Machining state estimation method and machining state estimation system

Abstract

To provide a processing state estimation method and a processing state estimation system that can accurately estimate a processing state.SOLUTION: A processing state estimation system 1 includes an observation device 20 and an estimation arithmetic device 30. The estimation arithmetic device 30 includes: a signal acquisition unit 31 for acquiring an AE signal S from the observation device 20; a frequency band effective value arithmetic unit 34 for calculating a respective effective value J of a plurality of frequency bands for the AE signal S; and a surface roughness estimation unit 36 for estimating surface roughness as a processing state, on the basis of an interrelation between the effective value J and the processing state.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to a machining state estimation method and a machining state estimation system.

Background Art

[0002] Generally, when machining a workpiece using a tool of a machine tool, there is a risk that the tool deteriorates and the machining accuracy decreases. Therefore, it is extremely important to manage the deterioration of the tool and appropriately perform tool correction to maintain the machining accuracy of the workpiece. For this reason, conventionally, Patent Document 1 and Patent Document 2 disclose a technique of acquiring an acoustic emission signal (hereinafter referred to as an "AE signal") that represents acoustic emission (AE) having frequency characteristics generated during machining of a workpiece by a tool, and estimating the machining state when the tool machines the workpiece.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0004] According to the above-described conventional technology, it is possible to acquire an AE signal acquired during machining of a workpiece by a tool and estimate the machining state by analyzing the acquired AE signal. Thereby, it is possible to set the timing for correcting the tool based on the estimated machining state.

[0005] By the way, in the above-described conventional technology, disturbances and the like generated during processing and the like are superimposed on the acquired AE signals, and the estimation accuracy of the processing state cannot be sufficiently ensured. When correcting a tool, time and cost required for the correction are necessary, so it is important to improve the estimation accuracy when estimating the processing state. Therefore, when estimating the processing state using AE signals, it is desirable to improve the estimation accuracy of the processing state by using information highly related to the processing state included in the acquired AE signals.

[0006] An object of the present invention is to provide a processing state estimation method and a processing state estimation system capable of accurately estimating a processing state using highly related information.

Means for Solving the Problems

[0007] (Processing State Estimation Method) One aspect of the present invention is Applied to a machine tool that processes a workpiece with a tool, the The tool the A processing state estimation method for estimating a processing state when a workpiece is processed, the Due to the the During the processing of the workpiece by the tool, observe the acoustic emission that occurs with frequency characteristics, The observed the Calculate the effective value of each of a plurality of frequency bands for the acoustic emission signal representing the acoustic emission, Performing multiple regression analysis using, as a parameter, the statistical calculation amount selected from among the statistical calculation amounts including the effective value for each frequency band obtained by performing statistical calculations on the acoustic emission signal, to obtain a correlation relationship between the effective value and the surface roughness of the workpiece as the processing state The said processing state the surface roughness of the workpiece as To estimate and In the multiple regression analysis, it includes first multiple regression analysis and second multiple regression analysis. In the first multiple regression analysis, perform multiple regression analysis including at least the effective value for each frequency band and the surface roughness. In the second multiple regression analysis, based on the analysis result obtained by performing multiple regression analysis by the first multiple regression analysis, perform multiple regression analysis including the processing efficiency, the effective value of the frequency band selected based on the statistical calculation amount, and the surface roughness. In the first multiple regression analysis and the second multiple regression analysis, for the surface roughness, divide it into ranges defined by a preset threshold value and perform multiple regression analysis. In the estimation of the surface roughness as the processing state, using the relationship with the surface roughness as the objective variable, which is the correlation relationship obtained by performing multiple regression analysis in the second multiple regression analysis, and at least the effective value and the surface roughness as the explanatory variables, estimate the surface roughness. There is a processing state estimation method.

[0008] (Processing State Estimation System) Another aspect of the present invention is Applied to a machine tool for machining a workpiece with a tool, the wherein the tool the is a machining state estimation system for estimating a machining state when the tool machines the workpiece, the During machining of the workpiece by the tool, the it has an observation device for observing acoustic emission generated with frequency characteristics, and the a signal acquisition unit for acquiring an acoustic emission signal representing the acoustic emission observed from the observation device, the and the an effective value calculation unit for calculating the effective value of each of a plurality of frequency bands for the acoustic emission signal acquired by the signal acquisition unit, the and the an estimation unit for estimating the machining state based on the correlation between the effective value calculated by the effective value calculation unit and the the effective value and the the machining state, the and an estimation calculation device having the same, equipped with 、 The estimation unit includes a multiple regression analysis unit that performs multiple regression analysis using, as a parameter, the statistical calculation amount selected from among the statistical calculation amounts including the effective value for each frequency band obtained by performing statistical calculations on the acoustic emission signal, to obtain the correlation relationship and estimate the surface roughness of the workpiece as the processing state. The multiple regression analysis unit includes a first multiple regression analysis unit and a second multiple regression analysis unit. The first multiple regression analysis unit performs multiple regression analysis including at least the effective value for each frequency band and the surface roughness. The second multiple regression analysis unit performs multiple regression analysis including the processing efficiency, the effective value of the frequency band selected based on the statistical calculation amount, and the surface roughness, based on the analysis result obtained by performing multiple regression analysis by the first multiple regression analysis unit. The first multiple regression analysis unit and the second multiple regression analysis unit divide the surface roughness into ranges defined by a preset threshold value and perform multiple regression analysis. In the machining state estimation system, the estimation unit estimates the surface roughness using a relationship in which the surface roughness, which is a correlation obtained by the second multiple regression analysis unit performing multiple regression analysis, is the target variable, and at least the effective value and the actual roughness value of the surface roughness are the explanatory variables.

[0009] According to these, the machining state when the tool machines the workpiece can be accurately estimated using the effective value of the frequency band of the AE signal. Further, the machining state estimation system can also accurately estimate the deterioration state of the tool of the machine tool for machining the workpiece using the estimated machining state. Thereby, in the machine tool, good machining accuracy can be maintained and the tool can be corrected at an appropriate timing.

[0010] Therefore, it is possible to prevent the tool of the machine tool from being excessively corrected or the correction from being delayed. As a result, the correction time required to correct the tool can be shortened, and the cost for correcting the tool can be reduced.

Brief Description of the Drawings

[0011]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0012] (1. Machine Tool to Which the Machining State Estimation System Is Applied) The machining state estimation system estimates the machining state when a workpiece is machined by a tool of a machine tool. Examples of the machine tool for machining the workpiece include various cutting devices for performing cutting (for example, gear cutting devices, machining centers, etc.), and various grinding devices for performing grinding (for example, cylindrical grinding machines, cam grinding machines, surface grinding machines, etc.).

[0013] In this example, as a machine tool, a grinding device that grinds a workpiece through grinding processes such as rough grinding, finish grinding, micro grinding, and spark out is exemplified. Here, examples of workpieces to be ground by the grinding device include, for example, simple shaft-shaped members, crankshafts, camshafts, flat plates, etc. Note that, as the workpiece in this example, as will be described later, the case where a simple shaft-shaped member with a circular cross-section is ground by a cylindrical grinding machine is exemplified.

[0014] (2. Outline of the Configuration of the Machining State Estimation System 1) Next, the outline of the configuration of the machining state estimation system 1 will be described with reference to FIG. 1. The machining state estimation system 1 includes at least one grinding device 10, an observation device 20, and at least one estimation calculation device 30. The grinding device 10 may be one unit, or as shown in FIG. 1, may be a plurality of units.

[0015] The machining state estimation system 1 in this example exemplifies the case of including a plurality of grinding devices 10. And the machining state estimation system 1 in this example takes as an example the case where an observation device 20 is provided for each grinding device 10 and includes a plurality of estimation calculation devices 30 provided for each of the grinding devices 10. Note that, in this example, the estimation calculation device 30 is provided one-to-one for each grinding device 10 and functions as a so-called edge computer, enabling high-speed calculation processing.

[0016] The grinding device 10 grinds the workpiece W through grinding processes such as rough grinding, finish grinding, micro grinding, and spark out. The grinding device 10 is provided with at least an observation device 20 for observing AE during the grinding of the grinding surface W1 of the workpiece W.

[0017] The observation device 20 acquires the AE signal S observed during the grinding of the grinding surface W1 of the workpiece W by the grinding device 10. For this reason, the observation device 20 in this example includes an AE sensor 21 for detecting the AE signal. And the observation device 20 in this example outputs the AE signal acquired by the AE sensor 21 to be detected to the estimation calculation device 30.

[0018] The estimation calculation device 30 time-divisionally divides the AE signal S output from the observation device 20 and divides it into a plurality of frequency bands, and calculates the effective value in each frequency band. Then, the estimation calculation device 30 calculates various statistical calculation amounts using the effective values in the plurality of frequency bands, and estimates the surface roughness of the ground surface W1 by performing multiple regression analysis on the parameters selected from among the various statistical calculation amounts.

[0019] Here, examples of the various statistical calculation amounts statistically calculated by the estimation calculation device 30, that is, the parameters used for multiple regression analysis, include, for example, peak value, average value, standard deviation, kurtosis, and skewness. Further, as other various statistical calculation amounts (parameters), the effective value (substantially equal to the standard deviation in the AC waveform such as the AE signal), the crest factor (the so-called crest factor, which is the value obtained by dividing the peak value by the effective value), and the waveform factor (the value obtained by dividing the effective value by the average value) can be exemplified. Also, the correlation obtained by performing multiple regression analysis has the surface roughness of the ground surface W1 to be estimated as the target variable, and represents the correlation with the effective value obtained from the AE signal S and the surface roughness (actual roughness value) actually measured in advance by the roughness meter 2 experimentally as the explanatory variables.

[0020] (3. Details of the Configuration of the Processing State Estimation System 1) The configuration of the processing state estimation system 1 will be described in more detail with reference to FIG. 1. The processing state estimation system 1 includes a plurality of grinding devices 10, an observation device 20 provided in each grinding device 10, and an estimation calculation device 30 provided in each grinding device 10.

[0021] The grinding device 10 mainly includes a grinding wheel 11 that grinds the ground surface W1 of the workpiece W using a grinding wheel T, and a control device 12 that controls the grinding wheel 11. In this example, as the grinding wheel 11, as shown in FIG. 2, a cylindrical grinding wheel 40 of a grinding wheel table traverse type is taken as an example.

[0022] The cylindrical grinding machine 40 is a machine tool for grinding the outer peripheral surface, i.e., the grinding surface W1, of the workpiece W which is a shaft-shaped member. As shown in FIG. 2, the cylindrical grinding machine 40 mainly includes a bed 41, a spindle headstock 42, a center rest 43, a traverse base 44, a grinding wheel base 45, a grinding wheel 46 (grinding wheel T), a sizing device 47, and a grinding wheel dressing device 48.

[0023] The bed 41 is fixed on the installation surface. The spindle headstock 42 is provided on the upper surface of the bed 41, on the front side in the X-axis direction (the lower side in FIG. 2) and one end side in the Z-axis direction (the left side in FIG. 2). The spindle headstock 42 rotatably supports the workpiece W about the Z-axis. The workpiece W is rotated by the drive of a motor 42a provided on the spindle headstock 42.

[0024] The center rest 43 is provided on the upper surface of the bed 41 at a position facing the spindle headstock 42 in the Z-axis direction, i.e., on the front side in the X-axis direction (the lower side in FIG. 2) and the other end side in the Z-axis direction (the right side in FIG. 2). Thereby, the workpiece W is rotatably supported at both ends by the spindle headstock 42 and the center rest 43.

[0025] 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 moves by the drive of a motor 44a provided on the bed 41. The grinding wheel base 45 is provided on the upper surface of the traverse base 44 so as to be movable in the X-axis direction. The grinding wheel base 45 moves by the drive of a motor 45a provided on the traverse base 44. The grinding wheel 46 (tool T) is rotatably supported by the grinding wheel base 45. The grinding wheel 46 rotates by the drive of a motor 46a provided on the grinding wheel base 45. The grinding wheel 46 is configured by fixing a plurality of abrasive grains with a bonding material.

[0026] The sizing device 47 measures the dimension (diameter) of the workpiece W. The sizing device 47 is provided on the upper surface of the bed 41 so as to be movable in the Z-axis direction. The position of the sizing device 47 in the Z-axis direction is controlled by a feed mechanism 47a provided on the bed 41.

[0027] The grinding wheel dressing device 48 corrects the shape of the grinding wheel 46. That is, the grinding wheel dressing device 48 is a device for truing and dressing the grinding wheel 46. Here, truing is a shape correction operation. When deterioration occurs in the grinding wheel 46 due to grinding, such as wear, abrasion, grain dropout, or grain crushing, etc., it is an operation to shape the grinding wheel 46 according to the shape of the workpiece W, and an operation to remove the runout of the grinding wheel 46 due to uneven wear. Dressing is an operation to true (condition) the grinding wheel 46, an operation to adjust the protrusion amount of the abrasive grains, and an operation to create the cutting edges of the abrasive grains. That is, dressing is an operation to correct clogging, blockage, spillage, etc., and is usually performed after truing.

[0028] The control device 12 includes a CNC device, a PLC device, etc., and is provided in the cylindrical grinding machine 40 (grinding machine 11) in this example. The control device 12 controls each motor 42a, 44a, 45a, 46a, etc. in the cylindrical grinding machine 40 according to the NC program generated based on the operation command data such as the shape of the workpiece W, the processing conditions, the shape of the grinding wheel 46 (tool T), and the supply timing information of the coolant. That is, when the operation command data is input, the control device 12 generates an NC program based on the operation command data.

[0029] Thereby, the control device 12 controls the operation of the cylindrical grinding machine 40 (grinding machine 11) to perform the grinding process on the grinding surface W1 of the workpiece W. Here, the control device 12 performs the grinding process until the grinding surface W1 of the workpiece W reaches the finished shape based on the dimension (diameter) of the workpiece W measured by the sizing device 47. For this reason, the NC program generated by the control device 12 is generated according to the grinding process content, that is, rough grinding, finish grinding, fine grinding, spark out, etc.

[0030] The control device 12 is provided on the bed 41 and can communicate with the estimation calculation device 30 via the interface 13 (see FIG. 1). Accordingly, the control device 12 in this example controls each of the motors 42a, 44a, 45a, 46a and the grinding wheel dressing device 48, etc. at the timing when the correction signal R output from the estimation calculation device 30 is acquired, thereby correcting (truing and / or dressing) the grinding wheel 46. Here, as will be described later, the estimation calculation device 30 in this example outputs the correction signal R to the control device 12 based on the estimated machining state of the ground surface W1, that is, the surface roughness.

[0031] As shown in FIGS. 2 and 3, the observation device 20 includes an AE sensor 21. In this example, the AE sensor 21 is provided, for example, on the center rest 43 of the cylindrical grinding machine 40 (grinding machine 11). However, the installation location of the AE sensor 21 is not limited to the center rest 43 and can also be provided on the constituent members (such as the headstock 42 and the grinding wheel table 45) of the cylindrical grinding machine 40 (grinding machine 11). The AE sensor 21 acquires an AE signal S observable in a state where the cylindrical grinding machine 40 (grinding machine 11) is grinding the ground surface W1 of the workpiece W at a predetermined sampling rate. Accordingly, the observation device 20 outputs the AE signal S acquired by the AE sensor 21 to the estimation calculation device 30.

[0032] As shown in FIG. 4, the estimation calculation device 30 includes a signal acquisition unit 31, a signal division unit 32, a signal normalization unit 33, a frequency band effective value calculation unit 34, a multiple regression analysis unit 35, a surface roughness estimation unit 36, a determination unit 37, and an output unit 38.

[0033] As shown in FIG. 5, the signal acquisition unit 31 acquires the AE signal S acquired by the AE sensor 21 at a predetermined sampling rate. Note that the AE signal shown in FIG. 5 acquired by the signal acquisition unit 31 is a raw waveform without being subjected to filter processing or the like, and is a waveform observed for a series of grinding processes, that is, rough grinding, finish grinding, fine grinding, and spark out.

[0034] The signal splitting unit 32 splits the AE signal S acquired by the signal acquisition unit 31 at predetermined time intervals to generate a split signal S1.

[0035] The signal normalization unit 33 normalizes the split signal S1 generated by the signal splitting unit 32. That is, the signal normalization unit 33 normalizes the split signal S1 with an average of "0" and a variance of "1", for example. Note that the signal normalization unit 33 can normalize the split signal S1 using, for example, the formula represented by X = (x - xa) / σ.

[0036] Here, "X" in the above formula represents the signal after normalization, "x" represents the signal before normalization, "xa" represents the average value of the signal, and "σ" represents the standard deviation of the signal. Therefore, the signal normalization unit 33 can perform statistical calculations on various statistical calculation quantities (for example, peak value, kurtosis, skewness, wave height ratio (= peak value / rms value), waveform ratio (= rms value / average value), etc.) including at least the average value and the standard deviation for the split signal S1.

[0037] The effective value calculation unit 34 for each frequency band calculates the effective value J for each of a plurality of frequency bands for the split signal S1 normalized by the signal normalization unit 33. Note that the effective value J is calculated according to a well-known calculation formula and represents the average magnitude (intensity) of the signal time (for example, a predetermined time interval), and is also referred to as the rms (root mean square) value. The effective value calculation unit 34 for each frequency band calculates the effective value J for each frequency band of, for example, several 100 kHz (preferably around 100 KHz) for the split signal S1. Further, the effective value calculation unit 34 for each frequency band calculates the time average value Ja of the effective value J calculated for each frequency band for the split signal S1.

[0038] Here, as shown in FIGS. 6 and 7, the effective value J calculated for each frequency band by the effective value calculation unit 34 for each frequency band can be associated with, for example, the surface roughness Sj actually measured by the roughness meter 2 experimentally.

[0039] That is, the time-averaged value Ja of the effective value J shown in FIG. 6 and FIG. 7 which is an enlarged view of the vertical axis of FIG. 6 is calculated for each of a plurality of frequency bands corresponding to the surface roughness Sj associated with the grinding conditions. And the time-averaged value Ja indicated by the black circles in the figure represents the value in the low-frequency band (for example, from 300 kHz to 400 kHz), and the time-averaged value Ja indicated by the black squares represents the value in the high-frequency band (for example, from 1.1 MHz to 1.2 MHz). Incidentally, in FIGS. 6 and 7, the time-averaged value Ja indicated by the dashed circles represents the value of each of the other frequency bands.

[0040] The multiple regression analysis unit 35 performs multiple regression analysis including at least the time-averaged value Ja (or the effective value J) calculated by the frequency band effective value calculation unit 34 and the surface roughness Sj measured experimentally. Specifically, in this example, the multiple regression analysis unit 35 performs multiple regression analysis on the time-averaged value Ja (or the effective value J) in the corresponding frequency band of the AE sensor 21. For example, the multiple regression analysis unit 35 performs multiple regression analysis using each statistical calculation amount (for example, standard deviation, kurtosis, etc.) in the entire corresponding frequency range of the AE sensor 21 as a parameter, and for example, from the p-value obtained for the change in the actual roughness value Sj of the ground surface W1 measured experimentally in advance by the roughness meter 2, selects a parameter showing a good correlation. Next, multiple regression analysis is performed using the selected parameter to estimate the surface roughness after grinding.

[0041] Therefore, the multiple regression analysis unit 35 of this example includes a first multiple regression analysis unit 351 and a second multiple regression analysis unit 352. The first multiple regression analysis unit 351 performs multiple regression analysis including at least the time-averaged value Ja (or the effective value J) of a plurality of frequency bands calculated by the frequency band effective value calculation unit 34 and the surface roughness Sj for each grinding process. The second multiple regression analysis unit 352 performs multiple regression analysis including the grinding efficiency Z' of the grinding process and the time-averaged value Ja (or the effective value J) of the frequency band selected based on the p-value of various statistical calculation amounts (parameters) and the surface roughness Sj based on the analysis results of all the processes multiple regression analyzed by the first multiple regression analysis unit 351.

[0042] Here, the multiple regression analysis unit 35 (the first multiple regression analysis unit 351 and the second multiple regression analysis unit 352) performs multiple regression analysis on the surface roughness Sj by dividing it into ranges defined by a preset threshold value Sjs, specifically, ranges where Sj < Sjs and Sj ≥ Sjs. This improves the accuracy of the multiple regression analysis.

[0043] However, the second multiple regression analysis unit 352 is not limited to specifying the frequency band (i.e., the parameter corresponding to the frequency band) based on the p-value. For example, the second multiple regression analysis unit 352 can also specify a frequency band having the same tendency as the change tendency of the surface roughness Sj grasped empirically or experimentally.

[0044] The surface roughness estimation unit 36 estimates the surface roughness Ss of the ground surface W1 based on the multiple regression analysis result based on the relative relationship between frequency bands (i.e., between the low-frequency band and the high-frequency band) by the multiple regression analysis unit 35 (i.e., the first multiple regression analysis unit 351 and the second multiple regression analysis unit 352). That is, the surface roughness estimation unit 36 estimates the surface roughness Ss using the correlation obtained by the second multiple regression analysis unit 352 performing multiple regression analysis, that is, an expression having the surface roughness Ss as the objective variable and at least the time average value Ja (or the effective value J) and the surface roughness Sj which is the actual roughness value as the explanatory variables.

[0045] Here, the estimation accuracy of the surface roughness Ss estimated by the surface roughness estimation unit 36 will be described with reference to FIG. 8. FIG. 8 shows a comparison between the surface roughness Ss (represented by white squares) estimated by the surface roughness estimation unit 36 and the surface roughness Sj (represented by embossed squares) actually measured by the roughness meter 2 for 27 evaluation samples.

[0046] In addition, in FIG. 8, the surface roughness estimated values and the actually measured values in three types of grinding efficiencies of evaluation samples 1-4, evaluation samples 5-20, and evaluation samples 21-27 are illustrated. As is also clear from FIG. 8, the surface roughness Ss estimated by the surface roughness estimation unit 36 agrees well with the magnitude and change tendency of the surface roughness Sj (actual roughness value).

[0047] As shown in FIG. 4, the determination unit 37 determines whether or not the surface roughness Ss of the estimated grinding surface W1 is equal to or greater than a reference surface roughness Sr preset for the grinding surface W1 for each grinding process (i.e., rough grinding, finish grinding, fine grinding, and spark out). That is, if the surface roughness Ss is equal to or greater than the reference surface roughness Sr, the determination unit 37 determines that the deterioration of the grinding wheel 46 (tool T) of the cylindrical grinding machine 40 (grinding machine 11) has progressed and the grinding surface W1 cannot be appropriately ground, that is, correction is necessary. On the other hand, if the surface roughness Ss is less than the reference surface roughness Sr, the determination unit 37 determines that the deterioration of the grinding wheel 46 (tool T) of the cylindrical grinding machine 40 (grinding machine 11) has not progressed and the grinding surface W1 can be appropriately ground, that is, correction is unnecessary.

[0048] The output unit 38 outputs a correction signal R to the control device 12 according to the determination result by the determination unit 37. That is, when the determination unit 37 determines that the grinding wheel 46 (tool T) needs to be corrected, the output unit 38 outputs a correction signal R to the control device 12. Also, when the determination unit 37 determines that the grinding wheel 46 (tool T) does not need to be corrected, the output unit 38 does not output a correction signal R to the control device 12.

[0049] Accordingly, in the control device 12 of the cylindrical grinding machine 40 (grinding machine 11), when the deterioration of the grinding wheel 46 (tool T) has progressed, that is, when the grinding wheel 46 (tool T) needs to be corrected, as shown in FIG. 1, the correction signal R is acquired from the estimation calculation device 30 via the interface 13. Then, the control device 12 controls the grinding wheel correction device 48 to correct the grinding wheel 46. That is, the grinding wheel correction device 48 performs truing and / or dressing on the grinding wheel 46. Thereby, the grinding wheel 46 can be corrected at an appropriate timing and an appropriate number of times. Therefore, the time required for correction can be shortened and the cost required for correction can also be reduced.

[0050] As can be understood from the above description, according to the machining state estimation system 1 of this example, the AE signal S is discriminated into a plurality of frequency bands, and by performing a statistical operation using the effective value J in each frequency band, based on the relative relationship between the frequency bands, the surface roughness Ss, which is the machining state on the ground surface W1 of the workpiece W, can be accurately estimated. Then, the machining state estimation system 1 can determine the progress of deterioration of the grinding wheel carriage 46 (tool T) of the cylindrical grinding machine 40 (grinding machine 11), which is the grinding device 10 for grinding the ground surface W1, using the estimated surface roughness Ss, and output a correction signal R. Thereby, the control device 12 of the grinding device 10 can control the grinding wheel carriage correction device 48 at the timing when the correction signal R is acquired, and can correct the grinding wheel carriage 46 by truing and / or dressing.

[0051] Therefore, it is possible to prevent the grinding wheel carriage 46 (tool T) of the cylindrical grinding machine 40 (grinding machine 11) from being excessively corrected or the correction from being delayed. Thereby, the correction time required for correcting the grinding wheel carriage 46 can be shortened, and the cost for correcting the grinding wheel carriage 46 can be reduced.

[0052] (5. Other alternative examples) In the above-described example and the first alternative example, the machining state estimation system 1 is configured such that the grinding device 10 includes a cylindrical grinding machine 40 as the grinding machine 11. And the machining state estimation system 1 is configured to estimate the machining state, that is, the surface roughness Ss, of the ground surface W1 of the workpiece W, which is a shaft-like member ground by the cylindrical grinding machine 40.

[0053] Also, in the above-described example and the first alternative example, the machining state estimation system 1 is configured to estimate the machining state of the workpiece W machined by the grinding device 10. However, the machining state estimation system 1 can also estimate the machining state of the workpiece W machined by a cutting device.

[0054] Furthermore, the machining state is not limited to surface roughness. For example, with respect to the shape (outer contour, inner contour, step, etc.) formed by the tool T contacting the workpiece W, the estimation calculation device 30 of the machining state estimation system 1 can accurately estimate the shape and the like as the machining state using the AE signal S.

[0055] Also, in the above-described example, the estimation calculation device 30 outputs a correction signal R for correcting the tool T based on the estimated surface roughness Ss. However, the estimation calculation device 30 can also output a signal for appropriately changing the machining conditions under which the machine tool processes the workpiece W using the tool T based on the estimated surface roughness Ss. Thereby, in the machine tool, by acquiring the output signal, the machining conditions can be appropriately changed to accurately machine the workpiece W.

[0056] Furthermore, in the above-described example, the observation device 20 includes the AE sensor 21 provided on the center rest 43, which is a component of the cylindrical grinding machine 40 (grinding machine 11) that is the machine tool. When the grinding wheel carriage 46 (tool T) contacts the workpiece W to grind the grinding surface W1 of the workpiece W, the AE signal S generated is acquired. Then, using the acquired AE signal S, the surface roughness Ss as the machining state of the grinding surface W1, and thus the machining state of the grinding wheel carriage 46 (tool T) is estimated.

[0057] However, the AE observed by the observation device 20 is not limited to acquiring the AE signal S propagated from the components of the cylindrical grinding machine 40 (grinding machine 11) including the center rest 43. The observation device 20 can also estimate the surface roughness Ss as the machining state of the grinding surface W1, in other words, the machining state by the grinding wheel carriage 46 (tool T), based on the AE signal acquired via components other than those of the cylindrical grinding machine 40 (grinding machine 11).

Explanation of Reference Numerals

[0058] 1…Machining state estimation system, 2…Roughness meter, 10…Grinding device, 11…Grinding machine, 12…Control device, 13…Interface, 20…Observation device, 21…AE sensor, 30…Estimation calculation device, 31…Signal acquisition unit, 32…Signal division unit, 33…Signal normalization unit, 34…Frequency band effective value calculation unit, 35…Multiple regression analysis unit, 351…First multiple regression analysis unit, 352…Second multiple regression analysis unit, 36…Estimation unit, 37…Judgment unit, 38…Output unit, 40…Cylindrical grinding machine, 41…Bed, 42…Spindle headstock, 42a…Motor, 43…Center rest, 44…Cross slide, 44a…Motor, 45…Wheelhead, 45a…Motor, 46…Grinding wheel, 46a…Motor, 47…Sizing device, 47a…Feed mechanism, 48…Grinding wheel dressing device, 49…Coolant device, J…Effective value, Ja…Average value, R…Correction signal, S…AE signal, S1…Divided signal, Sjs…Threshold value, T…Tool, W…Workpiece, W1…Grinding surface

Claims

1. A machining state estimation method applied to a machine tool for machining a workpiece with a tool, the method estimating a machining state when the tool machines the workpiece, comprising: observing acoustic emissions generated with frequency characteristics as the workpiece is machined by the tool; calculating effective values for each of a plurality of frequency bands for an acoustic emission signal representing the observed acoustic emissions; performing multiple regression analysis using, as a parameter, the statistical calculation amount selected from among the statistical calculation amounts including the effective values for each frequency band obtained by performing statistical calculation on the acoustic emission signal, thereby obtaining a correlation between the effective values and the surface roughness of the workpiece as the machining state, and estimating the surface roughness of the workpiece as the machining state; the multiple regression analysis includes first multiple regression analysis and second multiple regression analysis; in the first multiple regression analysis, performing multiple regression analysis including at least the effective value for each frequency band and the surface roughness; in the second multiple regression analysis, based on the analysis result obtained by performing multiple regression analysis in the first multiple regression analysis, performing multiple regression analysis including the machining efficiency, the effective value of the frequency band selected based on the statistical calculation amount, and the surface roughness; in the first multiple regression analysis and the second multiple regression analysis, performing multiple regression analysis by dividing the surface roughness into ranges defined by a preset threshold value; in the estimation of the surface roughness as the machining state, using a relationship with the surface roughness, which is the correlation obtained by performing multiple regression analysis in the second multiple regression analysis, as the objective variable, and at least the effective value and the surface roughness value, which is the actual roughness value, as the explanatory variables, to estimate the surface roughness. A machining state estimation method.

2. Based on the estimated machining state, setting a timing for correcting the tool, or changing machining conditions under which the machine tool machines the workpiece using the tool. The machining state estimation method according to claim 1.

3. The machine tool is a grinding device having a grinding wheel as the tool, and grinding the workpiece with the grinding wheel. The machining state estimation method according to claim 1 or 2.

4. The machining state is estimated as surface roughness. The machining state estimation method according to any one of claims 1 to 3.

5. A machining state estimation system applied to a machine tool that machines a workpiece with a tool, and estimates a machining state when the tool machines the workpiece, comprising: an observation device that observes acoustic emissions generated with frequency characteristics as the workpiece is machined by the tool; a signal acquisition unit that acquires an acoustic emission signal representing the acoustic emissions observed by the observation device; an effective value calculation unit that calculates the effective value of each of a plurality of frequency bands for the acoustic emission signal acquired by the signal acquisition unit; an estimation unit that estimates the machining state based on the correlation between the effective value calculated by the effective value calculation unit and the machining state; and an estimation calculation device having the same; is provided, the estimation unit includes a multiple regression analysis unit that performs multiple regression analysis using, as a parameter, the statistical calculation amount selected from among the statistical calculation amounts including the effective value for each frequency band obtained by performing statistical calculation on the acoustic emission signal, and obtains the correlation to estimate the surface roughness of the workpiece as the machining state; the multiple regression analysis unit includes a first multiple regression analysis unit and a second multiple regression analysis unit; the first multiple regression analysis unit performs multiple regression analysis including at least the effective value for each frequency band and the surface roughness; the second multiple regression analysis unit performs multiple regression analysis including the machining efficiency, the effective value of the frequency band selected based on the statistical calculation amount, and the surface roughness, based on the analysis result obtained by performing multiple regression analysis by the first multiple regression analysis unit; the first multiple regression analysis unit and the second multiple regression analysis unit perform multiple regression analysis by dividing the surface roughness into ranges defined by a preset threshold value; the estimation unit estimates the surface roughness using a relationship in which the surface roughness, which is the correlation obtained by the second multiple regression analysis unit performing multiple regression analysis, is the objective variable, and at least the effective value and the surface roughness value, which are the actual roughness values, are the explanatory variables. A machining state estimation system.

Images