JP4860444B2 - Abnormality detection method in cutting - Google Patents

Description

The present invention relates to a cutting related to abnormality detection how in the processing, abnormality detection how to detect the abnormal working like monitors, especially vibration during cutting as needed.

  Cutting is a work that is a work material such as metal, cut with a cutting tool such as a drill, bite, or miller into a predetermined shape, and is used to manufacture molds and parts for modern industrial products. It is indispensable. As described above, since the workpiece is cut by the cutting tool in the cutting process, the cutting tool is gradually worn out by using for a long time. In addition, cutting tools may be damaged during cutting due to fatigue or sudden overload. If the wear of the cutting tool exceeds the allowable range or breaks during the cutting process, if the cutting process is continued as it is, the dimensional accuracy and surface roughness of the workpiece will deteriorate, resulting in product defects. However, it is not preferable because of the load.

  For this reason, many processing abnormality detection apparatuses that detect an abnormality during cutting more accurately and quickly have been studied and developed. For example, in the invention disclosed in the following [Patent Document 1], first, data such as vibration value, temperature, current value, etc. over the entire one-cycle operation from the start to the end of normal cutting. Is obtained in advance to create reference waveform data. Then, a threshold value (upper and lower limit values) from the start to the end of cutting is set based on the reference waveform data. This threshold value is corrected by data acquired at the start of cutting, and then compared with each data acquired during cutting. When the data during the cutting process exceeds the threshold value, it recognizes that an abnormality has occurred and issues an alarm.

  In the invention disclosed in [Patent Document 2] below, active power waveform data during normal cutting is acquired in advance. Then, after the obtained active power waveform data is divided into a plurality of regions, a threshold value (alarm set value) is set for each divided region based on the active power waveform data. And when the active power waveform data acquired at the time of cutting exceeds this threshold value, it recognizes that abnormality has occurred and issues an alarm.

JP-A-6-201398

JP 2006-82154 A

  However, even if the same material is cut with the same cutting tool, the data acquired at the time of cutting varies from day to day depending on the degree of wear of the cutting tool, how it is mounted, the lot of the workpiece, the state of the cutting machine, etc. Is. In the inventions disclosed in [Patent Document 1] and [Patent Document 2], the threshold value to be set is corrected for each cutting process, or the reference data used for setting the threshold value is acquired a plurality of times. Measures such as averaging are taken. However, even if these countermeasures are taken, the invention disclosed in [Patent Document 1] and [Patent Document 2] in which threshold values are set in advance as long as there is a variation in the data at the time of cutting is not abnormal. If the detection accuracy is increased, erroneous detection occurs, and if the detection accuracy is lowered, there is a possibility that a detection failure may occur.

  Further, in the inventions disclosed in [Patent Document 1] and [Patent Document 2], when the cutting process is delayed or progressed for a certain time from a predetermined time, the set threshold value and the data at the actual cutting process are There is a possibility that a time lag will occur between the two and false detection will occur frequently.

The present invention has been made in view of the above circumstances, is less affected by variations in data acquired during cutting, and does not cause a time shift between the data during cutting and a threshold value. and to provide an abnormality detecting how in precision machining.

The present invention
(1) In an abnormality detection method for detecting abnormality during cutting from vibration data Sb generated during cutting,
Acquiring idling vibration data generated before cutting;
Calculating an idling amplitude value E based on an absolute value of the idling vibration data;
Calculating an idling amplitude upper limit F based on the idling amplitude value E;
Obtaining vibration data Sb generated during cutting;
Calculating amplitude data Sb1 based on the absolute value of the vibration data Sb;
Calculating a short-term average amplitude value A by performing a moving average process on the amplitude data Sb1 at predetermined time intervals t;
Calculating a long-term average amplitude value B by performing a moving average process on the amplitude data Sb1 at a time t ′ interval longer than the time t interval;
Calculating a threshold based on the long-term average amplitude value B;
When an abnormality is detected by comparing the short-term average amplitude value A with the threshold value and the short-term average amplitude value A exceeds the idling amplitude upper limit F , an abnormality detection signal Sa is output, and the short-term average amplitude Invalidating the abnormality detection by the threshold when the value A is below the idling amplitude upper limit F ;
The problem is solved by providing a method for detecting an abnormality in a cutting process characterized by comprising:
(2) The abnormality detection signal Sa is output when the short-term average amplitude value A exceeds the threshold continuously for a predetermined time Ta and Tc and the short-term average amplitude value A exceeds the idling amplitude upper limit value F. The problem is solved by providing a method for detecting an abnormality in the cutting process described in (1) above.
(3) In addition, when the short-term average amplitude value A falls below the idling amplitude upper limit F for a predetermined time Te,
While resetting the long-term average amplitude value B and the short-term average amplitude value A,
If the short-term average amplitude value A is below the idling amplitude upper limit F even after a predetermined time has elapsed after resetting,
By providing the abnormality detection method in the cutting process according to the above (1) or (2), wherein the abnormality detection signal Sa is output, the above problem is solved .

Abnormality detecting how the cutting according to the present invention, from the order said hands,
Even if there is a variation in the vibration data acquired during the cutting process, a threshold value corresponding to the variation is calculated, so that it is possible to detect an abnormality during the cutting process with high accuracy. Further, since the short-term average amplitude and the threshold value, which are used as comparative controls, are calculated based on amplitude data obtained from the same vibration data, there is no time deviation and no erroneous detection associated therewith.

Be described with reference to the accompanying drawings, embodiments of the abnormality detecting how the cutting according to the present invention. FIG. 1 is a schematic diagram showing a configuration of a machining abnormality detection apparatus to which an abnormality detection method according to the present invention is applied . FIG. 2 is a diagram illustrating a method for calculating a short-term average amplitude value and a long-term average amplitude value according to the present invention. 3 and 4 are diagrams for explaining an abnormality detection method in cutting according to the present invention .

A machining abnormality detection device 50 to which the abnormality detection method according to the present invention shown in FIG. 1 is applied acquires vibration data Sb of the cutting tool 14 from vibration data acquisition means 18 installed in the cutting tool 14 of the cutting machine 10, and the A short-term average amplitude value, a long-term average amplitude value, and upper and lower limit values are calculated by performing arithmetic processing described later on the vibration data Sb. When the short-term average amplitude value satisfies a predetermined condition to be described later, the abnormality detection signal Sa is output to the alarm output means 16 or the control unit 11 of the cutting machine 10 or both. When the abnormality detection signal Sa is input to the alarm output means 16, the alarm output means 16 notifies the operator or the like of the occurrence of abnormality by a predetermined alarm sound, sound, lighting of an alarm lamp, or the like. Further, when the abnormality detection signal Sa is input to the control unit 11 of the cutting machine 10, the control unit 11 performs predetermined control such as deceleration stop and immediate stop of the cutting machine 10.

  The cutting machine 10 is attached with a workpiece 12 as a work material and a cutting tool 14 such as a drill, a cutting tool, and a milling cutter. The cutting machine 10 cuts the workpiece 12 into a predetermined shape by rotating or moving the workpiece 12 and / or the cutting tool 14. At this time, vibration is generated in the cutting tool 14, the cutting machine 10, and the workpiece 12 due to the cutting tool 14 cutting the workpiece 12. The vibration data acquisition means 18 such as an acceleration sensor acquires the vibration generated in the cutting tool 14 as vibration data Sb that is an electrical signal and outputs the vibration data Sb to the machining abnormality detection device 50. The vibration data acquisition means 18 does not necessarily have to be installed on the cutting tool 14, but is installed on the cutting machine 10 or the workpiece 12 side, and the machining abnormality detection device uses the vibration of the cutting machine 10 or the workpiece 12 as vibration data Sb. 50 may be output.

Next, a method for calculating the short-term average amplitude value and the long-term average amplitude value of the abnormality detection method according to the present invention will be described. FIG. 2 is a diagram showing, in time series, vibration data Sb, amplitude data Sb1, short-term average amplitude value, long-term average amplitude value, and the like when time 5t has elapsed. Note that FIG. 2 uses an example in which the time t ′ for calculating the long-term average amplitude value B is three times the time t.

  As shown in FIG. 2, the vibration data Sb of the cutting tool 14 input to the machining abnormality detection device 50 is data distributed positively and negatively around 0, and preferably has a data number of about 12,500 per second. Since the vibration data Sb is an electric signal, it is actually a value such as a voltage. However, since this electric signal indirectly represents the amplitude of vibration of the cutting tool 14, it will be referred to as the amplitude of the cutting tool 14. I will do it.

  As a first step, vibration data Sb is input to the machining abnormality detection device 50. Then, as the next step, the machining abnormality detection device 50 calculates the absolute value of the displacement of the vibration data Sb and sets it as the absolute value data Sb ′ in which the negative value of the vibration data Sb is converted to a positive value, and then for a certain period of time. The absolute value data Sb ′ within or a predetermined number of data is averaged to obtain amplitude data Sb1. In FIG. 2, two absolute value data Sb 'are averaged to obtain amplitude data Sb1, but in reality, the amplitude data Sb1 is obtained by averaging thousands of absolute value data Sb'. In addition to the above average value, the amplitude data Sb1 may be a predetermined number of data or the maximum value, the center value, the mode value, etc. of the absolute value data Sb 'within a certain time.

When the amplitude data Sb1 is obtained, as a next step, the machining abnormality detection device 50 calculates a short-term average amplitude value A by taking a moving average of the obtained amplitude data Sb1 at a predetermined time t. At the same time, a long-term average amplitude value B is calculated by taking a moving average at a predetermined time t ′ longer than the time t. That is, when the time 3t has elapsed, the value of the short-term average amplitude value A becomes the value A _(3t) that is the average value of the amplitude data Sb1 during the time 2t to 3t, and the value of the long-term average amplitude value B changes from A value B _(3t) which is an average value of the amplitude data Sb1 between them is obtained. When the time 5t has elapsed, the value of the short-term average amplitude value A becomes the value A _(5t) that is the average value of the amplitude data Sb1 during the time 4t to _5t , and the value of the long-term average amplitude value B is the time 2t to 5t. A value B _(5t) which is an average value of the amplitude data Sb1 between them is obtained. Note that these moving average processes are continuously performed whenever new amplitude data Sb1 is acquired. The time t for calculating the short-term average amplitude value A and the time t ′ for calculating the long-term average amplitude value B are preferably input in units of several seconds at the time of initial setting of the machining abnormality detection device 50, and basically cut. It does not change during operation of the processing machine 10.

Next, the abnormality detecting method of this invention FIG. 3 will be described with reference to FIG. FIG. 3 shows time-series amplitude values such as short-term average amplitude value A and long-term average amplitude value B, threshold values, and the like. FIG. 3 shows an example in which three upper limit values C, a second upper limit value C ′ larger than the upper limit value C, and a lower limit value D are provided as threshold values.

First, in the abnormality detection method of the present invention, when the long-term average amplitude value B is calculated, as the next step, an upper limit value C, a second upper limit value C ′, and a lower limit value D, which are threshold values for recognizing the occurrence of an abnormality, are obtained. Calculation is based on the value of the long-term average amplitude value B. For this reason, the upper limit value C, the second upper limit value C ′, and the lower limit value D, which are threshold values, are updated as needed as the long-term average amplitude value B is updated.

  As a method for calculating the upper limit value C, the second upper limit value C ′, and the lower limit value D, there are a method in which a predetermined value is added to or subtracted from the long-term average amplitude value B, and a method in which a predetermined value is multiplied. That is, a method in which a predetermined value is added to the long-term average amplitude value B to obtain an upper limit value C and a second upper limit value C ′, and a predetermined value is subtracted from the long-term average amplitude value B to obtain a lower limit value D; When it is desired to set the amplitude value B to a predetermined value, for example, the lower limit value D is 80% of the long-term average amplitude value B, a value obtained by multiplying the long-term average amplitude value B by 0.8 is calculated as the lower limit value D. Is 120% of the long-term average amplitude value B, and the second upper limit value C ′ is 150% of the long-term average amplitude value B, the values obtained by multiplying the long-term average amplitude value B by 1.2 and 1.5, respectively. There is a method of calculating the upper limit value C and the second upper limit value C ′. In particular, the method of calculating the threshold value by multiplying the long-term average amplitude value B by a predetermined value increases detection accuracy because the range of the upper limit value C and the lower limit value D increases and decreases as the value of the long-term average amplitude value B increases and decreases. The threshold value calculation method is more preferable.

  Note that the above threshold value calculation methods can be appropriately combined as necessary. For example, the upper limit value C is calculated by multiplication, and the lower limit value D and the second upper limit value C ′ are calculated by addition / subtraction. Is also possible. Further, the input of a value for calculating the threshold is basically performed at the time of initial setting of the machining abnormality detection device 50.

When the threshold value is calculated as described above, as a next step, the short-term average amplitude value A is compared with the threshold value as shown below, and the abnormality detection signal Sa is output when a predetermined condition is satisfied.

  Consider a case where the vibration of the cutting tool 14 becomes larger than usual due to the wear of the cutting tool 14 exceeding an allowable amount during cutting, and the amplitude data Sb1 increases. In such a case, the short-term average amplitude value A having a short interval of time t for taking the moving average immediately reflects the change of the amplitude data Sb1 and shows an increasing tendency. However, the long-term average amplitude value B having a long interval of the time t ′ at which the moving average is taken does not immediately reflect the change of the amplitude data Sb1, and shows an increasing tendency after a certain time delay. Accordingly, the upper limit value C calculated based on the long-term average amplitude value B also shows an increasing tendency after being delayed by the same time as the long-term average amplitude value B. At this time, by appropriately setting a value for calculating the upper limit value C, a state in which the short-term average amplitude value A exceeds the upper limit value C occurs as shown in the region a in FIG. The machining abnormality detection device 50 recognizes that an abnormality has occurred when such a state continues for a predetermined time Ta, and outputs an abnormality detection signal Sa to the alarm output means 16, the control unit 11, and the like.

  Further, when the amplitude data Sb1 from the cutting tool 14 increases instantaneously due to a loss of the cutting tool 14 at the time of cutting, the short-term average amplitude value A reflects a change in the amplitude data Sb1 and rapidly increases. Indicates. However, since the second upper limit value C ′ calculated based on the long-term average amplitude value B having a long interval of the time t ′ at which the moving average is taken does not immediately reflect the change in the amplitude data Sb1, the point b in FIG. As shown, a state occurs in which the short-term average amplitude value A exceeds the second upper limit value C ′. When such a state occurs, the machining abnormality detection device 50 recognizes that an abnormality has occurred, and instantaneously outputs an abnormality detection signal Sa to the alarm output means 16, the control unit 11, and the like.

  Further, consider a case where some abnormality occurs during the cutting process, the vibration of the cutting tool 14 becomes smaller than usual, and the amplitude data Sb1 decreases. Also in such a case, the short-term average amplitude value A immediately reflects the change in the amplitude data Sb1 and shows a decreasing tendency, but the long-term average amplitude value B and the lower limit value D calculated based on the long-term average amplitude value B are Immediately, the change in the amplitude data Sb1 is not reflected, and a decreasing tendency is shown after a certain time delay. Therefore, by appropriately setting a value for calculating the lower limit value D, a state in which the short-term average amplitude value A is lower than the lower limit value D occurs as shown in a region c in FIG. The machining abnormality detection device 50 recognizes that an abnormality has occurred when such a state continues for a predetermined time Tc, and outputs an abnormality detection signal Sa to the alarm output means 16, the control unit 11, and the like.

  The abnormality detection signals Sa output when the above three abnormalities are detected may be the same, or may be different depending on the abnormality to be detected. If the configuration is such that a different abnormality detection signal Sa is output depending on the abnormality to be detected, the alarm output means 16, the control unit 11 and the like that receive it determine the abnormality detection signal Sa. When the abnormality detection signal Sa continues for more than Ta, the cutting machine 10 is stopped while decelerating with an alarm sound, and when the short-term average amplitude value A exceeds the second upper limit value C ′, the abnormality detection signal Sa cuts with an alarm sound. Appropriate measures can be taken according to the abnormality that has occurred, such as the processing machine 10 being immediately stopped and the abnormality detection signal Sa when the region c continues for a time Tc or longer, only the alarm sound. Further, since the amplitude data Sb1 is an average value, even if a sudden and abnormal fluctuation occurs in the vibration data Sb due to noise or the like, this is not erroneously detected as a machining abnormality.

  Even if the second upper limit value C ′ is not provided in the above threshold value, the second lower limit value is smaller than the lower limit value D that outputs the abnormality detection signal Sa instantaneously when the short-term average amplitude value A falls below. May be added. In addition, the abnormality detection judgment regarding the upper limit value C and the lower limit value D is output immediately when the short-term average amplitude value A exceeds the upper limit value C and the lower limit value D, similarly to the second upper limit value C ′. You may do it. Furthermore, if necessary, the third and fourth upper and lower limit values may be added, or one of the upper limit value and the lower limit value may be eliminated.

  In FIG. 3, three abnormality determinations occur continuously. Actually, however, when the abnormality detection signal Sa is output, appropriate measures are taken automatically or artificially and the abnormal state is left unattended. Cutting is not continued.

Moreover, idle amplitude upper limit threshold value of the abnormality detection method in cutting of the present invention the upper limit value C in Figure 3, the second upper limit value C ', in addition to the lower limit value D, which is calculated from the idling amplitude value E shown in FIG. 4 A value F is provided. The idling amplitude upper limit value F is used for setting a region in which a threshold abnormality determination based on the long-term average amplitude value B is not performed .

  The idling amplitude value E shown in FIG. 4 is obtained by obtaining the idling vibration data of the cutting tool 14 when the workpiece 12 and the cutting tool 14 are not in contact, that is, when idling is not performed, and calculating the absolute value thereof. Then, this is a value obtained by averaging this within a certain time. Further, the idling amplitude upper limit value F is calculated by adding or multiplying the idling amplitude value E by a predetermined value so as to be larger than the idling amplitude value E. It is preferable that the value is twice the value. These idling amplitude upper limit value F and idling amplitude value E are automatically acquired before the cutting of the cutting machine 10 is started, or are automatically or manually set at the time of initial setting of the machining abnormality detecting device 50, and the short-term average amplitude is set. Unlike the value A, the long-term average amplitude value B, etc., it is not updated while the cutting machine 10 is in operation.

Here, let us consider a case where the workpiece 12 and the cutting tool 14 are brought into a non-contact state due to the cutting of the workpiece 12 by the cutting machine 10 being completed. When the workpiece 12 and the cutting tool 14 are not in contact with each other, the amplitude data Sb1 decreases rapidly. As the amplitude data Sb1 rapidly decreases, the short-term average amplitude value A also decreases rapidly as shown after the point e in FIG. 4 and approaches the idling amplitude value E beyond the idling amplitude upper limit F. Stable in state. At this time, when the short-term average amplitude value A falls below the idling amplitude upper limit value F, the threshold abnormality determination based on the long-term average amplitude value B is not performed. Accordingly, the time Tc used for recognizing the abnormality of the lower limit value D is changed from the time when the workpiece 12 and the cutting tool 14 are in the non-contact state, so that the short-term average amplitude value A becomes equal to the idling amplitude upper limit value F in FIG. By setting the time longer than the time up to the point d, the abnormality detection based on the lower limit D when the workpiece 12 and the cutting tool 14 are not in contact with each other can be invalidated. Therefore, according to the onset bright, it is possible to prevent detecting a non-contact state between the workpiece 12 and the cutting tool 14 as an abnormal machining.

  Further, in actual work, after the cutting by the cutting machine 10 is completed, replacement work such as replacement of members such as the workpiece 12 and the cutting tool 14 and change of a cutting program to the cutting machine 10 is performed automatically or manually. Then, cutting by the cutting machine 10 is started again. In particular, when the above replacement work is performed automatically, the time required for the replacement work can be estimated to some extent. For this reason, when the time during which the short-term average amplitude value A is less than the idling amplitude upper limit value F exceeds the predetermined time Te, at a time when the cutting process is resumed (point f in FIG. 4), If the configuration is such that the short-term average amplitude value A, the long-term average amplitude value B, and the respective threshold values at the time of the previous cutting are reset, the cutting can be performed even when the machining abnormality detection device 50 continuously performs the cutting. It is possible to reduce the erroneous detection of processing abnormality at the start of processing and to perform cutting smoothly. Note that the reset instruction of the machining abnormality detection device 50 can be manually performed by pressing a reset switch provided in the machining abnormality detection device 50 instead of being automatic as described above.

  Further, if the short-term average amplitude value A is below the idling amplitude upper limit F even though the predetermined time Tf has elapsed after the reset operation, the machining abnormality detection device 50 recognizes that the replacement work is not properly performed. The abnormality detection signal Sa can also be output to the alarm output means 16, the control unit 11, and the like. The counting of the time Tf may be started at the time of resetting, or may be started simultaneously with the counting of the time Te as shown in FIG.

  Note that the reset instruction by the time Te and the abnormality detection by the time Tf are not essential. Further, when the workpiece 12 and the cutting tool 14 are repeatedly in contact state and non-contact state temporarily according to the cutting program, an extra reset operation is omitted by setting the time Te longer, and more efficiently. Cutting can be performed.

As described above, the abnormality detecting how the cutting according to the present invention, a threshold used for determining the short-term average amplitude value A and the processing abnormality is optionally calculated based on the amplitude data Sb1 obtained during cutting Therefore, even if variations occur in the vibration data Sb and the amplitude data Sb1, the detection of the machining abnormality is not affected. Similarly, there is no time lag between the short-term average amplitude value A and the threshold value.

  Furthermore, if the initial settings such as the time interval for taking the moving average of the short-term average amplitude value A and the long-term average amplitude value B, the value for calculating the threshold value, etc. have been optimized, it has been conventionally performed for each cutting process. It is not necessary to acquire data before cutting necessary for threshold setting, and cutting can be performed extremely efficiently.

From the above, according to the abnormality detection how the cutting according to the present invention, it is possible to detect an abnormality in cutting the less erroneous detection precision.

  The initial setting of the machining abnormality detection device 50 can be manually input, or a value recorded in advance in a memory can be selected and input. Further, it may be automatically set by linking with a cutting program input to the cutting machine 10. In addition, the present invention can be modified and implemented without departing from the gist of the present invention.

It is the schematic which shows the structure of the processing abnormality detection apparatus to which the abnormality detection method which concerns on this invention is applied . It is a figure explaining the calculation method of the short-term average amplitude value and long-term average amplitude value which concern on this invention. It is a figure explaining the abnormality detection method which concerns on this invention. It is a diagram for explaining an abnormality detecting how according to the present invention.

14 Cutting tools
18 Vibration data acquisition means
50 Processing abnormality detection device
Sb vibration data
Sb1 amplitude data
Sa abnormality detection signal
A Short-term average amplitude value
B Long-term average amplitude value
C Upper limit
C 'second upper limit
D Lower limit
E Idling amplitude value
F idling amplitude upper limit

Claims (3)

In the abnormality detection method for detecting abnormalities during cutting from vibration data generated during cutting,
Acquiring idling vibration data generated before cutting;
Calculating an idling amplitude value based on an absolute value of the idling vibration data;
Calculating an idle amplitude upper limit based on the idle amplitude value;
Obtaining vibration data generated during cutting;
Calculating amplitude data based on the absolute value of the vibration data;
Calculating a short-term average amplitude value by performing a moving average process on the amplitude data at predetermined time intervals;
Calculating a long-term average amplitude value by performing a moving average process on the amplitude data at a time interval longer than the time interval;
Calculating a threshold based on the long-term average amplitude value;
When an abnormality is detected by comparing the short-term average amplitude value and the threshold value and the short-term average amplitude value exceeds the idling amplitude upper limit value , an anomaly detection signal is output, and the short-term average amplitude value is the idling Invalidating the abnormality detection by the threshold when the amplitude is below the upper limit value ;
A method for detecting an abnormality in cutting, characterized by comprising: 2. The cutting process according to claim 1, wherein an abnormality detection signal is output when the short-term average amplitude value continuously exceeds a threshold value for a predetermined time and the short-term average amplitude value exceeds the idling amplitude upper limit value. Anomaly detection method. When the short-term average amplitude value continues below the upper limit value of the idling amplitude for a predetermined time,
While resetting the long-term average amplitude value and the short-term average amplitude value,
If the short-term average amplitude value is below the upper limit of the idling amplitude even after a predetermined time has elapsed after resetting,
The abnormality detection method in cutting processing according to claim 1 or 2, wherein an abnormality detection signal is output.

Images