WO2021107139A1

WO2021107139A1 - Grinding system, correction amount estimation device, computer program, and grinding method

Info

Publication number: WO2021107139A1
Application number: PCT/JP2020/044359
Authority: WO
Inventors: 剛横矢
Original assignee: 株式会社安川電機
Priority date: 2019-11-27
Filing date: 2020-11-27
Publication date: 2021-06-03
Also published as: CN114616076A; JP7294448B2; JPWO2021107139A1

Abstract

A grinding system (100) comprising a reaction force related information acquisition unit that acquires reaction force related information about the pressing reaction force of a grinding tool during grinding processing, a difference calculation unit that acquires the difference between two pieces of the reaction force related information acquired in different repetitive executions of repetitive processing on the same part of the workpiece, a correction amount estimation unit that estimates a correction amount in the pressing direction of the grinding tool on the basis of the difference, and a processing control unit that controls the position of the grinding tool during processing on the basis of a target position and the correction amount.

Description

Grinding system, correction amount estimation device, computer program and grinding method

The present invention relates to a grinding system, a correction amount estimation device, a computer program, and a grinding method.

In Patent Document 1, the polishing processing condition data indicating the processing condition of the polishing process is observed as a state variable representing the current state of the environment, and based on the state variable, the wear of the polishing tool with respect to the processing condition of the polishing process. A polishing tool wear amount prediction device including a machine learning device that performs learning or prediction using a learning model that models a quantity is described.

Japanese Unexamined Patent Publication No. 2019-139755

Grinding of a workpiece by a grinding machine is performed by repeatedly pressing a grinding tool against a machining site until a desired machining result is obtained. As the machining progresses, not only the workpiece but also the grinding tool wears at the same time. Therefore, in order to ensure the machining accuracy, the workpiece after machining is measured to determine the necessity and conditions of reworking. Is necessary, and it is difficult to improve productivity.

The invention described in Patent Document 1 predicts the wear of the polishing tool at this point by using a machine learning device, and estimates the wear of the polishing tool based on various machining conditions, but the amount of wear of the tool is It is difficult to accurately estimate the amount of tool wear according to the present invention because it strongly depends on the state of the workpiece.

The present invention has been made in view of such circumstances, and an object of the present invention is to obtain a correction amount in a pressing direction that reflects the state of a workpiece in a grinding system.

The invention disclosed in this application in order to solve the above problems has various aspects, and the outline of typical ones of these aspects is as follows.

The grinding system according to one aspect of the present invention has a reaction force-related information acquisition unit that acquires reaction force-related information regarding the pressing reaction force of a grinding tool during grinding, and a reaction force-related information acquisition unit that acquires reaction force-related information regarding the pressing reaction force of a grinding tool during grinding. A difference calculation unit that calculates the difference between the two acquired reaction force-related information, a correction amount estimation unit that estimates a correction amount in the pressing direction of the grinding tool based on the difference, a target position, and the correction. It has a machining control unit that controls the position of the grinding tool at the time of machining based on the amount.

Further, in the grinding system according to another aspect of the present invention, the correction amount estimation unit may input the difference into the machine learning model to obtain the correction amount.

Further, the grinding system according to another aspect of the present invention has a cumulative correction amount calculation unit for calculating a cumulative correction amount which is a cumulative value of the correction amount for the grinding wheel of the grinding tool, and the machining control unit. May control the position of the grinding tool at the time of machining based on the target position and the cumulative correction amount when machining different workpieces.

Further, in the grinding system according to another aspect of the present invention, the machining target portion of the workpiece may be classified, and the correction amount estimation unit may estimate the correction amount for each classification.

Further, in the grinding system according to another aspect of the present invention, the grinding tool is supported by the motor control support mechanism, and the reaction force related information may be the external force torque of the motor in the pressing direction of the grinding tool. ..

Further, in the grinding system according to another aspect of the present invention, the reaction force related information is input to the second machine learning model, and the estimated number of repeated machining times at the same location of the workpiece is estimated. May have a part.

Further, the correction amount estimation device according to one aspect of the present invention is a difference in reaction force-related information regarding the pressing reaction force of the grinding tool during the grinding process, which is acquired at different repetitive times for the repetitive machining at the same place of the workpiece. It has a difference calculation unit for calculating the difference, and a correction amount estimation unit for estimating the correction amount in the pressing direction of the grinding tool based on the difference.

Further, the computer program according to one aspect of the present invention uses the computer to obtain reaction force-related information regarding the pressing reaction force of the grinding tool during the grinding process of 2 acquired at different repeated times for the repeated machining at the same location of the workpiece. It functions as a correction amount estimation device having a difference calculation unit for calculating the difference and a correction amount estimation unit for estimating the correction amount in the pressing direction of the grinding tool based on the difference.

Further, in the grinding method according to one aspect of the present invention, reaction force-related information regarding the pressing reaction force of the grinding tool during grinding is acquired, and the repeated machining at the same location of the workpiece is obtained at different repeated times. The difference in the reaction force related information is calculated, the correction amount in the pressing direction of the grinding tool is estimated based on the difference, and the target position and the position of the grinding tool at the time of machining are controlled based on the correction amount. To do.

It is the schematic which shows the whole structure of the example of the grinding system which concerns on a preferred embodiment of this invention. It is a block diagram which shows the functional structure of a grinding system. It is a figure which shows the example of the torque command value which is the reaction force related information obtained in the nth grinding process and the n + 1th grinding process, and the difference thereof. It is a figure which shows the example of the torque command value which is the reaction force related information obtained in the nth grinding process and the n + 1th grinding process, and the difference thereof. It is a block diagram which shows the functional structure of the learning apparatus of the machine learning model using a grinding system. It is a figure which shows the outline of the structure for estimating the estimated number of repeated machining in a grinding system. It is a figure which shows the outline of the system configuration. It is a figure which shows the work to be polished to be verified. It is a figure which shows the robot composition of the verification system. It is a table which shows the specifications of a disc grinding machine and a grindstone. It is a figure which shows the outline of the polishing operation. It is a figure which shows the appearance of the welding bead every 2 times of polishing. It is a diagram illustrating an example of the data of the disturbance torque tau _d in the

polishing operation

2,4,6,8,10 time. It is a table which shows the result estimated using the evaluation data from the NN model generated by learning. It is a table which shows the estimation result when the polishing operation of the same grindstone was performed 10 times in succession to the 11th to 15th workpieces. It is a block diagram which shows the functional structure of the grinding system which includes the iteration number estimation part.

FIG. 1 is a schematic view showing an overall configuration of an example of a grinding system 100 according to a preferred embodiment of the present invention.

The grinding system 100 includes a grinding tool 1 and a support mechanism 3 that supports the grinding tool 1 so as to be relatively movable with respect to the workpiece 2 and enables grinding of the workpiece 2. Further, FIG. 1 shows a support base 4 for supporting the workpiece 2 being machined and a controller 5 for controlling the support mechanism 3.

Here, the grinding tool 1 is a disc grinder in this embodiment, and has a grinder main body 6 and a grinding wheel 7 attached to the tip thereof. Further, the support mechanism 3 uses a general multi-axis industrial robot, and a grinding tool 1 is attached as an end effector thereof. The controller 5 is a robot controller.

Further, since the workpiece 2 is a member obtained by bead-welding a steel plate and the weld bead 8 is raised in a ridge shape, the grinding system 100 grinds the unnecessary raised portion of the weld bead 8 to the surface. Illustrates those whose processing purpose is to remove the swelling so that it becomes smooth.

Note that the grinding system 100 shown in FIG. 1 is an example, and there is no problem in the gist of the present invention even if the specific configuration of each part is replaced with another configuration. For example, although the grinding tool 1 is shown in FIG. 1 as using a generally commercially available disc grinder, the grinding tool 1 may be designed and manufactured exclusively for the grinding system 100. Further, although the support mechanism 3 is shown here as using a vertical articulated robot, various mechanisms such as a SCARA robot, a Cartesian robot, and a gantry mechanism may be used in addition to the support mechanism 3, and the grinding tool 1 is supported. Not only the side but also the support base 4 that supports the workpiece 2 may be movable.

Further, the controller 5 is not limited to a robot controller, but may be any device such as a servo controller, PLC (Programmable Logic Controller), or a PC for automatically controlling the support mechanism 3. It may be a combination of a plurality of devices described above. The workpiece 2 is not limited as long as it is an object that requires grinding, and in addition to surface grinding such as removal of unnecessary parts of the weld bead 8 shown in the figure, the inner surface of the object to be rotated or the inner surface of the object to be rotated or It may require various grinding processes such as external grinding and cutting.

FIG. 2 is a block diagram showing the functional configuration of the grinding system 100. The controller 5 is provided with a machining control unit 50, and by controlling the position of the support mechanism 3 by program control, the grinding tool 1 supported by the support mechanism 3 is desired for the workpiece 2. Move along the trajectory of. In the present embodiment, since the support mechanism 3 is an electric motor control support mechanism driven by an electric motor, the machining control unit 50 includes a motor controller for driving the electric motor, and each of the support mechanisms 3 includes a motor controller. The motor is servo-controlled.

The support mechanism 3 is provided with a reaction force-related information acquisition unit 30, which acquires reaction force-related information regarding the pressing reaction force of the grinding tool 1 during grinding. Here, the reaction force-related information means information that can indicate the pressing reaction force of the grinding tool 1 directly or by some conversion, and may be the pressing reaction force itself or other information. May be good. The pressing reaction force itself can be obtained, for example, by attaching the grinding tool 1 to the support mechanism 3 via a load cell and directly measuring the grinding tool 1. As other information, in the grinding system 100 according to the present embodiment, the reaction force torque acting on the motor of the support mechanism 3 in the pressing direction of the grinding tool 1, that is, the torque command value is used as the reaction force related information. There is. Adopting motor control information, particularly current command value or torque command value, as reaction force related information is economical because the support mechanism 3 does not require a special additional configuration for acquiring reaction force related information. .. Hereinafter, in the present embodiment, the torque command value of the motor of the support mechanism 3 will be used as the reaction force related information.

The controller 5 is provided with a correction unit 51 in addition to the processing control unit 50, and outputs a cumulative correction amount to the processing control unit 50 in the present embodiment. Here, the meaning of the correction by the correction unit 51 will be described before the detailed description of the configuration of the correction unit 51.

In the grinding process using the grinding tool 1, wear of the grinding wheel 7 is unavoidable as the processing progresses. Since the wear of the grinding wheel 7 means the retreat of the machining point of the grinding process, when the grinding tool 1 is ground by the position control by the support mechanism 3, the machining point is accompanied by the progress of the wear of the grinding wheel 7. However, it becomes impossible to process the desired shape of the workpiece 2. Therefore, in order to perform accurate grinding, it is necessary to estimate the amount of wear of the grinding wheel 7 by some method and add a correction corresponding to it to the position control by the support mechanism 3.

Measuring the amount of wear of the grinding wheel 7 with high accuracy each time it is processed makes the grinding system 100 complicated and costly, such as installation and adjustment of a measuring machine, and maintenance management becomes difficult. On the other hand, even if an attempt is made to estimate the amount of wear of the grinding wheel 7, the amount of wear of the grinding wheel 7 largely depends on the amount of grinding of the workpiece 2 in the grinding process. It changes, and it is difficult to estimate this only by the conditions on the grinding system 100 side.

Therefore, in the grinding system 100 according to the present embodiment, the correction unit 51 is provided with a difference calculation unit 52, and the difference between the reaction force-related information of 2 acquired at different repetitive times for the repetitive machining at the same location of the workpiece 2 is obtained. calculate.

Generally, in the grinding process, it is rare that the desired shape of the workpiece 2 can be obtained by one grinding, so the same portion is ground a plurality of times along the same trajectory until the desired shape is obtained.

FIG. 3 shows the torque command value which is the reaction force related information obtained in the nth grinding process and the n + 1th grinding process in the repeated machining of the workpiece 2 at the same location in the grinding system 100, and the torque command value thereof. It is a figure which shows the example of the difference. The upper graph of FIG. 3 shows the nth torque command value with a solid line, the n + 1th torque command value is shown with a broken line, and the lower graph shows the n + 1th torque command value from the nth torque command value. The difference obtained by subtracting the torque command value is shown by the solid line.

Since the torque command value reflects the pressing reaction force when the grinding tool 1 is pressed against the workpiece 2 during grinding, the larger the portion requiring processing, the stronger the pressing reaction force is generated on the support mechanism 3. Therefore, the waveform of the torque command value reflects the shape of the portion of the work portion 2 that requires machining. Then, as the grinding process progresses, the portion requiring processing is scraped and its size becomes smaller, so that the torque command value in the subsequent repeated times becomes smaller than the torque command value in the previous repeated times. There is a tendency.

Then, the difference in the torque command value acquired in the different iterations reflects the difference in the shape of the portion requiring machining in the previous iterations and the shape of the portion requiring machining in the later iterations. It is nothing but a reflection of the amount of change in the shape of the part that requires processing between the first and subsequent iterations. In the example shown in FIG. 3, the difference shown in the lower graph is associated with the amount of change in the shape of the portion requiring machining caused by the nth machining, that is, the amount of grinding by the grinding tool 1.

Therefore, in the grinding system 100 according to the present embodiment, based on this difference, the correction amount estimation unit 53 provided in the correction unit 51 further estimates the correction amount in the pressing direction of the grinding system 100.

The algorithm for estimating the correction amount may be any algorithm as long as the correction amount can be reasonably estimated based on the difference. Further, the correction amount may be any correction amount necessary for performing the grinding process with high accuracy, including the deviation of the processing point of the grinding tool 1 due to the wear of the grinding wheel 7. In the present embodiment, the moving distance of the machining point due to the wear of the grinding wheel 7 (hereinafter, simply referred to as "wear amount") is used as the correction amount, but in addition to this moving distance, the support mechanism due to the reaction force during machining is used. The amount including the amount of mechanical change such as the deflection compensation of 3 may be used as the correction amount.

Then, as already described, it is considered that there is a correlation between the difference shown in the lower graph of FIG. 3 and the required correction amount, but a highly accurate conversion relationship between the two was discovered. It's not always easy to do. Therefore, the correction amount estimation unit of the grinding system 100 according to the present embodiment uses a machine learning model in which the relationship between the difference and the wear amount has been learned in advance, and inputs the difference into the machine learning model. As a result, a correction amount is obtained as an output.

The architecture of this machine learning model is not particularly limited, but since the difference as an input value is time series data, it is preferable to use an RNN (recurrent neural network). The learning of the machine learning model will be described later.

The correction amount obtained in this way indicates the amount of wear of the grinding wheel 7 caused by the grinding process between the two different repetitions for which the torque command value was acquired. Since the wear of the grinding wheel 7 is accumulated each time the grinding wheel is processed, the correction for the position control of the support mechanism 3 is finally based on the accumulated value of the correction amount.

Therefore, the cumulative correction amount calculation unit 54 provided in the correction unit 51 accumulates the correction amount each time the correction amount estimation unit 53 estimates the correction amount, calculates and holds the cumulative correction amount, and performs such cumulative correction. The amount is output to the machining control unit 50.

Since this cumulative correction amount is an amount corresponding to the amount of wear of the individual grinding wheel 7 of the grinding tool 1, it is held as an amount for the grinding wheel 7 in use, and when the grinding wheel 7 is replaced, the amount is retained. The value is reset.

The machining control unit 10 receives the cumulative correction amount obtained as described above from the correction unit 51, and at the time of grinding, the grinding tool is based on the cumulative correction amount in addition to the preset target position. Control the position of 1. More specifically, the cumulative correction amount is added to the coordinates in the pressing direction of the target position of the grinding tool 1. Therefore, it can be said that the machining control unit 50 controls the position of the grinding tool 1 at the time of machining based on the target position and the correction amount estimated by the correction amount estimation unit 53.

Further, since the cumulative correction amount is the amount for the grinding wheel 7 in use, naturally, even when the grinding process for one workpiece 2 is completed and replaced with another different workpiece 2. It will continue to be used unless the grinding wheel 7 is replaced. Therefore, when machining a different workpiece 2, the machining control unit 50 controls the position of the grinding tool 1 at the time of machining based on the target position and the cumulative correction amount.

When the correction unit 51 is configured as an independent device in the controller 5 of the grinding system 100 described above, this can be considered as a correction amount estimation device. In that case, the correction amount estimation device is an independent device having the difference calculation unit 52, the correction amount estimation unit 53, and the cumulative correction amount calculation unit 54 described above.

The correction amount estimation device may be designed as a dedicated device or may be realized by using a general computer. FIG. 5 is a diagram showing a configuration of a general computer 11 that can be used as a correction amount estimation device. The computer 11 includes a CPU (Central Processing Unit) 11a as a processor, a RAM (Random Access Memory) 11b as a memory, an external storage device 11c, a GC (Graphics Controller) 11d, and

input devices

11e and 11f (Inpur / Output) 306. The data bus 11g is connected so that electric signals can be exchanged with each other. The hardware configuration of the computer 11 shown here is an example, and other configurations may be used.

The external storage device 11c is a device that can statically record information such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). Further, the signal from the GC 11d is output to a monitor 11h such as a CRT (Cathode Ray Tube) or a so-called flat panel display in which the user visually recognizes the image, and is displayed as an image. The input device 11e is one or more devices such as a keyboard, mouse, and touch panel for the user to input information, and the I / O 11f is one or more interfaces for the computer 11 to exchange information with an external device. Is. The I / O 11f may include various ports for wired connection and a controller for wireless connection.

The computer program for making the computer 11 function as the machine learning data generation device 1 and the machine learning device 2 is stored in the external storage device 11c, read into the RAM 11b as needed, and executed by the CPU 11a. That is, the RAM 11b stores the code for realizing the various functions shown as the correction unit 51 in FIG. 2 by being executed by the CPU 11a. Even if such a computer program is recorded and provided on an appropriate computer-readable information recording medium such as an appropriate optical disk, magneto-optical disk, or flash memory, the computer program is provided via an external information communication line such as the Internet via the I / O 11f. May be provided.

FIG. 5 shows the nth repetitive machining at the same location of the workpiece 2 when the length of the machining target for grinding is long or the shape of the machining target portion changes significantly during machining. It is a figure which shows the example of the torque command value which is the reaction force related information obtained in the grinding process and the n + 1th grinding process, and the difference thereof.

In this case, the torque command value obtained in the same repeated round is divided into a plurality of sections, and in the example shown in FIG. 5, the section a, the section b, and the section c are divided into three sections, and similarly, the nth time. The difference between the torque command value and the n + 1th torque command value is also divided into the same three sections. Then, the difference for each division may be input to the correction amount estimation unit 53, and as a result, the correction amount for each division may be obtained.

The final cumulative correction amount is the cumulative value of the correction amount for each category obtained in this way. In FIG. 4, each division is performed with respect to time, but since the grinding process is performed by the grinding tool 1 that moves on the workpiece 2 with the passage of time, this division is performed at the machining target location of the workpiece 2. It is nothing but a division of. By appropriately classifying the machining target parts of the workpiece 2 and estimating the correction amount for each classification in this way, the correction amount can be obtained with high accuracy when the machining conditions fluctuate significantly during the grinding process. Be done.

For example, when the machining conditions are significantly different, such as the size of the weld bead 8 being significantly different between the front half and the rear half of the machining target portion of the workpiece 2, it is better than estimating the correction amount collectively. This is because the correction amount can be estimated more accurately by estimating the correction amount separately for the first half and the second half, which are considered to have substantially the same processing conditions. The processing target portion of the work 2 may be classified at regular intervals, or may be classified at arbitrary intervals according to the properties of the work 2.

As explained above, the learning of the machine learning model used in the correction amount estimation unit 53 is any method as long as the machine learning model can be learned so as to input the difference and output the correction amount. It is also good and is not particularly limited. Therefore, an example of such a learning method will be described below.

FIG. 6 is a block diagram showing a functional configuration of a learning device 101 of a machine learning model using a grinding machine. In the figure, the same reference numerals are given to the configurations equivalent to those of the grinding system 100 shown in FIG. 2, and the duplicated description thereof will be omitted.

Similar to the grinding system 100, the machine learning device 101 is attached to a support mechanism 3 to which a grinding tool 1 is attached and having a reaction force related information acquisition unit 30, and a controller 5 having a machining control unit 50 that controls the support mechanism 3. In addition, the learning device 9 and the cumulative correction amount measuring device 10 are provided.

In this example, the cumulative correction amount measuring device 10 is a measuring device that measures the amount of wear of the grinding wheel 7 of the grinding tool 1 attached to the support mechanism 3. The amount of wear is measured by measuring the surface position of the grinding wheel 7 with an arbitrary sensor, for example, a non-contact laser sensor or an arbitrary contact sensor, or by driving the support mechanism 3 to drive the grinding tool 1 in advance. This may be done by pressing against the installed reference surface and detecting the coordinates of the support mechanism 3 at the time when the grinding wheel 7 and the reference surface come into contact with each other. That is, the cumulative correction amount measuring device 10 may be an independent device, or may be configured by utilizing one or a plurality of existing devices of the machine learning device 101.

At this time, the amount of wear of the grinding wheel 7 is measured, that is, the cumulative value of the wear caused by the grinding process so far. Therefore, in this example, the cumulative correction amount is measured.

Therefore, when performing repetitive machining on an arbitrary workpiece 2, the reaction force-related information acquisition unit 30 acquires the measured value of the reaction force-related information at the time of machining at any repetitive times, and also at any repetitive times. The actual measured value of the cumulative correction amount can be obtained.

The learning device 9 has a difference calculation unit 90, a machine learning model 91, and a correction amount calculation unit 92. The difference calculation unit 90 calculates the difference between the reaction force-related information acquired by the reaction force-related information acquisition unit 30, here, by inputting the torque command value, the difference between the two reaction force-related information acquired in different iterative times. can do.

Further, the correction amount calculation unit 92 can actually calculate the correction amount generated by the processing during the repetition times from the difference between the measured values of the cumulative correction amounts of 2 acquired in the different repetition times.

Therefore, the learner 9 learns the machine learning model 91 by repeatedly learning the machine learning model 91 using the difference obtained based on the actual measurement as input data and the correction amount obtained based on the actual measurement as the correct answer data. Can be made to.

The hardware that realizes the learner 9 is not particularly limited. The learner 9 may be realized as a part of the controller 5, or for example, the general computer 11 shown in FIG. 4 may be used as the learner 9. Further, in the example of the machine learning device 101 shown in FIG. 6, the learning device 9 is directly connected to the instruction mechanism 3 and the cumulative correction amount measuring device 10, and machine learning is performed every time the support mechanism 3 is driven to execute the grinding process. The configuration in which the model 91 is trained is shown, but in addition to this, a large number of reaction force-related information and actual measurement values of the cumulative correction amount are acquired and accumulated in advance, and later, a learning device 9 independently configured is used. It may be used to perform learning collectively.

Further, the grinding system 100 may have a configuration for estimating the estimated number of repeated machining, which is an estimated value of the number of repeated machining required to complete the machining when performing repeated machining at the same location of the workpiece 2. Hereinafter, the configuration for estimating the estimated number of repeated machining in the grinding system 100 will be described.

<System overview>
FIG. 7 shows an outline of the system configuration. In this system, the control torque data collected by the robot while performing polishing work is used to generate a model that estimates the state of the work target. A disc grinder for polishing is attached to the robot's minions. A damper mechanism is provided between the disc grinder and the hand, and even if it comes into contact, a certain amount of external force is absorbed. The robot controller (hereinafter referred to as RC) collects work data using the MotoPlus (Note) application and saves it in the USB memory. Using the saved data, learning is performed with the machine learning software in the PC. For the learning model, a neural network (hereinafter referred to as NN) consisting of four layers including the input layer and the output layer is used. The learned NN model is downloaded to RC via a USB memory. The RC inputs the control torque data at the time of work to this NN, and estimates the work state based on the output result.
(Note) MotoPlus: A function to develop application software that operates inside RC

<Estimation method of working condition>
In a general sensorless hand force estimation method, the hand force F uses the Jacobian matrix J, _{where τ d} is the disturbance torque that is the difference between the control torque estimated from the operating trajectory and the actual control torque. , Can be calculated by the number 1.

Hand force estimation

Since the contact work state S including the tool is a system in which F is input, if the system function is G, the number is 2.

Work state estimation with manual force as input

Here, it is replaced with a system function having a _{disturbance torque τ d} as an input as shown in Equation 3.

Work state estimation with disturbance torque as input

<Technical evaluation>
Evaluate the work state estimation using this technology. The estimated working state is the state of achievement of polishing.

<Work to be worked>
The work to be polished to be verified is shown in FIG. This involves welding work on an iron plate with a thickness of 4 mm and arbitrarily attaching a welding bead, and the target is polishing of this welding bead. The length of the weld bead is about 155 mm and the height is about 2.5 mm.

<Verification system configuration>
The robot configuration of the verification system is shown in FIG. A 6-axis vertical articulated robot MOTOMAN-GP7 (hereinafter referred to as GP7) is used as the robot, and YRC1000 is used as the RC. Attach the electric disc grinder G10SH5 manufactured by HiKOKI to the tip of GP7. This disc grinder can be polished by pressing the rotating surface by attaching a grindstone. A resinoid flexible toishi manufactured by HiKOKI is used as the grindstone. The welding bead to be polished is polished while being fixed to a table installed in front of the robot. The specifications of the disc grinding machine and the grindstone are shown in Tables 1 and 2 of FIG. 10, respectively.

<Target for polishing work>
FIG. 11 is an outline of the polishing operation. Polishing is performed by pulling the hand toward the front of the robot. This operation has been taught in advance, and the work uses only the three axes (second, third, and fifth axes) of the GP7 axes.

Actually polish and collect data. Polishing operation is performed 10 times for one weld bead. Experience has shown that when a new grindstone is used, polishing is approximately completed in these 10 polishing operations. Also, when the polishing of 10 workpieces is completed in order to obtain data with a new grindstone, the grindstone is replaced. FIG. 12 shows the appearance of the weld bead every two times of polishing. Every time it is polished, the weld bead is approaching flat, and after 10 times of polishing, the polishing is achieved.

<Evaluation result>
A total of 500 data were prepared by polishing 50 workpieces 10 times. Of these, a total of 50 data obtained by polishing the fifth work after replacing with a new grindstone was used as evaluation data, and learning was performed using the remaining 450 data. Figure 13 is an example of a data of the disturbance torque tau _d in the polishing

operation

2,4,6,8,10 time.

Table 3 of FIG. 14 shows the results estimated using the evaluation data from the NN model generated by learning. This is a table called a mixing matrix, which shows the estimated number of labels for each correct label (in this case, the number of polishings). The gray frame on the diagonal is the number that can be accurately estimated for each label. As a result, of the 50 data, 49 data can be estimated within the correct label or ± 1 time range, and it can be said that the estimation is approximately accurate.

The following are assumed as the causes of failure in the inference result. The number of polishings was used as the label estimated this time, but in reality, the grindstone deteriorated with each polishing, and even with the same number of polishings, there was a difference in the quality of polishing, and it was mistaken for a larger or smaller number of polishings. Actually, the estimation result when the polishing operation of the same grindstone is continued up to the 11th to 15th workpieces and the polishing operation is performed 10 times is as shown in Table 4 of FIG. It can be seen that it is presumed that the polishing achievement status of In other words, it can be determined that the grindstone is not sufficiently polished due to deterioration.

That is, it can be said that it can be applied to estimate the polishing state each time the polishing operation is performed, and when it is estimated to be the 10th polishing operation, stop the polishing operation to automatically end the polishing operation with an appropriate number of times.

With the configuration described above, the estimated number of iterations estimation unit 55 for estimating the estimated number of iterations can be configured. FIG. 16 is a block diagram showing a functional configuration of the grinding system 100 including the repetition number estimation unit 55. Since each configuration of the grinding system 100 according to this example is basically the same as that shown in FIG. 2, the same reference numerals are given to the same configurations, and duplicate description will be omitted. The only difference from that shown in FIG. 2 is that the controller 5 is provided with the estimated number of iterations estimation unit 55.

In this example, the estimated number of iterations estimation unit 55 estimates the estimated number of iterations at the same location of the workpiece 2, so that the obtained estimated number of iterations is used to estimate the remaining time required for the current machining. It can be useful for determining the timing of replacing the grinding wheel 7.

Claims

A reaction force-related information acquisition unit that acquires reaction force-related information related to the pressing reaction force of a grinding tool during grinding,
A difference calculation unit that acquires the difference between the two reaction force-related information acquired at different repetitive times for repetitive machining at the same location of the workpiece, and
A correction amount estimation unit that estimates a correction amount in the pressing direction of the grinding tool based on the difference,
A machining control unit that controls the position of the grinding tool at the time of machining based on the target position and the correction amount.
Grinding system with.
The grinding system according to claim 1, wherein the correction amount estimation unit inputs the difference into a machine learning model to obtain the correction amount.
It has a cumulative correction amount calculation unit that calculates a cumulative correction amount that is a cumulative value of the correction amount for the grinding wheel of the grinding tool.
The grinding system according to claim 1 or 2, wherein the machining control unit controls the position of the grinding tool at the time of machining based on the target position and the cumulative correction amount when machining different workpieces.
The grinding system according to any one of claims 1 to 3, wherein the processing target portion of the workpiece is classified, and the correction amount estimation unit estimates the correction amount for each classification.
The grinding system according to any one of claims 1 to 4, wherein the grinding tool is supported by an electric motor control support mechanism, and the reaction force related information is an external force torque of the electric motor in a pressing direction of the grinding tool.
The item according to any one of claims 1 to 5, which has an estimated repetitive machining number estimation unit that inputs the reaction force-related information into the second machine learning model and estimates the estimated repetitive machining number at the same location of the workpiece. Grinding system.
A difference calculation unit that acquires the difference in reaction force-related information regarding the pressing reaction force of the grinding tool during the grinding process, which was acquired at different repetitive times for the repetitive machining at the same location of the workpiece.
A correction amount estimation unit that estimates a correction amount in the pressing direction of the grinding tool based on the difference,
Correction amount estimation device having.
Computer,
A difference calculation unit that acquires the difference in reaction force-related information regarding the pressing reaction force of the grinding tool during the grinding process, which was acquired at different repetitive times for the repetitive machining at the same location of the workpiece.
A correction amount estimation unit that estimates a correction amount in the pressing direction of the grinding tool based on the difference,
A computer program for functioning as a correction amount estimation device.
Obtain reaction force related information regarding the pressing reaction force of the grinding tool during grinding,
The difference between the two reaction force-related information acquired at different repetitive times for repetitive machining at the same location of the workpiece is acquired.
Based on the difference, the correction amount in the pressing direction of the grinding tool is estimated.
The position of the grinding tool at the time of machining is controlled based on the target position and the correction amount.
Grinding method.