (1b).

[0038] Next, the determination unit 331 compares the calculated abnormality level with the threshold value set in step S10 to determine whether the welding state is stable or not (step S15). The determination unit 331 determines that the welding state is unstable if the abnormality level is equal to or higher than the threshold value, and determines that the welding state is stable if the abnormality level is less than the threshold value. If it is determined that the welding state is stable (step S15; Yes), the determination unit 331 outputs the determination result "stable" to the output unit 34. The output unit 34 displays the welding parameters on the display device 40 based on the determination result "stable" (step S16). As illustrated in the upper and lower screens 41 of FIG. 4, the output unit 34 displays that the welding state is stable and the latest parameters transmitted in real time from the data logger 20. The screen 41 illustrated in FIG. 4 is an example. For example, if the welding parameters include several tens of parameters, all of the parameters cannot be displayed on one screen, so all of the parameters may be displayed by switching the screen at a predetermined time interval.

[0039] When the welding state is determined to be unstable (step S15; No), the determination unit 331 outputs the determination result "unstable" to the output unit 34 and instructs the analysis unit 332 to identify the cause of instability and analyze how to stabilize the welding state. The analysis unit 332 identifies the cause of instability (step S17). The analysis unit 332 focuses on the distance (such as (k1-x1) in formula (1a)) for each parameter calculated when calculating the degree of abnormality, and selects the parameters in descending order of distance, up to the number set in step S11. For example, when the top two are set to be selected in step S11, the analysis unit 332 identifies two parameters in descending order of distance from among parameters 1 to 3 (parameters 1 and 2 are assumed to be identified). Next, the analysis unit 332 calculates an improvement direction for stabilizing the welding state (reducing the degree of abnormality) for the parameters identified in step S17 (step S18). For example, the analysis unit 332 calculates (k1-x1) in the direction toward stabilization of parameter 1, and if the calculation result is negative, the value of parameter 1 is decreased, and if the calculation result is positive, the value of parameter 1 is increased. The analysis unit 332 also calculates (k2-x2) for parameter 2, and if the calculation result is negative, the value of parameter 2 is decreased, and if the calculation result is positive, the value of parameter 2 is increased. The analysis unit 332 outputs to the output unit 34 the identified parameter causing instability and information on the direction of improvement for that parameter.

[0040] Next, the output unit 34 displays guidance for improvement on the display device 40 (step S19). For example, as exemplified by screen 41 in the middle part of Fig. 4, the output unit 34 displays a message indicating that the welding state has become unstable and a message on how to stabilize the welding state with respect to the parameters identified in step S17.

[0041] Next, the control unit 33 records the determination result and the abnormality level value in step S15 and the analysis results in steps S17 to S18 in the memory unit 35 in association with the welding parameters and the identification information of the welder 60 recorded in step S13 (step S20).

[0042] Until the welding work by the welder 60 is completed (step S21; No), the welding support system 100 repeats the processes of steps S13 to S20. When the welding work is completed (step S21; Yes), the control unit 33 displays the work result of this welding work on the display device 40 through the output unit 34 based on the record of step S20 (step S22). For example, the control unit 33 may calculate the number of times the welding work was determined to be unstable, the type of parameter identified as the destabilizing factor at that time and the deviation direction (opposite to the improvement direction), the time required from the determination of instability to stabilization, and the like, and display them on the display device 40. An example of the work result displayed on the display device 40 is shown in FIG. 6. The welder 60 can grasp the tendency of his own work by referring to the work result, and use it for future work.

[0043] Furthermore, when the work history during welding by multiple welders 60 (information recorded in step S20) is accumulated in memory unit 35, control unit 33 can create work result information for each welder 60 as exemplified in FIG. 6 and output it to display device 40 or an electronic file. For example, a welding work manager designates identification information for one or more welders 60 and instructs welding support device 30 to display the work results of the designated welders 60. Then, control unit 33 creates work result information for the designated welders 60 (for example, FIG. 6) and outputs the work result information. The welding work manager can refer to the work result information for each welder 60 to evaluate the tendency and quality of the welding work of each welder 60 and use it for improvement.

[0044] (effect) As described above, according to the welding support system 100 of the present embodiment, the welding state (abnormality degree) is evaluated based on the evaluation model M for evaluating the welding stability, and the result is displayed, thereby realizing the improvement of productivity by stabilizing the welding quality and the reduction of rework (repair, reinspection) when a welding defect occurs. Specifically, when the welding state is stable, the sensor measurement results (each parameter of the welding parameters) of the welding conditions are displayed on the screen 41. This allows the welder 60 to proceed with the work while understanding the current welding conditions. In addition, when the welding state becomes unstable, the fact that the welding state is unstable is displayed on the screen 41, and the cause and a method for reducing the abnormality degree (a method for improving the welding conditions) are presented. The welder 60 can stabilize the welding state by making a judgment based on the presented method and changing the welding conditions. By continuing stable welding, as a result, welding defects can be reduced, and the stabilization and improvement of welding quality, the improvement of productivity due to these, the reduction of rework (repair, reinspection) when a welding defect occurs, and the reduction of costs can be realized.

[0045] FIG. 7 is a diagram illustrating an example of a hardware configuration of the welding support system according to the embodiment. The computer 900 includes a CPU 901, a main storage device 902, an auxiliary storage device 903, an input / output interface 904, and a communication interface 905. The data logger 20, the welding support device 30, and the model construction device 50 are implemented in the computer 900. The above-mentioned functions are stored in the auxiliary storage device 903 in the form of a program. The CPU 901 reads the program from the auxiliary storage device 903, loads it in the main storage device 902, and executes the above-mentioned processing according to the program. The CPU 901 also secures a storage area in the main storage device 902 according to the program. The CPU 901 also secures a storage area in the auxiliary storage device 903 for storing data being processed according to the program.

[0046] A program for realizing all or part of the functions of the welding support device 30 and the model construction device 50 may be recorded in a computer-readable recording medium, and the program recorded in the recording medium may be read into a computer system and executed to perform processing by each functional unit. The term "computer system" as used herein includes hardware such as an OS and peripheral devices. In addition, if a WWW system is used, the term "computer system" also includes a homepage providing environment (or display environment). In addition, the term "computer-readable recording medium" refers to a portable medium such as a CD, DVD, or USB, and a storage device such as a hard disk built into a computer system. In addition, when the program is distributed to the computer 900 through a communication line, the computer 900 that receives the program may deploy the program in the main storage device 902 and execute the above processing. In addition, the above program may be for realizing part of the above-mentioned functions, and may further be capable of realizing the above-mentioned functions in combination with a program already recorded in the computer system.

[0047] As described above, several embodiments according to the present disclosure have been described, but all of these embodiments are presented as examples and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, substitutions, and modifications can be made without departing from the gist of the invention. These embodiments and their modifications are included in the scope of the invention and its equivalents as described in the claims, as well as in the scope and gist of the invention.

[0048] <Additional Notes> The welding support method, the welding support system, and the program described in each embodiment can be understood, for example, as follows.

[0049] (1) A welding assistance method according to a first aspect includes the steps of acquiring values ​​of welding parameters during welding for the welding parameters indicating the state of welding (S13), evaluating the degree of abnormality of the welding parameters based on an evaluation model M for evaluating the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters (S14, S15), identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold (S17), calculating a direction of improvement for the value of the parameter (S18), and displaying information to assist welding related to the welding parameters together with the evaluation result of the degree of abnormality (S16, S19). By determining the welding condition during welding and presenting welding work support information (latest welding parameter values ​​or improvement information) that helps stabilize the welding condition, it is possible to improve welding quality and reduce rework.

[0050] (2) A welding support method according to a second aspect is the welding support method of (1), wherein, when the degree of abnormality is less than a threshold value (S15; No), in the display step, values ​​of the welding parameters acquired during welding are displayed as information for supporting the welding (S16). When the welding state is stable, the welder can carry out the welding work while checking the current welding parameter values ​​in the stable state.

[0051] (3) A welding support method according to a third aspect is the welding support method of (1) to (2), and if the degree of abnormality is equal to or greater than a threshold value (S15; Yes), in the display step, the parameter to be improved and the direction of improvement are displayed as information to support the welding (S19). The welder can check the displayed parameters to be improved and the direction of improvement, and then adjust the welding conditions at his / her own discretion to stabilize the welding condition.

[0052] (4) A welding support method according to a fourth aspect is the welding support method of (1) to (3), further comprising a step (S11) of setting how many of the parameters to be improved are to be identified, and in the calculating steps (S17, S18), a distance between a value of each parameter included in the welding parameters (point x in FIG. 2A1) and a normal value of each of the parameters based on the evaluation model (point k in FIG. 2A1) is calculated, and the number of the parameters set in the setting step is identified in descending order of the distance. This makes it possible to identify the parameter that is causing the increase in the degree of abnormality.

[0053] (5) A welding support method according to a fifth aspect is the welding support method of (4), wherein in the calculating steps (S17, S18), a direction for reducing the distance between the measured value of the identified parameter and a normal value of the parameter based on the evaluation model is calculated as the improvement direction. This makes it possible to calculate how to reduce the degree of abnormality for the parameter that is the cause of an increase in the degree of abnormality.

[0054] (6) A welding support method according to a sixth aspect is the welding support method of (1) to (5), wherein the evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, and in the evaluating step, a density ratio between the probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal values ​​is calculated as a degree of abnormality, and in the calculating step, a direction in which an average value of the probability density distribution of the evaluation target values ​​approaches an average value of the probability density distribution of the normal values ​​is calculated as the improvement direction. This makes it possible to determine the degree of abnormality based on the probability density distribution and calculate how to reduce the degree of abnormality.

[0055] (7) A welding support method according to a seventh aspect is the welding support method according to any one of (1) to (6), wherein the evaluation model is a model that reproduces normal values ​​constructed so as to output normal values ​​of each parameter included in the welding parameters when the normal values ​​of the parameters are input, and the evaluating step calculates, as an abnormality degree, a reproducibility indicating to what extent a value output by the evaluation model when a value of each parameter included in the welding parameters is input to the evaluation model reproduces the input parameter value, and the calculating step calculates, as the improvement direction, a direction in which the reproducibility is improved (a direction in which the value of the reproducibility is brought closer to 0). This makes it possible to determine the degree of abnormality based on the reproducibility of the evaluation model, and to calculate how to reduce the degree of abnormality.

[0056] (8) A welding support method according to an eighth aspect is a welding support method according to any one of (1) to (7), further comprising the steps of: recording the evaluation results in chronological order for each welder (S20); and displaying the evaluation results for a specific welder based on the evaluation results recorded for each welder (S21, see paragraph 0040). This makes it possible to understand the welding habits and tendencies of each welder and use this information to improve welding quality.

[0057] (9) A welding assistance method according to a ninth aspect is a welding assistance method according to any one of (1) to (8), further comprising a step (S10) of learning values ​​of the welding parameters measured when normal welding is performed and creating the evaluation model. This allows the evaluation model M to be constructed.

[0058] (10) The welding support system 100 according to the tenth aspect includes a means for acquiring values ​​of welding parameters during welding (data acquisition unit 31) for the welding parameters indicating the state of welding, a means for evaluating the degree of abnormality of the welding parameters based on an evaluation model M for evaluating the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters (judgment unit 331), a means for identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value (analysis unit 332) and a means for displaying information for supporting welding regarding the welding parameters together with the evaluation result of the degree of abnormality (output unit 34).

[0059] (11) A program according to an eleventh aspect causes a computer 900 to execute the steps of acquiring values ​​of welding parameters during welding, the welding parameters indicating a welding state, evaluating the degree of abnormality of the welding parameters based on an evaluation model for evaluating the degree of abnormality of the welding parameters and the acquired values ​​of the welding parameters, identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold, and calculating a direction of improvement for the value of the parameter, and displaying information to support welding related to the welding parameters together with the evaluation result of the degree of abnormality. [Explanation of symbols]

[0060] 10. Welding machine 11. Welding rod 20. Data logger 30. Welding support device 31 Data acquisition section 32 Settings reception section 33 Control section 331...judgment section 332...Analysis Department 34 Output section 35...Storage section 40...Display device 50...Model building device 60. Welder 100 Welding support system 900...Computer 901···CPU 902...Main memory 903...Auxiliary storage device 904 Input / Output Interface 905 Communication Interface

Claims

1. For a welding parameter indicative of a state of the weld, obtaining a value of the welding parameter during the weld; Evaluating the degree of abnormality of the welding parameter based on an evaluation model for evaluating the degree of abnormality of the welding parameter and the acquired value of the welding parameter; a step of identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction of improvement of the parameter value; displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, In the step of evaluating, a density ratio between a probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal value is calculated as an abnormality degree; In the calculating step, a direction in which an average value of the probability density distribution of the evaluation target value approaches an average value of the probability density distribution of the normal value is calculated as the improvement direction. Welding assistance methods.

2. For a welding parameter indicative of a state of the weld, obtaining a value of the welding parameter during the weld; Evaluating the degree of abnormality of the welding parameter based on an evaluation model for evaluating the degree of abnormality of the welding parameter and the acquired value of the welding parameter; a step of identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction of improvement of the parameter value; displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a model that reproduces the normal values ​​and is constructed so as to output the normal values ​​when normal values ​​of each parameter included in the welding parameters are input, In the step of evaluating, a reproducibility indicating to what extent a value outputted from the evaluation model when a value of each parameter included in the welding parameters is inputted into the evaluation model reproduces the inputted parameter value is calculated as an abnormality degree; In the calculating step, a direction in which the reproducibility is improved is calculated as the improvement direction. Welding assistance methods.

3. When the degree of abnormality is less than a threshold value, in the display step, values ​​of the welding parameters acquired during welding are displayed as information for assisting the welding. The welding support method according to claim 1 or 2.

4. When the degree of abnormality is equal to or greater than a threshold, the display step displays the parameter to be improved and the direction of improvement as information for supporting the welding. The welding support method according to any one of claims 1 to 3.

5. The method further includes a step of setting how many parameters to be improved are to be identified; In the calculating step, a distance between a value of each parameter included in the welding parameters and a normal value of each parameter based on the evaluation model is calculated, and the number of the parameters set in the setting step is identified in descending order of the distance. The welding support method according to any one of claims 1 to 4.

6. In the calculating step, a direction for reducing a distance between the specified measured value of the parameter and a normal value of the parameter based on the evaluation model is calculated as the improvement direction. The welding assistance method according to claim 5.

7. A step of recording the evaluation results in chronological order for each welder; Displaying the evaluation result for a specific welder based on the evaluation results recorded for each welder; The welding support method according to any one of claims 1 to 6, further comprising:

8. A step of learning values ​​of the welding parameters measured when normal welding is performed to create the evaluation model; The welding support method according to any one of claims 1 to 7, further comprising:

9. A means for acquiring values ​​of welding parameters indicative of a welding state during welding; a means for evaluating a degree of abnormality of the welding parameter based on an evaluation model for evaluating a degree of abnormality of the welding parameter and the acquired value of the welding parameter; a means for identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction in which the parameter value should be improved; a means for displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, the evaluating means calculates a density ratio between a probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal value as an abnormality degree; the calculating means calculates, as the improvement direction, a direction in which an average value of the probability density distribution of the evaluation target value approaches an average value of the probability density distribution of the normal value; Welding support system.

10. A means for acquiring values ​​of welding parameters indicative of a welding state during welding; a means for evaluating a degree of abnormality of the welding parameter based on an evaluation model for evaluating a degree of abnormality of the welding parameter and the acquired value of the welding parameter; a means for identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction in which the parameter value should be improved; a means for displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a model that reproduces the normal values ​​and is constructed so as to output the normal values ​​when normal values ​​of each parameter included in the welding parameters are input, the evaluating means calculates, as an anomaly degree, a reproducibility indicating to what extent a value output by the evaluation model when a value of each parameter included in the welding parameters is input to the evaluation model reproduces the input parameter value; The calculation means calculates a direction in which the reproducibility is improved as the improvement direction. Welding support system.

11. On the computer, For a welding parameter indicative of a state of the weld, obtaining a value of the welding parameter during the weld; Evaluating the degree of abnormality of the welding parameter based on an evaluation model for evaluating the degree of abnormality of the welding parameter and the acquired value of the welding parameter; a step of identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction of improvement of the parameter value; displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a probability density distribution of normal values ​​of each parameter included in the welding parameters, In the step of evaluating, a density ratio between a probability density distribution of an evaluation target value of each parameter included in the welding parameters and a probability density distribution of the normal value is calculated as an abnormality degree; a process of calculating, as the improvement direction, a direction in which an average value of a probability density distribution of the evaluation target value approaches an average value of a probability density distribution of the normal value, in the calculating step; A program that executes the following.

12. On the computer, For a welding parameter indicative of a state of the weld, obtaining a value of the welding parameter during the weld; Evaluating the degree of abnormality of the welding parameter based on an evaluation model for evaluating the degree of abnormality of the welding parameter and the acquired value of the welding parameter; a step of identifying a parameter to be improved among the welding parameters when the degree of abnormality is equal to or greater than a threshold value, and calculating a direction of improvement of the parameter value; displaying information for supporting welding related to the welding parameters together with the evaluation result of the abnormality degree; having the evaluation model is a model that reproduces the normal values ​​and is constructed so as to output the normal values ​​when normal values ​​of each parameter included in the welding parameters are input, In the step of evaluating, a reproducibility indicating to what extent a value outputted from the evaluation model when a value of each parameter included in the welding parameters is inputted into the evaluation model reproduces the inputted parameter value is calculated as an abnormality degree; In the calculating step, a process of calculating a direction in which the reproducibility is improved as the improvement direction; A program that executes the following.

Images