CN114616076A - Grinding system, correction amount estimation device, computer program, and grinding method - Google Patents
Grinding system, correction amount estimation device, computer program, and grinding method Download PDFInfo
- Publication number
- CN114616076A CN114616076A CN202080073652.1A CN202080073652A CN114616076A CN 114616076 A CN114616076 A CN 114616076A CN 202080073652 A CN202080073652 A CN 202080073652A CN 114616076 A CN114616076 A CN 114616076A
- Authority
- CN
- China
- Prior art keywords
- grinding
- correction amount
- reaction force
- machining
- related information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000012937 correction Methods 0.000 title claims abstract description 116
- 238000000034 method Methods 0.000 title claims description 21
- 238000004590 computer program Methods 0.000 title claims description 7
- 238000003754 machining Methods 0.000 claims abstract description 70
- 238000006243 chemical reaction Methods 0.000 claims abstract description 55
- 238000003825 pressing Methods 0.000 claims abstract description 33
- 238000004364 calculation method Methods 0.000 claims abstract description 15
- 230000007246 mechanism Effects 0.000 claims description 34
- 238000010801 machine learning Methods 0.000 claims description 26
- 238000012545 processing Methods 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 7
- 238000005498 polishing Methods 0.000 description 32
- 238000010586 diagram Methods 0.000 description 18
- 239000011324 bead Substances 0.000 description 10
- 230000008569 process Effects 0.000 description 8
- 230000001186 cumulative effect Effects 0.000 description 7
- 238000005259 measurement Methods 0.000 description 6
- 238000011156 evaluation Methods 0.000 description 5
- 230000001276 controlling effect Effects 0.000 description 4
- 238000005299 abrasion Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000002596 correlated effect Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 239000012636 effector Substances 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007429 general method Methods 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 229910052742 iron Inorganic materials 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 239000011347 resin Substances 0.000 description 1
- 229920005989 resin Polymers 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000003466 welding Methods 0.000 description 1
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B24—GRINDING; POLISHING
- B24B—MACHINES, DEVICES, OR PROCESSES FOR GRINDING OR POLISHING; DRESSING OR CONDITIONING OF ABRADING SURFACES; FEEDING OF GRINDING, POLISHING, OR LAPPING AGENTS
- B24B27/00—Other grinding machines or devices
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B24—GRINDING; POLISHING
- B24B—MACHINES, DEVICES, OR PROCESSES FOR GRINDING OR POLISHING; DRESSING OR CONDITIONING OF ABRADING SURFACES; FEEDING OF GRINDING, POLISHING, OR LAPPING AGENTS
- B24B49/00—Measuring or gauging equipment for controlling the feed movement of the grinding tool or work; Arrangements of indicating or measuring equipment, e.g. for indicating the start of the grinding operation
- B24B49/16—Measuring or gauging equipment for controlling the feed movement of the grinding tool or work; Arrangements of indicating or measuring equipment, e.g. for indicating the start of the grinding operation taking regard of the load
Landscapes
- Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Constituent Portions Of Griding Lathes, Driving, Sensing And Control (AREA)
- Numerical Control (AREA)
Abstract
A grinding system (100) having: a reaction force related information acquisition unit that acquires reaction force related information relating to a pressing reaction force of a grinding tool during grinding; a difference value calculation unit that acquires a difference value between the two pieces of reaction force related information acquired at different repetition times in repeated machining of the same portion of the workpiece; a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference value; and a machining control unit that controls a position of the grinding tool during machining based on the target position and the correction amount.
Description
Technical Field
The invention relates to a grinding system, a correction amount estimation device, a computer program and a grinding method.
Background
Prior art documents
Patent document
Patent document 1: japanese patent laid-open publication No. 2019-139755.
Disclosure of Invention
Problems to be solved by the invention
Grinding of a workpiece by a grinding machine is performed by repeatedly pressing and grinding a grinding tool against a machining site until a desired machining result is obtained. Further, since not only the workpiece but also the grinding tool is worn as the machining progresses, it is necessary to measure the machined workpiece to determine whether or not re-machining and conditions are required in order to ensure machining accuracy, and it is difficult to improve productivity.
The invention described in patent document 1 is an invention relating to this point of predicting the wear of a polishing tool by a machine learning device, and the wear of the polishing tool is estimated based on various machining conditions, but the amount of wear of the tool depends heavily on the state of the workpiece, and therefore it is difficult to accurately estimate the amount of wear of the tool by the invention.
The present invention has been made in view of the above circumstances, and an object thereof is to obtain a correction amount of a pressing direction reflecting a state of a workpiece in a grinding system.
Means for solving the problems
In order to solve the above problems, the invention disclosed in the present application has various aspects, and a summary of representative aspects among these aspects is as follows.
One aspect of the present invention relates to a grinding system having: a reaction force related information acquisition unit that acquires reaction force related information relating to a pressing reaction force of a grinding tool during grinding; a difference calculation unit that acquires a difference between the two pieces of reaction-force-related information, the difference being acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece; a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference value; and a machining control unit that controls a position of the grinding tool during machining based on the target position and the correction amount.
In the grinding system according to another aspect of the present invention, the correction amount estimation unit may input the difference value to a machine learning model to obtain the correction amount.
In addition, the grinding system according to another aspect of the present invention may further include: and an accumulated correction amount calculation unit that calculates an accumulated correction amount that is an accumulated value of the correction amount for the grinding wheel of the grinding tool, wherein the machining control unit controls the position of the grinding tool during machining based on a target position and the accumulated correction amount when machining a different workpiece.
In the grinding system according to another aspect of the present invention, the grinding system may divide a portion of the workpiece to be processed, and the correction amount estimating unit may estimate the correction amount for each division.
In the grinding system according to another aspect of the present invention, the grinding tool may be supported by a motor control support mechanism, and the reaction force related information may be an external force torque of the motor in a pressing direction of the grinding tool.
In the grinding system according to another aspect of the present invention, the grinding system may further include an estimated number of iterative processing times estimating unit that inputs the reaction force related information to a second machine learning model and estimates an estimated number of iterative processing times of the same portion of the workpiece.
In addition, a correction amount estimation device according to an aspect of the present invention includes: a difference value calculation unit that acquires a difference value between two pieces of reaction force-related information, which are acquired in different numbers of repetitions during repeated machining of the same portion of the workpiece, and a pressing reaction force of the grinding tool during the grinding machining; and a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference.
A computer program according to an aspect of the present invention causes a computer to function as a correction amount estimation device including: a difference value calculation unit that acquires a difference value between two pieces of reaction force-related information, which are acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece, and a pressing reaction force of the grinding tool during grinding; and a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference.
In addition, an aspect of the present invention relates to a grinding method including: the method includes acquiring reaction force related information related to a pressing reaction force of a grinding tool during grinding, calculating a difference between the two pieces of reaction force related information acquired at different numbers of repetitions during repeated machining of the same portion of a workpiece, estimating a correction amount in a pressing direction of the grinding tool based on the difference, and controlling a position of the grinding tool during machining based on a target position and the correction amount.
Drawings
Fig. 1 is a schematic diagram showing an overall configuration of an example of a grinding system according to a preferred embodiment of the present invention;
fig. 2 is a block diagram showing a functional configuration of the grinding system;
fig. 3 is a diagram showing an example of torque command values and their difference values obtained in the nth grinding and the (n + 1) th grinding as reaction force related information;
fig. 4 is a diagram showing an example of torque command values and their difference values obtained in the nth grinding and the (n + 1) th grinding as reaction force related information;
fig. 5 is a block diagram showing a functional configuration of a learning device using a machine learning model of a grinding system;
fig. 6 is a diagram showing an outline of a configuration for estimating the number of iterative processing in the grinding system;
fig. 7 is a diagram showing an outline of a system configuration;
fig. 8 is a diagram showing an abrasive target workpiece to be verified;
fig. 9 is a diagram showing a robot structure of the authentication system;
FIG. 10 is a table showing specifications of a disc grinder and a grinding wheel;
fig. 11 is a view showing an outline of a grinding operation;
fig. 12 is a view showing the appearance of a weld bead every 2 grinding times;
FIG. 13 shows the disturbance torques τ of the 2 nd, 4 th, 6 th, 8 th and 10 th polishing operationsdA graph of an example of the data of (1);
fig. 14 is a table showing the results of estimation using evaluation data from the NN model generated by learning;
fig. 15 is a table showing the estimated results when 10 polishing operations are performed while continuing the polishing operation of the same grinding wheel to the 11 th to 15 th workpieces;
fig. 16 is a block diagram showing a functional configuration of a grinding system including a repetition number estimation unit.
Detailed Description
Fig. 1 is a schematic diagram 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, and the support mechanism 3 supports the grinding tool 1 to be relatively movable with respect to a workpiece 2 and can grind the workpiece 2. Fig. 1 shows a support table 4 for supporting a workpiece 2 being processed and a controller 5 for controlling the support mechanism 3.
Here, the grinding tool 1 is a disc grinder (disc grinder) in the present embodiment, and includes a grinder body 6 and a grinding wheel 7 attached to a tip end thereof. In addition, a general multi-axis industrial robot is used as the support mechanism 3, and the grinding tool 1 is attached as an end effector thereof. The controller 5 is a robot controller.
Here, since the workpiece 2 is a member formed by overlaying a steel plate and the weld bead 8 is raised in a ridge shape, the grinding system 100 illustrated is intended to remove unnecessary raised portions of the weld bead 8 by grinding so as to smooth the surface thereof.
The grinding system 100 shown in fig. 1 is an example, and the gist of the present invention is not affected 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 a tool using a generally commercially available disk grinder, the grinding tool 1 dedicated to the grinding system 100 may be designed and manufactured. The support mechanism 3 is shown here as a structure using a vertical articulated robot, but various mechanisms such as a SCARA robot, an orthogonal robot, and a gantry mechanism may be used, and the support table 4 supporting the workpiece 2 may be movable, not only on the side supporting the grinding tool 1.
The Controller 5 may be not only a robot Controller but also any device for automatically controlling the support mechanism 3, such as a servo Controller, a Programmable Logic Controller (PLC), or a Personal Computer (PC), or may be a combination of a plurality of these devices. The object of the workpiece 2 is not limited to any one as long as it is an object to be ground, and may be an object requiring various kinds of grinding such as grinding and cutting of the inner surface or the outer surface of the rotating object, in addition to the plane grinding for removing unnecessary portions of the weld bead 8 shown in the figure.
Fig. 2 is a block diagram showing a functional configuration of the grinding system 100. The controller 5 is provided with a machining control unit 50 for moving the grinding tool 1 supported by the support mechanism 3 in a desired trajectory with respect to the workpiece 2 by performing position control based on program control on the support mechanism 3. In the present embodiment, since the support mechanism 3 is a motor control support mechanism driven by an electric motor, the machining control unit 50 includes a motor controller for driving the electric motor, and servo-controls each electric motor included in the support mechanism 3.
The support mechanism 3 is provided with a reaction force related information acquisition unit 30 that acquires reaction force related information relating to the pressing reaction force of the grinding tool 1 during grinding. Here, the reaction force-related information is information that can directly or by some conversion represent the pressing reaction force of the grinding tool 1, and may be the pressing reaction force itself or other information. The pressing reaction force itself is, for example, obtained by: the grinding tool 1 is attached to the support mechanism 3 via a load cell, and the pressing reaction force and the like are directly measured. As another information, in the grinding system 100 according to the present embodiment, a reaction torque acting on the motor of the support mechanism 3 in the pressing direction of the grinding tool 1, that is, a torque command value is used as the reaction-related information. If the control information of the electric motor, particularly the current command value and the torque command value, is adopted as the reaction force related information, it is not necessary to add a configuration for acquiring the reaction force related information to the support mechanism 3, and therefore, it is economical. In the present embodiment, a torque command value of the electric motor using the support mechanism 3 is assumed as the reaction force related information.
The controller 5 is provided with a correction unit 51 in addition to the machining control unit 50, and outputs an accumulated correction amount to the machining control unit 50 in the present embodiment. Here, before explaining the configuration of the correction unit 51 in detail, the meaning of the correction by the correction unit 51 will be explained.
In the grinding process using the grinding tool 1, abrasion of the grinding wheel 7 is inevitable as the process progresses. Further, since the abrasion of the grinding wheel 7 means the retreat of the machining point of the grinding process, when the grinding process is performed by controlling the position of the grinding tool 1 by the support mechanism 3, the machining point is shifted as the abrasion of the grinding wheel 7 progresses, and the desired shape of the workpiece 2 cannot be machined. 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 apply a correction commensurate therewith to the position control of the support mechanism 3.
The wear amount of the grinding wheel 7 is measured with high accuracy every time of machining, which makes the grinding system 100 complicated in installation and adjustment of a measuring machine, increases cost, and makes it difficult to perform maintenance and management. On the other hand, even if the amount of wear of the grinding wheel 7 is estimated, the amount of wear of the grinding wheel 7 greatly depends on the amount of grinding of the workpiece 2 during grinding, and therefore, the amount of wear greatly varies depending on the shape of the portion to be machined of the workpiece 2, and it is difficult to estimate the amount of wear only based on 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 the difference calculation unit 52 for calculating the difference between the 2 pieces of reaction force-related information acquired in the different numbers of repetitions of the repeated machining of the same portion of the workpiece 2.
In general, in grinding, since a desired shape of the workpiece 2 is rarely obtained by one grinding, the same portion is ground a plurality of times along the same trajectory until the desired shape is obtained.
Fig. 3 is a diagram showing an example of torque command values and their differences as reaction force related information obtained in the nth grinding and the (n + 1) th grinding in the grinding system 100 in the repeated machining of the same portion of the workpiece 2. In the upper graph of fig. 3, the nth torque command value is indicated by a solid line, and the (n + 1) th torque command value is indicated by a broken line, and in the lower graph of fig. 3, the difference obtained by subtracting the (n + 1) th torque command value from the nth torque command value is indicated by a 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 to be machined, the stronger the pressing reaction force generated on the support mechanism 3. Therefore, the waveform of the torque command value reflects the shape of the portion of the machining unit 2 that needs to be machined. Further, as the grinding process progresses, the portion to be processed is ground and the size thereof becomes smaller, and therefore the torque command value of the subsequent number of repetitions tends to be smaller than the torque command value of the previous number of repetitions.
Further, the difference in torque command values obtained at different numbers of iterations reflects: the difference between the shape of the portion to be machined in the preceding number of iterations and the shape of the portion to be machined in the subsequent number of iterations reflects only the amount of change in the shape of the portion to be machined during the period from the preceding number of iterations to the subsequent number of iterations. In the example shown in fig. 3, the difference shown in the following graph is correlated with the amount of change in the shape of the portion to be machined, which is generated 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 the 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 correction amount estimation algorithm may be any algorithm as long as it can reasonably estimate the correction amount based on the difference. The correction amount may be any correction amount as long as it is necessary to perform grinding with high accuracy including a shift of the machining point of the grinding tool 1 due to wear of the grinding wheel 7. In the present embodiment, the movement 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 the movement distance, an amount including mechanical variations such as deflection compensation of the support mechanism 3 due to a reaction force during machining may be used as the correction amount.
As described above, it is considered that the difference value and the necessary correction amount are correlated with each other as shown in the lower graph of fig. 3, but it is not easy to find a highly accurate conversion relationship between the two. Therefore, the correction amount estimation unit of the grinding system 100 according to the present embodiment obtains the correction amount as an output by inputting the difference value to a machine learning model in which the relationship between the difference value and the wear amount is learned in advance, using the machine learning model.
The architecture of the machine learning model is not particularly limited, but a structure based on an RNN (recurrent neural network) is preferable because the difference values as input values are time-series data. The learning with respect to the machine learning model is described later.
The correction amount thus obtained represents: the amount of wear of the grinding wheel 7 due to the grinding process between 2 different repetitions of the torque command value is obtained. Since the wear of the grinding wheel 7 is accumulated every time the grinding wheel is machined, the final correction of the position control of the support mechanism 3 is a correction based on the accumulated value of the correction amount.
Therefore, in the accumulated correction amount calculation unit 54 provided in the correction unit 51, every time the correction amount estimation unit 53 estimates the correction amount, the accumulated correction amount is accumulated and calculated and held, and the accumulated correction amount is output to the processing control unit 50.
The cumulative correction amount is an amount corresponding to the individual wear amount of the grinding wheel 7 of the grinding tool 1, and therefore is held as an amount for the grinding wheel 7 in use, and this value is reset when the grinding wheel 7 is replaced.
The machining control unit 10 receives the accumulated correction amount obtained as described above from the correction unit 51, and controls the position of the grinding tool 1 based on the accumulated correction amount in addition to the target position programmed in advance during grinding. More specifically, the accumulated correction amount is added to the coordinates of the target position of the grinding tool 1 in the pressing direction. Therefore, it can be said that the machining control unit 50 controls the position of the grinding tool 1 during machining based on the target position and the correction amount estimated by the correction amount estimation unit 53.
Since the integrated correction amount is an amount related to the grinding wheel 7 in use, it is needless to say that even when grinding of a certain workpiece 2 is completed and the workpiece 2 is replaced with a different workpiece 2, the integrated correction amount can be continuously used as long as the grinding wheel 7 is not replaced. Therefore, when machining a different workpiece 2, the machining control unit 50 controls the position of the grinding tool 1 during machining based on the target position and the accumulated correction amount.
In the controller 5 of the grinding system 100 described above, when the correction unit 51 is configured as an independent device, it can be considered as a correction amount estimation device. In this case, the correction amount estimation means is an independent apparatus having the difference calculation section 52, the correction amount estimation section 53, and the accumulated correction amount calculation section 54 described above.
The correction amount estimation device may be designed as a dedicated device, or may be implemented using a general computer. Fig. 5 is a diagram showing the configuration of a general computer 11 that can be used as the correction amount estimation device. The computer 11 is connected to 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 (input/Output) 306, which are capable of exchanging electrical signals with each other via a data bus 11 g. The hardware configuration of the computer 11 shown here is an example, and may be other configurations.
The external storage device 11c is a device capable of statically recording information, such as an HDD (Hard Disk Drive), SSD (Solid State Drive), or the like. The signal from the GC11d is output to a monitor 11h such as a CRT (Cathode Ray Tube) or a so-called flat panel display, on which a user visually recognizes an image, and displayed as an image. The input device 11e is one or more devices for a user to input information, such as a keyboard, a mouse, a touch panel, etc., and the I/O11f is one or more interfaces for the computer 11 to exchange information with external devices. The I/O11f may include various ports for wired connections and a controller for wireless connections.
Computer programs for causing the computer 11 to function as the machine learning data generation device 1 and the machine learning device 2 are stored in the external storage device 11c, read out to the RAM 11b as needed, and executed by the CPU 11 a. That is, the RAM 11b stores: codes of various functions shown as the correction section 51 of fig. 2 are realized by being executed by the CPU 11 a. The computer program may be recorded on a suitable computer-readable information recording medium such as a suitable optical disc, magneto-optical disc, or flash memory, or may be provided via an I/O11f via an information communication line such as the external internet.
Fig. 5 is a diagram showing an example of torque command values and their difference values as reaction force related information obtained in the n-th grinding and the n + 1-th grinding in the repeated machining of the same portion of the workpiece 2 in the case where the length of the machining target to be ground is long or the shape of the machining target portion is greatly changed during the machining.
In this case, the torque command values obtained in the same number of repetitions are divided into 3 sections, i.e., a section a, a section b, and a section c in the example shown in fig. 5, and similarly, the difference between the nth torque command value and the (n + 1) th torque command value is divided into 3 sections. Then, the difference value for each division is input to the correction amount estimation section 53, and as a result, the correction amount for each division is obtained.
The final accumulated correction amount becomes the accumulated value of the correction amount for each division thus obtained. In fig. 4, each division is performed with respect to time, but since the grinding is performed by the grinding tool 1 moving over the workpiece 2 with the passage of time, the division does not mean to divide the machining target portion of the workpiece 2. If the machining target portion of the workpiece 2 is divided appropriately in this manner and the correction amount is estimated for each division, the correction amount can be obtained with high accuracy in a case where the machining condition greatly varies during the grinding process.
For example, when the machining conditions are greatly different such as the size of the weld bead 8 being significantly different between the first half and the second half of the part to be machined of the workpiece 2, the correction amount can be estimated more accurately by estimating the correction amount in the first half and the second half which are considered to have substantially equal machining conditions, as compared with estimating the correction amount by summing them. The division of the machining target portion of the workpiece 2 may be performed at regular intervals, or may be performed at arbitrary intervals according to the properties of the workpiece 2.
The learning of the machine learning model used in the correction amount estimation unit 53 is not particularly limited, and any learning method of the machine learning model may be used as long as it is a method capable of learning the machine learning model in which the correction amount can be input and output with the difference value as described above. Therefore, an example of the learning method is described below.
Fig. 6 is a block diagram showing a functional configuration of the learning device 101 using a machine learning model of a grinding machine. In the figure, the same components as those of the grinding system 100 shown in fig. 2 are denoted by the same reference numerals, and redundant description thereof will be omitted.
The machine learning apparatus 101 is configured to: in addition to the supporting mechanism 3 and the controller 5 which are similar to the grinding system 100, the learning device 9 and the cumulative correction amount measuring device 10 are provided, the supporting mechanism 3 is provided with the grinding tool 1 and the reaction force related information acquiring unit 30, and the controller 5 is provided with the machining control unit 50 which controls the supporting mechanism 3.
In this example, the cumulative correction amount measuring instrument 10 is a measuring instrument that measures the amount of wear of the grinding wheel 7 of the grinding tool 1 attached to the support mechanism 3. The measurement of the wear amount was performed by the following method: the surface position of the grinding wheel 7 is measured by an arbitrary sensor, for example, a noncontact laser sensor or an arbitrary contact sensor, or the support mechanism 3 is driven to press the grinding tool 1 against a preset reference surface, and the coordinates of the support mechanism 3 at the time when the grinding wheel 7 comes into contact with the reference surface are detected. That is, the integrated correction amount measuring device 10 may be an independent device, or may be configured by using 1 or more devices of the machine learning device 101 in the related art.
Since the wear amount of the grinding wheel 7, that is, the cumulative value of the wear by the grinding process up to this time is measured, only the cumulative correction amount is actually measured in this example.
Therefore, when repeatedly machining an arbitrary workpiece 2, the actual measurement value of the reaction force related information at the time of machining in an arbitrary number of repetitions is acquired by the reaction force related information acquisition unit 30, and similarly, the actual measurement value of the accumulated correction amount can be obtained for the arbitrary number of repetitions.
The learner 9 has a difference value calculating section 90, a machine learning model 91, and a correction amount calculating section 92. The difference calculation unit 90 can calculate the difference between two pieces of reaction-force-related information acquired in different numbers of repetitions by inputting the reaction-force-related information acquired by the reaction-force-related-information acquisition unit 30, which is a torque command value here.
The correction amount calculation unit 92 can actually calculate the correction amount generated by the processing between the repetition times, based on the difference between the actually measured values of the two accumulated correction amounts acquired in the different repetition times.
Therefore, the learning unit 9 can make the machine learning model 91 learn by repeating the learning of the machine learning model 91 using the difference value obtained based on the actual measurement as input data and the correction amount obtained based on the actual measurement as correct data.
The hardware for implementing the learner 9 is not particularly limited. The learner 9 may be implemented as a part of the controller 5, or, for example, a general computer 11 shown in fig. 4 may be used as the learner 9. In the example of the machine learning device 101 shown in fig. 6, the learning unit 9 is directly connected to the support mechanism 3 and the cumulative correction amount measuring device 10, and the learning of the machine learning model 91 is performed every time the support mechanism 3 is driven to perform grinding, but in addition to this, a plurality of actual measurement values of the reaction force-related information and the cumulative correction amount may be acquired and accumulated in advance, and thereafter, the learning may be performed collectively by using the learning unit 9 which is configured independently.
The grinding system 100 may have the following configuration: when the same portion of the workpiece 2 is repeatedly machined, the estimated number of times of repeated machining is estimated, and the estimated number of times of repeated machining is an estimated value of the number of times of repeated machining required until the machining is completed. Hereinafter, the grinding system 100 is described as a configuration for estimating the estimated number of times of repeated machining.
< summary of the System >
Fig. 7 shows an outline of the system configuration. In this system, model generation for estimating the state of a work object is performed using control torque data collected while a robot performs a grinding work. A disc grinder for performing a grinding operation is attached to a fingertip of the robot. A vibration damping mechanism is provided between the disc grinder and the fingertip, and absorbs a certain degree of external force even if the disc grinder is in contact with the fingertip. The robot controller (hereinafter referred to as RC) collects data at the time of a job by MotoPlus application and stores the data in the USB memory. The stored data is used to learn by machine learning software in the PC. A neural network (hereinafter referred to as NN) including four layers including an input layer and an output layer is used as the learning model. And downloading the learned NN model to the RC through the USB memory. The RC inputs control torque data during operation to the NN, and estimates the operation state based on the output result.
(Note) Motoplus: developing functions of application software operating inside RC
< method for estimating work State >
In a general method for estimating fingertip force without a sensor, when a disturbance torque is set to taudIn this case, the fingertip force F can be calculated by equation 1 using the jacobian matrix J, and the disturbance torque is the difference between the control torque estimated from the operation trajectory and the actual control torque.
[ mathematical formula 1]
F＝(JT)-1τd
Fingertip force estimation
Since the state S of the contact operation of the inner pack tool is a system to which F is input, if the system function is G, equation 2 is obtained.
[ mathematical formula 2]
S＝G(F)
Job state estimation with fingertip force as input
Here, the disturbance torque τ is replaced with the disturbance torque τ as shown in equation 3dSet as the system function of the input.
[ mathematical formula 3]
Operating state estimation with disturbance torque as input
If it is to be
The NN model can be used to determine the input/output system by machine learning. A method of performing machine learning by learning weights of NN using a conventional error inverse propagation algorithm.
< technical evaluation >
The job state estimation using the present technology is evaluated. The estimated operating state is assumed to be a polishing completion state.
< work target workpiece >
The verified polishing target workpiece is shown in fig. 8. This is a welding operation performed on an iron plate having a thickness of 4mm, and a weld bead is arbitrarily applied and ground. The length of the weld bead is about 155mm and the height is about 2.5 mm.
< construction of verification System >
Fig. 9 shows a robot structure of the authentication system. The robot adopts a 6-axis vertical multi-joint robot MOTOMAN-GP7 (hereinafter, GP7), and the RC adopts YRC 1000. An electric disk grinder G10SH5 manufactured by HiKOKI was mounted on the top end of GP 7. The disk grinder can grind a rotating surface by pressing the rotating surface with a grinding wheel attached thereto. A resin Flexible Grinding wheel (Reginoid Flexible Grinding wheel) manufactured by HiKOKI corporation was used for the Grinding wheel. The weld bead as the object of polishing is fixed to a base provided in front of the robot, and the polishing operation is performed. The specifications of the disc grinder and the grinding wheel are shown in table 1 and table 2 of fig. 10, respectively.
< object of polishing work >
Fig. 11 is a schematic diagram of the polishing operation. The grinding is performed by the action of pulling the fingertip to the near side of the robot. This operation is previously taught, and is a task using only 3 axes (2 nd, 3 rd, and 5 th axes) among the axes using the GP 7.
Data were collected by actually performing the grinding. The grinding operation was performed 10 times for 1 pass. It has been empirically found that when a new grinding wheel is used, the grinding is substantially completed in the 10 grinding operations. After polishing of 10 workpieces is completed to acquire data of a new grinding wheel, the grinding wheel is replaced. Fig. 12 shows the appearance of the weld bead every 2 grinding times. Each polishing was performed in a state where the bead became flat and the polishing was completed by 10 times of polishing.
< evaluation results >
A total of 500 pieces of data were prepared, which were obtained by polishing 50 pieces of work 10 times. The total of 50 data obtained by grinding the 5 th workpiece from which the grinding wheel was replaced with a new one was used as evaluation data, and the remaining 450 data were learned. FIG. 13 shows the disturbance torque τ of the 2 nd, 4 th, 6 th, 8 th and 10 th polishing operationsdAn example of the data of (1).
The results estimated from the NN model generated by learning using the evaluation data are shown in table 3 of fig. 14. This represents the number of labels estimated for each positive label (in this case, the number of grinds) in a table called the mixing matrix. The gray boxes on the diagonal line represent the number of labels that can be accurately estimated. As a result, it can be said that 49 data out of the 50 data can be estimated within the range of the positive label or ± 1 time, and can be estimated with substantially high accuracy.
As a cause of failure in the estimation result, the following can be assumed. As a label for this estimation, the number of grinding times is used, but in practice, each time grinding is performed, the grinding wheel deteriorates, and even with the same number of grinding times, the quality of grinding varies, and the number of grinding times is mistaken for a larger number or smaller number. Actually, the estimated results when the polishing work with the same grindstone was continued until the 11 th to 15 th workpieces were subjected to the polishing operation 10 times are shown in table 4 of fig. 15, and it is understood that the polishing completion state of the workpieces was estimated to be not reached the more the polishing was continued with the same grindstone. In other words, it can be determined that sufficient grinding cannot be performed due to deterioration of the grinding wheel.
That is, it can be said that the polishing state is estimated every 1 polishing operation, and if it is estimated that the polishing operation is stopped for the 10 th polishing operation, the polishing operation can be automatically terminated by an appropriate number of times of polishing.
With the above-described configuration, the estimated number of repetitions estimating unit 55 that estimates the estimated number of repetitions of machining can be configured. Fig. 16 is a block diagram showing a functional configuration of the grinding system 100 including the repetition number estimating unit 55. The respective configurations of the grinding system 100 according to the present example are basically the same as those shown in fig. 2, and therefore the same components are denoted by the same reference numerals and redundant description thereof is omitted. The only difference from that shown in fig. 2 is that the controller 5 is provided with an estimated number of repetitions estimating unit 55.
The estimated number of repetitions estimation unit 55 estimates the estimated number of repetitions of the same portion of the workpiece 2 in this example, and therefore the estimated number of repetitions obtained can contribute to estimation of the remaining time required for the current machining and determination of the timing for replacing the grinding wheel 7.
Claims (9)
1. A grinding system having:
a reaction force related information acquisition unit that acquires reaction force related information relating to a pressing reaction force of a grinding tool during grinding;
a difference value calculation unit that acquires a difference value between the two pieces of reaction-force-related information, the difference value being a difference value between the two pieces of reaction-force-related information acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece;
a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference value; and
and a machining control unit that controls a position of the grinding tool during machining based on the target position and the correction amount.
2. The grinding system of claim 1,
the correction amount estimation unit obtains the correction amount by inputting the difference value to a machine learning model.
3. A grinding system according to claim 1 or 2,
having an accumulated correction amount calculation section that calculates an accumulated correction amount that is an accumulated value of the correction amount for a grinding wheel of the grinding tool,
the machining control unit controls a position of the grinding tool during machining based on a target position and the accumulated correction amount when machining a different workpiece.
4. A grinding system according to any one of claims 1 to 3,
the grinding system divides a portion of the workpiece to be processed, and the correction amount estimation unit estimates the correction amount for each division.
5. A grinding system according to any one of claims 1 to 4,
the grinding tool is supported by a motor control support mechanism,
the reaction force related information is an external force torque of the motor in a pressing direction of the grinding tool.
6. A grinding system according to any of claims 1 to 5 having:
and an estimated number-of-iterations estimating unit that inputs the reaction force related information to a second machine learning model and estimates an estimated number of iterations of the same portion of the workpiece.
7. A correction amount estimation device has:
a difference value calculation unit that acquires a difference value between two pieces of reaction force-related information, which are acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece, and a pressing reaction force of the grinding tool during grinding; and
a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference.
8. A computer program for causing a computer to function as a correction amount estimation device, the correction amount estimation device comprising:
a difference value calculation unit that acquires a difference value between two pieces of reaction force-related information, which are acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece, and a pressing reaction force of the grinding tool during grinding; and
a correction amount estimation section that estimates a correction amount in a pressing direction of the grinding tool based on the difference.
9. A method of grinding comprising:
reaction force related information relating to a pressing reaction force of a grinding tool in grinding processing is acquired,
acquiring a difference value between two pieces of the reaction force related information acquired in different numbers of repetitions in repeated machining of the same portion of the workpiece,
estimating a correction amount in a pressing direction of the grinding tool based on the difference,
controlling a position of the grinding tool at the time of machining based on the target position and the correction amount.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US201962940912P | 2019-11-27 | 2019-11-27 | |
| US62/940,912 | 2019-11-27 | ||
| PCT/JP2020/044359 WO2021107139A1 (en) | 2019-11-27 | 2020-11-27 | Grinding system, correction amount estimation device, computer program, and grinding method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114616076A true CN114616076A (en) | 2022-06-10 |
Family
ID=76129657
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202080073652.1A Pending CN114616076A (en) | 2019-11-27 | 2020-11-27 | Grinding system, correction amount estimation device, computer program, and grinding method |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP7294448B2 (en) |
| CN (1) | CN114616076A (en) |
| WO (1) | WO2021107139A1 (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117742239B (en) * | 2024-02-19 | 2024-05-14 | 南京超颖新能源科技有限公司 | Vertical correction system and correction method for machine tool |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4967127A (en) * | 1988-06-30 | 1990-10-30 | Toyota Jidosha Kabushiki Kaisha | Robot with controlled tool tracking displacement |
| US5126645A (en) * | 1989-09-30 | 1992-06-30 | Kabushiki Kaisha Toshiba | Grinder robot |
| JP2017148908A (en) * | 2016-02-25 | 2017-08-31 | セイコーエプソン株式会社 | Control device, robot, and robot system |
| JP2018206286A (en) * | 2017-06-09 | 2018-12-27 | 川崎重工業株式会社 | Operation prediction system and operation prediction method |
| JP2019162712A (en) * | 2018-03-20 | 2019-09-26 | ファナック株式会社 | Control device, machine learning device and system |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH074765B2 (en) * | 1986-03-03 | 1995-01-25 | 長尾 高明 | Curved surface processing equipment |
| JPS6347058A (en) * | 1986-08-13 | 1988-02-27 | Daikin Ind Ltd | Control device for robot |
| US4886529A (en) * | 1988-07-08 | 1989-12-12 | Showa Precision Machinery Co., Ltd. | Polishing robot and polishing method using the same |
| JPH07205022A (en) * | 1993-12-30 | 1995-08-08 | Hitachi Constr Mach Co Ltd | Correcting method for grinding wheel wear in force control robot |
| JP2011125986A (en) * | 2009-12-21 | 2011-06-30 | Saito-Seiki Co Ltd | Surface grinding device of square block |
| JP6136337B2 (en) * | 2013-02-15 | 2017-05-31 | 株式会社Ihi | Attitude detection control apparatus, polishing apparatus, and attitude detection control method |
| JP6347058B2 (en) | 2014-03-11 | 2018-06-27 | 産業振興株式会社 | Spacer loading device |
| JP6088583B2 (en) * | 2015-06-08 | 2017-03-01 | ファナック株式会社 | Robot controller with robot and force display function |
| WO2020032109A1 (en) | 2018-08-08 | 2020-02-13 | 株式会社クレハ | Molded resin object, method for producing molded resin object, and sterilization method |
-
2020
- 2020-11-27 WO PCT/JP2020/044359 patent/WO2021107139A1/en active Application Filing
- 2020-11-27 JP JP2021561572A patent/JP7294448B2/en active Active
- 2020-11-27 CN CN202080073652.1A patent/CN114616076A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4967127A (en) * | 1988-06-30 | 1990-10-30 | Toyota Jidosha Kabushiki Kaisha | Robot with controlled tool tracking displacement |
| US5126645A (en) * | 1989-09-30 | 1992-06-30 | Kabushiki Kaisha Toshiba | Grinder robot |
| JP2017148908A (en) * | 2016-02-25 | 2017-08-31 | セイコーエプソン株式会社 | Control device, robot, and robot system |
| JP2018206286A (en) * | 2017-06-09 | 2018-12-27 | 川崎重工業株式会社 | Operation prediction system and operation prediction method |
| JP2019162712A (en) * | 2018-03-20 | 2019-09-26 | ファナック株式会社 | Control device, machine learning device and system |
Also Published As
| Publication number | Publication date |
|---|---|
| JP7294448B2 (en) | 2023-06-20 |
| JPWO2021107139A1 (en) | 2021-06-03 |
| WO2021107139A1 (en) | 2021-06-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP1834738B1 (en) | Teaching position correcting apparatus and teaching position correction method | |
| CN110116371B (en) | Abrasive tool wear amount prediction device, machine learning device, and system | |
| JP6836577B2 (en) | Abrasive tool wear prediction device, machine learning device and system | |
| CN104608114B (en) | Robot system and method for producing to-be-processed material | |
| JP2021058975A (en) | Surface finishing device | |
| EP3508313B1 (en) | Control method of positioning control device and positioning control device | |
| JP6781242B2 (en) | Controls, machine learning devices and systems | |
| US11707842B2 (en) | Robot system and coordinate conversion method | |
| CN112676925B (en) | Method, device and equipment for grinding and polishing blade of aero-engine and storage medium | |
| Fajardo-Pruna et al. | Analysis of a single-edge micro cutting process in a hybrid parallel-serial machine tool | |
| US7512457B2 (en) | In-process non-contact measuring systems and methods for automated lapping systems | |
| CN114616076A (en) | Grinding system, correction amount estimation device, computer program, and grinding method | |
| US20150336201A1 (en) | Spot welding system and spot welding method | |
| CN110174871B (en) | Control device, machine learning device, and system | |
| CN110154043A (en) | The robot system and its control method of study control are carried out based on processing result | |
| Verbert et al. | Obtainable accuracies and compensation strategies for robot supported SPIF | |
| JP2002346899A (en) | Curve surface working method and working device | |
| CN112743509A (en) | Control method and computing device | |
| JP2005088106A (en) | Method and apparatus for calculating working data | |
| Phan et al. | Instrumentation of a grinding tool for capturing dynamic interactions with the workpiece | |
| CN112105489B (en) | Abnormality detection device and abnormality detection method | |
| US11833695B2 (en) | Diagnosis of state of device | |
| Sufian | Geometrical Error Analysis and Correction in Robotic Grinding | |
| JPH0719982A (en) | Inspection method for force sensor | |
| WO2022154021A1 (en) | Gap detection device and gap detection method for robot joint |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |