Breakage determination device and computer-readable storage medium

Abstract

This breakage determination device: obtains samples of the value of the load applied to a shaft of a machine tool; determines whether the samples were detected during a cutting time period or detected during a non-cutting time period; calculates a probability density function of samples during the non-cutting time period; calculates a feature amount of samples during the cutting time period; and determines breakage of a tool on the basis of the probability of occurrence of the feature amount in the probability density function for the non-cutting time period.

Description

Breakage detection device and computer-readable storage medium

[0001] The present disclosure relates to a breakage determination device and a computer-readable storage medium.

[0002] Conventionally, in a device for detecting or predicting breakage of a tool used in a machine tool, there is a technology in which a feature amount of load during a cutting time period is multiplied by a coefficient to calculate a fluctuating threshold value and determine breakage. For example, Patent Document 1 discloses such a technology.

[0003] Japanese Patent Application Laid-Open No. 2004-130407

[0004] When load is used to detect breakage, trial machining is required, so breakage cannot be determined from the first machining.

[0005] In the field of numerical control devices, it is desirable to make breakage detection more efficient.

[0006] The breakage detection device includes a data acquisition unit that acquires the value of the load applied to the axis of the machine tool as a sample, a determination unit that determines whether the sample acquired by the data acquisition unit is a sample detected during a cutting time period or a sample detected during a non-cutting time period, a probability calculation unit that calculates a probability density function of the sample during the non-cutting time period, a feature calculation unit that calculates a feature of the sample during the cutting time period, and a breakage detection unit that determines tool breakage based on the probability of the feature occurring in the probability density function during the non-cutting time period.

[0007] 1 is a block diagram of a breakage determination device. FIG. 2 is a graph showing changes in commands during cutting. FIG. 3 is a graph of samples during non-cutting time periods and cutting time periods under normal conditions. FIG. 4 is a graph showing the relationship between the probability density function and feature values ​​during non-cutting time periods. FIG. 5 is a graph of samples during non-cutting time periods and cutting time periods when the tool breaks. FIG. 6 is a graph showing the relationship between the probability density function and feature values ​​during non-cutting time periods. FIG. 7 is a flowchart explaining the operation of a breakage determination device. FIG. 8 is a graph of load under normal conditions. FIG. 9 is a graph of load when the tool breaks before the cutting time period. FIG. 10 is a graph of the probability density function of samples during the non-cutting section. FIG. 11 is a graph showing the relationship between the probability density function and feature values. FIG. 12 is a graph of load when the tool breaks during the cutting time period. FIG. 13 is a graph showing the relationship between the probability density function and feature values. FIG. 14 is a block diagram of a modified example of a breakage determination device. FIG. 15 is a hardware configuration diagram of a breakage determination device.

[0008] The following describes the breakage detection device 100. The breakage detection device 100 is realized by an information processing device such as a numerical control device or a PC (personal computer).

[0009] 1 is a block diagram of a breakage determination device 100. The breakage determination device 100 includes a data acquisition unit 10, a determination unit 11, a probability calculation unit 12, a feature calculation unit 13, and a breakage detection unit 14.

[0010] The data acquisition unit 10 acquires at least one of the loads of the spindle and the feed axis. Hereinafter, the load value will be referred to as a sample. The judgment unit 11 judges whether the sample acquired by the data acquisition unit 10 was detected during a cutting time period or a non-cutting time period. A cutting time period is a time period during which the tool is cutting the workpiece. A non-cutting time period is a time period during which the tool is not cutting the workpiece. One method of judgment is to detect a signal. For example, during a cutting time period, a "cutting in progress" command signal is output from the numerical control device. Figure 2 shows an example of a "cutting in progress" command signal. The upper graph in Figure 2 shows the load value, and the lower graph in Figure 2 shows the "cutting in progress" command signal. When the numerical control device analyzes a machining program and executes a cutting command such as G01, the "cutting in progress" command signal turns "ON." This signal can be used to determine whether the time period is a cutting time period or a non-cutting time period.

[0011] The probability calculation unit 12 calculates a probability density function of samples acquired during non-cutting time periods. FIG. 3 is a graph showing changes in samples during normal times (when the tool is not broken). The feature calculation unit 13 calculates feature values ​​for samples acquired during cutting time periods. Feature values ​​are representative values ​​of samples acquired during cutting time periods. Feature values ​​include average values, median values, maximum values, minimum values, and modes. Feature values ​​may also be moving averages, moving median values, moving maximum values, moving minimum values, and moving modes. The type of feature value is not limited.

[0012] The breakage detection unit 14 compares the probability density function of the samples during the non-cutting time period with the feature values ​​of the samples during the cutting time period. Since the load value during the cutting time period is unlikely to appear during the non-cutting time period, if the probability of the feature value of the sample during the cutting time period occurring is sufficiently small, it is determined that the tool is not broken. For example, point A in Figure 4 indicates the value of the feature value during the cutting time period under normal conditions. If the tool is not broken, the probability that the feature value occurs during the non-cutting time period is infinitesimally close to zero. The breakage detection unit 14 substitutes the value of the feature value into the probability density function, and if the probability of the feature value occurring is sufficiently small, it is determined that the tool is not broken.

[0013] FIG. 5 shows an example in which a tool is broken. In this example, the sample value during the cutting time period is low. Point A in FIG. 6 indicates the value of the feature during the cutting time period after the tool breakage. When the tool is broken, there is a high possibility that the value of the feature will occur during the non-cutting time period. The breakage detection unit 14 substitutes the value of the feature into a probability density function, and if the probability of the value of the feature occurring is sufficiently high, it determines that the tool is broken.

[0014] The load waveform changes depending on the cause of breakage, material, type of tool, timing of breakage (during cutting or not during cutting), etc. Therefore, multiple feature values ​​are prepared to respond to various waveform changes.

[0015] 7 is a flowchart illustrating the operation of the breakage determination device 100. The numerical control device is started, setup is completed, and machining is started (step S1). The numerical control device moves the axis at rapid traverse and rotates the spindle at the cutting start position. The data acquisition unit 10 acquires the spindle load value as a sample. The judgment unit 11 judges whether the sample acquired by the data acquisition unit 10 is a sample detected during a cutting time period or a sample detected during a non-cutting time period. The data acquisition unit 10 acquires the sample during the non-cutting time period (step S2).

[0016] The probability calculation unit 12 calculates a probability density function of samples from the non-cutting time period (step S3). The data acquisition unit 10 acquires samples from the cutting time period (step S4). The feature calculation unit 13 calculates feature quantities of the samples from the cutting time period (step S5).

[0017] The breakage detection unit 14 substitutes the feature value into a probability density function to calculate the probability that the feature value occurs during the non-cutting time period (step S6). The breakage detection unit 14 compares the probability with a threshold value (which may be zero), and if the probability is sufficiently low, determines that the tool is normal. If the probability is sufficiently high, determines that the tool is broken (step S7).

[0018] As described above, the breakage detection device 100 of this embodiment calculates the probability density distribution of samples during the cutting time period and the feature values ​​during the non-cutting time period, calculates the probability that the feature values ​​will take on values ​​during the non-cutting time period, and determines that there is no breakage if the probability that the feature values ​​during the cutting time period will occur during the non-cutting time period is sufficiently low.The breakage detection device 100 of this embodiment determines breakage without trial machining, eliminating the need for trial machining and making the breakage detection work more efficient.

[0019] (Specific Example 1) In specific example 1, a method for determining tool breakage before the cutting period will be described. Fig. 8 shows a load waveform under normal conditions, and Fig. 9 shows a load waveform when the tool breaks before the cutting period. As shown in Fig. 8, the load under normal conditions rises at the start of cutting, remains approximately constant during cutting, and returns to the value of the non-cutting period when cutting ends, maintaining an approximately constant value during the non-cutting period. When the tool breaks before the cutting period, as shown in Fig. 9, there is no change in the load between the non-cutting period and the cutting period, and the load maintains the same value as during the non-cutting period.

[0020] In this example, the feature value is the maximum value of the samples during the cutting time period. The breakage detection unit 14 determines that there is no breakage if the probability of obtaining the feature value during the cutting time period during the non-cutting time period is approximately 0.03% or less. Figure 10 shows the probability density function of the samples during the non-cutting time period. In this example, the probability density function is a normal distribution with a mean μ (= 0.043) and a variance σ (= 0.043). The breakage detection unit 14 calculates the "3σ interval (99.7 percentile)" for the non-cutting time period. The "3σ interval" for the non-cutting time period is an interval between ±3σ centered on the mean value μ. The value of the "3σ interval" occurs with a probability of 99.7%. In Figure 10, it is [-0.086, 0.172]. The breakage detection unit 14 determines breakage based on whether the feature value is included in the "3σ interval." As shown in Figure 8, if the tool is not broken, the feature value (maximum value of the sample in the cutting section) is "2.714", which is not included in the "3σ section", and the probability that the feature value will occur in the non-cutting time period is sufficiently low. In other words, the probability that the feature value "2.714" will occur in the non-cutting time period is lower than 0.03%. Therefore, it can be determined that the tool is not broken.

[0021] As shown in Figure 9, if the tool breaks before the cutting period, the load during the cutting period will be approximately the same value as during the non-cutting period. If the feature value during the cutting period (the maximum value of the sample during the cutting period) is "0.047," as shown in Figure 11, the feature value "0.047" is included in the "3σ interval," so the probability that the feature value during the cutting period (the maximum value of the sample during the cutting period) will occur during the non-cutting period is sufficiently high. In other words, the probability that the feature value "0.047" will occur during the non-cutting period is higher than 0.03%. Therefore, it can be determined that the tool is broken.

[0022] (Specific Example 2) In specific example 2, a method for determining tool breakage during a cutting time period will be described. FIG. 12 shows the load waveform when a tool breaks during a cutting time period. When a tool breaks during a cutting time period, as shown in FIG. 12, the load rises at the start of cutting, then gradually increases, then rises sharply just before breakage, and returns to the value in a non-cutting time period after breakage. Note that the load waveform under normal conditions is the same as in FIG. 8. Therefore, the probability density distribution during a non-cutting time period is also the same as in FIG. 10.

[0023] In this example, the feature value is the minimum value of the samples during the cutting time period. The breakage detection unit 14 determines that the tool is not broken if the probability that the feature value during the cutting time period is obtained during the non-cutting time period is 0.03% or less. If the tool breaks during the cutting time period, the feature value (minimum value of the samples during the cutting time period) will be "0.042." Since the feature value "0.042" is included in the "3σ interval" as shown in Figure 13, the probability that the feature value during the cutting time period (minimum value of the samples during the cutting time period) will occur during the non-cutting time period is sufficiently high. In other words, the probability that the feature value "0.042" will occur is higher than 0.03%. Therefore, it can be determined that the tool is broken.

[0024] As described above, breakage was detected under different conditions by using the maximum value of the samples during the cutting time period as the feature value in Specific Example 1, and the minimum value of the samples during the cutting time period as the feature value in Specific Example 2. By providing a plurality of feature values, the breakage determination device 100 of this embodiment can detect breakage under various conditions.

[0025] (Modification) A modification of the breakage detection device 100 combines two types of detection methods. The breakage detection device 100 in Fig. 14 is configured by adding a change amount unit 15, a second probability calculation unit 16, and a second breakage detection unit 17 to the breakage detection device 100 in Fig. 1.

[0026] The change amount unit 15 calculates the change amount of the samples acquired during the cutting time period. The change amount of the sample is the increase or decrease value of the sample. The change amount unit 15 calculates the difference between the samples by subtracting the value of the sample at a certain point in time from the value of the sample before or after that point.

[0027] The second probability calculation unit 16 calculates a probability density function of the change amount. The second probability calculation unit 16 performs preprocessing. In the preprocessing, some of the change amounts are excluded from the samples. The second probability calculation unit 16 excludes values ​​with large absolute values. In other words, values ​​with large changes are excluded regardless of whether they are positive or negative. The exclusion criteria may be the order of the magnitude of the change, or a threshold value. The second breakage detection unit 17 uses the probability density function of the change amount of the sample to determine the probability of occurrence of the excluded value, and determines that the tool has broken if the probability of occurrence of the excluded value is sufficiently small. In other words, after excluding some of the change amounts, if the probability density function calculated from the remaining change amounts shows a sufficiently small probability of occurrence of the excluded value, the second breakage detection unit 17 determines that the tool has broken.

[0028] In this modification, breakage can be determined using both the slope and the sample value, which broadens the range of breakage determination. Also, by providing multiple feature amounts, breakage can be detected under various circumstances.

[0029] The hardware configuration of the breakage detection device 100 to which the present disclosure is applied will be described below. Fig. 15 is a hardware configuration diagram of the breakage detection device 100. As shown in Fig. 15, the breakage detection device 100 includes a CPU 111 that controls the entire breakage detection device 100, a ROM 112 that records programs and data, and a RAM 113 for temporarily expanding data. The CPU 111 reads out a system program recorded in the ROM 112 via a bus and determines whether a breakage has occurred in accordance with the system program.

[0030] The nonvolatile memory 114 is backed up by, for example, a battery (not shown), and the stored state is maintained even when the power to the breakage detection device 100 is turned off. The nonvolatile memory 114 stores various data such as programs read from the external device 120 via the interfaces 115, 118, and 119 and operation inputs input via the input unit 30. The nonvolatile memory 114 may store programs and data for executing the breakage detection device 100 of this embodiment.

[0031] The interface 115 is an interface for connecting the breakage detection device 100 to an external device 120 such as an adapter. Programs, various parameters, etc. are loaded from the external device 120. The interface 118 is an interface for connecting the breakage detection device 100 to a display unit 70 such as a liquid crystal display. The display unit 70 displays various data loaded into memory, data obtained as a result of executing programs, etc. The interface 119 is an interface for connecting the breakage detection device 100 to an input unit 30 such as a keyboard or pointing device. The input unit 30 passes commands, data, etc. based on operations by an operator to the CPU 111 via the interface 119.

[0032] Although the present disclosure has been described in detail, the present disclosure is not limited to the individual embodiments described above. Various additions, substitutions, modifications, partial deletions, etc. are possible in these embodiments without departing from the gist of the present disclosure or the gist of the present disclosure derived from the claims and their equivalents. Furthermore, these embodiments can also be implemented in combination. For example, in the above-described embodiments, the order of each operation and the order of each process are shown as examples and are not limited to these.

[0033] The following supplementary notes are further disclosed regarding the above-described embodiment and modified examples. (Supplementary Note 1) A breakage determination device (100) includes a data acquisition unit (10) that acquires samples of load values ​​applied to a machine tool shaft, a determination unit (11) that determines whether the samples acquired by the data acquisition unit are samples detected during a cutting time period or a non-cutting time period, a probability calculation unit (12) that calculates a probability density function of the samples during the non-cutting time period, a feature calculation unit (13) that calculates feature values ​​of the samples during the cutting time period, and a breakage detection unit (14) that determines tool breakage based on the probability of occurrence of the feature value in the probability density function for the non-cutting time period. (Supplementary Note 2) The determination unit (11) determines whether the time period is a cutting time period or a non-cutting time period based on a command signal that controls the machine tool. (Supplementary Note 3) The breakage detection unit (14) determines that the tool is not broken if the probability of occurrence of the feature value is sufficiently small. (Supplementary Note 4) The feature calculation unit (13) calculates a plurality of feature values, and the breakage detection unit uses the plurality of feature values ​​to detect breakage under different conditions. (Supplementary Note 5) The feature value is a maximum value of samples in a cutting section. (Supplementary Note 6) The feature value is a minimum value of samples in a cutting section. (Supplementary Note 7) The computer-readable storage medium (112, 113, 114) stores instructions that cause one or more processors (111) to execute processing to acquire a value of a load applied to a shaft of a machine tool as a sample, determine whether the sample is a sample detected during a cutting time period or a sample detected during a non-cutting time period, calculate a probability density function of the sample during the non-cutting time period, calculate feature values ​​of the sample during the cutting time period, and determine tool breakage based on the probability of occurrence of the feature value in the probability density function for the non-cutting time period.

[0034] REFERENCE SIGNS LIST 100 Breakage determination device 10 Data acquisition unit 11 Determination unit 12 Probability calculation unit 13 Feature calculation unit 14 Breakage detection unit 111 CPU 112 ROM 113 RAM 114 Non-volatile memory

Claims

1. A breakage determination device comprising: a data acquisition unit that acquires, as a sample, a value of a load applied to a shaft of a machine tool; a determination unit that determines whether the sample acquired by the data acquisition unit is a sample detected in a cutting time zone or a sample detected in a non-cutting time zone; a probability calculation unit that calculates a probability density function of the samples in the non-cutting time zone; a feature amount calculation unit that calculates a feature amount of the samples in the cutting time zone; and a breakage detection unit that determines breakage of a tool based on a probability of occurrence of the feature amount in the probability density function of the non-cutting time zone.

2. The breakage determination device according to claim 1, wherein the determination unit determines a cutting time zone and a non-cutting time zone based on a command signal for controlling the machine tool.

3. The breakage determination device according to claim 1, wherein the breakage detection unit determines that the tool has not broken when the probability of occurrence of the feature amount is sufficiently small.

4. The breakage determination device according to claim 1, wherein the feature amount calculation unit calculates a plurality of feature amounts, and the breakage detection unit detects breakage under different situations using the plurality of feature amounts.

5. The breakage determination device according to claim 1, wherein the feature amount is a maximum value of samples in a cutting section.

6. The breakage determination device according to claim 1, wherein the feature amount is a minimum value of samples in a cutting section.

7. A computer-readable storage medium storing instructions for causing one or more processors to execute a process of acquiring, as a sample, a value of a load applied to a shaft of a machine tool, determining whether the sample is a sample detected in a cutting time zone or a sample detected in a non-cutting time zone, calculating a probability density function of the samples in the non-cutting time zone, calculating a feature amount of the samples in the cutting time zone, and determining breakage of a tool based on a probability of occurrence of the feature amount in the probability density function of the non-cutting time zone.