JP2021074842A - Processing quality prediction system - Google Patents

Abstract

To provide a processing quality prediction system that can predict a depth of an affected layer generated in a work-piece, with higher accuracy.SOLUTION: A processing quality prediction system 1 comprises: a grinder main body 11 that performs grinding processing to a work-piece W with a grinding wheel T; a detector 13d that detects observable state data in the grinder main body 11 during grinding processing of the work-piece W; and a prediction part 73 that predicts a depth of an affected layer generated in the work-piece W on the basis of the state data detected by the detector 13d. The detector 13d includes a heat flow sensor that detects as the state data heat radiation data obtained from a position different from a grinding point Pa on a surface of at least either of the grinding wheel T and the work-piece W.SELECTED DRAWING: Figure 4

Description

The present invention relates to a processing quality prediction system.

In the grinding process of a workpiece by a grindstone, it is required that the grinding quality of the workpiece satisfies a predetermined condition. For example, it is required that the geographic feature does not have a work-altered layer, that the surface texture (for example, surface roughness) of the geographic feature is within a predetermined value, and that the geographic feature does not have a chatter pattern. Be done.

The operator inspects the work piece after grinding to confirm whether or not the grinding quality satisfies the predetermined condition, and if the work piece meets the predetermined condition, the work piece is judged to be a good product. Here, Patent Document 1 describes that it is determined whether or not a work-altered layer is formed on a workpiece based on a grinding load measured during grinding.

Further, in the grinding process of a workpiece by a grindstone, in order to maintain the sharpness of the grindstone, the surface of the grindstone is trued (used in the sense of including dressing). If the sharpness of the grindstone is reduced, the quality of the workpiece may be reduced. Therefore, the truing is performed every time the number of ground workpieces reaches a predetermined number, and the predetermined number is determined so that the quality of the workpieces does not deteriorate. However, since it is determined by the operator, there is a risk that grinding will be continued even though the sharpness has deteriorated, and there is a risk that the quality of the workpiece will deteriorate.

Therefore, in Patent Document 2, the vibration of the spindle head is detected by the vibration detector attached to the spindle head, and the amplitude of the spindle reaches a preset value according to the grinding accuracy of the grinding surface of the workpiece. Later, it is described that the grinding operation is stopped and the grindstone is dressed.

By the way, in recent years, artificial intelligence has been rapidly developed with the improvement of the processing speed of a computer. For example, Patent Document 3 describes that laser processing condition data is generated by machine learning.

Japanese Unexamined Patent Publication No. 2013-129028 JP-A-2002-307304 JP-A-2017-164801

Here, even if a work-altered layer is generated in the work, if the depth of the work-altered layer can be grasped, the grinding processing conditions for preventing the work-altered layer from being generated in the next work are determined. can do.

The reason why the work-altered layer is formed on the workpiece is that the grinding point temperature due to the grinding process is high. Here, it is known that the higher the grinding load, the higher the grinding point temperature. However, the grinding point temperature becomes high not only when the grinding load is high but also when the cooling capacity of the coolant is low, for example. Therefore, it is not possible to grasp the depth of the work-altered layer with high accuracy only by the grinding load.

As a means for measuring the grinding point temperature, it is conceivable to use a non-contact temperature measuring device such as a thermography for measuring the grinding point temperature. However, grinding is generally performed while supplying coolant, and the temperature measuring device can only measure the temperature of the coolant circulating near the grinding point, and it is easy to measure the grinding point temperature. is not it. As a result, it is not possible to grasp the depth of the work-altered layer by using the temperature measuring device.

An object of the present invention is to provide a machining quality prediction system capable of predicting the depth of a machining alteration layer generated in a workpiece with higher accuracy.

The processing quality prediction system includes a grinding machine body that grinds a workpiece with a grindstone, a detector that detects observable state data in the grinding machine body during grinding of the workpiece, and the detector. It is provided with a prediction unit that predicts the depth of the work-altered layer generated in the workpiece based on the detected state data. The detector includes a heat flow sensor that detects heat dissipation data from a position different from the grinding point on at least one surface of the grindstone and the workpiece as the state data.

Here, the heat generated at the grinding point by the grinding process is transferred to the workpiece, the grindstone, the chips, and the coolant. Then, the heat flow sensor detects heat dissipation data from a position different from the grinding point on at least one surface of the grindstone and the workpiece. That is, the heat flow sensor does not directly detect the temperature at the grinding point. However, since the heat generated at the grinding point is transferred to the work piece and the grindstone, the heat remains on the surface of the work piece and the grindstone even at a position different from the grinding point. Therefore, the heat flow sensor can detect heat dissipation data corresponding to the heat generated at the grinding point even at a position different from the grinding point on at least one surface of the grindstone and the workpiece. Then, the heat dissipation data and the depth of the processed alteration layer have a predetermined relationship. Therefore, by using the heat dissipation data detected by the heat flow sensor, the depth of the work-affected layer can be predicted with high accuracy.

It is a figure which shows the structure of the processing quality prediction system. It is a top view which shows the grinding machine. It is a figure which shows the arrangement of a grindstone, a work, and a heat flow sensor, and is an enlarged view seen from the axial direction of a grindstone and a work. It is a figure which shows the 1st example of the functional block composition of the processing quality prediction system. It is a graph which shows the heat dissipation data by a heat flow sensor, and the vibration data generated in the support shaft of a grindstone. It is a graph which shows the relationship between the heat radiation data (heat flux) by a heat flow sensor, and the depth of a work alteration layer. It is a figure which shows the 3rd example of the functional block composition of the processing quality prediction system.

(1. Configuration of processing quality prediction system 1)
The hardware configuration of the processing quality prediction system 1 will be described with reference to FIG. The processing quality prediction system 1 includes at least one grinding machine 10 and one arithmetic unit 20. The grinding machine 10 may be targeted at one unit or a plurality of grinding machines 10. In this example, the processing quality prediction system 1 includes a case where one grinding machine 10 is provided. In this example, the processing quality prediction system 1 further includes a display device 30.

The grinding machine 10 includes at least a grinding machine main body 11 that grinds a workpiece W using a grindstone T, a control device 12, a detector 13, and an interface 14. The grinding machine main body 11 has a grindstone wheel T, supports the workpiece W, and has a configuration in which the grindstone wheel T and the workpiece W are relatively moved. The grinding machine main body 11 grinds the workpiece W with the grindstone T.

The control device 12 controls the grinding machine main body 11. The control device 12 includes a CNC device, a PLC device, and the like. The detector 13 detects observable state data in the grinding machine main body 11 during the grinding process of the workpiece W. The detector 13 includes at least a heat flow sensor. The interface 14 is a device that enables communication with the outside with the grinding machine main body 11, the control device 12, and the detector 13.

The arithmetic unit 20 predicts the processing quality of the workpiece W by applying machine learning using the state data detected by the detector 13. In particular, the arithmetic unit 20 predicts the state of the work-altered layer as the work quality of the work W. Further, the arithmetic unit 20 can also predict the chattering state, roundness and surface roughness as the processing quality of the workpiece W.

The arithmetic unit 20 is arranged at a position close to the grinding machine 10 and functions as a so-called edge computer. The arithmetic unit 20 includes a processor 21, a storage device 22, an interface 23, and the like. The arithmetic unit 20 is communicably connected to the grinding machine 10.

(2. Example of grinding machine 10)
As an example of the grinding machine 10, a cylindrical grinding machine will be described with reference to FIG. The grinding machine 10 is a machine tool for grinding the workpiece W. As the grinding machine 10, a grinding machine having various configurations such as a cylindrical grinding machine and a cam grinding machine can be applied. In this example, the grinding machine 10 is a grindstone traverse type cylindrical grinding machine as an example. However, a table traverse type can also be applied to the grinding machine 10.

The grinding machine 10 includes a grinding machine main body 11, a control device 12, a detector 13, and an interface 14 (shown in FIG. 1). The grinding machine main body 11 mainly includes a bed 41, a spindle base 42, a tailstock 43, a traverse base 44, a grindstone base 45, a grindstone T, a sizing device 46, a grindstone correction device 47, and a coolant nozzle 48.

The bed 41 is fixed on the installation surface. The headstock 42 is provided on the upper surface of the bed 41 on the front side in the X-axis direction (lower side in FIG. 2) and one end side in the Z-axis direction (left side in FIG. 2). The headstock 42 rotatably supports the workpiece W around the Z axis. The workpiece W is rotated by driving a motor 42a provided on the headstock 42. The tailstock 43 is located on the upper surface of the bed 41 at a position facing the headstock 42 in the Z-axis direction, that is, the front side in the X-axis direction (lower side in FIG. 2) and the other end side in the Z-axis direction (the lower side in FIG. 2). It is provided on the right side of FIG. 2). That is, the spindle base 42 and the tailstock 43 rotatably support both ends of the workpiece W.

The traverse base 44 is provided on the upper surface of the bed 41 so as to be movable in the Z-axis direction. The traverse base 44 is moved by driving a motor 44a provided on the bed 41. The grindstone base 45 is provided on the upper surface of the traverse base 44 so as to be movable in the X-axis direction. The grindstone base 45 is moved by driving a motor 45a provided on the traverse base 44. The grindstone wheel T is rotatably supported by the grindstone base 45. The grindstone wheel T is rotated by driving a motor 45b provided on the grindstone base 45. The grindstone T is configured by fixing a plurality of abrasive grains with a bond material.

The sizing device 46 measures the dimension (diameter) of the workpiece W. The grindstone correction device 47 corrects the shape of the grindstone T. The grindstone correction device 47 is a device for performing trueing (including dressing) of the grindstone T. Further, the grindstone correction device 47 also has a function of measuring the dimensions (diameter) of the grindstone T.

Here, the trueing is a reshaping work, such as a work of forming the grindstone T according to the shape of the workpiece W when the grindstone T is worn by grinding, a work of removing the runout of the grindstone T due to one-sided wear, and the like. Is. Dressing is a work of sharpening (sharpening), adjusting the amount of protrusion of abrasive grains, and creating a cutting edge of abrasive grains. Dressing is the work of correcting fouling, clogging, spillage, etc., and is usually performed after truing. However, since trueing and dressing may be carried out without any particular distinction, they are referred to as trueing in the present specification and are used in the sense of including dressing.

The coolant nozzle 48 is provided in the vicinity of the grindstone wheel T in the grindstone base 45. The coolant nozzle 48 supplies the coolant to the grinding point of the workpiece W by the grindstone T. The coolant supplied from the coolant nozzle 48 to the grinding point is recovered, cooled to a predetermined temperature, and supplied again from the coolant nozzle 48 to the grinding point.

The control device 12 controls each drive device based on an NC program generated based on operation command data such as the shape of the workpiece W, the machining conditions, the shape of the grindstone T, and the coolant supply timing information. That is, the control device 12 inputs the operation command data, generates an NC program based on the operation command data, and uses the motors 42a, 44a, 45a, 45b, the coolant device (not shown), and the like based on the NC program. Grinding of the workpiece W is performed by controlling. In particular, the control device 12 grinds the workpiece W until it has a finished shape based on the outer diameter of the workpiece W measured by the sizing device 46. Further, the control device 12 corrects the grindstone T (truing and dressing) by controlling the motors 44a, 45a, 45b, the grindstone repair device 47, and the like at the timing of correcting the grindstone T. ..

Further, the grinding machine 10 includes a plurality of detectors 13a-13g that detect observable state data in the grinding machine main body 11 during the grinding process of the workpiece W. However, the grinding machine 10 may include at least the detector 13d, which is a heat flow sensor, among the plurality of detectors 13a-13g.

The detector 13a is a rotational power sensor that detects the rotational power data of the motor 42a that rotationally drives the workpiece W as state data. The detector 13b is an axial power sensor that detects axial power data for moving the grindstone base 45 in the X-axis direction as state data. The detector 13c is a rotational power sensor that detects the rotational power data of the motor 45b that rotationally drives the grindstone T as state data.

The detector 13d is a heat flow sensor that detects heat dissipation data from a position different from the grinding point on at least one surface of the grindstone T and the workpiece W as state data. In this example, the detector 13d as the heat flow sensor is arranged so as to face the surface of the grindstone T, and the case of detecting the heat radiation data from the surface of the grindstone T will be taken as an example. Here, the heat flow sensor can detect the size of the heat flux and the direction of the heat flow. The heat dissipation data is a value corresponding to the size of the heat flux. When the surface of the grindstone T is hot, heat is dissipated from the surface of the grindstone T, so that the detector 13d as a heat flow sensor has the size of the heat flux radiated from the grindstone T (heat dissipation). Data) can be detected. The direction of the heat flow here is only the direction from the surface of the grindstone T toward the detector 13d.

The detector 13e is a vibration sensor that detects vibration data generated on a support shaft (support device) that supports the grindstone T as state data. The detector 13f is a vibration sensor that detects vibration data generated on the spindle (support device) of the headstock 42 that supports the workpiece W or the tailstock shaft (support device) of the tailstock 43 as state data. As the vibration sensor as the detectors 13e and 13f, an acceleration sensor, an AE (acoustic emission) sensor or the like can be used. The detector 13g is a processing sound sensor that detects sound data generated by grinding as state data. The processing sound sensor as the detector 13g can be arranged at any position within the processing area.

(3. Position of heat flow sensor 13d)
The position of the heat flow sensor 13d will be described with reference to FIG. In this example, the heat flow sensor 13d is arranged so as to face the surface of the grindstone wheel T, and detects heat dissipation data from the surface of the grindstone wheel T.

In FIG. 3, the grindstone wheel T rotates clockwise, and the workpiece W rotates counterclockwise, for example. A grinding point Pa in which the side surface of the grindstone T and the side surface of the workpiece W come into contact with each other is formed. In the grindstone T, Pa is the grinding point, Pb is the lower end, and Pc is the upper end. That is, in this example, in the rotation direction of the grindstone T, the predetermined position on the outer peripheral surface of the grindstone T moves in the order of the grinding point Pa (side position on the workpiece W side), the lower end Pb, the upper end Pc, and the grinding point Pa. It becomes the direction to do. Further, in the workpiece W, Pa is the grinding point, Pd is the lower end, and Pe is the upper end. That is, in this example, in the rotation direction of the workpiece W, the predetermined position on the outer peripheral surface of the workpiece W moves in the order of the grinding point Pa (side position on the grindstone T side), the lower end Pd, the upper end Pe, and the grinding point Pa. It becomes the direction to do.

Further, the coolant nozzle 48 supplies the coolant to the grinding point Pa of the workpiece W by the grindstone T. In particular, as shown in FIG. 3, the grindstone T grinds the workpiece W while the coolant nozzle 48 supplies coolant from a position above the grinding point Pa toward the grinding point Pa.

The heat flow sensor 13d is arranged on the outer peripheral surface of the grindstone T so as to face a position different from the grinding point Pa. Specifically, the heat flow sensor 13d is arranged so as to face a position above the grinding point Pa on the outer peripheral surface of the grindstone wheel T. Therefore, the position of the grindstone T on which the heat flow sensor 13d faces is a position where it is difficult for the coolant to be directly applied. As a result, since the heat flow sensor 13d is less affected by the coolant, it can detect heat dissipation data on the surface of the grindstone T.

Further, by arranging as described above, the heat flow sensor 13d is located at a position deviated by 180 ° or more from the grinding point Pa in the rotation direction of the grindstone T. Therefore, after the coolant is applied to the vicinity of the grinding point Pa of the grindstone T, the coolant is separated from the surface of the grindstone T by the centrifugal force accompanying the rotation of the grindstone T. Therefore, the amount of coolant adhering to the surface of the grindstone T with which the heat flow sensor 13d faces is reduced. As a result, the heat flow sensor 13d is less affected by the coolant, so that the heat dissipation data on the surface of the grindstone T can be detected with higher accuracy.

In particular, the heat flow sensor 13d is arranged so as to face the position between the grinding point Pa and the upper end Pc of the grindstone T on the outer peripheral surface of the grindstone T. When the grindstone T is in the rotation direction described above, by arranging the heat flow sensor 13d at the position, the coolant can be more effectively dispersed by the centrifugal force accompanying the rotation of the grindstone T. Therefore, the heat flow sensor 13d is less affected by the coolant, and can detect the heat dissipation data on the surface of the grindstone T with higher accuracy.

Here, in the above, the heat flow sensor 13d is arranged at a position facing the surface of the grindstone T. In addition to this, the heat flow sensor 13d may be arranged at a position facing the surface of the workpiece W to detect heat dissipation data from the surface of the workpiece W. However, if the outer diameter of the work W is small, there is a possibility that coolant has adhered to the entire surface of the work W, so the heat dissipation data from the work W should be detected with high accuracy. May not be possible. On the other hand, since the grindstone T generally has a large diameter, if the outer diameter of the workpiece W is small, the heat flow sensor 13d should be arranged at a position facing the surface of the grindstone T as described above. Good.

When the heat flow sensor 13d is arranged at a position facing the surface of the workpiece W, it is preferable to arrange the heat flow sensor 13d at a position substantially the same as the position with respect to the grindstone T described above. The heat flow sensor 13d may be arranged so as to face a position above the grinding point Pa on the outer peripheral surface of the workpiece W. In particular, the heat flow sensor 13d may be arranged so as to face the position between the grinding point Pa and the upper end Pe of the workpiece W on the outer peripheral surface of the workpiece W.

(4. First example of functional block configuration of processing quality prediction system 1)
The functional block configuration of the processing quality prediction system 1 will be described with reference to FIGS. 4 to 6. The processing quality prediction system 1 includes a plurality of detectors 13a-13g, an arithmetic unit 20, and a display device 30. As described above, the detector 13d is a heat flow sensor, and detects heat dissipation data as observable state data in the grinding machine main body 11 during grinding of the workpiece W.

Further, the processing quality prediction system 1 further includes an example in which the processing quality inspection device 50 is further provided. The processing quality inspection device 50 is an device for inspecting the target processing quality, and may be a device different from the grinding machine 10 or a built-in device of the grinding machine 10. The processing quality inspection device 50 is a device that detects the depth of the processing alteration layer. For example, the device 50 can be applied with a detection device using an eddy current or the like.

The arithmetic unit 20 includes a preparatory processing unit 60 and a prediction arithmetic unit 70. The preparatory processing apparatus 60 performs a process of generating relational data showing the relation between the heat radiation data as the state data and the depth data of the work-altered layer. The prediction arithmetic unit 70 performs arithmetic processing for predicting the depth of the processing alteration layer as the processing quality using the relational data. The preparation processing device 60 and the prediction calculation device 70 may form independent devices or one device.

The preparatory processing device 60 is based on the heat dissipation data as the state data detected by the detector 13d and the depth data of the machining alteration layer as the machining quality obtained by the machining quality inspection device 50. Generate relational data for predicting the depth of the work-altered layer as the work quality. The preparation processing device 60 includes a preparation data set acquisition unit 61, a preparation data set storage unit 62, and a relational data generation unit 63.

The preparation data set acquisition unit 61 acquires the preparation data set for generating the relational data. Here, the preparation data set acquisition unit 61 includes a state data acquisition unit 61a, a feature amount calculation unit 61b, and a processing alteration layer depth data acquisition unit 61c. The state data acquisition unit 61a acquires heat dissipation data as state data detected by the heat flow sensor as the detector 13d during the grinding process of the workpiece W.

Here, the behavior of the heat dissipation data during the grinding process is as shown in FIG. In FIG. 6, vibration data generated on the support shaft supporting the grindstone T is also shown. The vibration data shows a high value at the same time when the grinding process is started, and shows a low value at the same time when the grinding process is finished. As with the vibration data, the heat dissipation data also shows a high value at the same time when the grinding process is started, and shows a low value at the same time when the grinding process is finished. As described above, it can be seen that the heat dissipation data has very good followability to the grinding process. From this, it can be seen that the heat dissipation data is affected by the heat generated by the grinding process. It is considered that this good followability improves the prediction accuracy of the state of the work-altered layer.

The feature amount calculation unit 61b calculates a plurality of feature amounts for the heat dissipation data acquired by the state data acquisition unit 61a. Here, the feature amount is an index for generating the relational data, and for example, various statistics in the heat dissipation data can be used. For example, the feature amount is selected from the maximum value, the minimum value, the average value, the variance, the standard deviation, the skewness, the kurtosis, the median value, and the like in the heat dissipation data. Further, the feature amount may be selected from a statistic for the data obtained by differentiating the state data, a statistic for the data obtained by frequency analysis of the state data, and the like. The feature amount calculation unit 61b may calculate all the above-mentioned feature amounts, or may calculate a part of the feature amounts. In particular, in this example, only the maximum value in the heat dissipation data is used as the feature amount. The processing alteration layer depth data acquisition unit 61c acquires the processing alteration layer depth data acquired by the processing quality inspection device 50 in association with the target workpiece W.

The preparation data set storage unit 62 stores the preparation data set acquired by the preparation data set acquisition unit 61 described above, that is, the feature amount related to the heat dissipation data and the depth data of the work-altered layer in association with the workpiece W. ..

The relational data generation unit 63 uses the preparation data set stored in the preparation data set storage unit 62. The maximum value of the heat flux size (heat dissipation data) and the depth of the work-affected layer have the relationship shown in FIG. 7. That is, the maximum value of the size of the heat flux (heat dissipation data) and the depth of the work-altered layer have a linear relationship. Specifically, the larger the maximum value of the heat flux (heat dissipation data), the deeper the depth of the work-altered layer.

The prediction arithmetic unit 70 predicts the depth of the work-altered layer of the workpiece W based on the heat dissipation data detected during the grinding process. The prediction calculation device 70 includes a relational data storage unit 71, a prediction data acquisition unit 72, a prediction unit 73, a display output unit 74, a geographic feature post-processing unit 75, and an improvement processing unit 76.

The relationship data storage unit 71 stores the relationship data (shown in FIG. 7) generated by the relationship data generation unit 63. The prediction data acquisition unit 72 acquires prediction data during the grinding process. The prediction data acquisition unit 72 includes a state data acquisition unit 72a and a feature amount calculation unit 72b. The state data acquisition unit 72a acquires the state data (heat dissipation data) detected by the detector 13d (heat flow sensor) during the grinding process.

The feature amount calculation unit 72b calculates the feature amount related to the detected heat dissipation data, that is, the feature amount used for generating the relational data. For example, the feature amount calculation unit 72b uses the maximum value of the heat dissipation data detected by the heat flow sensor 13d as the feature amount.

The prediction unit 73 predicts the depth of the processing alteration layer based on the relational data stored in the relational data storage unit 71 and the data acquired by the prediction data acquisition unit 72. That is, the prediction unit 73 calculates the depth of the processing alteration layer corresponding to the heat dissipation data acquired by the prediction data acquisition unit 72 in the relational data. In this example, the prediction unit 73 calculates the depth of the processing alteration layer corresponding to the maximum value of the heat dissipation data in the relational data.

The display output unit 74 outputs the predicted depth of the work-altered layer to the display device 30. When the depth of the work alteration layer predicted by the prediction unit 73 is equal to or greater than a predetermined depth, the geographic feature post-processing unit 75 selects, discards, and discards the target geographic feature W with respect to the post-processing device 80. Have one of the detailed inspections performed.

The improvement processing unit 76 causes the control device 12 to change the grinding processing conditions or execute the grinding of the grindstone T according to the depth of the processing alteration layer predicted by the prediction unit 73. Changes in the grinding conditions include, for example, slowing the feed rate (speed in the cutting direction) of rough grinding and increasing the allowance for finish grinding (fine grinding, fine grinding). In particular, the improvement processing unit 76 may change the grinding conditions according to the predicted depth of the work-altered layer. At that time, the improvement processing unit 76 determines that the feed rate of rough grinding is slowed down and the allowance for finish grinding is increased in consideration of the cycle time, depending on the predicted depth of the work-altered layer. It is good to decide.

(8. Effect)
The heat generated at the grinding point by the grinding process is transferred to the workpiece, grindstone, chips, and coolant. Then, the heat flow sensor 13d detects heat dissipation data from a position different from the grinding point Pa on at least one surface of the grindstone T and the workpiece W. That is, the heat flow sensor 13d does not directly detect the temperature at the grinding point Pa. However, since the heat generated at the grinding point Pa is transferred to the workpiece W and the grindstone T, the heat remains even on the surface of the workpiece W and the grindstone T even at a position different from the grinding point Pa. ing. Therefore, the heat flow sensor 13d can detect heat dissipation data corresponding to the heat generated at the grinding point Pa even at a position different from the grinding point Pa on at least one surface of the grindstone T and the workpiece W. .. Then, the heat dissipation data and the depth of the processed alteration layer have a predetermined relationship. Therefore, by using the heat dissipation data detected by the heat flow sensor 13d, the depth of the work-altered layer can be predicted with high accuracy.

(9. Second example of functional block configuration of processing quality prediction system 1)
In the above first example, the feature amount calculation units 61b and 72b calculate the maximum value of the heat dissipation data as the feature amount, and the relational data generation unit 63 determines the relationship between the maximum value of the heat dissipation data and the depth of the processed alteration layer. The relational data to be represented was generated. Then, the prediction unit 73 predicted the depth of the work-altered layer based on the relational data and the maximum value of the heat dissipation data.

In addition to this, the relational data may represent the relationship between a plurality of features related to heat dissipation data and the depth of the processed alteration layer. For the plurality of features in this case, for example, the maximum value of heat dissipation data, the maximum value of the first derivative, and the like may be applied. That is, the relational data is three-dimensional data representing the relation between the maximum value of the heat dissipation data, the maximum value of the first derivative of the heat dissipation data, and the depth of the processed alteration layer. Of course, the relational data may be four-dimensional or higher data representing the relationship between the feature quantity of 3 or more and the depth of the processed alteration layer.

In this case, the feature amount calculation units 61b and 72b calculate a plurality of feature amounts to be applied. For example, the feature amount calculation units 61b and 72b calculate the maximum value of the heat dissipation data and the maximum value of the first derivative of the heat dissipation data. Then, the prediction unit 73 has a plurality of feature amounts based on the plurality of feature amounts calculated by the feature amount calculation unit 72b of the prediction data acquisition unit 72 and the relational data stored in the relational data storage unit 71. Calculate the depth of the work-altered layer corresponding to the value of.

Therefore, by using a plurality of feature quantities, the depth of the work-altered layer can be calculated with high accuracy. For example, it is considered that the larger the maximum value of the heat dissipation data, the deeper the depth of the processed alteration layer, and the larger the maximum value of the first derivative of the heat dissipation data, the deeper the depth of the processed alteration layer. In this way, by using the maximum value of the heat dissipation data and the maximum value of the first derivative of the heat dissipation data, the depth of the work-altered layer can be calculated with high accuracy.

(10. Third example of functional block configuration of processing quality prediction system 1)
In the first and second examples above, the case of generating the relational data showing the relation between the depth of the processed alteration layer and the feature amount related to the heat dissipation data is given as an example. Relationship data is not easy to generate when the number of features is large. Therefore, as a third example, machine learning using a large number of features is applied. Note that the same reference numerals are given to the configurations in which substantially the same processing as that of the first example is performed, and detailed description thereof will be omitted.

The processing quality prediction system 1 includes a plurality of detectors 13a-13g, an arithmetic unit 20, and a display device 30. As described above, the detectors 13a-13g detect observable state data in the grinding machine main body 11 during the grinding process of the workpiece W.

The arithmetic unit 20 includes a learning processing device 160 and a prediction arithmetic unit 170. The learning processing device 160 performs a process of generating a trained model by performing machine learning. That is, the learning processing device 160 functions as a learning phase in machine learning. The prediction arithmetic unit 170 uses the trained model to perform arithmetic processing for predicting the depth of the processing alteration layer as the processing quality. That is, the prediction arithmetic unit 170 functions as an inference phase in machine learning. The learning processing device 160 and the prediction calculation device 170 may form independent devices or may form one device.

The learning processing device 160 is based on the state data detected by the detector 13a-13g and the depth data of the machining alteration layer as the machining quality obtained by the machining quality inspection device 50, and the machining quality of the workpiece W. Generate a trained model for predicting the depth of the geographic feature layer. The learning processing device 160 includes a training data set acquisition unit 161, a training data set storage unit 162, and a model generation unit 163.

The training data set acquisition unit 161 acquires a training data set for performing machine learning. Here, the training data set acquisition unit 161 includes a state data acquisition unit 161a, a feature amount calculation unit 161b, and a processing alteration layer depth data acquisition unit 61c. The state data acquisition unit 161a acquires the state data detected by each of the detectors 13a to 13g during the grinding process of the workpiece W. That is, the state data acquisition unit 161a acquires the data detected by the other detectors 13a-13c and 13e-13g in addition to the heat dissipation data detected by the heat flow sensor 13d.

The feature amount calculation unit 161b calculates a plurality of feature amounts for the state data acquired by the state data acquisition unit 161a. Here, as the feature amount, various statistics in the state data are used. For example, the feature amount is the maximum value, the minimum value, the average value, the variance, the standard deviation, the skewness, the kurtosis, the median value, etc. in the state data. Further, the feature amount may include a statistic for the data obtained by differentiating the state data, a statistic for the data obtained by frequency analysis of the state data, and the like. The feature amount calculation unit 161b may calculate all the above-mentioned feature amounts, or may calculate a part of the feature amounts.

For example, when eight kinds of statistics are applied to three kinds of original data, differential data, and frequency analysis data as the types of feature quantities, there are 24 kinds of feature quantities. When calculating 24 types of features for each of the 7 types of detectors 13a-13g, there are a total of 168 types of features.

That is, the training data set acquisition unit 161 acquires a training data set including the explanatory variables and the objective variables by using a plurality of feature quantities for each of the plurality of state data as explanatory variables and the processing quality data as the objective variable. The training data set storage unit 162 stores the training data set acquired by the training data set acquisition unit 161.

The model generation unit 163 performs machine learning using the training data set stored in the training data set storage unit 162. Specifically, the model generation unit 163 uses a plurality of feature quantities (for example, 168 types of feature quantities) for each of the plurality of state data as explanatory variables, and machine learning using the depth data of the machining alteration layer as the objective variable. I do. That is, the model generation unit 163 performs machine learning using the explanatory variables and the objective variables as training data sets. Then, the model generation unit 163 generates a trained model for predicting the depth of the processed alteration layer.

Regression may be used as the machine learning method. For example, decision trees, linear regression, ridge regression, lasso regression, elastic nets, random forests and the like are useful. In particular, by using these methods, it is possible to grasp the degree of influence on a plurality of feature quantities. The machine learning method can be applied to other than regression.

The prediction arithmetic unit 170 predicts the machining quality of the workpiece W based on the state data detected during the grinding process. The prediction calculation device 170 includes a model storage unit 171, a prediction data acquisition unit 172, a prediction unit 173, a display output unit 74, a geographic feature post-processing unit 75, and an improvement processing unit 76.

The model storage unit 171 stores the trained model generated by the model generation unit 163. The prediction data acquisition unit 172 acquires prediction data during the grinding process. The prediction data acquisition unit 172 includes a state data acquisition unit 172a and a feature amount calculation unit 172b. The state data acquisition unit 172a acquires the state data detected by the plurality of detectors 13a-13g during the grinding process. The feature amount calculation unit 172b calculates the feature amount related to the detected plurality of state data. The feature amount calculation unit 172b performs the same processing as the feature amount calculation unit 161b in the learning processing device 160. The prediction unit 173 predicts the depth of the processing alteration layer based on the trained model stored in the model storage unit 171 and the data acquired by the prediction data acquisition unit 172. The depth of the processing alteration layer predicted by the prediction unit 173 is used for processing by the display output unit 74, the aftertreatment device 80, and the improvement processing unit 76.

1: Machining quality prediction system, 10: Grinding machine, 11: Grinding machine body, 12: Control device, 13, 13a-13g: Detector, 20: Computational device, 30: Display device, 45: Grinding table, 48: Coolant Nozzle, 50: Processing quality inspection device, 60: Preparation processing device, 61: Preparation data set acquisition unit, 61a: State data acquisition unit, 61b: Feature amount calculation unit, 61c: Depth data acquisition unit of processing alteration layer, 62 : Preparation data set storage unit, 63: Relationship data generation unit, 70: Prediction calculation device, 71: Relationship data storage unit, 72: Prediction data acquisition unit, 72a: State data acquisition unit, 72b: Feature amount calculation unit, 73 : Prediction unit, 74: Display output unit, 75: Work post-processing unit, 76: Improvement processing unit, 80: Post-processing device, 160: Learning processing device, 161: Training data set acquisition unit, 161a: State data acquisition unit , 161b: Feature amount calculation unit, 162: Training data set storage unit, 163: Model generation unit, 170: Prediction calculation device, 171: Model storage unit, 172: Prediction data acquisition unit, 172a: State data acquisition unit, 172b : Feature amount calculation unit, 173: Prediction unit, Pa: Grinding point, Pb: Lower end of grind wheel, Pc: Upper end of grind wheel, Pd: Lower end of workpiece, Pe: Upper end of workpiece, T: Grind wheel, W : Workpiece

Claims (12)

The main body of the grinding machine that grinds the workpiece with a grindstone,
A detector that detects observable state data in the grinding machine body during grinding of the workpiece, and
A prediction unit that predicts the depth of the work-altered layer generated in the geographic feature based on the state data detected by the detector.
With
The detector is a processing quality prediction system including a heat flow sensor that detects heat radiation data from a position different from a grinding point on at least one surface of the grindstone and the workpiece as the state data. The processing quality prediction system according to claim 1, wherein the heat flow sensor is arranged so as to face a position different from the grinding point on the outer peripheral surface of at least one of the grindstone and the workpiece. The grinding machine main body performs the grinding process while supplying coolant from a position above the grinding point toward the grinding point.
The processing quality prediction system according to claim 2, wherein the heat flow sensor is arranged so as to face a position above the grinding point on the outer peripheral surface of at least one of the grindstone and the workpiece. The rotation direction of the grindstone is a direction in which a predetermined position on the surface of the grindstone moves in the order of the side, the lower end, the upper end, and the grinding point of the grindstone where the grinding point is the grinding point.
The processing quality prediction system according to claim 3, wherein the heat flow sensor is arranged so as to face a position above the grinding point on the outer peripheral surface of the grindstone. The processing quality prediction system according to claim 4, wherein the heat flow sensor is arranged so as to face a position between the grinding point and the upper end of the grindstone on the outer peripheral surface of the grindstone. The rotation direction of the workpiece is a direction in which a predetermined position on the surface of the workpiece moves in the order of the side, the lower end, the upper end, and the grinding point of the workpiece whose grinding point is the grinding point.
The processing quality prediction system according to claim 3, wherein the heat flow sensor is arranged so as to face a position above the grinding point on the outer peripheral surface of the workpiece. The processing quality prediction system further
A relational data storage unit for storing relational data between the heat dissipation data and the depth of the processed alteration layer is provided.
Any one of claims 1-6, wherein the prediction unit predicts the depth of the processed alteration layer by calculating the depth of the processed altered layer corresponding to the heat radiation data in the related data. Processing quality prediction system described in. The relational data storage unit stores the relational data between a plurality of feature quantities related to the heat dissipation data and the depth of the processed alteration layer.
The seventh aspect of claim 7, wherein the prediction unit predicts the depth of the processed alteration layer by calculating the depth of the processed altered layer corresponding to a plurality of feature quantities related to the heat radiation data in the related data. Processing quality prediction system. The processing quality prediction system further
A model generation unit that generates a trained model by performing machine learning using the state data as an explanatory variable and the depth of the processed alteration layer as an objective variable and using the explanatory variable and a training data set including the objective variable. When,
A trained model storage unit that stores the trained model generated by the model generation unit, and a trained model storage unit.
With
The processing quality prediction system according to any one of claims 1 to 6, wherein the prediction unit predicts the depth of the processing alteration layer using the learned model and the state data. The detector further
A rotational power sensor that detects the rotational power data of at least one of the grindstone and the workpiece as the state data, and
An axial power sensor that detects axial power data for axially moving the grindstone and the workpiece in the cutting direction as the state data, and
A processing sound sensor that detects sound data generated by the grinding process as the state data, and
A vibration sensor that detects vibration data generated in the grindstone and a support device that supports at least one of the workpieces as the state data, and
The processing quality prediction system according to claim 9, further comprising at least one of the above. The processing quality prediction system further
A work post-processing unit for selecting, discarding, or performing a detailed inspection of the target work when the depth of the work-altered layer predicted by the prediction unit is equal to or greater than a predetermined depth is provided. The processing quality prediction system according to any one of claims 1-10. The processing quality prediction system further
Any one of claims 1-11, comprising an improvement processing unit that changes the grinding processing conditions or executes the truing of the grindstone according to the depth of the processing alteration layer predicted by the prediction unit. Processing quality prediction system described in the section.

Images