JP2019101497A - Consumption amount adjustment device, welding system, consumption amount adjustment method and computer program - Google Patents

Abstract

To provide a consumption adjusting device capable of reducing consumption of consumables consumed in a welding process.SOLUTION: A consumption adjustment device for adjusting consumption of consumables in a welding process, includes: an acquisition unit for acquiring welding data indicating a welding state during or after a welding process; and a consumption adjustment unit that adjusts the consumption so that the consumption of the consumable is reduced based on the welding data acquired by the acquisition unit.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a consumption adjustment device, a welding system, a consumption adjustment method, and a computer program for adjusting consumption of consumables in a welding process.

  One of the welding methods is a consumable electrode type gas shielded arc welding method. Gas shielded arc welding is a method of generating an arc between a welding wire fed to a base metal and a base metal and welding the base metal by the heat of the arc. In order to prevent the oxidation of the base material, welding is performed while injecting a shield gas around the welding portion.

  On the other hand, there is a technique of acquiring and graphing welding monitor data detected during a welding process, for example, welding current. In addition, there is a technique of analyzing image data captured after a welding process or appearance data of a welded portion measured by a laser to determine whether the welding result is good or bad. Furthermore, there is a technology that automatically sets welding conditions by machine learning using image data obtained by imaging the weld site, appearance data of beads obtained by processing the image data, spatter generation data, etc. For example, Patent Document 1).

JP, 2017-30014, A

  The shielding gas and welding wire used in the consumable electrode type gas shielded arc welding method are consumables. If the consumption of consumables is too low, welding results will be poor, so the consumption of consumables is usually set at a high level. Once the consumption of consumables is set, the set value is unlikely to be changed thereafter. For this reason, inherently, consumables tend to be consumed more than the amount required for welding, and cost increase and environmental load have been problems.

  An object of the present invention is to provide a consumption adjustment device, a welding system, a consumption adjustment method and a computer program capable of reducing consumption of consumables consumed in a welding process.

  The consumption adjustment device according to the present invention is a consumption adjustment device that adjusts consumption of consumables in a welding process, and includes an acquisition unit that acquires welding data indicating a welding state during or after the welding process; And a consumption adjusting unit that adjusts the consumption so that the consumption of the consumable is reduced based on the welding data acquired by the acquiring unit.

  According to this aspect, the acquisition unit acquires welding data, and the consumption adjustment unit adjusts the consumption based on the acquired welding data. The welding data is information indicating the state of welding, and includes information that contributes to the determination as to whether or not the consumption can be reduced, or whether the consumption should be increased. The consumption adjustment unit can adjust the consumption by using the welding data so that the consumption of the consumable can be reduced without deteriorating the welding result. Thus, the consumption of consumables consumed in the welding process can be reduced.

  The consumption adjustment device according to the present invention includes a quality determination unit that determines the quality of the welding result based on the welding data acquired by the acquisition unit, and the consumption adjustment unit determines that the quality determination unit is good. When it is determined, the consumption amount decreases, and when the quality determination unit determines no, the increase or decrease of the consumption amount is determined so that the consumption amount increases.

  According to this aspect, when the welding result is good, the consumption adjustment device reduces the consumption of consumables because the consumables may be excessive. If the welding result is not good, the consumption adjustment device may increase the consumption of consumables because there may be a shortage of consumables. By this adjustment process, the consumption of consumables can be reduced so as not to deteriorate the welding result as much as possible.

  In the consumption adjustment device according to the present invention, the consumption adjustment unit reduces the consumption amount related to the consumable item, and as a result, when the welding result changes from a good state to a defective state, the consumption amount before reduction decreases. The adjustment is confirmed at step S., and the determined consumption is stored in the storage unit.

According to this aspect, the consumption of the consumable can be minimized, and the adjustment is determined with the minimum consumption. Note that the minimum consumption is not necessarily the logical minimum consumption. The minimum consumption means the consumption before the consumption reduction when the welding result is changed from a good state to a defective state when welding is performed with a reduction in consumption of consumables.
After that, it is possible to adjust immediately to the minimum consumption based on the consumption stored in the storage unit.

  In the consumption adjustment device according to the present invention, the quality determination unit outputs neural data indicating, when the welding data is input, data indicating the quality of the welding result related to the welding process when the welding data is obtained. It has a good / bad decision neural network that has trained the network.

  According to this aspect, the quality determination neural network is, for example, a learned deep layer neural network, and the quality of the welding result can be appropriately determined. The type of the neural network is not particularly limited. It may be appropriately selected according to the characteristics of welding data, such as CNN (Convolutional Neural Network), RNN (Recurrent Neural Network), and LSTM (Long Short-Term Memory).

  In the consumption adjustment device according to the present invention, the consumption adjustment unit may output, when the welding data is input, data indicating an increase or decrease in the consumption that can reduce the consumption of the consumable item. A consumption adjustment neural network trained neural network is provided.

  According to this aspect, the consumption adjustment neural network is, for example, a learned deep layer neural network, and can appropriately adjust the increase or decrease of the consumables. The type of the neural network is not particularly limited. It may be appropriately selected according to the characteristics of welding data, such as CNN, RNN, LSTM, etc.

  In the consumption adjustment device according to the present invention, the consumption adjustment neural network outputs data indicating an increase / decrease amount of the consumption.

  According to this aspect, the consumption adjustment neural network can output an adjustable increase / decrease amount of the consumption, not whether to reduce the consumption of the consumable. For example, the consumption adjustment neural network may output a large decrease if the welding result is very stable, and may output a small decrease if the welding result is good but unstable. it can. Therefore, the consumption of consumables can be reduced more quickly.

  The consumption adjustment device according to the present invention determines whether the welding result is good or bad based on the welding data acquired by the acquisition unit, and the quality judgment unit obtained after increasing or decreasing the consumption amount. And a learning processing unit for learning the consumption adjustment neural network based on the determination result.

  According to this aspect, the consumption adjustment neural network performs learning using data indicating welding results when the consumption of the consumable item is increased or decreased. Therefore, the consumption of consumables can be reduced more effectively so that the welding result does not deteriorate.

  In the consumption adjustment device according to the present invention, when the learning processing unit determines that the quality determination unit is good, the consumption amount decreases, and when the quality determination unit determines that the consumption is not, the consumption amount increases. To train the consumption adjustment neural network.

  According to this aspect, the consumption adjustment neural network can be learned in the direction in which the consumption of the consumables is reduced. The learning can minimize the consumption of consumables.

  In the consumption adjustment device according to the present invention, the learning processing unit learns the consumption adjustment neural network such that the consumption is maintained when the welding result is in an intermediate state between good and bad.

  According to this aspect, when the welding result is an intermediate state between good and bad, the consumption adjustment neural network is trained so that the consumption is maintained. In the intermediate state, although the welding result is relatively good, if the consumption is further reduced, the welding result may be deteriorated. The learning can minimize the consumption of consumables and stabilize the welding result in a good state.

  In the consumption adjustment device according to the present invention, the quality determination unit outputs neural data indicating, when the welding data is input, data indicating the quality of the welding result related to the welding process when the welding data is obtained. It has a good / bad decision neural network that has trained the network.

  According to this aspect, the quality determination neural network is, for example, a learned deep layer neural network, and the quality of the welding result can be appropriately determined. By using the pass / fail judgment result of the pass / fail judgment neural network, the consumption adjustment neural network can be learned more effectively.

  In the consumption adjustment device according to the present invention, the consumption adjustment neural network includes substantially the same network configuration as all or part of the quality determination neural network.

  According to this aspect, the consumption adjustment neural network includes substantially the same neuron configuration as all or part of the quality determination neural network. For example, a part of the consumption adjustment neural network has the same or substantially the same middle layer and weighting factor as all or part of the pass / fail judgment neural network. Since the determination of the quality of the welding result and the determination of the increase or decrease of the consumption of the consumables have a feature in common, the quality determination neural network can be diverted to the consumption adjustment neural network. That is, the initial value of the weighting factor of the consumption adjustment neural network can be set to a more appropriate value. Therefore, even if learning data for learning increase or decrease of consumables is insufficient, if it is possible to prepare sufficient learning data of welding data and data indicating the quality of welding results, the weight coefficient of the consumption adjustment neural network The initial value of can be set appropriately, and the consumption adjustment neural network can be learned more efficiently. Needless to say, the network structures of the consumption adjustment neural network and the quality determination neural network may be identical.

  In the consumption adjustment device according to the present invention, the welding data includes the welding current and welding voltage detected during the welding process, the feeding speed of the welding wire, the short circuit condition, the welding noise collected during the welding process, and It includes data indicating at least one of the images of the weld site imaged after the welding process.

  According to this aspect, the welding current and welding voltage detected during the welding process, the feeding speed of the welding wire, the short circuit condition, the welding noise collected during the welding process, and the welding site imaged after the welding process Data indicative of at least one of the images can be used to determine the increase or decrease of the consumable item.

  In the consumption adjustment device according to the present invention, the consumable includes shield gas consumed in gas shield arc welding, or welding wire consumed in a welding process.

  According to this aspect, the consumption of shield gas or welding wire can be reduced.

  A welding system according to the present invention includes any one of the consumption adjustment devices described above, a welding robot that holds a welding torch, and a welding power supply that supplies a welding current to the welding torch.

  According to this aspect, the welding system including the welding robot and the welding power source can reduce the consumption of consumables consumed in the welding process. The consumption adjustment device may be provided inside the welding robot and the welding power source, or may be provided inside the control device for controlling the operation of the welding robot and the welding power source, or the welding robot, the welding power source and the control It may be separately provided outside the device. In addition, the consumption adjustment device may be a server, and the control device or the welding power source may be configured to communicate with the server to reduce consumption of consumables.

  The consumption adjustment method according to the present invention is a consumption adjustment method for adjusting consumption of consumables in a welding process, which acquires welding data indicating a welding state during or after the welding process, and acquired welding Based on the data, the consumption is adjusted such that the consumption of the consumable is reduced.

  According to this aspect, the consumption of consumables consumed in the welding process can be reduced. The consumption adjustment method may be automatically performed by a welding power source, a control device, etc. constituting the welding system, or a worker may connect the consumption adjustment device to the welding system to adjust the consumption. You may carry out.

  A computer program according to the present invention is a computer program for causing a computer to adjust the consumption of consumables in a welding process, and the computer acquires welding data indicating a welding state during or after the welding process. And based on the acquired welding data, the process which adjusts the said consumption is performed so that the consumption of the said consumable item may be reduced.

  According to this aspect, the computer can function as the consumption adjustment device.

  According to the present invention, the consumption of consumables consumed in the welding process can be reduced.

1 is a schematic view showing an arc welding system according to Embodiment 1. FIG. FIG. 1 is a block diagram showing a consumption adjustment device according to a first embodiment. FIG. 1 is a functional block diagram showing a consumption adjustment device according to a first embodiment. 5 is a flowchart showing an increase / decrease adjustment method according to the first embodiment. FIG. 6 is a functional block diagram showing a consumption adjustment device according to a second embodiment. It is a conceptual diagram which shows the network configuration of a consumption adjustment part. FIG. 13 is a functional block diagram showing a consumption adjustment device according to a third embodiment.

Hereinafter, the present invention will be described in detail based on the drawings showing the embodiments. In addition, at least part of the embodiments described below may be arbitrarily combined.
(Embodiment 1)
FIG. 1 is a schematic view showing an arc welding system according to a first embodiment. The arc welding system according to the present embodiment is a consumable electrode type gas shielded arc welder, and includes a welding robot 1, a welding power source 2, a control device 3, an imaging device 4 and a consumption adjustment device 5. The consumption adjustment device 5 is provided in the control device 3.

  The welding robot 1 performs arc welding of the base material A automatically. The welding robot 1 is provided with a base fixed to an appropriate place on the floor surface. A plurality of arms are rotatably connected to the base via a shaft, and a welding torch 11 is held at the tip of the arm. In addition, a wire feeding device 12 is provided at an appropriate position of the arm. A motor is provided at the connecting portion of each arm, and each arm is pivoted about the shaft by the rotational driving force of the motor. The rotation of the motor is controlled by the controller 3. The control device 3 can move the welding torch 11 up, down, front, back, left, and right relative to the base material A by rotating each arm. Further, an encoder for outputting a signal indicating the rotational position of the arm to the control device 3 is provided at the connecting portion of each arm, and the control device 3 can control the welding torch 11 based on the signal output from the encoder. Recognize the position of

The welding torch 11 is made of a conductive material such as a copper alloy and has a cylindrical contact tip for guiding the welding wire W to the base material A to be welded and supplying a welding current necessary for generating an arc. The welding current is supplied from the welding power source 2. The welding wire W is supplied to the welding torch 11 by a wire feeding device 12 from a wire source not shown. The welding wire W is, for example, a solid wire and functions as a consumable electrode.
The contact tip contacts the welding wire W passing therethrough and supplies a welding current to the welding wire W. Further, the welding torch 11 has a hollow cylindrical shape surrounding the contact tip, and has a nozzle for injecting shield gas from the opening at the tip to the base material A. The shielding gas is for preventing the oxidation of the base material A and the welding wire W melted by the arc. The shield gas is, for example, a mixed gas of carbon dioxide gas, carbon dioxide gas and argon gas, an inert gas such as argon, or the like. Shielding gas is supplied from the welding power source 2.

  The welding power supply 2 includes a power supply unit 21, a wire feeding control unit 22, a shield gas supply unit 23 and a detection unit 24. The power supply unit 21 is connected to the contact tip of the welding torch 11 and the base material A via a feed cable, and supplies a welding current. The wire feeding control unit 22 controls the feeding speed of the welding wire W by the wire feeding device 12. The feed speed of the welding wire W corresponds to the value of the welding current. The shield gas supply unit 23 supplies a shield gas to the welding torch 11. The detection unit 24 includes a current detection unit that detects a welding current flowing in an arc during the welding process, and a voltage detection unit that detects a voltage applied to the welding torch 11 and the base material A. The power supply unit 21 includes a power supply circuit that outputs a DC current that is PWM-controlled based on the welding current and the welding voltage detected by the detection unit 24, a signal processing circuit, and the like. The welding power source 2 also outputs welding monitor data indicating the welding state in the welding process to the control device 3. The welding monitor data is, for example, welding current data or welding voltage data indicating welding current or welding voltage detected during the welding process. In addition, even if welding speed data indicating the feeding speed of welding wire W, short circuit condition data indicating a short circuit condition, and welding sound data obtained by collecting sound with a microphone not shown are output to control device 3 as welding monitor data good.

  After the welding process, the imaging device 4 images a welding portion of the base material A, and outputs image data obtained by imaging to the control device 3.

  The control device 3 controls the operation of the welding robot 1 and outputs welding conditions such as welding current, welding voltage, feeding speed of the welding wire W, supply amount of shield gas to the welding power source 2 to the welding power source 2 The operation of welding power source 2 is controlled. The control device 3 stores various welding conditions for the material of the base material A, the type of the groove, and the like. The welding conditions stored in the control device 3 are not necessarily optimal, and consumption of consumables in the welding process, such as shielding gas and welding wire W, is minimized so that the welding result does not deteriorate. It is adjusted by the consumption adjustment device 5.

  FIG. 2 is a block diagram showing the consumption adjustment device 5 according to the first embodiment. The consumption adjustment device 5 includes a control unit 50 that controls the operation of each component of the consumption adjustment device 5. The control unit 50 is connected to an input unit 50a, an output unit 50b, and a storage unit 50c.

  The storage unit 50c is a nonvolatile memory such as an EEPROM (Electrically Erasable Programmable ROM), a flash memory, or the like. The storage unit 50c stores a computer program 50d for minimizing consumption of consumables consumed in the welding process within a range where the welding result does not deteriorate.

  The control unit 50 is a computer having a processor such as a central processing unit (CPU), a graphics processing unit (GPU) or a multi-core CPU, a read only memory (ROM), a random access memory (RAM), an input / output interface, etc. Are connected to an input unit 50a, an output unit 50b, and a storage unit 50c. The control unit 50 executes the computer program 50 d stored in the storage unit 50 c to implement the consumption adjustment method for minimizing the consumption of consumables, and causes the computer to function as the consumption adjustment device 5.

  The input unit 50 a is connected to the welding power source 2 and the imaging device 4. The welding monitor data output from the welding power source 2 and the image data output from the imaging device 4 are input to the input unit 50a. The welding monitor data is time-series data indicating, for example, welding current, welding voltage, feeding speed of the welding wire W, short circuit condition, welding noise and the like. The image data is data representing the appearance of the bead.

  The output unit 50 b is connected to the welding robot 1 and the welding power source 2. Control unit 50 outputs control data for controlling the consumption amount of consumables to welding power source 2. The data may be data instructing the increase or decrease of the consumption, or may be data indicating the consumption after the increase or decrease adjustment.

  FIG. 3 is a functional block diagram showing the consumption adjustment device 5 according to the first embodiment. The consumption adjustment device 5 includes, as functional blocks, a welding monitor data acquisition unit 51a, an image data acquisition unit 51b, a first pass / fail judgment unit 52a, a second pass / fail judgment unit 52b, a pass / fail comprehensive judgment unit 54, and a consumption A quantity adjustment unit 55, a consumption control unit 56, and a minimum consumption storage unit 57 are provided.

  The welding monitor data acquisition unit 51a acquires welding monitor data output from the welding power source 2, and outputs the acquired welding monitor data to the first pass / fail judgment unit 52a.

  The image data acquisition unit 51b acquires the image data output from the imaging device 4, and outputs the acquired image data to the second quality determination unit 52b.

The first pass / fail judgment unit 52a outputs a pass / fail judgment RNN (Recurrent Neural Network) 53a that outputs the data indicating the pass / fail of the welding result according to the welding process when the welding monitor data is obtained when the welding monitor data is input. Prepare. The quality determination RNN 53a is, for example, a learned recursive neural network.
The pass / fail judgment RNN 53a includes, for example, a first neuron that outputs data indicating the probability that the welding result is good, and a second neuron that outputs data indicating the probability that the welding result is bad in the output layer. In this case, the data indicating the quality is data output from the first and second neurons.
The quality determination RNN 53a may include, in the output layer, neurons that output the quality of the welding result in binary. In this case, the data indicating the quality is binary data output from the neuron.
Furthermore, the quality determination RNN 53a may include, in the output layer, a neuron that outputs an analog value indicating the degree of quality of the welding result.
The quality judgment RNN 53a is made to learn by providing welding monitor data (input data) and data (teacher data) indicating the quality of the welding result corresponding to the welding data as learning data to the recursive deep neural network before learning. Just do it.
The structure, such as the number of layers in the middle layer of the quality judgment RNN 53a and the number of neurons in each layer, is not particularly limited. Further, the quality judgment RNN 53a does not necessarily have to be a recursive neural network, and may be constituted by other types of neural networks.

The second pass / fail judgment unit 52b includes a pass / fail judgment CNN (Convolutional Neural Network) 53b that outputs data indicating the pass / fail of the welding result according to the welding process when the image data is obtained. The pass / fail judgment CNN 53b is a trained convolutional neural network.
The pass / fail determination CNN 53b includes, for example, a third neuron that outputs data indicating the probability that the welding result is good and a third neuron that outputs data indicating the probability that the welding result is bad in the output layer. In this case, the data indicating the quality is data output from the third and fourth neurons.
In addition, the pass / fail judgment CNN 53b may include, in the output layer, neurons that output the pass / fail of the welding result in binary. In this case, the data indicating the quality is binary data output from the neuron.
Furthermore, the good / bad determination CNN 53b may include, in the output layer, a neuron that outputs an analog value indicating the degree of the good or bad of the welding result.
The pass / fail judgment CNN 53b may be performed by giving image data (input data) and data (teacher data) indicating pass / fail of the welding result corresponding to the welding data as learning data to the convolutional neural network before learning .
The structure, such as the number of intermediate layers of the quality judgment CNN 53b and the number of neurons in each layer, is not particularly limited. The pass / fail judgment CNN 53b need not necessarily be a convolutional neural network, and may be composed of other types of neural networks.

Based on the data outputted from the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b, the pass / fail comprehensive judgment unit 54 judges pass / fail of the welding result and outputs the judgment result to the consumption adjustment unit 55.
For example, the good or bad comprehensive judgment unit 54 judges the data output from the first and second neurons of the good or bad judgment RNN 53a and the data outputted from the third and fourth neurons of the good or bad judgment CNN 53b. Specifically, by comparing the sum of the values of data output from the first neuron and the third neuron with the sum of the values of data output from the second and fourth neurons, the quality of the welding result is evaluated. It is good to judge. Also, the values of data output from each neuron may be weighted and compared.
Further, in the case where the binary data is output from the pass / fail judgment RNN 53a and the pass / fail judgment CNN 53b, the pass / fail comprehensive judgment unit 54 indicates that both the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b are good. When the data shown is output, it is determined as good, and when one of the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b is outputting data indicating that it is defective, it is judged as failure. In addition, the method of comprehensive determination is an example, and when one of the first pass / fail decision unit 52a and the second pass / fail decision unit 52b outputs data indicating that it is good, it is configured to decide that it is good. It is good.

The consumption adjustment unit 55 reduces the consumption of consumables when the judgment result of the good or bad comprehensive judgment unit 54 is good, and decreases the consumption of the consumable when the judgment is bad. Is determined, and the determination result of the increase or decrease is output to the consumption control unit 56. The determination result is, for example, binary data, and the consumption adjustment unit 55 outputs, to the consumption control unit 56, either data indicating that the consumption is increased or data indicating that the consumption is decreased.
In addition, as a result of the consumption adjustment unit 55 making a decision to reduce the consumption related to the consumable item, when the welding result changes from a good state to a defective state, the consumption before reduction is stored in the minimum consumption memory part Make it memorize in 57. The minimum consumption may be a value corresponding to the adjustment amount which is the difference between the consumption of the consumable as a standard value and the consumption at the time of minimization. It may be a corresponding value.

  The consumption control unit 56 controls the consumption of the consumable by outputting control data for controlling the consumption of the consumable to the welding power source 2 based on the determination of the consumption adjustment unit 55. However, when the minimum consumption storage unit 57 stores the minimum consumption, the consumption control unit 56 controls the consumption of the consumable based on the consumption stored in the minimum consumption storage 57. .

Next, the processing procedure of the control unit 50 related to the adjustment of the consumption amount of consumables will be described.
FIG. 4 is a flowchart showing an increase / decrease adjustment method according to the first embodiment. The control unit 50 repeatedly executes, for example, the following process for each welding process. Control unit 50 determines whether or not storage unit 50c stores the minimum consumption amount (step S11). When it is determined that the minimum consumption amount is stored (step S11: YES), the control unit 50 controls the consumption amount of consumables based on the minimum consumption amount stored in the storage unit 50c (step S12). For example, the control unit 50 controls the consumption amount by outputting control data indicating the minimum consumption amount to the welding power source 2. Of course, the control unit 50 may control the consumption amount by outputting control data indicating an increase or decrease amount to set the consumption amount to the minimum consumption amount to the welding power source 2.

  When it is determined that the storage unit 50c does not store the minimum consumption (step S11: NO), the control unit 50 acquires welding monitor data (step S13), and acquires image data (step S14). And control part 50 judges the quality of a welding result based on acquired welding monitor data and image data (Step S15). For example, the control unit 50 determines the quality of the welding result using the learned quality determination RNN 53a and the quality determination CNN 53b.

  Next, when it is determined that the welding result is good (step S15: YES), the control unit 50 reduces the consumption of the consumable item (step S16). If it is determined that the welding result is defective (step S15: NO), it is determined whether the welding result has changed from a good state to a defective state as a result of reducing consumables (step S17). Needless to say, the change to the defective state is not limited to the configuration that is determined based on only one welding result, but also includes a configuration that uses two or more welding results to determine. For example, it may be determined that the welding state has changed to a defective state when the welding result is in a bad state at a certain rate or more in ten times. When it is determined that the welding result does not change from a good state to a defective state (step S17: NO), the control unit 50 increases the consumption amount of consumables (step S18). The control unit 50 that has finished the process of step S16 or step S18 controls the consumption based on the determination result (step S19). Specifically, control unit 50 controls the amount of consumption by outputting control data indicating the amount of consumption after the adjustment process to welding power source 2. Of course, the control unit 50 may control the amount of consumption by outputting control data indicating the amount of increase or decrease of the amount of consumption to the welding power source 2.

  As a result of reducing the consumption of consumables, when it is determined that the welding result has changed from a good state to a defective state (step S17: YES), the control unit 50 reduces the consumption of the consumables before consumption (Step S20), the consumption before reduction is stored as the minimum consumption in the storage unit 50c (step S21), and the process returns to step S12.

  According to the consumption adjustment device 5 thus configured, the welding system, the consumption adjustment method, and the computer program 50d, the consumption of consumables consumed in the welding process can be reduced without deteriorating the welding result. it can.

  Further, since the minimized consumption amount is stored in the storage unit 50c, thereafter, the consumption adjustment device 5 can quickly control welding by minimizing the consumption amount of consumables.

In the first embodiment, an example in which the consumption adjustment device 5 includes the learned good / bad judgment RNN 53a and the good / bad judgment CNN 53b has been described, but the neural network of the first good / bad judgment unit 52a and the second good / good judgment unit 52b is defined. Various parameters may be configured to be downloaded from an external server and updated. The parameters are, for example, information including the number of layers in the middle layer, the number of neurons in each layer, the weighting factor of each neuron, the type of activation function, and the like. In addition, the consumption adjustment device 5 stores a flag indicating whether or not to allow the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b to reflect the downloaded various parameters, and the flag indicates permission In this case, the neural network of the quality determination RNN 53a and the quality determination CNN 53b may be updated using the downloaded parameter.
Moreover, when multiple welding systems provided with the consumption adjustment apparatus 5 are installed in the factory, the consumption adjustment apparatus 5 of each welding system may exchange the said parameter as needed.
Furthermore, the consumption adjustment device 5 may be configured as a cloud server. The welding power source 2 or the control device 3 may request the server to adjust the consumption of consumables, receive the amount of adjustment transmitted from the server according to the request, and adjust the consumption of consumables.
Furthermore, the consumption adjustment device 5 may be provided in the welding power source 2. Further, the consumption adjustment device 5 may be implemented as a dedicated device for adjusting the consumption of consumables. The operator can connect the dedicated device to the welding system to automatically adjust the consumption of consumables.

  Furthermore, although the example in which the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b include the pass / fail judgment RNN 53a and the pass / fail judgment CNN 53b has been described, either or both of the decision units pass / fail the welding result without using the neural network. May be determined. For example, the first pass / fail determination unit 52a may determine pass / fail by a simple determination process that compares the welding current value with a predetermined threshold value. In addition, the second pass / fail judgment unit 52b extracts the predetermined feature amount from the image data, and determines whether the pass / fail is a simple judgment processing that compares the presence / absence of the feature amount, the number of the feature amount, etc. good. Furthermore, it is not necessary to provide both the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b, and either one may be provided. In this case, the good or bad comprehensive judgment unit 54 is unnecessary.

Second Embodiment
FIG. 5 is a functional block diagram showing the consumption adjustment device 205 according to the second embodiment. The consumption adjustment device 205 according to the second embodiment includes the welding monitor data acquisition unit 51a, the image data acquisition unit 51b, the first pass / fail judgment unit 52a, the second pass / fail judgment unit 52b, and the pass / fail judgment unit 254 similarly to the first embodiment. , A consumption adjustment unit 255 and a consumption control unit 56, and further includes a learning processing unit 259.

  In the consumption adjustment device 205, the welding system, the consumption adjustment method, and the computer program 50d according to the second embodiment, the consumption adjustment unit 55 and the minimum consumption storage unit 57 of the first embodiment are configured by a deep neural network. Since the point and the output of the good / not comprehensive judgment unit 54 are different from those of the first embodiment, the difference will be mainly described below. The other configurations and effects are the same as those of the embodiment, and therefore the corresponding portions are denoted by the same reference numerals and the detailed description will be omitted.

  The welding monitor data acquisition unit 51a acquires welding monitor data output from the welding power source 2, and outputs the acquired welding monitor data to the first quality determination unit 52a and the consumption adjustment unit 255.

  The image data acquisition unit 51 b acquires the image data output from the imaging device 4, and outputs the acquired image data to the second quality determination unit 52 b and the consumption adjustment unit 255.

  The overall good or bad judgment unit 254 of the second embodiment outputs not the binary data indicating good or bad but data indicating the probability that the welding result is good and data indicating the probability that the welding result is bad to the learning processing unit 259 Do. For example, the probability that the welding result is good may be calculated using the value of the data output from the first neuron of the pass / fail judgment RNN 53a and the value of the data output from the third neuron of the pass / fail judgment CNN 53b. Similarly, using the value of the data output from the second neuron of the pass / fail judgment RNN 53a and the value of the data output from the fourth neuron of the pass / fail judgment CNN 53b, the probability that the welding result is defective may be calculated. .

The consumption adjustment unit 255 includes a consumption adjustment NN 258 that outputs data indicating increase or decrease in consumable consumption that can be reduced when welding monitor data and image data are input. Consumption adjustment NN 258 is a learned deep neural network.
The consumption adjustment NN 258 includes, for example, in the output layer, a plurality of neurons that output data indicating the probability that adjustment using the increase / decrease amount is preferable for each increase / decrease amount of consumption.
In addition, the consumption adjustment NN 258 may be configured to include, in the output layer, a neuron that outputs data indicating an increase or decrease in the consumption. Furthermore, the consumption adjustment NN 258 may be configured to include, in the output layer, neurons that output binary data indicating an increase or decrease in consumption. Hereinafter, in the second embodiment, it is assumed that the consumption adjustment NN 258 is not binary data, and outputs data indicating the reducible increase / decrease amount with probability.

FIG. 6 is a conceptual view showing a network configuration of the consumption adjustment unit 255. As shown in FIG. The consumption adjustment NN 258 of the consumption adjustment unit 255 includes a welding state recognition network unit 258 a, an appearance state recognition network unit 258 b, and an increase / decrease determination network unit 258 c.
The welding state recognition network unit 258a is a neural network that receives welding monitor data, recognizes a welding state in the welding process, and outputs data corresponding to the state. When the welding monitor data is a welding current, the welding state recognition network unit 258a can recognize a change state of the welding current. The welding state recognition network unit 258a may have, for example, the same neural network structure as the first quality determination unit 52a except for the output layer. The output layer comprises a plurality of neurons, preferably three or more neurons. In addition, it is preferable to set the weighting factor of each neuron constituting the first pass / fail judgment unit 52a as the initial value of the weighting factor before learning. The consumption adjustment unit 255 can be learned more efficiently.
The appearance state recognition network unit 258b is a neural network that receives image data, recognizes a state of a welded portion after welding, and outputs data according to the state. The appearance state recognition network unit 258b may have, for example, the same neural network structure as the second quality determination unit 52b except for the output layer. The output layer comprises a plurality of neurons, preferably three or more neurons. Further, it is preferable to set the weighting factor of each neuron constituting the second pass / fail judgment unit 52b as the initial value of the weighting factor before learning. The consumption adjustment NN 258 can be learned more efficiently.
The increase / decrease determination network unit 258c is learned to receive data output from the welding state recognition network unit 258a and the appearance state recognition network unit 258b and to output data indicating increase / decrease amount of consumables that can be reduced. Neural network. The neural network is preferably configured by a deep neural network including a plurality of intermediate layers.
The neural network configuration of the consumption adjustment unit 255 is an example, and may be configured as one neural network, or a plurality of neural networks may be combined.

The learning processing unit 259 learns data indicating the quality of the welding result when the consumption amount of consumables is increased or decreased based on the input data and the welding monitor data and the image data input to the consumption adjusting unit 255. It is a processing unit that learns the consumption adjustment NN 258 as data.
Specifically, according to the determination result of the good or bad comprehensive determination unit 254, the learning processing unit 259 reduces the consumption if the welding result is good, and increases the consumption if the welding result is not good. If the result is in an intermediate state of good and bad, let the consumption adjustment NN 258 learn so that the consumption is maintained.
A good welding result means, for example, a state in which the probability that the welding result is good is 50% or more and the probability that the welding result is bad is less than 50%. A poor weld result means, for example, a state in which the probability that the weld result is good is less than 50% and the probability that the weld result is bad is 50% or more. The threshold of 50% is an example, and may be a value larger than 50%.
The welding result is in an intermediate state, for example, when both the probability that the welding result is good and the probability that the welding result is bad are 50% or more, or both are less than 50%. Moreover, when the said threshold value is larger than 50%, for example, when it is 60%, the condition where the probability that the welding result is favorable and inferior is between 40%-60% is also an intermediate state. Note that such an intermediate state is an example. The intermediate state is a state in which the welding result may be deteriorated if the consumption of consumables is further reduced.
As described above, by learning the consumption adjustment NN 258, the consumption of consumables can be minimized without deteriorating the welding result.
In the initial learning stage of the consumption amount NN 258, in the case of an intermediate state, the consumption amount may not be maintained, and may be appropriately increased or decreased.

Another example of the learning processing method will be described. The learning processing unit 259 learns the consumption adjustment NN 258 so that the consumption decreases when the pass / fail comprehensive determination unit 254 determines that it is good, and the consumption increases when the pass / fail comprehensive determination unit 254 determines that it is not Let That is, when the welding result for the output data is good, the learning processing unit 259 changes the weighting factor of the neural network so as to further reduce the increase / decrease amount of the consumption which is the output data. When the welding result for the output data is not good, the learning processing unit 259 changes the weighting factor of the neural network so as to increase the increase / decrease amount of the consumption which is the output data. By repeating this learning process, consumption of consumables can be minimized.
In addition, as a result of the learning processing unit 259 outputting data indicating that the consumption amount related to the consumable item is reduced, the welding result changes from a good state to a defective state or a change to such a defective state occurs. Change the learning method if repeated several times. The learning processing unit 259 causes the consumption adjustment NN 258 to be learned so that the consumption is maintained when the good or bad comprehensive judgment unit 254 judges good, and the consumption increases when the good or bad comprehensive judgment unit 254 judges bad. .
As described above, by learning the consumption adjustment NN 258, the consumption of consumables can be minimized without deteriorating the welding result.

It should be noted that the learning of the consumption adjustment NN 258 may be performed at an appropriate timing such as when the welding system is installed, when the external environment changes, when the welding conditions are changed, when other setup changes are performed.
Further, in the second embodiment, an example in which the consumption adjustment NN 258 is learned has been described, but the learned consumption adjustment NN 258 may be provided so as to prevent further learning.

  According to the consumption adjustment device 205, the welding system, the consumption adjustment method, and the computer program 50d according to the second embodiment configured as described above, the consumption adjustment unit 255 configured by the deep layer neural network Since the configuration is to determine the increase and decrease, it is possible to control and minimize the consumption of consumables more appropriately without deteriorating the welding result.

  In addition, the consumption adjustment unit 255 can output an increase / decrease amount of the consumption to reduce the consumption of the consumable. For example, the consumption adjustment unit 255 may output a large reduction amount when the welding result is very stable, and may output a small reduction amount when the welding result is good but unstable. it can. Therefore, consumption of consumables can be minimized more quickly.

  Furthermore, the consumption adjustment part 255 can learn using the quality determination result of a welding result, and can adjust to the consumption adjustment part 255 suitable for the environment where the welding system was installed. Therefore, the consumption of consumables can be minimized according to the welding conditions and the external environment.

  Furthermore, the learning processing unit 259 can learn the consumption adjustment unit 255 in the direction in which the consumption of the consumable is reduced, and can output data that can minimize the consumption of the consumable. After that, the consumption adjustment neural network can be trained so that good welding results can be stably obtained. Therefore, it is possible to stabilize the welding result in a good condition and to minimize the consumption of consumables.

In the second embodiment, the consumption adjustment device 205 has been described as including the learned consumption adjustment NN 258. However, various parameters defining the neural network of the consumption adjustment unit 255 are downloaded from an external server and updated. It may be configured to The parameters are, for example, information including the number of layers in the middle layer, the number of neurons in each layer, the weighting factor of each neuron, the type of activation function, and the like. In addition, the consumption adjustment device 205 stores a flag indicating whether or not to allow the consumption adjustment unit 255 to reflect various downloaded parameters, and when the flag indicates permission, the downloaded parameter is displayed. It may be configured to be used to update the neural network of the consumption adjustment NN 258.
When a plurality of welding systems including the consumption adjustment device 205 are installed in the factory, the consumption adjustment device 205 of each welding system may exchange the above-mentioned parameters as needed.

  In addition, the consumption adjustment device 205 may be configured to upload various parameters that define the learned consumption adjustment NN 258 to an external server. The other consumption adjustment device 205 can update the consumption adjustment NN 258 using the parameters uploaded to the server.

  In the second embodiment, the consumption adjustment NN 258, and the first pass / fail judgment unit 52a and the second pass / fail judgment unit 52b have neural networks. However, one or both of the pass / fail judgment RNN 53a and the pass / fail judgment CNN 53b are neural. You may comprise so that the quality of a welding result may be determined, without using a network.

(Embodiment 3)
FIG. 7 is a functional block diagram showing the consumption adjustment device 305 according to the third embodiment. The consumption adjustment device 305, the welding system, the consumption adjustment method, and the computer program 50d according to the third embodiment are different from the second embodiment in the data input to the consumption adjustment unit 355. explain. The other configurations and effects are the same as those of the embodiment, and therefore the corresponding portions are denoted by the same reference numerals and the detailed description will be omitted.

  The consumption adjustment device 305 according to the third embodiment further includes a welding condition data acquisition unit 51c. The welding condition data acquisition unit 51c acquires welding condition data. The welding condition data includes, for example, information on the material of base material A, groove shape, welding current setting value, welding voltage setting value, welding speed setting value, frequency setting value when periodically changing the welding current, etc. .

  The consumption adjustment unit 355 is, based on the input welding monitor data and image data, and welding condition data, data indicating an increase or decrease in consumption that can reduce the consumption of consumables without deteriorating the welding result. A learned consumption adjustment NN 358 for outputting is provided. The learned consumption adjustment NN 358 according to the third embodiment can be further learned using the same good / low comprehensive judgment unit 354 and the learning processing unit 359 as in the second embodiment.

  According to the consumption adjustment device 305, the welding system, the consumption adjustment method, and the computer program 50d according to the third embodiment, since the consumption is adjusted in consideration of the welding conditions, the increase or decrease of consumables can be more effectively performed. It is possible to control

  The embodiment disclosed this time should be considered as illustrative in all points and not restrictive. The scope of the present invention is indicated not by the meaning described above but by the claims, and is intended to include all modifications within the meaning and scope equivalent to the claims.

DESCRIPTION OF SYMBOLS 1 welding robot 2 welding power supply 3 control device 4 imaging device 5, 205, 305 consumption adjustment device 11 welding torch 12 wire feeding device 21 power supply unit 22 wire feeding control unit 23 shield gas supply unit 24 detection unit 50 control unit 50 a Input unit 50b Output unit 50c Storage unit 50d Computer program 51a Welding monitor data acquisition unit 51b Image data acquisition unit 51c Welding condition data acquisition unit 52a First pass / fail judgment unit 52b Second pass / fail judgment unit 53a Pass / fail judgment RNN
53b pass / fail judgment CNN
54, 254, 354 comprehensive judgment unit 55, 255, 355 consumption adjustment unit 56 consumption control unit 57 minimum consumption storage unit 258 consumption adjustment NN
258a welding condition recognition network unit 258b appearance condition recognition network unit 258c increase / decrease decision network unit 259, 359 learning processing unit A base material W welding wire

Claims (16)

A consumption adjustment device for adjusting consumption of consumables in a welding process, comprising:
An acquisition unit for acquiring welding data indicating a welding state during or after the welding process;
And a consumption adjustment unit for adjusting the consumption so that the consumption of the consumable is reduced based on the welding data acquired by the acquisition unit. It has a quality determination unit that determines the quality of the welding result based on the welding data acquired by the acquisition unit,
The consumption adjustment unit
If the quality determination unit determines that the quality is good, the consumption amount decreases, and if the quality determination unit determines that the quality is not determined, the increase or decrease of the consumption amount is determined such that the consumption amount increases. Consumption adjustment device as described. The consumption adjustment unit
When the welding result changes from a good state to a defective state as a result of reducing the consumption amount related to the consumable item, the adjustment is determined with the consumption amount before reduction, and the determined consumption amount is stored in the storage unit. The consumption adjustment device according to claim 2 which makes it memorize. The quality determination unit
When the welding data is input, there is provided a quality judgment neural network in which a neural network is learned so as to output data indicating the quality of the welding result according to the welding process when the welding data is obtained. The consumption adjustment device according to Item 3. The consumption adjustment unit
A consumption adjustment neural network is provided which learns a neural network so as to output data indicating increase / decrease of the consumption that can reduce the consumption of the consumable when the welding data is input. Consumption adjustment device as described. The consumption adjustment device according to claim 5, wherein the consumption adjustment neural network outputs data indicating an increase / decrease amount of the consumption. A quality determination unit that determines the quality of the welding result based on the welding data acquired by the acquisition unit;
The consumption adjustment device according to claim 5 or 6, further comprising: a learning processing unit for learning the consumption adjustment neural network based on the determination result of the quality determination unit obtained after increasing or decreasing the consumption. . The learning processing unit
The consumption adjustment neural network is trained so that the consumption decreases when the quality judging unit judges that the consumption is good, and the consumption increases when the quality judging unit judges that it is not. Consumption adjustment device described in. The learning processing unit
The consumption adjustment device according to claim 8, wherein the consumption adjustment neural network is learned so that the consumption is maintained when the welding result is in an intermediate state between good and bad. The quality determination unit
8. The quality judgment neural network according to claim 7, further comprising: a quality judgment neural network for learning the neural network so as to output data indicating the quality of the welding result according to the welding process when the welding data is obtained. Item 10. The consumption adjustment device according to any one of items 1 to 9. The consumption adjustment device according to claim 10, wherein the consumption adjustment neural network includes a network configuration substantially the same as all or part of the quality determination neural network. The welding data is
At least one of the welding current and welding voltage detected during the welding process, the feeding speed of the welding wire, the short circuit condition, the welding noise collected during the welding process, and the image of the welding site imaged after the welding process The consumption adjustment device according to any one of claims 1 to 11, including data to be shown. The consumables are
The consumption adjustment device according to any one of claims 1 to 12, including a shielding gas consumed in gas shield arc welding or a welding wire consumed in a welding process. The consumption adjustment device according to any one of claims 1 to 13.
A welding robot holding a welding torch,
A welding power source for supplying a welding current to the welding torch. A consumption adjustment method for adjusting consumption of consumables in a welding process, comprising:
Obtain welding data indicating the welding condition during or after the welding process,
The consumption adjustment method which adjusts the said consumption so that the consumption of the said consumables may be reduced based on the acquired welding data. A computer program for causing a computer to adjust consumption of consumables in a welding process,
On the computer
Obtain welding data indicating the welding condition during or after the welding process,
A computer program for executing a process of adjusting the consumption amount so as to reduce the consumption amount of the consumable item based on the acquired welding data.

Images