Numerical control device, method for determining occurrence of chatter, and learning device

By using the drive command unit, the sensor signal processing unit, the timing signal generation unit and the phase difference operation unit in the working machine, the phase difference operation unit is used to generate phase difference information, and combined with the machine learning model, the rapid flutter judgment is realized, and the flutter judgment in the prior art is solved The problem of long time improves processing efficiency.

CN113646613BActive Publication Date: 2024-11-19MITSUBISHI ELECTRIC CORP
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202080025983.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2019-04-08
Filing Date
2020-01-22
Publication Date
2024-11-19
Estimated Expiration
2040-01-22

AI Technical Summary

Technical Problem

When the prior art determines flutter in the working machine, the time becomes longer, and the frequency analysis method is limited by the frequency resolution and data acquisition time, which affects the processing efficiency.

Method used

By using the drive command unit, the sensor signal processing unit, the timing signal generation unit, the phase difference calculation unit and the vibration determination unit, the phase difference information of various state quantities and no factor quantities is generated, and the flutter judgment is performed using the machine learning model to avoid Time limit for frequency resolution.

Benefits of technology

The time for flutter determination is shortened, the judgment efficiency is improved, and the flutter can be suppressed at a high speed without relying on processing conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113646613B_ABST
    Figure CN113646613B_ABST
Patent Text Reader

Abstract

The numerical control device (1) is characterized in that it has: a drive command unit (15) that gives operation commands to the main shaft (101) and feed shaft (102) of the working machine (100); a sensor signal processing unit (11) that generates a plurality of types of state quantities based on a sensor signal (200) that detects vibration of a tool or a processing object installed on the working machine (100); a timing signal generating unit (16) that periodically generates and outputs a timing signal (207) at a timing synchronized with the rotation of the main shaft (101); and a phase difference calculation unit. A unit (12) which generates a dimensionless quantity representing a plurality of types of state quantities in a state space for each timing signal (207) and calculates phase difference information (203) representing the difference between the phases of the dimensionless quantity; and a vibration determination unit (13) which includes an information observation unit which creates an inference data set based on the phase difference information (203) and an inference unit which has a trained inference model for machine learning that determines the presence or absence of flutter by inputting the inference data set, and inputs the inference data set to the inference unit to determine the occurrence of flutter.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a numerical control device for determining the occurrence of chatter in a working machine, a method for determining the occurrence of chatter, and a learning device. Background Art

[0002] There are machine tools that change the relative position between the object and the tool to remove the object into a desired shape. Machine tools such as milling machines and lathes load the tool or the object on the spindle and rotate the spindle to perform processing. During processing, vibrations called "chatter" sometimes occur. If chatter occurs, the precision of the finished surface deteriorates or it may cause damage to the tool.

[0003] Patent document 1 discloses a method in which vibration is measured twice at predetermined time intervals during machining, the second measurement result is compared with the first, and when the amplitude in the frequency region is large and the phase in the frequency region is different, it is determined that chatter has occurred.

[0004] Patent Document 1: Japanese Patent Application Publication No. 2013-7647 Summary of the invention

[0005] However, the technology described in the above-mentioned Patent Document 1 has the following problem: the time until determining whether or not chatter occurs during processing becomes longer. Specifically, in the technology described in Patent Document 1, frequency analysis is used when analyzing vibration. Frequency analysis such as FFT (Fast Fourier Transform) has a limitation in its calculation algorithm that the frequency resolution ΔF is inversely proportional to the data acquisition time T. Therefore, in order to obtain a sufficient frequency resolution ΔF when determining chatter, the data acquisition time becomes longer. In addition, in the technology described in Patent Document 1, it is necessary to perform two measurements at predetermined intervals, so the time for measuring the vibration during processing also becomes longer.

[0006] The present invention has been made in view of the above circumstances, and an object of the present invention is to obtain a numerical control device capable of shortening the time required for determining chatter vibration.

[0007] In order to solve the above-mentioned problems and achieve the purpose, the numerical control device of the present invention is characterized in that it has: a drive instruction unit, which gives operation instructions to the main shaft and feed axis of the working machine; a sensor signal processing unit, which generates multiple types of state quantities based on the sensor signal of the vibration of the tool or the processing object installed on the working machine; a timing signal generating unit, which periodically generates and outputs the timing signal at a timing synchronized with the rotation of the main shaft; a phase difference calculation unit, which generates a dimensionless quantity representing multiple types of state quantities in the state space for each timing signal, and calculates the phase difference information representing the difference between the phases of the dimensionless quantity; an information observation unit, which creates an inference data set based on the phase difference information; and a vibration judgment unit, which has an inference unit with a trained inference model that performs machine learning to determine the presence or absence of vibration by inputting the inference data set, and determines the occurrence of vibration by inputting the inference data set into the inference unit.

[0008] Effects of the Invention

[0009] According to the present invention, there is an effect that the time required for determining chatter vibration can be shortened. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 This is a diagram showing the functional structure of the numerical control device according to the first embodiment.

[0011] Figure 2 It is used to illustrate Figure 1 Graph showing filter characteristics of the filter processing unit.

[0012] Figure 3 It means by Figure 1 FIG. 1 is a diagram showing the relationship between a timing signal output by a timing signal generating unit and an angle command for a main shaft.

[0013] Figure 4 It means by Figure 1 FIG. 1 is a diagram of a dimensionless quantity calculated by the phase difference calculation unit shown.

[0014] Figure 5 This means that when a chatter is occurring, Figure 1 FIG. 1 is a diagram showing an example of transition of the phase difference calculated by the phase difference calculation unit shown.

[0015] Figure 6 It means that when no vibration occurs, Figure 1 FIG. 1 is a diagram showing an example of transition of the phase difference calculated by the phase difference calculation unit shown.

[0016] Figure 7 This is a diagram showing the experimental results when chatter vibration occurs, that is, the waveform of acceleration generated during machining.

[0017] Figure 8 It means according to Figure 7 A diagram showing the dimensionless phase difference obtained based on the experimental results shown.

[0018] Fig. 9 This is a diagram showing the experimental results when no chatter vibration occurs, that is, the waveform of the acceleration generated during machining.

[0019] Fig.10 It means according to Fig. 9 A diagram showing the dimensionless phase difference obtained based on the experimental results shown.

[0020] Fig.11 It is used for Figure 1 FIG. 1 is a diagram for explaining a first example of processing by the vibration determination unit shown in FIG.

[0021] Fig.12 It is used for Figure 1 FIG. 2 is a diagram for explaining a second example of the processing of the vibration determination unit shown in FIG.

[0022] Fig.13 This is a diagram showing the functional structure of a numerical control device according to the second embodiment.

[0023] Fig.14 It means by Fig.13 1 is a diagram showing a first example of the relationship between a timing signal output by a timing signal generating unit and an angle command for a main shaft.

[0024] Fig.15 It means by Fig.13 2 is a diagram showing a second example of the relationship between the timing signal output by the timing signal generating unit and the angle command of the main shaft.

[0025] Fig.16 This is a diagram showing the functional structure of a numerical control device according to the third embodiment.

[0026] Fig.17 This is a diagram showing the functional structure of a numerical control device according to a fourth embodiment.

[0027] Fig.18 This is a diagram showing the functional structure of a numerical control device according to the fifth embodiment.

[0028] Fig.19 This is a diagram showing the functional structure of a numerical control device according to the sixth embodiment.

[0029] Fig. 20 This is a diagram showing dedicated hardware for realizing the functions of the numerical control device according to the first to sixth embodiments.

[0030] Fig.21This is a diagram showing a configuration of a control circuit for realizing the functions of the numerical control device according to the first to sixth embodiments.

[0031] Fig. 22 It is used to Figure 1 FIG. 2 is a diagram showing a configuration example of a third example of realizing the processing of the vibration determination unit shown in FIG.

[0032] Fig.23 This is a diagram showing a structural example of an inference model using a neural network.

[0033] Fig.24 Yes Fig. 22 FIG. 2 is a diagram showing a functional configuration of a learning device for learning an estimation model used by a vibration determination unit. DETAILED DESCRIPTION

[0034] Hereinafter, a numerical control device, a chattering occurrence determination method, and a learning device according to an embodiment of the present invention will be described in detail with reference to the drawings. Note that the technical scope of the present invention is not limited to the present embodiment.

[0035] Implementation method 1.

[0036] Figure 1 1 is a diagram showing a functional configuration of the numerical control device 1 according to Embodiment 1. The numerical control device 1 gives an operation command 208 to the machine tool 100 , thereby numerically controlling the machine tool 100 .

[0037] The machine tool 100 has a spindle 101 and a feed shaft 102, which have motors driven by an operation command 208. In the present embodiment, a processing object is set on the spindle 101, and a tool is set on the feed shaft 102. In addition, the machine tool 100 outputs operation information 209 including at least the position, speed and motor current of the spindle 101 and the feed shaft 102 to the numerical control device 1. In addition, in the present embodiment, the number of blades of the tool is set to one.

[0038] The machine tool 100 is equipped with a sensor 103. The sensor 103 outputs a sensor signal 200 that detects the vibration of the tool or the object to be processed to the numerical control device 1. The sensor 103 is installed on a member of the spindle 101 or the feed shaft 102 of the machine tool 100. The position where the sensor 103 is installed can be any position where the vibration of the tool or the object to be processed can be detected, and it is preferably near the point where the tool and the object to be processed contact. The type of the sensor 103 can be any sensor that can detect the vibration of the tool or the object to be processed, and the sensor 103 is, for example, a displacement sensor, a velocity sensor, an acceleration sensor, an angular velocity sensor, etc. In addition, the sensor 103 may also be a force sensor that detects the cutting reaction force, or a microphone that detects the cutting sound during processing. Alternatively, instead of the sensor 103, at least one of the position, velocity, and motor current of the spindle 101 and the feed shaft 102 included in the motion information may be used to detect the vibration of the tool or the object to be processed generated during processing, and information indicating the detected vibration may be output as the sensor signal 200.

[0039] The numerical control device 1 includes a filter processing unit 10 , a sensor signal processing unit 11 , a phase difference calculation unit 12 , a vibration determination unit 13 , a command value correction unit 14 , a drive command unit 15 , and a timing signal generation unit 16 .

[0040] The drive command unit 15 receives the action information 209 from the machine tool 100, outputs the operation command 208 to the machine tool 100 based on the action information 209, and outputs the spindle operation command 206 to the timing signal generating unit 16. In addition, the drive command unit 15 corrects the operation command 208 based on the correction signal 205 output by the command value correction unit 14 described later. The operation command 208 includes at least one of the angle command and the speed command of the spindle 101 of the machine tool 100, and at least one of the position command and the speed command of the feed axis 102. In addition, the spindle operation command 206 includes at least one of the angle command and the speed command of the spindle 101.

[0041] The filter processing unit 10 filters the sensor signal 200 outputted from the machine tool 100 to generate a chatter component signal 201 , and outputs the generated chatter component signal 201 to the sensor signal processing unit 11 . The filter processing unit 10 removes the cutting component from the sensor signal 200 .

[0042] Figure 2 It is used to illustrate Figure 1 Graph showing filter characteristics of the filter processing unit 10. Figure 2 The horizontal axis represents frequency, and the vertical axis represents amplitude. Figure 2The frequency region shows the chatter component, cutting component, and filter passband. The cutting component is mainly a vibration component of a frequency obtained by multiplying the spindle speed by the number of cutting edges of the tool, and its harmonic component. The filter processing unit 10 removes the frequency band containing each cutting component by a predetermined bandwidth. For example, the predetermined bandwidth can be set to about several Hz near the cutting component. The filter processing unit 10 can be designed using a well-known bandpass filter, a notch filter, a combination of multiple bandpass filters, a combination of multiple notch filters, a comb filter, etc. By using Figure 2 By using the passband filter shown, the filter processing unit 10 can remove the cutting component from the sensor signal 200 and generate the chatter component signal 201 .

[0043] return Figure 1 Description of . The sensor signal processing unit 11 generates a plurality of types of state quantities based on the sensor signal 200, and outputs a state quantity signal 202 indicating the generated plurality of types of state quantities to the phase difference calculation unit 12. Specifically, the sensor signal processing unit 11 generates the state quantity signal 202 based on the flutter component signal 201 output by the filter processing unit 10. The state quantity signal 202 includes a first state quantity and a second state quantity. For example, the first state quantity is the flutter component signal 201, and the second state quantity is a state quantity obtained by performing a time differentiation on the flutter component signal 201 once. The sensor signal processing unit 11 simultaneously sets the flutter component signal 201 and the signal obtained by performing a time differentiation on the flutter component signal 201 once as a time series signal as the state quantity signal 202.

[0044] In addition, the difference between the times of the first state quantity and the second state quantity can be any combination as long as it is an odd number. For example, the first state quantity is a state quantity obtained by performing P times of time differentiation on the flutter component signal 201, and the second state quantity is a state quantity obtained by performing Q times of time differentiation. Here, P and Q are integers. In this case, the difference between P and Q can be an odd number. In addition, the dimensions of the first state quantity and the second state quantity can be dimensions obtained by time integration of the flutter component signal 201. As a specific example, the first state quantity can be acceleration, and the second state quantity can be velocity. Alternatively, the first state quantity can be jerk, which is a differential value of acceleration, and the second state quantity can be position. The sensor signal processing unit 11 can generate, for example, two or more state quantities among the sensor signal 200, the state quantity obtained by performing time differentiation on the sensor signal 200, and the state quantity obtained by performing time integration on the sensor signal.

[0045] When determining that the angle command for the spindle 101 has passed a predetermined angle based on the spindle operation command 206 outputted from the drive command unit 15, the timing signal generating unit 16 outputs a timing signal 207 to the phase difference calculating unit 12. The method of generating the timing signal 207 will be described in detail.

[0046] Figure 3 It means by Figure 1 1 is a diagram showing the relationship between the timing signal 207 output by the timing signal generating unit 16 and the angle command of the spindle 101. The angle command of the spindle 101 is included in the spindle operation command 206 output by the drive command unit 15. Figure 3 The middle angle command is a signal that takes a value from 0 to 360 degrees and returns to 0 degrees when it reaches 360 degrees. The timing signal generating unit 16 outputs a timing signal 207 each time the angle command of the spindle 101 passes through a specified angle ψ1. Thus, the timing signal 207 is periodically output at a timing synchronized with the rotation of the spindle 101 of the machine tool 100. If the angle ψ1 is fixed at 1 during one rotation of the spindle 101, it can be any angle, for example, it can be set to the angle ψ1 of the spindle 101 when the spindle 101 is oriented. Spindle orientation is a reference angle for stopping the spindle 101.

[0047] Alternatively, the timing signal generating unit 16 can also generate the timing signal 207 using the speed command of the spindle 101 included in the spindle operation command 206. In this case, the timing signal generating unit 16 sets the initial time t0 as the reference time, and outputs the timing signal 207 in units of time T1 calculated using the following formula (1) based on the speed command S (rpm) of the spindle 101. In addition, the timing signal generating unit 16 can also use a combination of the angle command and the speed command of the spindle 101. In addition, here, the timing signal generating unit 16 generates the timing signal 207 using the operation command of the spindle 101, but it is also possible to generate the timing signal 207 using the operation command of the feed axis 102 instead of the operation command of the spindle 101 or on the basis of the operation command of the spindle 101.

[0048] [Formula 1]

[0049]

[0050] The phase difference calculation unit 12 receives the timing signal 207 output by the timing signal generation unit 16 and the state quantity signal 202 output by the sensor signal processing unit 11. The phase difference calculation unit 12 generates dimensionless quantities representing multiple types of state quantities in the state space for each timing signal 207 based on the timing signal 207 and the state quantity signal 202, and calculates the phase difference information 203 representing the difference between the phases of the dimensionless quantities. The phase difference calculation unit 12 outputs the generated phase difference information 203 to the vibration determination unit 13.

[0051] The phase difference calculation unit 12 normalizes each of the multiple state quantities included in the state quantity signal 202. The phase difference calculation unit 12 normalizes each state quantity by dividing each state quantity by a predetermined maximum value of each state quantity. The maximum value used here can be the maximum value of each state quantity obtained by a prior processing experiment, or the maximum value of each state quantity obtained by a prior simulation.

[0052] It is assumed that the phase difference calculation unit 12 receives the timing signal 207 at time t1 and time t2. At this time, t2 is assumed to be a value greater than t1. The phase difference calculation unit 12 normalizes the state quantity signal 202 at time t1, and sets the normalized value as the first dimensionless quantity N1. At time t2, the phase difference calculation unit 12 normalizes the state quantity signal 202 in the same manner as at time t1, and sets the normalized value as the second dimensionless quantity N2.

[0053] Figure 4 It means by Figure 1 FIG. 1 is a diagram showing a dimensionless quantity calculated by the phase difference calculation unit 12. Figure 4 The first dimensionless quantity N1 and the second dimensionless quantity N2 calculated by the phase difference calculation unit 12 are expressed in a state space composed of the dimension of the first state quantity and the dimension of the second state quantity of the state quantity signal 202. The phase difference calculation unit 12 calculates the angle θ formed by the previous dimensionless quantity, i.e., the first dimensionless quantity N1, the origin of the state space, and the current dimensionless quantity, i.e., the second dimensionless quantity N2 as the phase difference between the first dimensionless quantity N1 and the second dimensionless quantity N2. The phase difference calculation unit 12 performs the above calculation for each timing signal 207, and calculates the phase difference between the latest dimensionless quantity and the previous dimensionless quantity. The phase difference calculation unit 12 outputs the phase difference information 203 indicating the calculated phase difference to the vibration determination unit 13.

[0054] Figure 5 This means that when a chatter is occurring, Figure 12 is a diagram showing an example of the transition of the phase difference calculated by the phase difference calculation unit 12 shown in FIG. 2 . The dimensionless quantities N10 to N13 are dimensionless quantities calculated at each time t10 to t13 when the timing signal 207 is received. At this time, it is assumed that t13>t12>t11>t10. In addition, the angle θ10 shows the angle formed by the dimensionless quantity N10, the origin of the state space, and the dimensionless quantity N11, the angle θ11 shows the angle formed by the dimensionless quantity N11, the origin of the state space, and the dimensionless quantity N12, and the angle θ12 shows the angle formed by the dimensionless quantity N12, the origin of the state space, and the dimensionless quantity N13.

[0055] When chatter is occurring, the loci of dimensionless quantities N10 to N13 in the state space become circular, and the angles between the dimensionless quantities become constant. Therefore, the angle θ at each time when the timing signal 207 is received becomes constant.

[0056] Figure 6 It means that when no vibration occurs, Figure 1 1 is a diagram showing an example of the transition of the phase difference calculated by the phase difference calculation unit 12 shown in FIG. Dimensionless quantities N20 to N23 are dimensionless quantities calculated at each time t20 to t23 when the timing signal 207 is received. At this time, it is assumed that t23>t22>t21>t20. In addition, angle θ20 shows the angle formed by the dimensionless quantity N20, the origin of the state space, and the dimensionless quantity N21, angle θ21 shows the angle formed by the dimensionless quantity N21, the origin of the state space, and the dimensionless quantity N22, and angle θ22 shows the angle formed by the dimensionless quantity N22, the origin of the state space, and the dimensionless quantity N23.

[0057] When chattering does not occur, the loci of the dimensionless quantities N20 to N23 in the state space and the angles between the dimensionless quantities become irregular. Therefore, the angle θ at each time when the timing signal 207 is received is not constant.

[0058] Therefore, the phase difference calculation unit 12 calculates the phase difference, which is an angle θ of a constant value, when chattering is occurring, and calculates the phase difference, which is an angle θ of an irregular value, when chattering is not occurring.

[0059] Here, the experimental results for confirming the difference in angle θ related to the presence or absence of chattering are described. In this experiment, a round bar of S45C, a carbon steel material, was set in the chuck of a common CNC (Computer Numerical Control) lathe, and the outer circumference of the round bar was processed by turning. In addition, a superhard tool was used as the tool, and an acceleration sensor was set on the tool to measure the acceleration during processing.

[0060] The machining conditions when chattering occurred were: spindle speed 2000 (RPM), feed per cutting edge 0.1 (mm), radial cut-in 1.0 (mm), and workpiece protrusion from the chuck 200 (mm).

[0061] Figure 7 This is a diagram showing the experimental results when chatter vibration occurs, that is, the waveform of acceleration generated during machining. Figure 8 It means according to Figure 7 In this experiment, the filter processing unit 10, the sensor signal processing unit 11, the timing signal generating unit 16 and the phase difference calculating unit 12 process the acceleration data output by the acceleration sensor as the sensor signal 200, thereby calculating Figure 8 The dimensionless quantity and angle θ are shown.

[0062] First, the filter processing unit 10 removes the cutting component composed of 2000 / 60=33.3 (Hz) and its harmonics from the acceleration data based on the spindle speed. Next, the sensor signal processing unit 11 calculates and outputs the state quantity signal 202 in which the first state quantity is set as acceleration and the second state quantity is set as speed. In addition, the second state quantity, i.e., speed, is obtained by integrating acceleration.

[0063] The timing signal generating unit 16 outputs the timing signal 207 each time the angle command of the spindle 101 passes through the spindle angle at the time of spindle orientation. The phase difference calculating unit 12 normalizes the state quantity signal 202 and calculates the dimensionless quantities N101 to N106 at each time t101 to t106 when each timing signal 207 is received. Here, t106>t105>t104>t103>t102>t101.

[0064] exist Figure 8 Dimensionless quantities N101 to N106 in the state space are shown. Angle θ101 shows the angle formed by dimensionless quantity N101, the origin of the state space, and dimensionless quantity N102. Angle θ102 shows the angle formed by dimensionless quantity N102, the origin of the state space, and dimensionless quantity N103. Angle θ103 shows the angle formed by dimensionless quantity N103, the origin of the state space, and dimensionless quantity N104. Angle θ104 shows the angle formed by dimensionless quantity N104, the origin of the state space, and dimensionless quantity N105. Angle θ105 shows the angle formed by dimensionless quantity N105, the origin of the state space, and dimensionless quantity N106. Angles θ101 to θ105 are constant. According to the above experimental results, it is experimentally shown that when chatter is occurring, the trajectory of dimensionless quantities in the state space becomes circular, and angles θ101 to θ105 at each time take constant values.

[0065] Next, the machining conditions when no chatter vibration occurs are: spindle speed 2000 (RPM), feed per cutting edge 0.1 (mm), radial cutting depth 1.0 (mm), and workpiece protrusion from the chuck 100 (mm).

[0066] Fig. 9 This is a diagram showing the experimental results when no chatter vibration occurs, that is, the waveform of the acceleration generated during machining. Fig.10 It means according to Fig. 9 In this experiment, the filter processing unit 10, the sensor signal processing unit 11, the timing signal generating unit 16 and the phase difference calculating unit 12 also process the acceleration data output by the acceleration sensor as the sensor signal 200, thereby calculating Fig.10 The dimensionless quantity and angle θ are shown.

[0067] First, the filter processing unit 10 removes the cutting component composed of 2000 / 60=33.3 (Hz) and its harmonics from the acceleration data based on the spindle speed. Next, the sensor signal processing unit 11 calculates and outputs the state quantity signal 202 in which the first state quantity is set as acceleration and the second state quantity is set as speed. In addition, the second state quantity, i.e., speed, is obtained by integrating acceleration.

[0068] The timing signal generating unit 16 outputs the timing signal 207 each time the angle command of the spindle 101 passes through the spindle angle at the time of spindle orientation. The phase difference calculating unit 12 normalizes the state quantity signal 202 and calculates the dimensionless quantities N201 to N206 at each time t201 to t206 when each timing signal 207 is received. Here, t206>t205>t204>t203>t202>t201.

[0069] exist Fig.10 Dimensionless quantities N201 to N206 in the state space are shown. Angle θ201 shows the angle formed by dimensionless quantity N201, the origin of the state space, and dimensionless quantity N202, angle θ202 shows the angle formed by dimensionless quantity N202, the origin of the state space, and dimensionless quantity N203, angle θ203 shows the angle formed by dimensionless quantity N203, the origin of the state space, and dimensionless quantity N204. Angle θ204 shows the angle formed by dimensionless quantity N204, the origin of the state space, and dimensionless quantity N205, and angle θ205 shows the angle formed by dimensionless quantity N205, the origin of the state space, and dimensionless quantity N206. According to the above experimental results, it is experimentally shown that when no chatter occurs, the trajectory of the dimensionless quantity in the state space is irregular, and the angles θ201 to θ205 at each time take non-constant values.

[0070] return Figure 1The vibration determination unit 13 receives the phase difference, that is, the angle θ, from the phase difference calculation unit 12, and determines the vibration using the following procedure. The vibration determination unit 13 determines whether the vibration occurs based on the phase difference information 203 calculated by the phase difference calculation unit 12. When it is determined that the vibration is occurring, the vibration determination unit 13 outputs the vibration determination information 204 to the command value correction unit 14. The details of the vibration determination unit 13 are described below.

[0071] The vibration determination unit 13 records the phase difference information 203 outputted from the phase difference calculation unit 12. This recording is repeated every time the phase difference information 203 is received, and the angle θ included in the phase difference information 203 is recorded as time series data.

[0072] In addition, the vibration determination unit 13 calculates the difference of the recorded angle θ, and when the difference of the angle θ is lower than a predetermined threshold value, it is determined that flutter is occurring, and the vibration determination information 204 is output to the command value correction unit 14. Here, the threshold used by the vibration determination unit 13 is described. In addition, it is assumed here that the output is performed from the phase difference calculation unit 12 in the range of ±180 degrees for the angle θ. When no flutter occurs, the angle θ takes an indefinite value in the range of ±180 degrees, so the difference of the angle θ takes a value of ±360 degrees. In addition, when flutter is occurring, the angle θ takes a constant value, so the difference of the angle θ becomes zero. However, the actual angle θ contains an error caused by a measurement error, etc. Therefore, if the threshold is set to a value sufficiently larger than the error of the angle θ, it is possible to determine whether there is flutter. For example, when the error of the angle θ is ±10 degrees, the threshold can be set to a value greater than 20 degrees, such as 30 degrees.

[0073] Alternatively, the vibration determination unit 13 calculates the standard deviation of the angle θ included in the plurality of phase difference information 203 that are repeatedly received, and determines that chatter vibration is occurring when the standard deviation of the angle θ is lower than a predetermined threshold value. Fig.11 It is used for Figure 1 The vibration determination unit 13 calculates the standard deviation using the angle θ obtained between time t3 and time t4 to determine whether there is chattering. Fig.11 In the first example shown, the standard deviation of the angle θ exceeds the threshold value of the standard deviation. Therefore, the vibration determination unit 13 can determine that no chattering occurs. Fig.12 It is used for Figure 1 FIG. 1 is a diagram for explaining a second example of the processing of the vibration determination unit 13 shown in FIG. Fig.12 In the second example shown, the standard deviation of the angle θ is lower than the threshold value of the standard deviation. Therefore, the vibration determination unit 13 can determine that chatter vibration is occurring.

[0074] The method for determining the threshold value of the standard deviation is described below. When chattering occurs, the angle θ takes a constant value, so the difference in the angle θ becomes zero. However, the actual angle θ has an error caused by measurement errors, etc. Therefore, by setting a value sufficiently larger than the error included in the angle θ as the threshold value, it is possible to determine whether chattering exists. For example, when the error of the angle θ is ±10 degrees, the threshold value of the standard deviation can be set to a value greater than 10 degrees, such as 20 degrees.

[0075] As another example, the vibration determination unit 13 may determine the presence or absence of chatter using the dispersion of the angle θ instead of the standard deviation of the angle θ included in the plurality of repeatedly received phase difference information 203. In this case as well, the dispersion threshold value may be determined in the same manner as in the case of the standard deviation.

[0076] Furthermore, as another example, the vibration determination unit 13 may determine the flutter using the difference between the maximum value and the minimum value of the angle θ. In this case, the vibration determination unit 13 calculates the difference between the maximum value and the minimum value of the angle θ received during a predetermined length, and when the calculated difference between the maximum value and the minimum value is less than a predetermined threshold value, it can be determined that flutter is occurring.

[0077] Furthermore, as another example, the vibration determination unit 13 may determine the presence of chatter vibration using an estimation model that has been previously machine-learned so as to determine the presence or absence of chatter vibration. Fig. 22 It is used to Figure 1 1 is a diagram showing a configuration example of a third example of the processing of the vibration determination unit 13 shown in FIG. In the third example, the vibration determination unit 13 includes an information observation unit 301 and an inference unit 302. The information observation unit 301 observes the angle θ included in the phase difference information 203 of a predetermined number of samples and generates an inference data set 303 as time series data. The inference unit 302 inputs the inference data set 303 generated by the information observation unit 301 to an inference model that has been machine-learned in advance so as to output the presence or absence of chatter, thereby outputting the vibration determination information 204 indicating the presence or absence of chatter.

[0078] The estimation unit 302 may use an estimation model of any algorithm. As an example, an estimation model using a neural network will be described. Fig.23 is a diagram showing a structural example of an inference model using a neural network. Fig.23 In the example shown, the neural network is composed of an input layer x1, x2, ..., xn having n neurons, an intermediate layer y1, y2, ..., ym having m neurons, and an output layer z1 having 1 neuron. Fig.23 Although an example in which the intermediate layer is one layer is shown, two or more intermediate layers may be provided.

[0079] The input layers x1, x2, ···, xn are connected to the intermediate layers y1, y2, ···, ym, and the intermediate layers y1, y2, ···, ym are connected to the output layer z1. In addition, Fig.23 The connection between the input layer and the intermediate layer shown is an example, and each input layer x1, x2, ..., xn can be connected to any intermediate layer y1, y2, ..., ym.

[0080] in the case of Fig.23 In the three-layer neural network shown in FIG. 1 , after multiple inputs are input to the input layer x1, x2, ..., xn, weights A1 to Aa are assigned to the values ​​and input to the intermediate layers y1, y2, ..., ym. The values ​​input to the intermediate layers y1, y2, ..., ym are further assigned weights B1 to Bb and input to the output layer z1, from which they are output. In addition, a and b are each natural numbers. Fig.23 In the example shown, b = m. The output result changes according to the values ​​of weights A1 to Aa and B1 to Bb.

[0081] Fig.24 Yes Fig. 22 FIG. 4 is a diagram showing a functional configuration of a learning device 400 for learning an estimation model used by the vibration determination unit 13 . The learning device 400 includes a data acquisition unit 401 and a learning processing unit 402 .

[0082] The data acquisition unit 401 acquires a learning data set 304 that associates the phase difference information 203 acquired in actual machining with chattering presence or absence information 305 indicating the presence or absence of chattering. The presence or absence of chattering can be indicated by using different numerical values ​​depending on whether chattering occurs. For example, a value of "1" can be used when chattering occurs, and a value of "0" can be used when chattering does not occur. In addition, the chattering presence or absence information 305 included in the learning data set 304 can use, for example, the result determined by evaluating the machined surface after machining.

[0083] The learning processing unit 402 performs learning using the phase difference information 203 included in the learning data set 304 as input data. The learning processing unit 402 performs so-called supervised learning so that the output of the inference model coincides with the numerical value of the chattering presence / absence information 305. The learning processing unit 402 adjusts the weights A1 to Aa and B1 to Bb so that the value output from the output layer when the phase difference information 203 is input to the input layer is close to the numerical value indicating the presence / absence of chattering, thereby learning the inference model that outputs an output value corresponding to the presence / absence of chattering when the phase difference information 203 is input.

[0084] The learning processing unit 402 can use the error back propagation method as a teacher learning method used when learning the inference model. In addition, in order to improve the generalization performance of the inference model, the learning processing unit 402 can use methods such as "dropout" that randomly excludes neurons during learning and "early stopping" that monitors errors and ends learning in advance. The learning processing unit 402 outputs the learned inference model as a trained inference model. The trained inference model output by the learning processing unit 402 can be Fig. 22 is used in the inference unit 302 shown.

[0085] In addition, although the above description describes the case where a neural network is used in the inference model, other well-known methods such as RNN (Recurrent Neural Network), LSTM (Long Short-Term Memory), SVM (Support Vector Machine), etc. may be used to construct the inference model.

[0086] When receiving vibration determination information 204 from vibration determination unit 13, command value correction unit 14 outputs correction signal 205 to drive instruction unit 15 to instruct change of operation instruction 208. Correction signal 205 is a signal for changing the speed instruction of spindle 101 at a predetermined ratio according to the current spindle speed.

[0087] Alternatively, the correction signal 205 may be a signal for changing the rotation speed of the spindle 101 so that the rotation speed of the spindle 101 matches an integer fraction of the resonance frequency of the mechanical structure calculated in advance by finite element analysis. In addition, the resonance frequency of the mechanical structure may be a value determined in advance by experiments instead of by finite element analysis.

[0088] Furthermore, the correction signal 205 may be a signal for changing the speed command value of the feed axis 102 at a predetermined ratio. Specifically, the correction signal 205 may be a signal for reducing the speed command value of the feed axis 102 so that the motor current of the feed axis 102 is less than a predetermined reference value.

[0089] As described above, according to the first embodiment, time series data of multiple types of state quantities are generated based on the sensor signal 200 that detects the vibration of the tool or the processing object installed in the machine tool 100. In addition, in the state space, dimensionless quantities representing multiple types of state quantities are generated for each timing signal 207, and phase difference information 203 representing the difference between the phases of the dimensionless quantities, that is, the angle θ, is calculated. Based on the fluctuation of the phase difference information 203, the occurrence of chatter in the machine tool 100 is determined.

[0090] The numerical control device 1 does not use frequency analysis represented by FFT when determining chatter, so it is not limited by the data acquisition time and can obtain sufficient frequency resolution when determining chatter. Therefore, the time required for determining chatter, that is, the time from the occurrence of chatter to the determination of the occurrence of chatter, can be shortened. Therefore, the time from the occurrence of chatter to the suppression of chatter is also shortened, and chatter can be suppressed at a high speed.

[0091] In addition, in the numerical control device 1, the chatter is determined based on the dimensionless quantity, that is, the phase difference information 203. As an existing method, for example, there is a method of determining the chatter based on the amplitude of the chatter. When the amplitude of the chatter is used, the magnitude of the amplitude varies depending on the machining conditions, so the threshold of the amplitude must be set according to the machining conditions. In contrast, in the numerical control device 1 that determines the chatter based on the dimensionless quantity, that is, the phase difference information 203, the chatter can be determined independently of the machining conditions, and there is no need to set a threshold according to the machining conditions.

[0092] Furthermore, the machine tool 100 according to the first embodiment is configured such that a processing object is set on the main spindle 101 , but similar effects can be obtained when a tool is set on the main spindle 101 , for example, in a milling machine or a lathe.

[0093] In addition, in the first embodiment, a configuration is used in which one sensor 103 is used, but the sensors 103 may be provided at a plurality of locations of the machine tool 100. In this case, the processing described in the first embodiment is performed on all the provided sensors 103, and the occurrence of chattering can be determined. When the sensors 103 are provided at a plurality of locations, even if chattering occurs at a plurality of locations during machining, the chattering can be determined and suppressed.

[0094] In addition, the sensor 103 provided in the working machine 100 may detect vibration using two or more sensors 103, and output the sensor signal 200 including two or more state quantities to the filter processing unit 10. In this case, the sensor signal processing unit 11 can set the two sensor outputs as the first state quantity and the second state quantity, respectively. By directly measuring the second state quantity by the sensor 103, it is possible to determine the chatter vibration without being affected by the quantization error associated with differentiation or integration.

[0095] Furthermore, in the first embodiment, the sensor signal processing unit 11 extracts two types of state quantities, but may extract three or more types of state quantities.

[0096] Implementation method 2.

[0097] In the first embodiment, an example in which the number of blades of the tool is one is described. In the second embodiment, a configuration for determining the occurrence of chatter vibration when the number of blades of the tool is two or more is described.

[0098] Fig.13 1 is a diagram showing a functional structure of a numerical control device 2 according to Embodiment 2. The numerical control device 2 includes a filter processing unit 10, a sensor signal processing unit 11, a phase difference calculation unit 12, a vibration determination unit 13, a command value correction unit 14, a drive command unit 15, a timing signal generation unit 16-1, and a tool information recording unit 17.

[0099] The numerical control device 2 has the same structure as the numerical control device 1 according to the first embodiment, and further includes a tool information recording unit 17 that outputs tool information 210, and a timing signal generating unit 16-1 instead of the timing signal generating unit 16 of the numerical control device 1. The following mainly describes the parts different from the numerical control device 1.

[0100] The tool information recording unit 17 records the tool information 210, which is information related to the tool installed in the machine tool 100, and outputs the recorded tool information 210 to the timing signal generating unit 16-1. The tool information 210 includes at least information indicating the number of blades of the tool. In addition, the tool information 210 may include information indicating the shape of the tool, such as the type of tool such as an end mill, a turning tool, a tool length, and a tool diameter. In addition, when the machine tool 100 is equipped with one turning tool, the tool information recording unit 17 outputs the tool information 210 indicating that the number of blades of the tool is one. In addition, when the machine tool 100 is equipped with a turning tool, when multiple tools are cutting simultaneously, the number of blades becomes the same as the number of tools. For example, when the machine tool 100 has a lower tool post and an upper tool post, and when machining is performed by installing one turning tool on each of the two tool posts, the tool information recording unit 17 outputs the tool information 210 indicating that the number of blades is two.

[0101] The timing signal generating unit 16 - 1 determines whether the spindle angle has passed a predetermined angle based on the spindle operation command 206 output by the drive command unit 15 and the tool information 210 output by the tool information recording unit 17 , and outputs a timing signal 207 to the phase difference calculating unit 12 at the timing when the spindle angle has passed.

[0102] The details of the processing of the timing signal generating unit 16 - 1 will be described. The following describes a case where a tool having a number α of blades is mounted on the spindle 101 . α is a natural number equal to or greater than 2.

[0103] First, the case where α=2 will be described. Fig.14 It means by Fig.13The diagram of the first example of the relationship between the timing signal 207 output by the timing signal generating unit 16-1 shown in the figure and the angle command of the spindle 101. The angle command is a signal that takes a value between 0 degrees and 360 degrees and returns to 0 degrees when it reaches 360 degrees. The timing signal 207 is output each time the angle command of the spindle 101 passes through a predetermined plurality of angles, namely, angle ψ1 and angle ψ2. That is, the timing signal 207 is output twice during one rotation of the spindle 101, the same number of times as the number of blades. If the angle ψ1 is fixed at 1 during one rotation of the spindle 101, it can be an arbitrary angle. For example, the angle ψ1 can be set as the spindle angle when the spindle is oriented. In addition, the difference between the angle ψ1 and the angle ψ2 is taken as the value obtained by dividing 360 degrees by 2 of the number of tool blades, which is ψ2-ψ1=180 degrees.

[0104] Next, a case where α is an arbitrary natural number will be described. Fig.15 It means by Fig.13 1 is a diagram showing a second example of the relationship between the timing signal 207 output by the timing signal generating unit 16 - 1 and the angle command of the main shaft 101. The timing signal 207 is a timing signal generated by the timing signal generating unit 16 - 1 and the angle command of the main shaft 101. β Β is a natural number greater than or equal to 2 and less than or equal to α. The timing signal 207 is output α times during one rotation of the spindle 101. If ψ1 is fixed at 1 during one rotation of the spindle 101, it can be any angle. There is a restriction between ψ1 and ψβ-1 as shown in the following formula (2). In other words, the timing signal generating unit 16-1 outputs the timing signal 207 each time the spindle 101 rotates by an angle obtained by dividing 360 degrees by the number of blades α.

[0105] [Formula 2]

[0106]

[0107] The timing signal generating unit 16-1 may generate the timing signal 207 using the speed command of the spindle 101 included in the spindle operation command 206. In this case, the timing signal generating unit 16-1 generates the timing signal 207 in units of time T2 calculated from the speed command S (rpm) of the spindle 101 using the following equation (3).

[0108] [Formula 3]

[0109]

[0110] As described above, according to embodiment 2, a timing signal 207 is generated based on the tool information 210. Therefore, when the number of blades of the tool installed on the working machine 100 is multiple, as in embodiment 1, the time required for determining the vibration can be shortened, that is, the time from the occurrence of the vibration to the determination of the occurrence of the vibration.

[0111] Furthermore, when the number of blades is 1, the numerical control device 2 can determine and suppress chatter vibration by the same processing as that of the numerical control device 1 .

[0112] Implementation method 3.

[0113] Fig.16 1 is a diagram showing the functional structure of the numerical control device 3 according to the third embodiment. The numerical control device 3 includes a filter processing unit 10, a sensor signal processing unit 11, a phase difference calculation unit 12, a vibration determination unit 13, a command value correction unit 14, a drive command unit 15, and a timing signal generation unit 16-2. The numerical control device 3 includes the timing signal generation unit 16-2 instead of the timing signal generation unit 16 of the numerical control device 1. The following mainly describes the parts different from the numerical control device 1.

[0114] The machine tool 100 has an angle information sensor 104 mounted on the spindle 101 in addition to the structure described in the first embodiment. The angle information sensor 104 is a sensor for measuring the spindle angle, and an encoder, a potentiometer sensor, etc. can be used, for example. The angle information 211 is a signal including at least the actual value of the spindle angle measured by the angle information sensor 104. The angle information sensor 104 outputs the generated angle information 211 to the numerical control device 3.

[0115] The timing signal generating unit 16 - 2 generates the timing signal 207 based on the measured value of the spindle angle included in the angle information 211 . The method by which the timing signal generating unit 16 - 2 generates the timing signal 207 is the same as that of the first embodiment, but uses the measured value of the spindle angle instead of the angle command of the spindle 101 .

[0116] In addition, the angle information sensor 104 may also be a sensor that detects the measured value of the spindle speed. In this case, the angle information sensor 104 is an angular velocity sensor. The timing signal generating unit 16-2 can generate the timing signal 207 based on the measured value of the spindle speed instead of the angle command of the spindle 101. The timing signal generating unit 16-2 uses the initial time t0 as the reference time to generate the timing signal 207 based on the measured value S of the spindle speed. ref (rpm) The timing signal 207 is output as a unit of time T3 calculated using the following equation (4).

[0117] [Formula 4]

[0118]

[0119] As described above, according to the third embodiment, the timing signal 207 is generated based on the actual measured value of the spindle angle. Therefore, the timing signal 207 can be generated without being affected by the control following error generated between the angle command of the spindle 101 and the actual angle. Therefore, even if the control following error occurs between the angle command of the spindle 101 and the actual angle, the numerical control device 3 can shorten the time required for the determination of chatter vibration, that is, the time from the occurrence of chatter vibration to the determination of the occurrence of chatter vibration, as in the first embodiment.

[0120] Furthermore, in Embodiment 3, an example is described in which the timing signal 207 is generated using the measured value of the spindle angle or the spindle speed, but the timing signal generating unit 16 - 2 may use both the spindle angle command and the measured value of the spindle angle, or may use both the spindle speed command and the measured value of the spindle speed.

[0121] Implementation method 4.

[0122] Fig.17 1 is a diagram showing the functional structure of the numerical control device 4 according to the fourth embodiment. The numerical control device 4 includes a filter processing unit 10, a sensor signal processing unit 11, a phase difference calculation unit 12, a vibration determination unit 13, a command value correction unit 14, a drive command unit 15, a timing signal generation unit 16-3, and a tool information recording unit 17. That is, the numerical control device 4 further includes a tool information recording unit 17 on the basis of the structure of the numerical control device 3 described in the third embodiment, and includes a timing signal generation unit 16-3 instead of the timing signal generation unit 16-2. Below, the parts different from the numerical control device 2 are mainly described.

[0123] The timing signal generating unit 16-3 generates the timing signal 207 based on the angle information 211 outputted from the angle information sensor 104 and the tool information 210 outputted from the tool information recording unit 17, and outputs the generated timing signal 207 to the phase difference calculating unit 12. The timing signal generating unit 16-3 can generate the timing signal 207 by the same method as the timing signal generating unit 16-1 of the second embodiment. In this case, the timing signal generating unit 16-3 generates the timing signal 207 using the actual measured value of the spindle angle instead of the angle command of the spindle 101.

[0124] As described above, according to embodiment 4, similarly to embodiment 3, based on the effect of being able to generate the timing signal 207 without being affected by the control tracking error generated between the angle command of the spindle 101 and the actual angle, similarly to embodiment 2, when the number of blades is multiple and also similarly to embodiment 1, the time required for determining the chatter can be shortened, that is, the time from the occurrence of the chatter to the determination of the occurrence of the chatter.

[0125] Implementation method 5.

[0126] Fig.18 1 is a diagram showing the functional structure of the numerical control device 5 according to the fifth embodiment. The numerical control device 5 includes a filter processing unit 10, a sensor signal processing unit 11, a phase difference calculation unit 12, a vibration determination unit 13, a command value correction unit 14, a drive command unit 15, and a timing signal generation unit 16-4. In addition, the numerical control device 5 can obtain the machining program information 212 including the spindle speed command 213 from the machining program 18 for controlling the machine tool 100. Below, the parts different from the numerical control device 1 are mainly described.

[0127] The spindle speed command 213 is output to the timing signal generating unit 16 - 4 , and the machining program information 212 is output to the drive command unit 15 . The drive command unit 15 can output the operation command 208 to the machine tool 100 according to the machining program information 212 .

[0128] The timing signal generating unit 16-4 generates the timing signal 207 based on the spindle speed command 213 extracted from the machining program information 212. The timing signal generating unit 16-4 outputs the timing signal 207 in units of time T4 calculated using the spindle speed command 213 and the following equation (5) with the initial time t0 as the reference time. cmd The spindle speed command 213 indicates the spindle speed.

[0129] [Formula 5]

[0130]

[0131] As described above, according to the fifth embodiment, the timing signal 207 can be generated based on the spindle speed command 213 included in the machining program 18. When the timing signal 207 is generated using the machining program 18, the time required for determining the chatter vibration, that is, the time from the occurrence of the chatter vibration to the determination of the occurrence of the chatter vibration, can be shortened as in the first embodiment.

[0132] Implementation method 6.

[0133] Fig.191 is a diagram showing the functional structure of the numerical control device 6 according to the sixth embodiment. The numerical control device 6 includes a filter processing unit 10, a sensor signal processing unit 11, a phase difference calculation unit 12, a vibration determination unit 13, a command value correction unit 14, a drive command unit 15, a timing signal generation unit 16-5, and a tool information recording unit 17. The numerical control device 6 has the structure of the numerical control device 5, and further includes a tool information recording unit 17, and has a timing signal generation unit 16-5 instead of the timing signal generation unit 16-4.

[0134] The timing signal generating unit 16 - 5 generates the timing signal 207 based on the tool information 210 and the spindle speed command 213 , and outputs the generated timing signal 207 to the phase difference calculating unit 12 .

[0135] The timing signal generating unit 16 - 5 uses the initial time t0 as the reference time and uses the following equation (6), the number of blades α, and the spindle speed S cmd The timing signal 207 is outputted based on the calculated time T5.

[0136] [Formula 6]

[0137]

[0138] As described above, according to the sixth embodiment, in addition to the effects of the fifth embodiment, even when the tool has a plurality of cutting edges, it is possible to determine the occurrence of chatter vibration.

[0139] Next, the hardware configuration of the numerical control devices 1 to 6 according to the first to sixth embodiments will be described. Each functional unit of the numerical control devices 1 to 6 is implemented by a processing circuit. These processing circuits may be implemented by dedicated hardware or may be a control circuit using a CPU (Central Processing Unit).

[0140] In the case where the above-mentioned processing circuits are implemented by dedicated hardware, they are realized by Fig. 20 The processing circuit 90 shown is implemented. Fig. 20 The figure shows dedicated hardware for realizing the functions of the numerical control devices 1 to 6 according to the embodiments 1 to 6. The processing circuit 90 is a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof.

[0141] When the above-mentioned processing circuit is implemented by using a control circuit of a CPU, the control circuit is, for example, Fig.21The control circuit 91 of the structure shown. Fig.21 1 is a diagram showing the structure of a control circuit 91 for realizing the functions of the numerical control devices 1 to 6 according to the first to sixth embodiments. Fig.21 As shown, the control circuit 91 has a processor 92 and a memory 93. The processor 92 is a CPU, which is also called a central processing unit, a processing unit, an arithmetic unit, a microprocessor, a microcomputer, a DSP (Digital Signal Processor), etc. The memory 93 is, for example, a non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), a flash memory, an EPROM (Erasable Programmable ROM), an EEPROM (registered trademark) (Electrically EPROM), a magnetic disk, a floppy disk, an optical disk, a compact disk, a mini disk, a DVD (Digital Versatile Disk), etc.

[0142] When the processing circuit is realized by the control circuit 91, the processor 92 reads and executes a program corresponding to the processing of each component stored in the memory 93. The memory 93 is also used as a temporary memory in each processing executed by the processor 92.

[0143] The structure shown in the above embodiment is an example of the content of the present invention, and can be combined with other well-known technologies, and the embodiments can be combined with each other. Part of the structure can be omitted or changed without departing from the scope of the gist.

[0144] Description of the label

[0145] 1, 2, 3, 4, 5, 6 numerical control device, 10 filter processing unit, 11 sensor signal processing unit, 12 phase difference calculation unit, 13 vibration determination unit, 14 command value correction unit, 15 drive command unit, 16, 16-1, 16-2, 16-3, 16-4, 16-5 timing signal generation unit, 17 tool information recording unit, 18 machining program, 90 processing circuit, 91 control circuit, 92 processor, 93 memory, 100 machine tool, 101 spindle, 102 feed axis, 103 sensor, 104 angle information sensor, 200 sensor device signal, 201 chatter component signal, 202 state quantity signal, 203 phase difference information, 204 vibration judgment information, 205 correction signal, 206 spindle operation instruction, 207 timing signal, 208 operation instruction, 209 action information, 210 tool information, 211 angle information, 212 machining program information, 213 spindle speed instruction, 301 information observation unit, 302 inference unit, 303 inference data set, 304 learning data set, 305 chatter presence or absence information, 400 learning device, 401 data acquisition unit, 402 learning processing unit.

Claims

1. A numerical control device, It is characterized in that have: A drive command unit that gives operation commands to the main shaft and feed shaft of the machine tool; a sensor signal processing unit that performs time differentiation on a chatter component signal obtained from a sensor signal that detects vibration of a tool or a workpiece mounted on the machine tool to generate a first state quantity and a second state quantity, wherein a difference between the order of the time differentiation of the first state quantity and the order of the time differentiation of the second state quantity is an odd number; a timing signal generating unit that periodically generates and outputs a timing signal at a timing synchronized with the rotation of the main shaft; a phase difference calculation unit that generates, for each of the timing signals, a dimensionless quantity representing the first state quantity and the second state quantity in a state space, and calculates phase difference information representing a difference between phases of the dimensionless quantities; and A vibration determination unit determines the occurrence of chatter vibration in the machine tool based on the phase difference information.

2. The numerical control device according to claim 1, It is characterized in that The invention further includes a command value correction unit that outputs a correction signal for changing the operation command to the drive command unit when the vibration determination unit determines that chatter vibration has occurred.

3. The numerical control device according to claim 1 or 2, It is characterized in that The invention further comprises a filter processing unit for removing, from the sensor signal, a vibration component of a frequency obtained by multiplying the number of blades of a tool mounted on the machine tool by the rotation speed of the spindle and a harmonic component of the vibration component. The sensor signal processing unit generates a plurality of types of state quantities based on the sensor signal output by the filter processing unit.

4. The numerical control device according to claim 1 or 2, It is characterized in that The machine tool further comprises a tool information recording unit for recording blade number information indicating the number of blades of the tool mounted on the machine tool. The timing signal generating unit outputs the timing signal every time the main shaft rotates by an angle obtained by dividing 360 degrees by the number of blades.

5. The numerical control device according to claim 1 or 2, It is characterized in that The phase difference calculation unit repeatedly calculates the phase difference information indicating an angle formed by a previous dimensionless quantity, an origin of the state space, and a current dimensionless quantity.

6. The numerical control device according to claim 1 or 2, It is characterized in that The vibration determination unit determines that chatter vibration has occurred when a difference between the plurality of repeatedly received phase difference information is smaller than a predetermined value.

7. The numerical control device according to claim 1 or 2, It is characterized in that The vibration determination unit determines that chatter vibration has occurred when a standard deviation of the plurality of repeatedly received phase difference information is lower than a predetermined value.

8. The numerical control device according to claim 1 or 2, It is characterized in that The vibration determination unit calculates a difference between a maximum value and a minimum value of the phase difference information received during a predetermined length, and determines that chatter vibration has occurred when the difference between the maximum value and the minimum value is less than a predetermined value.

9. The numerical control device according to claim 1 or 2, It is characterized in that The timing signal generating unit generates the timing signal based on at least one of an angle command of the main shaft and a speed command of the main shaft.

10. The numerical control device according to claim 1 or 2, It is characterized in that The timing signal generating unit generates the timing signal based on at least one of an actual angle and an actual speed of the main shaft.

11. The numerical control device according to claim 1 or 2, It is characterized in that The timing signal generating unit generates the timing signal based on a spindle speed command described in a machining program for controlling the machine tool.

12. A method for determining the occurrence of chattering, the method being executed by a numerical control device having a drive command unit for giving operation commands to a main shaft and a feed shaft of a machine tool, comprising the following steps: Performing time differentiation on a chatter component signal obtained from a sensor signal detecting vibration of a tool or a workpiece mounted on the machine tool to generate a first state quantity and a second state quantity, wherein the difference between the order of the time differentiation of the first state quantity and the order of the time differentiation of the second state quantity is an odd number; periodically generating a timing signal at a timing synchronized with the rotation of the main shaft; generating, for each of the timing signals, a dimensionless quantity representing the first state quantity and the second state quantity in a state space, and calculating phase difference information representing a difference between phases of the dimensionless quantities; and The occurrence of chatter vibration in the machine tool is determined based on the phase difference information.

13. A numerical control device, It is characterized in that have: A drive command unit that gives operation commands to the main shaft and feed shaft of the machine tool; a sensor signal processing unit that performs time differentiation on a chatter component signal obtained from a sensor signal that detects vibration of a tool or a workpiece mounted on the machine tool to generate a first state quantity and a second state quantity, wherein a difference between the order of the time differentiation of the first state quantity and the order of the time differentiation of the second state quantity is an odd number; a timing signal generating unit that periodically generates and outputs a timing signal at a timing synchronized with the rotation of the main shaft; a phase difference calculation unit that generates, for each of the timing signals, a dimensionless quantity representing the first state quantity and the second state quantity in a state space, and calculates phase difference information representing a difference between phases of the dimensionless quantities; and A vibration determination unit includes an information observation unit that creates an inference data set based on the phase difference information, and an inference unit that has a trained inference model for machine learning that determines the presence or absence of flutter by inputting the inference data set, and the occurrence of flutter is determined by inputting the inference data set into the inference unit.

14. A learning device, It is characterized in that have: A data acquisition unit that acquires phase difference information from a numerical control device and generates a learning data set that associates the phase difference information with the presence or absence of chatter in a working machine, the numerical control device comprising: a drive command unit that gives operation commands to a spindle and a feed axis of the working machine; a sensor signal processing unit that performs time differentiation on a chatter component signal obtained from a sensor signal that detects vibration of a tool or a processing object installed on the working machine to generate a first state quantity and a second state quantity, wherein the difference between the order of the time differentiation of the first state quantity and the order of the time differentiation of the second state quantity is an odd number; a timing signal generating unit that generates and outputs a timing signal periodically at a timing synchronized with the rotation of the spindle, generates a dimensionless quantity representing the first state quantity and the second state quantity in a state space for each timing signal, and a phase difference calculating unit that calculates phase difference information representing the difference between the phases of the dimensionless quantities; as well as The learning processing unit performs machine learning of the estimation model so that a numerical value corresponding to the presence or absence of chattering is output when the phase difference information is input.