Rotation deflection amount detection device of cylindrical workpiece

Abstract

To provide a rotation deflection amount detection device without requiring installation work and attachment / detachment, while detection accuracy is not affected by external disturbances caused by grinding liquid or the like.SOLUTION: According to a rotation deflection amount detection device (electronic control device 70, AE sensor 64), a rotation deflection amount SW of a cylindrical workpiece W is detected by a deflection amount calculation part 82 based on a signal intensity integral value SPI within a predetermined frequency bandwidth of an AE signal by using the AE signal as an elastic wave generated from a grind processing point with respect to the cylindrical workpiece W of a grind stone 42. Thus, detection accuracy of the rotation deflection amount SW is not affected by external disturbances caused by grinding liquid or the like, or installation work and attachment / detachment of a sensor head or the like are not required in the vicinity of the cylindrical workpiece W.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a device for detecting runout of a cylindrical workpiece, which detects the amount of runout of a cylindrical workpiece by utilizing elastic waves generated from a grinding point of a grinding wheel. [Background technology]

[0002] In a cylindrical grinding machine that grinds a cylindrical workpiece, it is desirable to detect the amount of rotational runout of the cylindrical workpiece during the grinding process, for example, in order to control the amount of cut so as to eliminate the amount of rotational runout of the cylindrical workpiece.

[0003] Patent Document 1 discloses a rotational runout detection device that uses a pair of contacts that clamp the cylindrical workpiece to detect the amount of rotational runout of the cylindrical workpiece during grinding. For example, the rotational runout detection device described in Patent Document 1 is such a device. However, this type of rotational runout detection device is of a type in which the pair of contacts are brought into contact with the surface of the cylindrical workpiece, and therefore there is a risk of scratches being generated on the surface of the cylindrical workpiece.

[0004] In response to this, a rotational runout detection device has been proposed that places a sensor head, which includes a light-emitting element that emits light onto the surface of a cylindrical workpiece and a light-receiving element that receives reflected light from the surface of the cylindrical workpiece, close to the surface of the cylindrical workpiece, performs FFT analysis on the received light amount data, and calculates the rotational runout of the cylindrical workpiece during grinding based on the magnitude of the amplitude of the signal within a predetermined low frequency range. For example, the rotational runout detection device described in Patent Document 2 is such an example.

[0005] The rotational runout detection device described in Patent Document 2 is of a non-contact type in which the sensor head does not come into contact with the surface of the cylindrical workpiece, and therefore there is no risk of scratching the surface of the cylindrical workpiece. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2013-111686 A [Patent Document 2] JP 2018-024079 A Summary of the Invention [Problem to be solved by the invention]

[0007] However, in the rotational runout detection device described in Patent Document 2, the sensor head needs to be placed close to the surface of the cylindrical workpiece, so that even if the grinding fluid is blown away by air, the remaining grinding fluid may disturb the received light amount data, resulting in insufficient detection accuracy. In addition, in order to install the sensor head of the rotational runout detection device close to the surface of the cylindrical workpiece, construction work for installing electrical piping and air piping is required, and there is also the problem that the installation and removal of the device after installation is complicated.

[0008] The present invention has been made in light of the above circumstances, and an object of the present invention is to provide a rotational runout detection device whose detection accuracy is not affected by disturbances such as grinding fluid, and which does not require installation work or removal. [Means for solving the problem]

[0009] The inventor of the present invention has conducted various studies in the light of the above circumstances, and has found that by successively calculating the integral value of the signal strength within a predetermined frequency band when the AE signal output from the AE sensor is frequency-analyzed, the time during which the signal strength exceeds an arbitrary threshold corresponds to the runout time of the cylindrical workpiece, and the product of the runout time and the cutting speed can be calculated as the runout amount of the cylindrical workpiece. The present invention has been made based on this finding.

[0010] In other words, the gist of the present invention is (1) a rotational runout detection device for a cylindrical workpiece that detects the rotational runout of the cylindrical workpiece by utilizing an AE signal, which is an elastic wave generated from a grinding point of a grinding wheel on the cylindrical workpiece, and includes: (2) a frequency analysis unit that sequentially performs frequency analysis on the AE signal; (3) a signal intensity integral calculation unit that sequentially calculates a signal intensity integral within a predetermined frequency band of the frequency spectrum obtained by the frequency analysis unit; and (4) a runout detection unit that detects the rotational runout of the cylindrical workpiece based on the signal intensity integral sequentially calculated by the signal intensity integral calculation unit from a predetermined relationship. Effect of the Invention

[0011] According to the rotational runout detection device for a cylindrical workpiece of the present invention, an AE signal, which is an elastic wave generated from the grinding point of the grinding wheel on the cylindrical workpiece, is utilized, and the runout calculation unit detects the rotational runout of the cylindrical workpiece based on the integrated value of the signal intensity of the AE signal within a predetermined frequency band. Therefore, the detection accuracy of the rotational runout is not affected by disturbances such as grinding fluid, and there is no need to install or remove a sensor head near the cylindrical workpiece.

[0012] Preferably, the signal intensity integral calculated by the signal intensity integral calculation unit is an integral of the signal intensity within a preset frequency range of the frequency spectrum of the AE signal frequency-analyzed by the frequency analysis unit. This allows the AE signal to be effectively reflected in the signal intensity integral, reduces the effect of noise, and improves the detection accuracy of the rotational runout amount. The preset frequency range is preferably a band in which signal intensity peaks are concentrated.

[0013] Preferably, the data length of the AE signal subjected to frequency analysis by the frequency analysis unit is equal to or shorter than 0.13 revolutions of the cylindrical workpiece, thereby making the detection error of the amount of rotational runout of the cylindrical workpiece sufficiently small.

[0014] Preferably, an integral analysis pitch, which is a plot interval of the signal intensity integral calculated by the signal intensity integral calculation unit, is equal to or less than ½ of the rotation period T of the cylindrical workpiece. This makes the integral analysis pitch equal to or less than the contact period of the cylindrical workpiece with the grinding wheel, thereby improving the detection accuracy of the amount of rotational runout.

[0015] Preferably, the runout detection unit includes a moving standard deviation calculation unit that calculates a moving standard deviation of the signal intensity integrals sequentially calculated by the signal intensity integral calculation unit, and calculates the rotational runout of the cylindrical workpiece based on the moving standard deviation of the signal intensity integrals calculated by the moving standard deviation calculation unit. In this way, by using the moving standard deviation, the calculation accuracy of the rotational runout of the cylindrical workpiece is improved without being affected by local fluctuations in the signal intensity integrals.

[0016] Preferably, the moving standard deviation calculation unit calculates the moving standard deviation of the signal intensity integral value SPI using a moving average time longer than one rotation time of the cylindrical workpiece, thereby making it possible to stably obtain the moving standard deviation regardless of pulsation of the signal intensity integral value SPI, thereby improving the detection accuracy of the amount of rotational runout.

[0017] Preferably, the runout detection unit includes a runout time calculation unit that calculates a runout time, which is the time when the moving standard deviation of the signal intensity integral value calculated by the moving standard deviation calculation unit exceeds a preset runout time judgment threshold, and a runout calculation unit that calculates the multiplication value of the runout time and the cutting speed of the grinding wheel as the rotational runout amount of the cylindrical workpiece. In this way, the rotational runout amount of the cylindrical workpiece can be obtained based on the actual runout time and the actual cutting speed of the grinding wheel from the multiplication equation of the runout time when the moving standard deviation of the signal intensity exceeds a preset runout time judgment threshold and the cutting speed of the grinding wheel, which is a preset relationship, so that the detection accuracy of the rotational runout amount is not affected by disturbances due to grinding fluid, etc.

[0018] Preferably, the preset shake time determination threshold is set as a percentage of the maximum value of the moving standard deviation. This allows the shake time to be accurately determined. Preferably, the shake time determination threshold is a value equal to or less than 50% of the maximum value of the moving standard deviation.

[0019] Preferably, the preset shake time determination threshold value is a value within a range of 40 to 80% of the maximum value of the moving standard deviation, thereby enabling the shake time to be determined accurately. [Brief description of the drawings]

[0020] [Figure 1] 1 is a plan view illustrating a configuration of a grinding machine equipped with a cylindrical workpiece rotation runout detection device according to an embodiment of the present invention; [Diagram 2] 2 is a cross-sectional view for explaining the configuration of a grindstone and its mounting structure used in the grinding apparatus of FIG. 1. [Diagram 3] FIG. 1 is a diagram showing the change over time in the load (power consumption of the grinding wheel spindle) of a grinding wheel drive motor that drives and rotates a cylindrical workpiece in a cylindrical grinding machine. [Figure 4] FIG. 2 is a diagram showing a time change in the signal intensity integral value SPI, which is the integral value of the signal intensity within a predetermined frequency band, in a spectrum obtained by frequency analysis of an AE signal generated during cylindrical grinding. [Diagram 5] 5 is a diagram showing a pulsating waveform of the signal intensity integral value SPI with the time axis of FIG. 4 expanded. [Figure 6] 6 is an explanatory diagram for calculating a deflection time Tf of a cylindrical workpiece W from the pulsation waveform of the signal intensity integral value SPI in FIG. 5. [Figure 7] 2 is a flowchart illustrating a main part of the control operation of the electronic control device of FIG. [Figure 8] 1 is a table showing a comparison of actual measured values ​​and analytical values ​​of rotational runout SW of a cylindrical workpiece obtained in Experimental Example 1. [Figure 9] 9 is a line graph comparing actual measured values ​​and analytical values ​​of rotational runout SW of the cylindrical workpiece of FIG. 8 . [Figure 10]13 is a table showing a comparison between the AE signal data length used in FFT analysis, obtained by changing the AE signal data length in Experimental Example 2, and the difference (error) between the value of the rotational runout amount SW in Experimental Example 1. [Figure 11] 11 is a line graph showing a comparison between the AE signal data length used in the FFT analysis in FIG. 10 and the difference (error) in the rotational shake amount SW. [Figure 12] FIG. 13 is a diagram showing the time change of the signal intensity integral value SPI in the case where the AE signal data length used in the FFT analysis is 0.016 sec (corresponding to 0.03 rotations of a cylindrical workpiece) in Experimental Example 2. [Figure 13] FIG. 13 is a diagram showing the change over time of the signal intensity integral value SPI in the case where the AE signal data length used in the FFT analysis is 0.5 sec (corresponding to 1.0 rotation of a cylindrical workpiece) in Experimental Example 2. [Figure 14] 13 is a table showing a comparison between the moving average time and the value of the rotational runout amount SW in Experimental Example 1, obtained by changing the moving average time in Experimental Example 3. [Figure 15] 15 is a line graph illustrating the difference (error) between the moving average time and the value of the rotational shake amount SW in Experimental Example 1 in comparison with FIG. 14. [Figure 16] 13 is a table showing analytical values ​​obtained by changing the integral analysis pitch in Experimental Example 4, and the differences between the analytical values ​​and the actually measured values. [Figure 17] 17 is a line graph showing the relationship between the integral analysis pitch in FIG. 16 and the difference between the analytical value and the actual measurement value. [Figure 18] 13 is a table showing the optimum runout time judgment threshold values ​​obtained for each of a plurality of cylindrical workpieces in Experimental Example 5. [Figure 19] 19 is a line graph showing the relationship between the multiple cylindrical workpieces in FIG. 18 and the optimum runout time determination threshold value obtained for each of them. [Figure 20] 13 is a line graph showing optimal vibration time determination thresholds set as ratios to the maximum value, obtained for each of a plurality of FFT analysis data lengths in Experimental Example 6. [Figure 21]13 is a line graph showing optimal shake time determination threshold values ​​set as magnifications obtained for each of a plurality of FFT analysis data lengths in Experimental Example 6. [Figure 22] 13 is a table showing the amount of rotational runout obtained by changing the analysis frequency band for each of a plurality of cylindrical workpieces in Experimental Example 7, and the difference between the amount of rotational runout and the actual measured value. [Figure 23] 23 is a line graph showing the difference between the amount of rotational runout and the actual measured value for each of the multiple cylindrical workpieces in FIG. 22. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0021] An embodiment of the present invention will be described in detail below with reference to the drawings. Note that in the following embodiment, the drawings are for explaining the main parts related to the invention, and dimensions, shapes, etc. are not necessarily drawn accurately. EXAMPLES

[0022] Fig. 1 is a diagram for explaining the configuration of a cylindrical grinding machine 12, which is a grinding device equipped with an AE sensor 64 that detects elastic waves generated in a grinding wheel 42 during grinding of a cylindrical workpiece W and outputs an AE signal. In Fig. 1, a bed 14 of the cylindrical grinding machine 12 includes an X-direction bed portion 14a elongated in the X direction and a Y-direction bed portion 14b elongated in the Y direction and protruding from the middle of the X-direction bed portion 14a in the Y direction from the longitudinal direction, and forms an inverted T shape as a whole in a plan view.

[0023] A pair of X-direction guide rails 18 for guiding the X-direction table 16 in the X-direction are fixed on the X-direction bed portion 14a. A spindle housing 24 is fixed on one end of the X-direction table 16, which rotatably supports a spindle 20 and has a spindle motor 22 fixed thereto for rotating the spindle 20. A tailstock housing 28 for supporting a tailstock center 26 is provided on the other end of the X-direction table 16. The tailstock housing 28 is fixed on a guide rail 30 fixed to the X-direction table 16 so that its position can be changed. As a result, a longitudinal cylindrical workpiece W, such as a steel bar, is supported by the spindle center 32 formed on the spindle 20 and the tailstock center 26 so as to be rotatable about an axis parallel to the X-direction.

[0024] An X-direction drive motor 34 is disposed at one end of the X-direction bed portion 14a. The X-direction table 16 is screwed into an X-direction ball screw 36 that is rotated by the X-direction drive motor 34. This allows the cylindrical workpiece W to be moved in the X direction during grinding.

[0025] A pair of Y-direction guide rails 40 for guiding a Y-direction table 38 elongated in the Y direction in the Y direction are fixed on the Y-direction bed section 14b. A grindstone shaft housing 46 that rotatably supports a grindstone shaft 44, one end of which has a grindstone 42 fixed thereto, around an axis C1 parallel to the X direction, and a grindstone drive motor 50 that rotates and drives the grindstone shaft 44 via a belt 48 are fixed on the Y-direction table 38. A Y-direction drive motor 52 is disposed on one end of the Y-direction bed section 14b. The Y-direction table 38 is screwed into a Y-direction ball screw 54 that is rotated and driven by the Y-direction drive motor 52. This causes the grindstone 42 to move in the cutting direction relative to the cylindrical workpiece W.

[0026] 2 shows a structure for fixing the grinding wheel 42 to the grinding wheel shaft 44. The grinding wheel 42 is fixed to the grinding wheel shaft 44 by being sandwiched between a fixing flange 56 fixed to the shaft end of the grinding wheel shaft 44 and a mounting flange 60 fixed to the fixing flange 56 by a fixing bolt 58. An annular groove 62 is formed on the outer periphery of the mounting flange 60, and the annular groove 62 contains an AE sensor 64, a preamplifier (not shown), an A / D converter, a transmission circuit, a constant voltage circuit, a battery, and the like for amplifying the AE signal, which is the output signal of the AE sensor 64, and transmitting it as a digital signal. The annular groove 62 is closed by a cover member 66 made of a material that allows radio waves to pass through, such as synthetic resin.

[0027] The A / D converter has high resolution and converts the AE signal into a digital signal at a sampling period of, for example, 10 μs (microseconds) or less, preferably 5 μs or less, and more preferably 1 μs or less. The shorter (faster) the sampling period of the A / D converter, the clearer the first frequency band B1 (described below) associated with grain spillage (abrasive grain breakage) and the second frequency band B2 (described below) associated with frictional vibration and elastic vibration caused by contact (rubbing) between the abrasive grains and the workpiece. In the following examples, a sampling period of 1 μs is used for the A / D converter.

[0028] The grinding wheel 42 is annular, and is made of general abrasive grains such as fused alumina-based abrasive grains, silicon carbide-based abrasive grains, ceramic abrasive grains, CBN abrasive grains, diamond abrasive grains, and other abrasive grains bound together by a binder such as a vitrified bond or metal bond. The AE sensor 64 detects AE waves (acoustic emissions), which are extremely high-frequency crushing vibrations in the ultrasonic region of, for example, 20 kHz or more that are generated when the abrasive grains contained in the grinding wheel 42 are crushed and propagate through the grinding wheel 42, through a mounting flange 60 that is in close contact with the grinding wheel 42, and outputs an AE signal, which is an analog electrical signal that represents the crushing vibration. The AE sensor 64 is configured, for example, to have a receiving plate (not shown) at one end that detects elastic waves, and to include a mechanical / electrical conversion element, such as a piezoelectric element, that converts the mechanical vibrations received by the receiving plate into an AE signal and outputs it.

[0029] Returning to FIG. 1, the electronic control device 70 is a so-called microcomputer, and detects the rotational runout SW of the cylindrical workpiece W being ground by processing the AE signal SAE, which is an input signal from the receiving circuit 68, according to a prestored program. The electronic control device 70 functions as a rotational runout detection device for the cylindrical workpiece W together with the AE sensor 64. The electronic control device 70 functionally includes a frequency analysis unit 72 that generates a frequency spectrum by performing frequency analysis of the AE signal SAE by FFT at a period of time equal to or shorter than 0.13 rotations of the cylindrical workpiece W being ground, a signal intensity integral value calculation unit 74 that sequentially calculates a value representing the signal intensity within a predetermined wavelength band including, for example, 25-45 kHz or 45-75 kHz from the frequency spectrum, for example, a signal intensity integral value (accumulated value) SPI, and a runout detection unit 76 that detects the rotational runout SW (mm) of the cylindrical workpiece W based on the signal intensity integral value SPI. If the rotation speed of the cylindrical workpiece W is N (rpm), the frequency analysis data length Tl (sec) in the frequency analysis unit 72 is set to a value equal to or less than (60 / N) / 8, and the integral analysis pitch Δt of the signal intensity integral value SPI is set to a value equal to or less than half the frequency analysis data length Tl (Tl / 2).

[0030] The runout detection unit 76 includes a moving standard deviation calculation unit 78 that sequentially calculates the moving standard deviation MSD of the signal intensity integral value SPI sequentially calculated by the signal intensity integral value calculation unit 74 using a moving average time longer than one rotation time of the cylindrical workpiece W, and calculates the rotational runout amount SW of the cylindrical workpiece W based on the moving standard deviation MSD of the signal intensity integral value SPI calculated by the moving standard deviation calculation unit 78.

[0031] The runout detection unit 76 includes a runout time calculation unit 80 that calculates a runout time Tf (sec), which is the time during which the moving standard deviation MSD of the integrated value SPI of the signal intensity calculated by the moving standard deviation calculation unit 78 continuously exceeds a runout time judgment threshold JD that is preset within a range of, for example, approximately 40% to 80% of the peak value (maximum value) of the moving standard deviation MSD, and a runout calculation unit 82 that calculates the multiplication value of the runout time Tf (sec) and the cutting speed CV (mm / min) of the grinding wheel from a pre-stored formula (SW=Tf×CV / 60) and sets the multiplication value as the rotational runout amount SW (mm) of the cylindrical workpiece W, and displays the rotational runout amount SW (mm) of the cylindrical workpiece W calculated by the runout calculation unit 82 on a display 84.

[0032] The moving standard deviation MSD is calculated from the following formula 1. In formula 1, n is the number of analysis data (integer) within the moving average time (moving section), x i is the i-th signal strength integral value SPI, x av is the average value of the signal strength integral value SPI within the moving average time.

number

[0033] FIG. 3 shows the change in power consumption (kW), which is the load of the grinding wheel drive motor 50 in cylindrical grinding. At the beginning of processing, the power consumption is the same as that at the time of non-processing, but it increases from the end of the initial processing period and saturates after about 8 seconds. FIG. 4 shows the change in the signal intensity integral value SPI in the frequency range of 45-75 kHz extracted from the frequency spectrum of the AE signal SAE. At the beginning of processing, it pulsates in synchronization with the rotation of the cylindrical workpiece W with an amplitude larger than that at the time of non-processing, and the pulsation converges at the end of the initial processing period and saturates near the peak value at the beginning of processing. FIG. 5 shows the change in the signal intensity integral value SPI at the beginning of processing in FIG. 4 on an expanded time axis. FIG. 6 shows the runout time Tf, which is the time when the pulsating signal intensity integral value SPI exceeds the runout time judgment threshold value JD.

[0034] FIG. 7 is a flow chart for explaining the main part of the control operation of the electronic control device 70. In FIG. 7, in step S1 (hereinafter, step will be omitted), AE signal data generated during cylindrical grinding is read. In S2 corresponding to the frequency analysis unit 72, the AE signal SAE read in S1 at a period of time of 0.13 revolutions or less of the cylindrical workpiece W being ground is frequency analyzed by FFT to generate a frequency spectrum. In S3 corresponding to the signal strength integral value calculation unit 74, the signal strength integral value (integrated value) SPI within a predetermined wavelength band including, for example, 25-45 kHz or 45-75 kHz of the frequency spectrum is sequentially calculated. In S4 corresponding to the moving standard deviation calculation unit 78, the moving standard deviation MSD of the signal strength integral value SPI sequentially calculated by S3 is sequentially calculated using a moving average time longer than one rotation time of the cylindrical workpiece W. In the following S5, it is determined whether cylindrical grinding has been completed or not. If the determination in S5 is negative, steps S1 and onwards are repeated, but if the determination in S5 is positive, steps S6 and onwards are executed.

[0035] In S6 corresponding to the runout time calculation unit 80, a runout time TF is calculated as the time during which the moving standard deviation MSD of the integrated value SPI of the signal strength calculated in S3 continuously exceeds a runout time judgment threshold JD preset to, for example, about 50% of the peak value (maximum value) of the moving standard deviation MSD. In S5, in S7 corresponding to the runout amount calculation unit 82, a multiplication value of the runout time Tf (sec) calculated in S4 and the cutting speed CV (mm / min) of the grinding wheel 42 is calculated from a pre-stored formula (SW=TF×CV / 60), and the multiplication value is set as the rotational runout amount SW (mm) of the cylindrical workpiece W. In S8, the rotational runout amount SW (mm) of the cylindrical workpiece W calculated in S7 is displayed on the display 84.

[0036] (Experimental Example 1) The inventors performed cylindrical grinding using eight cylindrical workpieces No. 1 to No. 8 in which rotational runout was intentionally generated using the cylindrical workpieces W obtained under the grinding conditions shown below. The rotational runout amounts SW of the eight cylindrical workpieces No. 1 to No. 8 were measured using the measurement conditions shown below, and calculated using the AE signal wave analysis conditions shown below. Grinding conditions Processing machine: Cylindrical grinding machine Grinding wheel specifications: CX 60 KV Grinding wheel outer diameter: 405mmφ Grinding wheel speed: 33mm / sec Workpiece peripheral speed: 0.33mm / sec Material of cylindrical workpiece: SUS-304 Outer diameter of cylindrical workpiece: 50mmφ Cutting speed: 0.382mmφ / min Grinding efficiency: Z'=0.5mm 2 / mm·sec Sparkout: 10rev. Grinding oil: SEC-700 (dilution ratio 50 times) - Measurement conditions for the amount of rotational runout of cylindrical workpiece W Measurement method: Contact measurement using a probe Measuring Instrument: Test Indicator AE signal analysis conditions FFT analysis data length: 0.016 sec. Integral analysis pitch Δt: 0.004 sec. Integration frequency range: 45~75kHz Moving standard deviation analysis time: 0.050 sec Threshold for determining the duration of the swing: 50% of the maximum value of the moving standard deviation Sampling period for grinding wheel power consumption: 0.1sec.

[0037] Next, the AE signal SAE obtained during cylindrical grinding of each of the cylindrical workpieces No. 1 to No. 8 was processed as described above, and the rotational runout amount SW for each of the cylindrical workpieces No. 1 to No. 8 was calculated from the AE signal SAE using the signal intensity integral value calculation unit 74 and the runout amount detection unit 76. For each of the eight cylindrical workpieces No. 1 to No. 8, the rotational runout amount SW analyzed (calculated) from the AE signal SAE using the signal intensity integral value calculation unit 74 and the runout amount detection unit 76 was compared with the actually measured rotational runout amount, and the results are shown in the table of FIG. 8 and the line graph of FIG. 9.

[0038] 8 and 9, the difference between the actual value and the analytical value was 8 μm or less. From this, as described above, it is understood that the rotational runout amount SW analyzed (calculated) from the AE signal SAE using the signal intensity integral value calculation unit 74 and the runout amount detection unit 76 does not differ significantly from the actual measurement value, and the rotational runout amount SW can be estimated instead of the actual measurement.

[0039] (Experimental Example 2) Next, the inventors processed the AE signals SAE obtained during cylindrical grinding of each of the cylindrical workpieces No. 1 to No. 8 as described above to calculate the rotational runout amounts SW for each of the cylindrical workpieces No. 1 to No. 8. The data lengths (FFT analysis data lengths) of the AE signals used in the frequency analysis (FFT) were seven types: 0.008 sec (equivalent to 0.015 rotations of the cylindrical workpiece), 0.016 sec (equivalent to 0.03 rotations of the cylindrical workpiece), 0.031 sec (equivalent to 0.06 rotations of the cylindrical workpiece), 0.063 sec (equivalent to 0.13 rotations of the cylindrical workpiece), 0.125 sec (equivalent to 0.25 rotations of the cylindrical workpiece), 0.25 sec (equivalent to 0.5 rotations of the cylindrical workpiece), and 0.5 sec (equivalent to one rotation of the cylindrical workpiece). The inventors calculated the rotational runout amounts SW when calculating each of the rotational runout amounts SW from the AE signals SAE using the signal intensity integral value calculation unit 74 and the runout amount detection unit 76. 10 and the line graph of FIG. 11 show, for each of cylindrical workpieces No. 1 to No. 8, different data lengths, the rotational runout amounts SW obtained using the data lengths, and the differences in the runout amounts SW from the values ​​in Experimental Example 1.

[0040] 10 and 11, when the FFT analysis data length exceeds 0.125 sec, the difference between the analytical value and the actual measurement value of the rotational runout amount SW is 11 μm or more, so this is an inappropriate analysis time.When the FFT analysis data length is 0.063 sec or less, the difference between the analytical value and the actual measurement value of the rotational runout amount SW is 8 μm or less, so it is desirable for the FFT analysis data length to be 0.063 sec (0.13 rotations as analysis time) or less.

[0041] In the AE wave analysis of cylindrical workpiece No. 3, Fig. 12 shows the time change of the signal intensity integral value SPI when the analysis data length is set to 0.016 sec. (equivalent to 0.03 rotation of the cylindrical workpiece), and Fig. 13 shows the time change of the signal intensity integral value SPI when the analysis data length is set to 0.5 sec. (equivalent to 1.0 rotation of the cylindrical workpiece). In Fig. 12, a large increase and decrease in the signal intensity integral value SPI of the AE wave caused by the rotational runout of the cylindrical workpiece can be seen, but in Fig. 13, such increase and decrease is small, so the moving standard deviation MSD is small, and as a result, it is thought that the rotational runout cannot be calculated appropriately. The integral analysis pitch Δt is the plot interval of the vibration intensity integral value SPI. The analysis flow is as follows: (a) FFT subjecting the AE signal wave; (b) integrating the signal strength of any frequency band among the FFT-analyzed signal strengths over a time period equal to the FFT analysis data length, and plotting the signal strength integral value as one point on a horizontal-axis graph; (c) integrating from a point advanced in time by an integral analysis pitch Δt over a time period equal to the FFT analysis data length, and plotting one point on a graph of horizontal-axis time vs. vertical-axis vibration intensity component value; and (d) repeating the above (a) to (c).

[0042] (Experimental Example 3) Next, the inventors processed the AE signals SAE obtained during cylindrical grinding of each of the cylindrical workpieces No. 1 to No. 8 as described above, and when calculating the rotational runout amounts SW for each of the cylindrical workpieces No. 1 to No. 8, five different moving average times of the moving standard deviation MSD were used: 0.25 sec (equivalent to 0.5 rotations of the cylindrical workpiece), 0.5 sec (equivalent to 1 rotation of the cylindrical workpiece), 1 sec (equivalent to 2 rotations of the cylindrical workpiece), 1.5 sec (equivalent to 3 rotations of the cylindrical workpiece), and 2 sec (equivalent to 4 rotations of the cylindrical workpiece), and calculated the rotational runout amounts SW when each rotational runout amount SW was calculated from the AE signal SAE using the signal intensity integral calculation unit 74 and the runout amount detection unit 76. The table in FIG. 14 and the line graph in FIG. 15 show, for each of cylindrical workpieces No. 1 to No. 8, different data lengths, the rotational runout amounts SW obtained using the data lengths, and the differences in the rotational runout amounts SW from the values ​​in Experimental Example 1.

[0043] As is clear from Figures 14 and 15, when the moving average time is shorter than one rotation time of the cylindrical workpiece, all the analytical values ​​are 1 μm, and the greater the runout of the actual cylindrical workpiece W, the greater the difference between the measured value and the analytical value of the rotation runout amount SW, making this an inappropriate moving average time. When the moving average time is longer than one rotation time of the cylindrical workpiece, the difference between the measured value and the analytical value of the rotation runout amount SW is small, so the moving average time needs to be longer than one rotation time of the cylindrical workpiece. On the other hand, the longer the moving average time is than one rotation time of the cylindrical workpiece, the greater the difference between the measured value and the analytical value of the rotation runout amount SW tends to be, so it is preferable for the moving average time to be a value equivalent to one rotation of the cylindrical workpiece.

[0044] (Experimental Example 4) Next, in order to investigate the relationship between the amount of rotational runout of the cylindrical workpiece W and the integral analysis pitch Δt, the inventors performed an FFT analysis of cylindrical workpiece No. 3 with an FFT analysis data length T1 of 0.016 sec (corresponding to 1 / 32 rotation) at six levels of integral analysis pitch Δt (T1 / 8 to 4T1) and calculated the difference between the analytical value at each level and the actual measured value (16 μm) of cylindrical workpiece No. 3. The obtained analytical values ​​(μm) and differences (μm) for the six levels are shown in the table of FIG. 16 and the line graph of FIG. 17.

[0045] 16 and 17, the greater the integral analysis pitch Δt is, the greater the deviation between the measured value and the analytical value is. For this reason, it is preferable to perform the analysis with an integral analysis pitch Δt of T1 / 2 or less.

[0046] (Experimental Example 5) Next, in the present invention, the runout time judgment threshold used in calculating the runout time is defined as a ratio to the maximum value of the moving standard deviation. In order to investigate the setting range of the runout time judgment threshold, the inventors changed the runout time judgment threshold in 1% increments from 0 to 100% for each of cylindrical workpieces No. 1 to No. 8, and determined the optimum threshold value that minimizes the difference between the calculated workpiece rotation runout amount analysis value and the actual measurement value (1 μm or less), and the obtained optimum threshold value is shown in the table of FIG. 18 and the line graph of FIG. 19 for each of cylindrical workpieces No. 1 to No. 8.

[0047] 18 and 19, the optimal threshold was in the range of 39 to 82% of the maximum value of the moving standard deviation. This revealed that it is preferable to set the shake time determination threshold in the range of 40 to 80% of the moving standard deviation.

[0048] (Experimental Example 6) Next, regarding the definition of the swing time determination threshold used when calculating the swing time, it is often defined as a specific value or a multiplication factor relative to a reference value. When the swing time determination threshold in this embodiment is defined using such a general method, the threshold becomes a specific moving standard deviation or a value obtained by multiplying the moving standard deviation during non-machining by a specific multiplication factor. The inventors have examined the suitability of such a definition below.

[0049] Figure 20 shows the change in the shake time judgment threshold with respect to the change in the FFT analysis data length T when the shake time judgment threshold is defined as a percentage of the maximum value of the moving standard deviation, and Figure 21 shows the change in the shake time judgment threshold with respect to the change in the FFT analysis data length T when the shake time judgment threshold is defined as a percentage of a certain specific moving standard deviation value (broken line connecting ● points) and when the shake time judgment threshold is defined as a value obtained by multiplying the moving standard deviation value at the unprocessed time by a certain magnification factor (broken line connecting ■ points).

[0050] In Fig. 20, the optimum ratio does not change even if the FFT analysis data length T changes, but in Fig. 21, the optimum ratio increases as the FFT analysis data length T increases. For this reason, it is preferable to estimate the optimum shake time determination threshold value as a ratio to the maximum value of the moving standard deviation.

[0051] (Experimental Example 7) Next, in order to investigate the relationship between the frequency bands in the frequency spectrum obtained by FFT analysis of the AE signal SAE and the rotational runout of the cylindrical workpiece W obtained by analyzing the signal in that frequency band, the inventors calculated the analytical values ​​of the rotational runout SW for each of cylindrical workpieces No. 1 to No. 8 and the differences between the analytical values ​​and the actual measured values, and compared them for each of four frequency bands (25 to 45 kHz, 45 to 75 kHz, 80 to 100 kHz, 100 to 120 kHz).

[0052] Figure 22 shows the amount of rotational runout obtained by changing the analysis frequency band for each of several cylindrical workpieces No. 1 to No. 8, and the difference between the amount of rotational runout and the actual measured value, and Figure 23 shows the difference between the amount of rotational runout and the actual measured value for each of the several cylindrical workpieces in Figure 22 in a line graph.

[0053] 22 and 23, the rotational runout SW for each of cylindrical workpieces No. 1 to No. 8 can be calculated with a difference from the actual measured value of 11 μm or less in any frequency band, so regeneration does not need to be limited to a specific frequency band. However, since vibrations in the high frequency band are generally more easily attenuated and have smaller intensity than vibrations in the low frequency band, it is desirable to analyze a frequency band in the range of 25 to 75 kHz when performing FFT analysis of the AE signal wave.

[0054] As described above, according to the rotational runout detection device of this embodiment (electronic control device 70, AE sensor 64), the AE signal, which is an elastic wave generated from the grinding point of the grinding wheel 42 on the cylindrical workpiece W, is utilized, and the rotational runout SW of the cylindrical workpiece W is detected by the runout calculation unit 82 based on the signal intensity integral value SPI within a predetermined frequency band of the AE signal. Therefore, the detection accuracy of the rotational runout SW is not affected by disturbances due to grinding fluid, etc., and there is no need to install or remove a sensor head, etc., near the cylindrical workpiece W.

[0055] In this embodiment, the runout detection unit 76 includes a moving standard deviation calculation unit 78 that calculates a moving standard deviation MSD of the signal intensity integral value SPI sequentially calculated by the signal intensity integral calculation unit 74, and calculates the rotational runout amount SW of the cylindrical workpiece W based on the moving standard deviation MSD of the signal intensity integral value SPI calculated by the moving standard deviation calculation unit 78. In this way, by using the moving standard deviation MSD, the calculation accuracy of the rotational runout amount SW of the cylindrical workpiece W is improved without being affected by local fluctuations in the signal intensity integral value SPI.

[0056] In this embodiment, the runout detection unit 76 includes a runout time calculation unit 80 that calculates a runout time Tf, which is the time when the moving standard deviation MSD of the signal intensity integral value SPI calculated by the moving standard deviation calculation unit 78 exceeds a preset runout time judgment threshold value JD, and a runout calculation unit 82 that calculates the multiplication value of the runout time Tf and the cutting speed CV of the grinding wheel as the rotational runout amount SW of the cylindrical workpiece W. In this manner, the runout amount of the cylindrical workpiece W is obtained based on the actual runout time Tf and the actual cutting speed CV value of the grinding wheel 42 from the multiplication equation of the runout time Tf when the moving standard deviation MSD of the signal intensity integral value SPI exceeds the preset runout time judgment threshold value JD and the cutting speed CV of the grinding wheel 42, which is a preset relationship. Therefore, the detection accuracy of the rotational runout amount SW is not affected by disturbances due to grinding fluid, etc.

[0057] In this embodiment, the preset shake time determination threshold value JD is 50% of the maximum value of the moving standard deviation MSD, thereby enabling the shake time to be determined accurately.

[0058] In this embodiment, the signal intensity integral calculation unit 74 sequentially calculates the signal intensity integral SPI within a predetermined frequency range, for example, a predetermined frequency band of 45 to 75 kHz, in which the signal intensity peaks are concentrated in the frequency spectrum of the AE signal. This allows the AE signal to be effectively reflected in the signal intensity integral SPI.

[0059] In this embodiment, the data length of the AE signal subjected to frequency analysis by the frequency analysis unit 72 is equal to or shorter than 0.13 rotations of the cylindrical workpiece W. This makes it possible to sufficiently reduce the detection error of the rotational runout amount SW of the cylindrical workpiece W.

[0060] In this embodiment, the moving standard deviation calculation unit 78 calculates the moving standard deviation MSD of the signal intensity integral value SPI by using a moving average time longer than one rotation time of the cylindrical workpiece W. This makes it possible to stably obtain the moving standard deviation MSD regardless of pulsation of the signal intensity integral value SPI, thereby improving the detection accuracy of the rotational runout amount SW.

[0061] Although one embodiment of the present invention has been described above with reference to the drawings, the present invention can be applied to other embodiments.

[0062] For example, in the above-described embodiment, the AE sensor 64 is provided on the mounting flange 60, but it may be provided on the fixed flange 18. Furthermore, if the grinding wheel 42 has a base metal, the AE sensor 64 may be provided within the base metal.

[0063] In the above-described embodiment, a storage battery was disposed within the annular groove 62 of the mounting flange 60, but instead of this, a contactless power supply device having a fixed-position power supply coil and a power receiving coil magnetically coupled to the same, a contactless power supply device having a solar cell that generates electricity upon receiving light from a floodlight device, or a contactless power supply device having an antenna device that generates electricity upon receiving microwaves may be provided.

[0064] It should be noted that the above is merely one embodiment of the present invention, and various modifications can be made to the present invention without departing from the spirit and scope of the present invention. [Explanation of symbols]

[0065] 12: Cylindrical grinding machine: 14: Bed, 14a: X-direction bed section, 14b: Y-direction bed section, 16: X-direction table, 18: X-direction guide rail, 20: Spindle, 22: Spindle motor, 24: Spindle housing, 26: Tailstock center, 28: Tailstock housing, 30: Guide rail, 32: Spindle center, 34: X-direction drive motor, 36: X-direction ball screw, 38: Y-direction table, 40: Y-direction guide rail, 42: Grinding wheel, 44: Grinding wheel shaft, 46: Grinding wheel shaft housing, 48: Belt, 50: Grinding wheel drive motor, 52: Y-direction drive motor, 54: Y-direction ball screw, 5 6: fixing flange, 58: fixing bolt, 60: mounting flange, 62: annular groove, 64: AE sensor (device for detecting rotational runout of cylindrical workpiece), 66: lid member, 68: receiving circuit, 70: electronic control device (device for detecting rotational runout of cylindrical workpiece), 72: frequency analysis section, 74: signal strength integral value calculation section, 76: runout amount detection section, 78: moving standard deviation calculation section, 80: runout time calculation section, 82: runout amount calculation section, 84: display, W: cylindrical workpiece, SPI: signal strength integral value, MSD: moving standard deviation, Tf: runout time, JD: runout time judgment threshold, SW: rotational runout amount, CV: cutting speed

Claims

1. A device for detecting a rotational runout of a cylindrical workpiece, the device detecting a rotational runout of the cylindrical workpiece by utilizing an AE signal, which is an elastic wave generated from a grinding point of a grinding wheel on the cylindrical workpiece, A frequency analysis unit that sequentially analyzes the frequency of the AE signal; a signal intensity integral value calculation unit that sequentially calculates an integral value of a signal intensity within a predetermined frequency band set in advance from the frequency spectrum obtained by the frequency analysis unit; and a runout amount detection unit that detects a rotation runout amount of the cylindrical workpiece based on the integral value of the signal strength sequentially calculated by the signal strength integral value calculation unit from a preset relationship. A rotational runout detection device for a cylindrical workpiece.

2. The signal strength integral value calculated by the signal strength integral value calculation unit is an integral value of signal strength within a preset frequency region of the frequency spectrum of the AE signal frequency-analyzed by the frequency analysis unit.

2. The apparatus for detecting runout of a cylindrical workpiece according to claim 1.

3. The data length of the AE signal that is frequency-analyzed by the frequency analysis unit is equal to or shorter than 0.13 revolutions of the cylindrical workpiece.

2. The apparatus for detecting runout of a cylindrical workpiece according to claim 1.

4. An integral analysis pitch, which is a plot interval of the integral value of the signal strength calculated by the signal strength integral value calculation unit, is equal to or less than 1 / 2 of a rotation period T of the cylindrical workpiece.

3. The apparatus for detecting the amount of rotational runout of a cylindrical workpiece according to claim 2.

5. The runout detection unit includes a moving standard deviation calculation unit that calculates a moving standard deviation of the integral value of the signal strength sequentially calculated by the signal strength integral value calculation unit, and calculates a rotational runout amount of the cylindrical workpiece based on the moving standard deviation of the integral value of the signal strength calculated by the moving standard deviation calculation unit.

3. The apparatus for detecting the amount of rotational runout of a cylindrical workpiece according to claim 2.

6. The moving standard deviation calculation unit calculates the moving standard deviation of the signal intensity integral value using a moving average time longer than one rotation time of the cylindrical workpiece.

6. An apparatus for detecting a rotational runout of a cylindrical workpiece according to claim 5.

7. The runout amount detection unit includes a runout time calculation unit that calculates a runout time, which is a time when the moving standard deviation of the signal intensity calculated by the moving standard deviation calculation unit exceeds a preset runout time determination threshold value, and a runout amount calculation unit that calculates a multiplication value of the runout time and a cutting speed of the grinding wheel as a rotational runout amount of the cylindrical workpiece.

6. An apparatus for detecting a rotational runout of a cylindrical workpiece according to claim 5.

8. The preset shake time determination threshold is set as a ratio to the maximum value of the moving standard deviation.

8. An apparatus for detecting a rotational runout of a cylindrical workpiece according to claim 7.

9. The preset shake time determination threshold is a value within a range of 40 to 80% of the maximum value of the moving standard deviation.

9. The apparatus for detecting the amount of rotational runout of a cylindrical workpiece according to claim 8.

Images