WO2020090030A1 - 数値制御装置、学習装置および学習方法 - Google Patents

数値制御装置、学習装置および学習方法 Download PDF

Info

Publication number
WO2020090030A1
WO2020090030A1 PCT/JP2018/040500 JP2018040500W WO2020090030A1 WO 2020090030 A1 WO2020090030 A1 WO 2020090030A1 JP 2018040500 W JP2018040500 W JP 2018040500W WO 2020090030 A1 WO2020090030 A1 WO 2020090030A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
machine tool
learning
temperature
thermal displacement
Prior art date
Application number
PCT/JP2018/040500
Other languages
English (en)
French (fr)
Japanese (ja)
Inventor
一樹 高幣
佐藤　剛
俊博 東
Original Assignee
三菱電機株式会社
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 三菱電機株式会社 filed Critical 三菱電機株式会社
Priority to DE112018008027.7T priority Critical patent/DE112018008027T5/de
Priority to JP2019516736A priority patent/JP6556413B1/ja
Priority to CN201880098973.XA priority patent/CN112912803A/zh
Priority to PCT/JP2018/040500 priority patent/WO2020090030A1/ja
Publication of WO2020090030A1 publication Critical patent/WO2020090030A1/ja

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/18—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form
    • G05B19/404—Numerical control [NC], i.e. automatically operating machines, in particular machine tools, e.g. in a manufacturing environment, so as to execute positioning, movement or co-ordinated operations by means of programme data in numerical form characterised by control arrangements for compensation, e.g. for backlash, overshoot, tool offset, tool wear, temperature, machine construction errors, load, inertia
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/30—Nc systems
    • G05B2219/33—Director till display
    • G05B2219/33034—Online learning, training
    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B2219/00—Program-control systems
    • G05B2219/30—Nc systems
    • G05B2219/49—Nc machine tool, till multiple
    • G05B2219/49219—Compensation temperature, thermal displacement

Definitions

  • the present invention relates to a numerical control device, a learning device, and a learning method for estimating a thermal displacement amount of a machine tool.
  • Machine tools are processing devices that perform removal processing and bending processing by applying force or energy to the work using tools.
  • Machine tools generally have multiple drive shafts.
  • the drive shaft is composed of, for example, a motor and one or more structures connected to the motor.
  • a machine tool includes a main shaft that is a drive shaft for rotating a tool or a work, and a feed shaft that is a drive shaft that positions a relative position between the tool and the work.
  • Each drive shaft of the machine tool is controlled by a numerical controller.
  • the numerical control device instructs the relative position of the tool with respect to the work to the drive shaft, and the drive shaft operates based on the command, whereby the tool and the work come into contact with each other and machining is realized.
  • thermal displacement occurs between the tool and the workpiece.
  • Thermal displacement is an error in machine position caused by changes in temperature of a machine tool. Examples of typical factors that give a temperature change to a machine tool include a change in ambient temperature of the machine tool and heat generation of a motor. Due to these factors, if the temperature distribution of the structure including the column and the spindle head becomes uneven, the structure is distorted, and the parallelism and squareness of the machine tool decrease. Further, the main shaft shaft, the ball screw, and the like expand due to temperature changes, causing thermal displacement.
  • ⁇ Thermal displacement causes machining error.
  • measures for suppressing the thermal displacement of the machine tool there are measures such as installing the machine tool in a temperature-controlled room so as not to change the temperature of the machine tool, or providing the machine tool with a cooling device. By taking these measures, the temperature of the machine tool can be kept constant irrespective of the inside and outside conditions of the machine tool, but it is necessary to prepare a temperature-controlled room for the machine tool or cooling by a cooling device. If there is a structure that requires, there is a problem that the thermal displacement due to thermal deformation of the structure cooled by the cooling device cannot be prevented.
  • Patent Document 1 a calculation formula representing the relationship between the operation state data of the machine tool and the thermal displacement amount is learned, the thermal displacement amount is calculated using this calculation formula and the operation state data, and the calculated thermal displacement amount is calculated. There has been proposed a technique for correcting the machine position of a machine tool using.
  • the amount of thermal displacement is calculated by repeatedly learning calculation formulas for various operating states and summing the amount of thermal displacement estimated in sampling time units within a predetermined period.
  • the operation state data in a unit of a constant period is always used as the learning data without being synchronized with the operation of the machine tool. If data with different trends of displacement amount with time is included, learning will not be performed correctly.
  • the tendency of temperature change inside and outside the machine tool that affects thermal displacement also changes.
  • the tendency of temperature change inside and outside the machine tool that affects the thermal displacement is referred to as a thermal tendency.
  • Specific examples of the change in the condition that causes the change in the thermal tendency include the case where the motor changes from the rotating state to the stationary state, the case where the coolant for cooling changes from the discharge state to the stop state, and the like.
  • the present invention has been made in view of the above, and an object of the present invention is to obtain a numerical control device capable of accurately learning the relationship between temperature and thermal displacement.
  • a numerical control device for controlling a machine tool, wherein the numerical control device has a relationship between a temperature of the machine tool and a thermal displacement amount of the machine tool.
  • the first time length which is the time length of the learning data used for learning, is determined based on the operation data indicating the content of the operation operation of the machine tool, and the machine tool of the machine tool is determined based on the determined first time length.
  • a data generation unit that generates learning data including temperature data indicating temperature, displacement data indicating displacement of the machine tool, and operation data.
  • the numerical controller further includes a learning unit that uses the learning data to learn the relationship between the temperature of the machine tool and the amount of thermal displacement of the machine tool to generate a learning model.
  • the numerical control device has the effect of being able to accurately learn the relationship between temperature and thermal displacement.
  • Block diagram showing a configuration example of an embodiment of a numerical control device according to the present invention Diagram showing a configuration example of a machine tool Diagram showing a configuration example of the processing circuit Diagram showing changes in the state of typical machine elements that affect the thermal tendency of machine tools
  • Neural network model diagram Flowchart showing an example of a learning processing procedure in the numerical control device Flowchart showing an example of the procedure for estimating the thermal displacement amount in the numerical control device
  • a numerical control device, a learning device, and a learning method according to an embodiment of the present invention will be described below in detail with reference to the drawings.
  • the present invention is not limited to this embodiment.
  • FIG. 1 is a block diagram showing a configuration example of an embodiment of a numerical controller according to the present invention.
  • the numerical control device 1 of the present embodiment can be connected to the machine tool 2.
  • FIG. 1 shows a state in which the numerical control device 1 is connected to the machine tool 2.
  • the numerical controller 1 controls the machine tool 2 based on the machining program 3.
  • the machining program 3 is a series of command statements given to the numerical controller 1 for the machine tool 2 to execute a desired operation.
  • FIG. 1 shows an example in which the machining program 3 is given from the outside of the numerical controller 1, the present invention is not limited to this, and the machining program may be held inside the numerical controller 1.
  • FIG. 2 is a diagram showing a configuration example of the machine tool 2.
  • the machine tool 2 is a processing device that performs cutting
  • the machine tool 2 to which the learning method of the present embodiment is applied is not limited to a processing device that performs cutting.
  • the number of drive shafts of the machine tool 2 is not limited to the example shown in FIG.
  • the machine tool 2 includes a main shaft 21, feed shafts 22-1 to 22-3, a coolant device 27, a cooling device 28, a temperature sensor 29, and a displacement sensor 30.
  • the main spindle 21 rotates based on a command from the numerical controller 1, and the tool attached to the main spindle 21 rotates together with the rotation of the main spindle 21, so that the workpiece fixed on the table (not shown) is rotated.
  • the feed shafts 22-1 to 22-3 operate based on a command from the numerical control device 1 so that the relative position between the tool and the work becomes the commanded position.
  • the spindle 21 includes a mechanical element 23, which is one or more structures, and a spindle motor 24.
  • the mechanical element 23 is, for example, a gear, a shaft, a tooling system, or the like.
  • the feed shaft 22-1 includes a feed shaft mechanism 25-1 and a feed shaft motor 26-1.
  • the feed shaft 22-2 includes a feed shaft mechanism 25-2 and a feed shaft motor 26-2.
  • the feed shaft 22-3 includes a feed shaft mechanism 25-3 and a feed shaft motor 26-3.
  • each of the feed shafts 22-1 to 22-3 will be referred to as a feed shaft 22 when not individually distinguished, and each of the feed shaft mechanisms 25-1 to 25-3 will be described as a feed shaft mechanism 25 unless individually distinguished.
  • each of the feed shaft motors 26-1 to 26-3 will be referred to as the feed shaft motor 26 unless individually distinguished.
  • the feed shaft mechanism 25 is, for example, a coupling, a ball screw, or a table.
  • the feed shaft 22-1 is a drive shaft that determines a position in the X-axis direction
  • the feed shaft 22-2 is a drive shaft that determines a position in the Y-axis direction
  • the feed shaft 22-3 is a position in the Z-axis direction. Is the drive shaft that determines.
  • the tool axis direction is made to coincide with the Z axis
  • the tool advancing direction in a plane perpendicular to the tool axis direction is the X axis
  • the X axis and the Z axis are perpendicular to each other.
  • the direction is defined as the Y-axis.
  • the cooling device 28 cools at least a part of the spindle motor 24, the feed shaft motors 26-1 to 26-3, the mechanical element 23, and the feed shaft mechanisms 25-1 to 25-3.
  • the coolant device 27 cools the processing part.
  • the processing unit is a processing area in the machine tool 2 in which the tool processes a work.
  • the temperature sensor 29 periodically detects the temperature of the machine tool 2 and outputs temperature data indicating the detected temperature to the numerical controller 1.
  • the detection cycle of the data of the temperature sensor 29 is called a temperature detection cycle.
  • one temperature sensor 29 is shown in FIG. 2, generally, there are a plurality of temperature sensors 29, and the temperature sensors 29 are installed at a plurality of locations inside and outside the machine tool.
  • the temperature sensor 29 detects the temperature, for example, one or more of the main shaft motor 24, the feed shaft motors 26-1 to 26-3, the mechanical element 23, and the feed shaft mechanisms 25-1 to 25-3. Structure, the tank (not shown) of the coolant device 27, the periphery of the machine tool 2, and the like.
  • the temperature detected by the temperature sensor 29 is not the temperature of the structure itself of the machine tool 2, but in this specification, in such a case.
  • the temperature including the machine tool 2 is also called.
  • the location where the temperature of the temperature sensor 29 is to be detected is, for example, a table, bed, column, spindle head, or the like.
  • a plurality of temperature sensors 29 may be installed for one structure.
  • the displacement sensor 30 detects the amount of displacement that occurs between the tool and the workpiece in the machining section of the machine tool 2, and outputs the detection result to the numerical controller 1 as thermal displacement data.
  • the displacement sensor 30 is a sensor capable of detecting displacement in at least one axis direction.
  • a plurality of displacement sensors 30 may be installed to detect displacements in a plurality of axial directions.
  • the machine tool 2 is provided with one or more sensors (not shown) for detecting the operating state of the machine tool 2, and the sensor uses the detection result of the operating state of the machine tool 2 as operating state data to control the numerical controller 1.
  • the operating state data is information including at least one of the position and speed of each motor of the main shaft motor 24 and the feed shaft motors 26-1 to 26-3 and the current.
  • the thermal displacement amount is a displacement amount generated between the tool and the work due to the structure of the machine tool 2 being distorted or elongated due to the influence of temperature changes inside and outside the machine tool 2.
  • Factors that cause a temperature change in the machine tool 2 are heat generated by driving each motor of the machine tool 2, frictional heat of each drive shaft of the machine tool 2, cooling by the coolant device 27, cutting heat generated by cutting, cooling by the cooling device 28. , Ambient temperature, etc.
  • the numerical controller 1 estimates the thermal displacement amount of the machine tool 2 with respect to the control command for controlling the machine tool 2 and adds a correction amount that cancels the estimated thermal displacement amount to the thermal command.
  • a driving operation command which is a command for correcting the displacement amount, is output to the machine tool 2. If the amount of thermal displacement cannot be accurately estimated, the numerical control device 1 will also have a reduced accuracy of correction with respect to the command, resulting in a processing error.
  • the numerical controller 1 according to the present embodiment generates learning data for each operation unit described later and learns the thermal displacement amount so that the thermal displacement amount can be accurately estimated.
  • the operation unit is a time period in which the thermal tendency of the machine tool 2 is constant. Details of the operation unit will be described later.
  • the configuration and operation of the numerical control device 1 according to the present embodiment will be described.
  • the numerical control device 1 includes a data collection unit 11, a learning unit 12, a control unit 13, a data selection unit 14, a thermal displacement estimation unit 15, and a thermal displacement correction unit 16.
  • the data collection unit 11 which is a data generation unit, uses the first time length, which is the time length of the learning data used for learning the relationship between the temperature of the machine tool 2 and the thermal displacement amount of the machine tool 2, as the operation data. Based on the first time length, learning data including temperature data, displacement data, and operation data is generated.
  • the operation data is data indicating the content of the operation operation of the machine tool 2, and is information including the analysis result of the machining program 3, the operation state data, and the control command.
  • the data collection unit 11 receives the temperature data output from the temperature sensor 29 of the machine tool 2 and the thermal displacement data output from the displacement sensor 30 of the machine tool 2.
  • the data collection unit 11 also receives the operation data from the control unit 13.
  • the control command is a command for the tool and the work to perform a desired operation by each motor of the machine tool 2.
  • the analysis result of the machining program 3 is information indicating the operation of the machine tool 2, including the rotation speed of the spindle motor 24 and the speed of the feed shaft motor 26.
  • the analysis result includes, for example, a rotation speed command (rotation speed command) to the main shaft 21, a position command to the feed shaft 22, a speed command to the feed shaft 22, a tool number, a coolant injection command, and Information indicating at least one of the coolant stop command is included.
  • the data collection unit 11 generates learning data by using the temperature data, the thermal displacement data and the operation data, and outputs the learning data to the learning unit 12. Specifically, the data collection unit 11 uses the information of the operation data to divide the temperature data and the thermal displacement data into operation operation units to generate divided data, and the number of sampling points in the divided data having different time widths is constant.
  • the learning data is generated by performing resampling so that the number of sampling points becomes. Details of resampling will be described later. Here, it is assumed that all the cycles of inputting each data to the data collecting unit 11 are the same, and the times corresponding to each data input to the data collecting unit 11 are all the same in this cycle.
  • the data collection unit 11 monitors the driving data, and when there is a change in the driving data that satisfies the condition of the break of the driving operation unit, which will be described later, the changing point is set as the end of the learning data, and then the learning data.
  • the input temperature data, thermal displacement data, and operation data are set as the beginning of the next learning data.
  • the data collection unit 11 may hold the temperature data, the thermal displacement data, and the operation data together with the time. This time is the time attached to the data when each of the temperature data, the thermal displacement data, and the operation data is attached to the data, and the data collection unit when the time is not attached to the data. 11 is the time when the data is received.
  • the learning unit 12 uses the learning data received from the data collection unit 11 to learn the relationship between the temperature of the machine tool 2 and the thermal displacement amount of the machine tool 2, generates a learning model, and generates the learning model. Output to the thermal displacement estimation unit 15.
  • the control unit 13 analyzes the command described in the machining program 3. Further, the control unit 13 receives the operation state data from the machine tool 2, and based on the operation state data and the command described in the machining program 3, each of the spindle motor 24 and the feed axis motors 26-1 to 26-3. Generates a control command for performing an operation corresponding to the command described in the machining program 3.
  • a general method can be used as a method of generating a control command relating to the operation of each motor of the main shaft motor 24 and the feed shaft motors 26-1 to 26-3, and a detailed description thereof will be omitted.
  • the control unit 13 outputs the generated control command to the thermal displacement correction unit 16. Further, the control unit 13 outputs the operation data, which is information including the analysis result of the machining program 3, the operation state data, and the control command, to the data collection unit 11 and the data selection unit 14.
  • the data selection unit 14 determines a second time length, which is the time length of the estimation data used for estimating the thermal displacement, based on the operation data, and based on the determined second time length, the temperature data and the operation are calculated. Data for estimation including data and is generated. Specifically, the data selection unit 14 receives and acquires temperature data from the temperature sensor 29 of the machine tool 2. Further, the data collection unit 11 acquires operation data from the control unit 13. The data selection unit 14 temporally divides the temperature data in a section of a driving action unit analyzed from the driving data, and outputs the divided temperature data to the thermal displacement estimation unit 15 as estimation data. At this time, similarly to the data collection unit 11, the data selection unit 14 resamples so that the number of sampling points in the divided data having different time widths becomes a constant number of sampling points to generate estimation data.
  • the thermal displacement estimation unit 15 estimates the amount of thermal displacement generated in the machine tool 2 using the learning model generated by the learning unit 12 and the estimation data received from the data selection unit 14. Specifically, the thermal displacement estimation unit 15 calculates the thermal displacement amount by inputting the estimation data input from the data selection unit 14 to the calculation formula indicating the learning model generated by the learning unit 12. .. In addition, the thermal displacement estimation unit 15 performs the process opposite to the resampling performed by the data selection unit 14 on the calculated thermal displacement amount, so that the thermal displacement estimation unit 15 outputs the thermal displacement amount output cycle to the temperature sensor. The thermal displacement amount after the process opposite to the resampling is output to the thermal displacement correction unit 16 in accordance with the temperature detection period of 29.
  • the thermal displacement correction unit 16 corrects the control command to the machine tool 2 with the correction amount calculated from the thermal displacement amount estimated by the thermal displacement estimation unit 15. Specifically, the thermal displacement correction unit 16 adds a correction amount that cancels the thermal displacement amount estimated by the thermal displacement estimation unit 15 to the control command generated by the control unit 13, and the correction amount is added.
  • the control command is output to the machine tool 2 as a driving operation command. Specifically, the thermal displacement correction unit 16 multiplies the amount of thermal displacement in each axial direction estimated by the thermal displacement estimation unit 15 by ⁇ 1 with respect to the position command in each axial direction included in the control command. Is added as a correction amount.
  • the data collection unit 11, the learning unit 12, the control unit 13, the data selection unit 14, the thermal displacement estimation unit 15, and the thermal displacement correction unit 16 illustrated in FIG. 1 are realized by a processing circuit.
  • the processing circuit may be a circuit including a processor or may be dedicated hardware.
  • FIG. 3 is a diagram illustrating a configuration example of the processing circuit.
  • the processing circuit 200 includes a processor 201 and a memory 202.
  • the processor 201 causes the memory 202 to operate. These are realized by reading and executing the program stored in. That is, when the data collection unit 11, the learning unit 12, the control unit 13, the data selection unit 14, the thermal displacement estimation unit 15, and the thermal displacement correction unit 16 are realized by the processing circuit 200 shown in FIG. , Is realized by using a program that is software.
  • the memory 202 is also used as a work area of the processor 201.
  • the processor 201 is a CPU (Central Processing Unit) or the like.
  • the memory 202 corresponds to, for example, a RAM (Random Access Memory), a ROM (Read Only Memory), a nonvolatile or volatile semiconductor memory such as a flash memory, a magnetic disk, or the like.
  • the processing circuit that realizes the data collection unit 11, the learning unit 12, the control unit 13, the data selection unit 14, the thermal displacement estimation unit 15, and the thermal displacement correction unit 16 is dedicated hardware
  • the processing circuit is, for example, an FPGA ( These are Field Programmable Gate Array) and ASIC (Application Specific Integrated Circuit).
  • the data collection unit 11, the learning unit 12, the control unit 13, the data selection unit 14, the thermal displacement estimation unit 15, and the thermal displacement correction unit 16 may be realized by combining a processing circuit including a processor and dedicated hardware. ..
  • the data collection unit 11, the learning unit 12, the control unit 13, the data selection unit 14, the thermal displacement estimation unit 15, and the thermal displacement correction unit 16 may be realized by a plurality of processing circuits.
  • FIG. 4 is a diagram showing changes in the states of typical machine elements that affect the thermal tendency of the machine tool 2.
  • FIG. 4 shows an example of a temporal transition between the operating state and the programmed operating state of each of the main shaft 21, the feed shaft 22, the coolant device 27, and the cooling device 28, which are the components of the machine tool 2.
  • the horizontal axis of FIG. 4 indicates time.
  • the program operating state indicates whether or not the program is operating.
  • the program operation is to operate the machine tool 2 under the control according to the machining program 3 by the numerical controller 1.
  • a program operation for performing a certain machining is performed by the machine tool 2, and when the program operation is completed, a worker performs work such as replacement of a work, and then a program for the next machining. Work is performed in a flow such as driving.
  • tool # 1 is mounted on the spindle 21, and the spindle motor 24 rotates at a rotation speed of 1000 rotations per minute.
  • a character represented by S and a number, such as “S1000”, indicates the rotation speed of the spindle motor 24, S indicates that the spindle motor 24, and the numerical value following S indicates the rotation speed. Is shown.
  • the rotation speed is indicated by the number of rotations, which is the number of rotations per minute. From time t1 to time t3, the shaft of the main shaft 21 expands due to heat generation of the main shaft motor 24 and friction heat of the bearing of the main shaft 21.
  • the spindle motor 24 is in a stationary state, so the spindle 21 itself does not generate heat.
  • the tool # 2 is attached to the spindle 21 and rotates at a rotation speed of 3000 rotations per minute.
  • the spindle 21 also generates heat from time t3 to t4, thermal displacement occurs, but the tendency of heat generation is different from time t1 to time t3 because of the different tool and different rotation speed from time t1 to time t3. Is different.
  • the thermal tendency of the main shaft 21 is not always constant, but varies depending on the mounted tool, the rotation speed, and the like.
  • the feed shaft 22, the coolant device 27, and the cooling device 28 also change the operating state, so that the thermal tendency differs depending on the operating action.
  • the feed shaft motor 26 rotates at a rotation speed of 100 revolutions per minute from the time t1 to the time t2, and the feed shaft motor 26 per minute from the time t2 to the time t3. It rotates at a rotation speed of 200 rotations.
  • a character represented by F and a number, such as “F100” indicates the rotation speed of the feed shaft motor 26, F is the feed shaft motor 26, and the numerical value following F is The rotation speed is shown.
  • the heat generation amount is different. Further, the temperature of the portion cooled by the coolant and its surroundings changes depending on whether or not the coolant device 27 is discharging the coolant. The temperature of the portion cooled by the cooling device 28 and its surroundings also change depending on whether or not the cooling device 28 is in operation.
  • the program operation is completed after the time t7, there is a possibility that setup change such as opening / closing of the door of the processing section and replacement of the work may be performed, and the thermal environment of the machine tool 2 is that the program operation is during the program operation. different. From the above, the tendency of thermal displacement generated in the machine tool 2 is different for each section from the time t1 to the time t2, the section from the time t2 to the time t3, ..., And the section where the program operation is stopped after the time t7. ..
  • the operation operation of the machine tool 2 such as the section from time t1 to time t2, the section from time t2 to time t3, ...
  • Each different section is defined as a driving operation unit.
  • the data collection unit 11 of the numerical control device 1 determines a section that is a driving operation unit based on the driving data received from the control unit 13. More specifically, the time at which the content of the driving action has changed is obtained from the driving data, and the section that is the driving action unit is determined by using this time as a delimiter. That is, the first time length, which is the time length of the learning data, is determined with the timing at which the operating state of at least one machine element of the machine tools 2 changes as a break. Then, the data collection unit 11 divides the temperature data, the thermal displacement data, and the operation data into section units that are operation operation units.
  • the operation data includes the analysis result of the machining program 3, the operation state data and the control command.
  • the rotation speed of the spindle motor 24 and the rotation speed of the feed shaft motor 26 are included in both the analysis result of the machining program 3, the control command, and the operation state data.
  • the rotation speed of the spindle motor 24 and the rotation speed of the feed shaft motor 26 can be obtained based on either the analysis result, the control command, or the operation state data.
  • the information indicating whether the coolant device 27 and the cooling device 28 are in operation is included in the operation state data.
  • the control command also includes commands relating to these operating states, so the data collecting unit 11 uses the control commands to determine the operating states of the coolant device 27 and the cooling device 28.
  • the operating states of the motors, the coolant device 27, and the cooling device 28 are included in the control command and also included in the operating state data as a result of the control. Only one of the command and the result may be used for learning, but by learning the command value together with the operating state data of the machine tool 2, it is expected to generate a learning model with higher performance.
  • FIG. 4 shows the state change for each type of the constituent elements of the machine tool 2
  • the machine tool 2 generally includes a plurality of feed shafts 22.
  • a plurality of cooling devices 28 may be provided according to the number of machine elements of the machine tool 2.
  • each data can be divided into driving operation units at the timing when the state of each component changes.
  • the length of the section that is the first time length is not constant for the data divided by section.
  • the time width Tw1 from time t1 to time t2 and the time width Tw2 from time t2 to time t3 are different.
  • the time width of the section is not constant, the number of sampling points of the data included in each section is also not constant. If the number of sampling points for each section is different, the learning process in the learning unit 12 becomes complicated.
  • the data collection unit 11 divides the temperature data and the thermal displacement divided into sections so that the learning data corresponding to each section has the same sampling points. Resampling the data to generate training data. That is, the data collection unit 11 resamples the temperature data and the displacement data to generate learning data so that the number of sampling points of the data forming the learning data is the same between the learning data.
  • FIG. 5 shows an example of re-sampling the temperature data in the section of time width Tw1 from time t1 to time t2
  • FIG. 6 resamples the temperature data in the section of time width Tw2 from time t2 to time t3. An example is shown.
  • the sampling score of the section from time t1 to time t2 shown in FIG. 5 is L1
  • the sampling score of the section from time t2 to time t3 shown in FIG. 6 is L2. Since the temperature data is detected at a constant cycle, the longer the time interval, the greater the number of sampling points of the temperature data included in the interval. Since the time width Tw2 is longer than the time width Tw1 as shown in FIGS. 4 to 6, L2 is a value larger than L1.
  • the data collection unit 11 performs resampling such that the sampling points in the section from time t1 to time t2 and the sampling points in the section from time t2 to time t3 are both L. I do.
  • a general method can be used as the resampling method, and the resampling method is not particularly limited.
  • the data collection unit 11 resamples the thermal displacement data as well as the temperature data.
  • the data collection unit 11 may resample the operation data as well, but since the operation data is used to determine the temperature data and the thermal displacement data for each content of the operation operation, the data is acquired in operation operation units. Does not need to be resampled.
  • the learning unit 12 uses the learning data received from the data collection unit 11 to generate a learning model regarding the thermal displacement amount.
  • the learning model is, for example, a mathematical polynomial shown in the following formula (1).
  • N is the number of the temperature sensor 29
  • L is the number of sampling points in one of the learning data
  • d x, d y, d z are each X, Y, Z axis direction Is the amount of thermal displacement of.
  • j represents the time discretized in units of sampling points and is represented by a number.
  • T i, j represents the temperature data of the i-th temperature sensor 29 at the time j
  • a i , b i , c. i , C 1 , C 2 , and C 3 are model parameters.
  • i indicates the number of the temperature sensor 29 for identifying the temperature sensor 29.
  • the learning unit 12 identifies the model parameter in the above formula (1) using the learning data.
  • a known identification method such as the least square method can be used.
  • the learning unit 12 divides the driving data into the same group having the same value of the data that determines the driving motion unit, and identifies the model parameter for each group using the corresponding learning data. Therefore, the learning model can be constructed for each content of the driving motion.
  • the mathematical polynomial that is the learning model is not limited to the above (1), and may be another mathematical expression.
  • FIG. 7 is a model diagram of the neural network.
  • FIG. 7 shows a network structure including an input layer that inputs temperature data, an output layer that outputs thermal displacement data, and one or more intermediate layers that propagate signals from the input layer to the output layer. ..
  • the input / output relationship of the nodes included in each layer is expressed by the following equation (2).
  • x m, k represents a signal input from the m-th node to the k-th node
  • y k represents a signal output from the k-th node.
  • w m, k is a weighting coefficient for the m-th node and the k-th node
  • b k is a bias for the k-th node
  • f is an activation function.
  • the activation function f for example, a sigmoid function or a normalized linear function can be used.
  • the neural network shown in FIG. 7 can learn the relationship between the temperature data included in the learning data and the thermal displacement amount by using the learning method based on the error back propagation method.
  • a convolutional neural network may be used.
  • a recurrent neural network may be used.
  • FIG. 8 is a flowchart showing an example of a learning processing procedure in the numerical controller 1.
  • the data collecting unit 11 of the numerical controller 1 collects temperature data, thermal displacement data, and operation data, and generates learning data (step S1). Specifically, the data collection unit 11 acquires temperature data and thermal displacement data from the machine tool 2 and operation data from the control unit 13. As described above, the data collection unit 11 divides each data in driving operation units and resamples the divided temperature data and thermal displacement data to generate learning data.
  • the learning unit 12 learns the relationship between the temperature data and the thermal displacement data from the learning data generated by the data collection unit 11 and generates a learning model (step S2). Through the above processing, a learning model that learns the relationship between the temperature data and the thermal displacement data is generated.
  • the numerical controller 1 executes the learning operation described above for various driving operations to build a learning model that learns the relationship between temperature data and thermal displacement data in various driving operations.
  • FIG. 9 is a flowchart showing an example of a thermal displacement amount estimation processing procedure in the numerical controller 1.
  • the data selection unit 14 of the numerical control device 1 acquires the temperature data and the operation data and generates estimation data (step S11). Specifically, the data selection unit 14 acquires temperature data from the machine tool 2 and operation data from the control unit 13. Then, similarly to the data collection unit 11, the data selection unit 14 determines the time length of the estimation data, which is the second time length, based on the operation data. The second time length is determined with the timing at which the operating state of at least one of the machine elements included in the machine tool 2 changes as a delimiter.
  • the data selection unit 14 divides the operation data and the temperature data by the operation operation unit, resamples the divided temperature data, and uses the resampled data and the divided operation data as estimation data, the thermal displacement estimation unit 15 Output to. That is, the data selection unit 14 resamples the temperature data and generates the estimation data so that the number of sampling points of the data forming the estimation data is the same between the estimation data.
  • the thermal displacement estimation unit 15 estimates the thermal displacement amount using the estimation data and the learning model (step S12).
  • the learning model is input from the learning unit 12 to the thermal displacement estimation unit 15.
  • the thermal displacement estimation unit 15 outputs the estimated thermal displacement amount to the thermal displacement correction unit 16.
  • the thermal displacement estimation unit 15 performs the process opposite to the resampling performed by the data selection unit 14 on the calculated thermal displacement amount so that the output cycle of the thermal displacement amount in the thermal displacement estimation unit 15 is calculated by the temperature sensor 29. Match the temperature detection cycle.
  • the thermal displacement estimation unit 15 adjusts the output cycle of the thermal displacement amount to the output cycle of the control command. The reverse process of resampling is performed.
  • the thermal displacement correction unit 16 adds a correction amount that cancels the thermal displacement amount to the control command (step S13). Specifically, the thermal displacement correction unit 16 adds a correction amount that cancels the thermal displacement amount to the control command received from the control unit 13, and adds the correction amount to the machine tool 2 as a driving operation command. Output.
  • the numerical control device 1 can perform the processing of correcting the thermal displacement amount with respect to the control command. Therefore, the numerical control device 1 can reduce a processing error caused by thermal displacement.
  • the data collecting unit 11 divides the temperature data and the thermal displacement data into driving operation units to generate learning data
  • the learning unit 12 performs learning. Learn the relationship between temperature and thermal displacement from the data.
  • the data selection unit 14 generates estimation data by dividing the temperature data into driving operation units
  • the thermal displacement estimation unit 15 uses the estimation data and the learning model. To estimate the amount of thermal displacement. Therefore, it is possible to generate a highly accurate learning model suitable for the content of the driving motion for each driving motion. That is, the numerical control device 1 according to the present embodiment can learn the relationship between the temperature and the thermal displacement amount with high accuracy, and thus can accurately estimate the thermal displacement amount.
  • the numerical control device 1 according to the present embodiment has a higher estimation accuracy of the thermal displacement amount as compared with the example in which the learning model is generated by using the data in a unit of a fixed period without considering the content of the driving operation. Can be improved. Therefore, the numerical control device 1 according to the present embodiment can correct the control command using the estimated value of the thermal displacement amount with high estimation accuracy, and can reduce the processing error.
  • the machine tool 2 configured to cut the work by rotating the tool has been described, but the machine tool to which the present invention is applicable is not limited to this.
  • a machine tool having a structure in which a tool is fixed and a work is rotated like a lathe can achieve the same effect as that of the present embodiment.
  • the relationship between the temperature and the thermal displacement amount is learned, and the thermal displacement amount is estimated using the detected temperature and the learning model.
  • the relationship between the temperature and at least one of the position and rotation speed of each drive shaft and the amount of thermal displacement is learned, and at least one of the temperature and the position and rotation speed of each drive shaft and the learning model are learned. May be used to estimate the thermal displacement amount.
  • the temperature sensor 29 and the displacement sensor 30 are components of the machine tool 2
  • at least a part of the temperature sensor 29 and the displacement sensor 30 is a component of the machine tool 2.
  • it may be a sensor provided later on the machine tool 2.
  • the numerical control device 1 is configured to generate the learning model, but the data collection unit 11 and the learning unit 12 may be provided in a learning device different from the numerical control device 1.
  • the learning device receives the operation data from the numerical control device 1 and the temperature data and the thermal displacement data from the machine tool 2.
  • the data collection unit 11 and the learning unit 12 in the learning device perform the same operation as in the above-described example.
  • the learning device outputs the learning model generated by the learning unit 12 to the numerical control device 1.
  • the learning unit 12 may be included in a learning device other than the numerical control device 1. In this case, the learning data is input to the learning device from the numerical control device 1, and the learning device outputs the learning model to the numerical control device 1.

Landscapes

  • Engineering & Computer Science (AREA)
  • Human Computer Interaction (AREA)
  • Manufacturing & Machinery (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Numerical Control (AREA)
  • Automatic Control Of Machine Tools (AREA)
PCT/JP2018/040500 2018-10-31 2018-10-31 数値制御装置、学習装置および学習方法 WO2020090030A1 (ja)

Priority Applications (4)

Application Number Priority Date Filing Date Title
DE112018008027.7T DE112018008027T5 (de) 2018-10-31 2018-10-31 Numerisches Steuerungsgerät, Lerngerät und Lernverfahren
JP2019516736A JP6556413B1 (ja) 2018-10-31 2018-10-31 数値制御装置、学習装置および学習方法
CN201880098973.XA CN112912803A (zh) 2018-10-31 2018-10-31 数控装置、学习装置及学习方法
PCT/JP2018/040500 WO2020090030A1 (ja) 2018-10-31 2018-10-31 数値制御装置、学習装置および学習方法

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2018/040500 WO2020090030A1 (ja) 2018-10-31 2018-10-31 数値制御装置、学習装置および学習方法

Publications (1)

Publication Number Publication Date
WO2020090030A1 true WO2020090030A1 (ja) 2020-05-07

Family

ID=67539869

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2018/040500 WO2020090030A1 (ja) 2018-10-31 2018-10-31 数値制御装置、学習装置および学習方法

Country Status (4)

Country Link
JP (1) JP6556413B1 (zh)
CN (1) CN112912803A (zh)
DE (1) DE112018008027T5 (zh)
WO (1) WO2020090030A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023228253A1 (ja) * 2022-05-23 2023-11-30 三菱電機株式会社 熱変位モデル学習装置、熱変位推定装置、加工システム、および加工方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003094291A (ja) * 2001-09-27 2003-04-03 Kanto Seiki Kk 工作機械の熱変形補正方法および装置
JP2016221593A (ja) * 2015-05-27 2016-12-28 ブラザー工業株式会社 工作機械、演算方法及びコンピュータプログラム
JP2018111145A (ja) * 2017-01-10 2018-07-19 ファナック株式会社 工作機械の機械学習装置および熱変位補正装置

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3413068B2 (ja) * 1997-08-19 2003-06-03 オークマ株式会社 工作機械の熱変位推定方法
JP4488564B2 (ja) * 1999-11-17 2010-06-23 関東精機株式会社 工作機械の熱変形抑制方法および温度制御装置
JP3405965B2 (ja) * 2000-07-06 2003-05-12 ファナック株式会社 工作機械の熱変位補正方法
JP2009214283A (ja) * 2008-03-13 2009-09-24 Brother Ind Ltd 工作機械の熱変位補正方法、熱変位補正装置及びその熱変位補正用プログラム
JP4917665B1 (ja) * 2010-11-11 2012-04-18 ファナック株式会社 工作機械の熱変位補正方法及び熱変位補正装置
EP3168001B1 (en) * 2014-10-29 2019-12-18 Yamazaki Mazak Corporation Machine tool equipped with device for changing setting of thermal displacement correction amount
JP6464220B2 (ja) * 2017-03-21 2019-02-06 ファナック株式会社 機械学習装置及び熱変位補正装置
JP6564412B2 (ja) * 2017-03-21 2019-08-21 ファナック株式会社 機械学習装置及び熱変位補正装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003094291A (ja) * 2001-09-27 2003-04-03 Kanto Seiki Kk 工作機械の熱変形補正方法および装置
JP2016221593A (ja) * 2015-05-27 2016-12-28 ブラザー工業株式会社 工作機械、演算方法及びコンピュータプログラム
JP2018111145A (ja) * 2017-01-10 2018-07-19 ファナック株式会社 工作機械の機械学習装置および熱変位補正装置

Also Published As

Publication number Publication date
DE112018008027T5 (de) 2021-06-17
CN112912803A (zh) 2021-06-04
JPWO2020090030A1 (ja) 2021-02-15
JP6556413B1 (ja) 2019-08-07

Similar Documents

Publication Publication Date Title
US9956660B2 (en) Controller for a machine tool to perform efficient warm-up control
JP6469065B2 (ja) 機械学習装置及び加工時間予測装置
US9945799B2 (en) Thermal displacement correction device for machine tool
JP5905158B2 (ja) 数値制御装置
EP3168001B1 (en) Machine tool equipped with device for changing setting of thermal displacement correction amount
EP1650620B1 (en) Method and apparatus for correcting thermal displacement of machine tool
JP4469325B2 (ja) 熱変位補正装置
JP6443311B2 (ja) 制御装置、制御プログラムおよび記録媒体
JP2018111145A (ja) 工作機械の機械学習装置および熱変位補正装置
JP5803261B2 (ja) 工作機械の熱変位補正方法および熱変位補正装置
JP2012254499A (ja) 工作機械の加工異常検知装置及び加工異常検知方法
CN111052015B (zh) 数控系统及电动机控制装置
JP2019146421A (ja) 故障予測装置及び機械学習装置
JP7303065B2 (ja) 機械学習装置、制御システム及び機械学習方法
WO2019244638A1 (ja) 位置決めシステム、監視装置、監視方法およびプログラム
JP2019152936A (ja) 工作機械の加工シミュレーション装置
JP2017127922A (ja) 工具交換装置の減速機の異常診断機能を有する工作機械
WO2020090030A1 (ja) 数値制御装置、学習装置および学習方法
JP2019150932A (ja) 衝突位置推定装置及び機械学習装置
JP2018043317A (ja) 工作機械
WO2023228253A1 (ja) 熱変位モデル学習装置、熱変位推定装置、加工システム、および加工方法
JP6378249B2 (ja) サーボ制御および機械運動の遅れを考慮する加工時間予測機能を有する数値制御装置
Tan et al. Calibration of single-axis nanopositioning cell subjected to thermal disturbance
JP2003271213A (ja) 移動軸監視装置、移動軸監視方法、移動軸監視プログラムおよびその移動軸監視プログラムを記録したコンピュータ読取り可能な記録媒体
WO2023238192A1 (ja) 熱変位補正装置及びプログラムを記録したコンピュータ読み取り可能な記録媒体

Legal Events

Date Code Title Description
ENP Entry into the national phase

Ref document number: 2019516736

Country of ref document: JP

Kind code of ref document: A

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18939029

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 18939029

Country of ref document: EP

Kind code of ref document: A1