Inverter noise removal device, moving mechanism, motor device, inverter noise removal method, and program

Abstract

To easily and properly attenuate an inverter noise.SOLUTION: An inverter noise removal device sequentially receives a measured value of acceleration from an acceleration sensor for measuring the acceleration of a shaft of an inverter-driven motor, a movement member moved by the motor, or a support member for supporting the shaft or the movement member. The inverter noise removal device comprises: a generation part that generates acceleration data in a first time domain from the received measured value of the acceleration; a Fourier transformation part that generates the acceleration data in a first frequency domain by Fourier-transforming the acceleration data in the first time domain; an equalization part that generates the acceleration data in a second frequency domain having the equal acceleration values across all frequencies within the data, by performing equalization-processing including dividing each acceleration value of the acceleration data in the first frequency domain by the acceleration value; and an inverse Fourier transformation part that generates the acceleration data in a second time domain by performing inverse Fourier transformation of the acceleration data in the second frequency domain.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to an inverter noise removal device, a movement mechanism, an inverter noise removal method, and a program. [Background technology]

[0002] Conventionally, inverters have been used to drive rotary motors. Generally, PWM (pulse width modulation) is used to control motor rotation using an inverter. The current generated by a PWM inverter pulsates periodically at the PWM switching frequency (carrier frequency), so a rotary motor supplied with this current generates an electromagnetic force pulsating at the carrier frequency. Therefore, the rotary motor vibrates at the carrier frequency, its harmonic frequencies, and their sideband frequencies. The vibrations of the rotary motor are transmitted to the moving member that is moved by the rotary motor (that is, rotated or moved linearly by the driving force of the rotary motor).

[0003] Meanwhile, technologies have been developed to diagnose damage to the rotating shaft of a rotary motor, moving parts moved by the rotary motor, or parts that support the rotating shaft or this moving part. For example, technology to diagnose damage to a bearing that supports a rotating shaft or a rotating shaft that is linked to the rotating shaft is based on the periodic shock, i.e., vibration, that occurs with rotation. Analysis of acceleration data from a damaged bearing reveals periodic shocks that depend on the rotation speed of the rotary motor. The same is true for moving parts other than gears or support parts that support this moving part.

[0004] However, if the vibrations of the rotary motor generated by the PWM inverter are added to the vibrations of the rotating shaft and moving parts, it may be impossible to perform an appropriate damage diagnosis. For example, if the PWM carrier frequency, the harmonic frequencies of the carrier frequency, or their sideband frequencies match the moving frequency of the moving parts (e.g., the rotation frequency of the rotating shaft, i.e., the bearing damage frequency), large vibrations will be generated at the moving frequency of the rotating shaft and moving parts, and a normal bearing may be judged to be abnormal. In this application, the vibration component of the rotary motor generated by the PWM inverter is referred to as "inverter noise."

[0005] For proper damage diagnosis, it is desirable to attenuate inverter noise in acceleration data. Various techniques have been proposed to remove inverter noise from acceleration, as disclosed in Patent Documents 1 and 2. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-116251 [Patent Document 2] Japanese Patent Application Laid-Open No. 2011-259624 Summary of the Invention [Problem to be solved by the invention]

[0007] Existing inverter noise reduction technologies require inputting the carrier frequency and / or sideband frequencies into the noise reduction device before performing the reduction. However, the carrier frequency and / or sideband frequencies may be unknown or may fluctuate.

[0008] Furthermore, existing inverter noise removal techniques require a large amount of computational processing, which is time-consuming. Furthermore, some techniques require extensive know-how to set thresholds for inverter noise removal. Furthermore, when reducing acceleration values ​​at the carrier frequency, its harmonic frequencies, and their sideband frequencies, if the moving frequency of the rotating shaft or moving member nearly coincides with any of these frequencies, the acceleration value at the moving frequency of the rotating shaft or moving member will also be reduced, making it impossible to properly diagnose damage.

[0009] Therefore, the present invention provides a technique that can easily and appropriately attenuate inverter noise. [Means for solving the problem]

[0010] According to one aspect of the present invention, there is provided an inverter noise elimination device comprising: a receiving unit that sequentially receives, from an acceleration sensor that measures the acceleration of a rotating shaft of a rotary motor driven by an inverter, a moving member moved by the rotary motor, or a support member that movably supports the rotating shaft or the moving member, acceleration measurements of the rotating shaft, the moving member, or the support member resulting from vibration of the rotating shaft, the moving member, or the support member; a generating unit that generates first time-domain acceleration data from the acceleration measurements received by the receiving unit; a Fourier transform unit that performs a Fourier transform on the first time-domain acceleration data generated by the generating unit to generate first frequency-domain acceleration data; an equalization unit that performs an equalization process, including dividing each acceleration value of the first frequency-domain acceleration data generated by the Fourier transform unit by the acceleration value, to generate second frequency-domain acceleration data having uniform acceleration values ​​across all frequencies in the data; and an inverse Fourier transform unit that performs an inverse Fourier transform on the second frequency-domain acceleration data generated by the equalization unit to generate second time-domain acceleration data. [Effects of the Invention]

[0011] In one aspect of the present invention, second frequency-domain acceleration data is generated by performing an equalization process, which includes dividing each acceleration value of first frequency-domain acceleration data generated by Fourier transforming first time-domain acceleration data by the acceleration value. The equalization process attenuates the highly periodic vibration components of inverter noise, while barely attenuating the vibration components of the rotating shaft, moving member, or support member itself. This is because the rotation of the rotary motor causes the rotating shaft or moving member to vibrate periodically, but the vibration period of the rotating shaft, moving member, or support member is less stable than the period of the inverter noise. The equalization process sets the amplitude of each term in the Fourier series related to the vibration component of the inverter noise to 1. In the second time-domain acceleration data generated by performing an inverse Fourier transform on the second frequency-domain acceleration data generated by the equalization process, the vibration component of the inverter noise is attenuated compared to the first time-domain acceleration data, while the vibration component of the rotating shaft, moving member, or support member remains almost intact. As a result, the vibration component of the rotating shaft, moving member, or support member is emphasized. In this way, inverter noise can be easily and appropriately attenuated by performing Fourier transform, equalization, and inverse Fourier transform. There is no need to input the carrier frequency and / or sideband frequency for inverter noise removal, and no threshold setting is required for inverter noise removal processing. This also reduces the amount of calculation processing required. Furthermore, even if the carrier frequency and / or sideband frequency fluctuates, inverter noise can be appropriately attenuated. Furthermore, even if the moving frequency of the rotating shaft or moving member is approximately equal to the carrier frequency, the harmonic frequency of the carrier frequency, or any of these sideband frequencies, the vibration period of the rotating shaft, moving member, or support member is unstable compared to the period of the inverter noise, so only the inverter noise can be appropriately attenuated. [Brief explanation of the drawings]

[0012] [Figure 1]1 is a diagram showing a moving mechanism provided with an inverter noise removal device according to an embodiment of the present invention; [Figure 2] 2 is a flowchart showing an example of the operation of a processor of the inverter noise removal device of FIG. 1. [Figure 3] 2 is a graph of time domain acceleration corresponding to a first time domain acceleration data set generated from acceleration measurements received by the inverter noise removal device of FIG. 1; [Figure 4] 1 is a graph of frequency domain acceleration corresponding to a first frequency domain acceleration data set generated by Fourier transforming the first time domain acceleration data set. [Figure 5] 1 is a graph of acceleration in the frequency domain corresponding to a second frequency domain acceleration data set generated by an equalization process from a first frequency domain acceleration data set obtained by a Fourier transform. [Figure 6] 10 is a graph of time domain acceleration corresponding to a second time domain acceleration data set generated by an inverse Fourier transform of the second frequency domain acceleration data set obtained by the equalization process. [Figure 7] 10 is a flowchart showing a modified example of the operation of the processor of the inverter noise removal device of FIG. [Figure 8] 10 is a flowchart showing another modified example of the operation of the processor of the inverter noise removing device of FIG. [Figure 9] 10 is a graph of frequency domain acceleration corresponding to a third frequency domain acceleration data set generated by enveloping and Fourier transforming the second time domain acceleration data set generated by an inverse Fourier transform. [Figure 10] 1 is a graph of frequency domain acceleration corresponding to a frequency domain acceleration data set generated by enveloping and Fourier transforming a time domain acceleration data set generated by inverse Fourier transforming a first frequency domain acceleration data set without equalization. [Figure 11] 10A and 10B are diagrams illustrating a part of a movement mechanism according to a modified example of the embodiment of the present invention. [Figure 12] FIG. 10 is a diagram showing a part of a movement mechanism according to another modified example of the embodiment of the present invention. [Figure 13] FIG. 10 is a diagram illustrating a motor device according to a modified example of an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0013] Various embodiments of the present invention will now be described with reference to the accompanying drawings, in which the drawings are not necessarily to scale and some features may be exaggerated or omitted.

[0014] 1, the moving mechanism provided with the inverter noise removal device according to the embodiment of the present invention is a gear reducer. The gear reducer includes a rotary motor 1, gears 2 and 3, a bearing 4, a housing 5, a ring 6, an acceleration sensor 7, a PWM inverter 8, and an inverter noise removal device 10.

[0015] The rotary motor 1 is supplied with a current generated by a PWM inverter 8. The gear 2 is fixed to a rotary shaft 2A connected to a rotary shaft 1A of the rotary motor 1, and rotates together with the rotary shafts 1A and 2A.

[0016] The rotating shaft 2A is rotatably supported by a bearing (support member) 4 and other bearings (not shown). The bearing 4 has an inner ring 4A and an outer ring 4B, with the inner ring 4A fixed to the end of the rotating shaft 2A and the outer ring 4B fixed to a housing 5 of the gear reducer. The illustrated bearing 4 is a ball bearing, but it may be another type of bearing.

[0017] Gear 3 is in mesh with gear 2, and a rotating shaft 3A to which gear 3 is fixed is supported by a bearing (not shown).

[0018] The ring 6 is fixed to the outer ring 4B of the bearing 4 and does not contact the rotating shafts 2A and 3A. An acceleration sensor 7 that measures the acceleration of the bearing 4 is attached to the ring 6. The acceleration sensor 7 is, for example, a MEMS (Micro Electro Mechanical Systems) acceleration sensor. The acceleration sensor 7 may be a three-axis acceleration sensor that measures acceleration in each of three axes (X-axis, Y-axis, and Z-axis).

[0019] As the rotating shaft 2A rotates, the inner ring 4A of the bearing 4 is subjected to vibrations at the rotational frequency of the rotating shaft 2A. The acceleration sensor 7 attached to the ring 6 fixed to the outer ring 4B of the bearing 4 also vibrates at the rotational frequency of the rotating shaft 2A. The acceleration sensor 7 measures the acceleration caused by the vibration of the bearing 4. If the bearing 4 is damaged, periodic impacts caused by partial flaking of the bearing 4 will be applied to the acceleration sensor 7, depending on the rotational frequency of the rotary motor 1.

[0020] Rotary motor 1 is driven by PWM inverter 8, and therefore vibrates at the PWM carrier frequency, harmonic frequencies of the carrier frequency, and their sideband frequencies. The vibrations of rotary motor 1 are transmitted to rotating shaft 2A, which is rotated by rotary motor 1, and ultimately to inner ring 4A of bearing 4. Acceleration sensor 7, which is attached to ring 6 fixed to outer ring 4B of bearing 4, is also affected by rotary motor 1 and vibrates.

[0021] The inverter noise elimination device 10 is used as a device for diagnosing damage to the bearing 4, and is capable of eliminating inverter noise from acceleration data measured by the acceleration sensor 7. The inverter noise elimination device 10 includes a computer 11, an input device 12, and a display 13.

[0022] The computer 11 may be, for example, a personal computer, and includes a processor 14, an acceleration interface 15, an image interface 16, a read only memory (ROM) 17, and a random access memory (RAM) 18.

[0023] The processor 14 is configured by, for example, one or more CPUs (central processing units) or MPUs (microprocessor units). The ROM 17 is a non-volatile memory that stores various computer programs for the operation of the processor 14. The RAM 18 is a volatile memory that is used as a temporary storage area such as the main memory or work area of ​​the processor 14. The processor 14 controls each part of the inverter noise removal device 10 in accordance with the computer programs stored in the ROM 17, using the RAM 18 as a work memory.

[0024] The input device 12 is configured by, for example, a keyboard and a mouse. A user of the inverter noise removal device 10 can use the input device 12 to give instructions to the computer 11.

[0025] The display 13 is connected to the processor 14 via an image interface 16. The display 13 displays various images (for example, images required for diagnosing bearing damage) generated by processing by the processor 14. The display 13 may be a touch panel, in which case the keyboard of the input device 12 can be omitted.

[0026] The acceleration interface 15 is connected to the acceleration sensor 7 by wire or wirelessly, and is also connected by wire to the processor 14. The acceleration interface 15 supplies the acceleration data supplied from the acceleration sensor 7 to the processor 14.

[0027] 2 is a flowchart showing an example of the operation of the processor 14. This operation is executed after the rotary motor 1 starts to be driven. The following describes acceleration about one of the three axes (X-axis, Y-axis, and Z-axis), but if the acceleration sensor 7 is a three-axis acceleration sensor, the operation of FIG. 2 is executed in parallel for acceleration about each axis. The operations of FIGS. 7 and 8, which will be described later, are also executed in parallel for acceleration about each axis if the acceleration sensor 7 is a three-axis acceleration sensor.

[0028] When the rotary motor 1 rotates, acceleration data is periodically supplied from the acceleration sensor 7 to the processor 14 via the acceleration interface 15. The acceleration data includes a measurement time and an acceleration value (measured value of acceleration).

[0029] In step S1, the processor 14 functions as a receiver and sequentially receives acceleration data from the acceleration sensor 7. Also in step S1, the processor 14 functions as a generator and accumulates the acceleration data to generate a first acceleration data set that indicates changes in acceleration values ​​over a certain period of time. Therefore, the first acceleration data set is first time-domain acceleration data related to acceleration caused by vibration of the bearing 4, and includes vibration components of the bearing 4 and vibration components of inverter noise.

[0030] FIG. 3 is a graph of time domain acceleration corresponding to the first time domain acceleration data set generated in step S1, and is an example of actual measurement data.

[0031] In step S2, the processor 14 functions as a Fourier transform unit and performs a Fourier transform on the first time-domain acceleration data to generate first frequency-domain acceleration data. The Fourier transform used in step S2 is a fast Fourier transform (FFT), but may be another Fourier transform such as a discrete Fourier transform.

[0032] FIG. 4 is a graph of acceleration in the frequency domain corresponding to the first frequency domain acceleration data set generated by the Fourier transform in step S2, and is an example of actual measurement data.

[0033] In step S3, the processor 14 functions as an equalization unit and performs an equalization process including dividing each acceleration value of the acceleration data in the first frequency domain generated by the Fourier transform by the acceleration value, thereby generating acceleration data in the second frequency domain.

[0034] In one example of the equalization process of step S3, each acceleration value in the first frequency domain acceleration data is simply divided by that acceleration value, so that the acceleration data in the second frequency domain has a uniform acceleration value of "1" across all frequencies in the data.

[0035] FIG. 5 is a graph of frequency-domain acceleration corresponding to the second frequency-domain acceleration data set generated by the equalization process in step S3, and is an example of actual measurement data. As is clear from FIG. 5, the second frequency-domain acceleration data has a uniform acceleration value of "1" across all frequencies within the data. As a result, high values ​​in the first frequency-domain acceleration data are reduced, and low values ​​in the first frequency-domain acceleration data are increased. In other words, inverter noise vibration components with high acceleration values ​​in the frequency spectrum are attenuated relative to bearing 4 vibration components with low acceleration values. Conversely, bearing 4 vibration components with low acceleration values ​​in the frequency spectrum are increased (emphasized) relative to inverter noise vibration components with high acceleration values. In this way, the equalization process does not eliminate the inverter noise and the vibration components of bearing 4, but only changes the magnitude of the acceleration relatively, so it is important to note that when the acceleration data is converted back into a time waveform by the inverse Fourier transform described below, it does not result in a featureless, flat acceleration value waveform (the terms in the Fourier series corresponding to the inverter noise and the vibration components of bearing 4 are not eliminated, but rather the amplitude of each term is simply set to "1" to make them equal).

[0036] Another example of the equalization process of step S3 may involve dividing each acceleration value in the first frequency domain acceleration data by that acceleration value and multiplying each division result by a positive number a. In this case, the second frequency domain acceleration data has a uniform acceleration value "a" across all frequencies in the data. As a result, high values ​​in the first frequency domain acceleration data are reduced and low values ​​in the first frequency domain acceleration data are increased.

[0037] In step S4, the processor 14 functions as an inverse Fourier transform unit and performs an inverse Fourier transform on the second frequency domain acceleration data generated by the equalization process to generate second time domain acceleration data. The inverse Fourier transform in step S4 uses the phase information obtained by the Fourier transform in step S2. The inverse Fourier transform used in step S4 is an inverse FFT, but may be another type of inverse Fourier transform depending on the type of Fourier transform used in step S2.

[0038] FIG. 6 is a graph of acceleration in the time domain corresponding to the second time domain acceleration data set generated by the inverse Fourier transform in step S4, and is an example of actual measurement data.

[0039] As is clear from a comparison of Fig. 3 and Fig. 6, vibration components that have large amplitudes in Fig. 3 have small amplitudes in Fig. 6. The reduced vibration components are vibration components of inverter noise, and the graph in Fig. 6 mainly represents acceleration related to the vibration components of bearing 4. This will become clearer from the comparison of Fig. 9 and Fig. 10 described later.

[0040] In step S5, the processor 14 displays a graph of the time domain acceleration (illustrated in FIG. 6) corresponding to the second time domain acceleration data set generated by the inverse Fourier transform in step S4 on the display 13. A user of the inverter noise removal device 10 can visually check the graph to determine whether or not the bearing 4 is damaged.

[0041] Thus, in this embodiment, second frequency domain acceleration data is generated by performing an equalization process, which includes dividing each acceleration value of the first frequency domain acceleration data generated by Fourier transforming the first time domain acceleration data by the acceleration value. The equalization process attenuates the highly periodic vibration component of inverter noise, while barely attenuating the vibration component of the bearing 4 itself. This is because the rotating shaft 2A moved by the rotary motor 1 and the bearing 4 supporting it vibrate periodically, but the vibration period of the bearing 4 is less stable than the period of the inverter noise. The equalization process sets the amplitude of each term in the Fourier series related to the vibration component of the inverter noise to 1. In the second time domain acceleration data generated by performing an inverse Fourier transform on the second frequency domain acceleration data generated by equalization, the vibration component of the inverter noise is attenuated compared to the first time domain acceleration data, while the vibration component of the bearing 4 remains almost unchanged. As a result, the vibration component of the bearing 4 is emphasized.

[0042] In this way, inverter noise can be easily and appropriately attenuated simply by performing Fourier transform, equalization, and inverse Fourier transform. There is no need to input the carrier frequency and / or sideband frequency for inverter noise removal, and no threshold setting is required for inverter noise removal processing. This also reduces the amount of calculation required. Furthermore, even if the carrier frequency and / or sideband frequency fluctuates, inverter noise can be appropriately attenuated. Furthermore, even if the rotational frequency of the rotating shaft 2A approximately matches the carrier frequency, the harmonic frequency of the carrier frequency, or any of these sideband frequencies, the vibration period of the bearing 4 supporting the rotating shaft 2A is unstable compared to the inverter noise period, so only the inverter noise can be appropriately attenuated.

[0043] Figure 7 is a flowchart showing a modified example of the operation of processor 14. The same reference numerals are used to indicate steps common to Figure 2, and a description of those steps will be omitted.

[0044] 7 proceeds to step S6 after step S4. In step S6, processor 14 functions as an RMS calculation unit and calculates the root mean square (RMS) of all acceleration values ​​in the second time-domain acceleration data set generated by the inverse Fourier transform in step S4.

[0045] In step S7, the processor 14 displays the calculated RMS on the display 13. A user of the inverter noise removal device 10 can visually check the RMS and determine whether or not there is damage to the bearing 4. Alternatively or in addition to this, the processor 14 may compare the RMS with a threshold value and display the comparison result on the display 13. A user of the inverter noise removal device 10 can visually check the comparison result and determine whether or not there is damage to the bearing 4.

[0046] Instead of calculating the RMS, the processor 14 may select another index (e.g., maximum amplitude, crest factor, kurtosis) of the second time-domain acceleration data set generated by the inverse Fourier transform of step S4. By displaying the other index (e.g., maximum amplitude, crest factor, kurtosis) on the display 13, a user of the inverter noise removal device 10 can visually check the other index (e.g., maximum amplitude, crest factor, kurtosis) to determine whether or not the bearing 4 is damaged. Alternatively or in addition, the processor 14 may compare the other index (e.g., maximum amplitude, crest factor, kurtosis) with a threshold value and display the comparison result on the display 13. A user of the inverter noise removal device 10 can visually check the comparison result to determine whether or not the bearing 4 is damaged.

[0047] Additionally, in step S7, processor 14 may cause display 13 to display a time domain acceleration graph (illustrated in FIG. 6) corresponding to the second time domain acceleration data set generated by the inverse Fourier transform of step S4.

[0048] Figure 8 is a flow chart showing another variation of the operation of processor 14. The same reference numerals are used to indicate steps common to Figure 2, and a description of those steps will be omitted.

[0049] The operation of FIG. 8 proceeds to step S8 after step S4. In step S8, processor 14 functions as an envelope processor and performs envelope processing on the second-time-domain acceleration data generated by the inverse Fourier transform of step S4 to generate third-time-domain acceleration data. In the envelope processing, processor 14 functions as an absolute value converter and converts each acceleration value of the second-time-domain acceleration data generated by the inverse Fourier transform of step S4 into an absolute value. That is, negative acceleration values ​​are converted to positive values ​​while leaving their absolute values ​​unchanged, and positive acceleration values ​​are left unchanged. In this way, the envelope processing includes absolute value conversion processing.

[0050] In step S9, the processor 14 again functions as a Fourier transform unit, and performs a Fourier transform on the third time domain acceleration data generated in the envelope processing in step S8 to generate third frequency domain acceleration data.

[0051] FIG. 9 is a graph of acceleration in the frequency domain corresponding to the third frequency domain acceleration data set generated by the Fourier transform in step S9, and is an example of actual measurement data.

[0052] In step S10, the processor 14 causes the display 13 to display a graph of acceleration in the frequency domain (as shown in FIG. 9) corresponding to the third frequency domain acceleration data set generated by the Fourier transform in step S9. A user of the inverter noise removal device 10 can visually check the graph to determine whether or not there is damage to the bearing 4. Furthermore, in step S10, the processor 14 may cause the display 13 to display a graph of acceleration in the time domain (as shown in FIG. 6) corresponding to the second time domain acceleration data set generated by the inverse Fourier transform in step S4.

[0053] The calculation of RMS (step S6) in FIG. 7 may be added to the operation in FIG.

[0054] FIG. 10 is a graph of frequency-domain acceleration corresponding to a frequency-domain acceleration data set generated by enveloping (S8) and Fourier transforming (S9) the time-domain acceleration data set generated by inverse Fourier transform (S4) without equalization (step S3) from the first frequency-domain acceleration data set (corresponding to FIG. 4) generated by Fourier transform in step S2, and is an example of actual measurement data. That is, the difference between FIG. 9 and FIG. 10 is that the data set of FIG. 9 generated by the operation of FIG. 8 was generated through equalization (step S3), whereas the data set of FIG. 10 according to the comparative example was not subjected to equalization. While both data sets are derived from the same data set of FIG. 3, the data set of FIG. 9 is significantly different from the data set of FIG. 10.

[0055] The graph in Figure 10 shows peaks at multiples of approximately 20 Hz. These peaks correspond to the vibration component of inverter noise with a carrier frequency of approximately 20 kHz.

[0056] On the other hand, the graph in Figure 9 shows peaks at multiples of approximately 25 Hz. These peaks correspond to the vibration components of bearing 4 that originate at the characteristic frequency (approximately 25 Hz) when flaking occurs in bearing 4. In this way, by performing equalization processing, the vibration components of the inverter noise are attenuated and the vibration components of bearing 4 are emphasized. In the time domain graph in Figure 6, the vibration components of the inverter noise are also attenuated and the vibration components of bearing 4 are emphasized. Therefore, by visually inspecting the graphs in Figure 6 or Figure 9 or by performing arithmetic processing (for example, RMS calculation or envelope processing), it is possible to determine whether or not bearing 4 is damaged.

[0057] The graph in Figure 10 shows peaks at multiples of approximately 20 Hz, which corresponds to the carrier frequency, while the graph in Figure 9 shows peaks at multiples of approximately 25 Hz, which corresponds to the characteristic frequency when flaking occurs in bearing 4. However, if the carrier frequency and the characteristic frequency match, as shown in Figure 10, there is a risk of incorrectly diagnosing bearing damage when the acceleration amplitude is large. However, by performing equalization processing, the vibration components of the inverter noise are attenuated. The graph in Figure 9, in which the bearing vibration components are emphasized, shows that when the acceleration amplitude is large, damage to the bearing is definitely present.

[0058] In this way, inverter noise can be easily and appropriately attenuated by simply performing Fourier transform, equalization, inverse Fourier transform, envelope processing, and a second Fourier transform. There is no need to input the carrier frequency and / or sideband frequency for inverter noise removal, and no threshold setting is required for inverter noise removal. This also reduces the amount of calculation required. Furthermore, even if the carrier frequency and / or sideband frequency fluctuates, inverter noise can be appropriately attenuated. Furthermore, even if the rotational frequency of the rotating shaft 2A approximately matches the carrier frequency, the harmonic frequency of the carrier frequency, or any of these sideband frequencies, the vibration period of the bearing 4 supporting the rotating shaft 2A is unstable compared to the inverter noise period, so only the inverter noise can be appropriately attenuated.

[0059] Although not shown, an angle sensor for measuring the angular position of the rotating shaft 1A may be provided, for example, near the bearing 4. In this case, the processor 14 may determine the angular position of the rotating shaft 1A at each time based on the measurement results of the angle sensor, and identify the damaged position of the bearing 4 based on the acceleration and angular position at each time.

[0060] The inverter noise removal device 10 according to the embodiment is used as a device for diagnosing damage to a bearing 4, but may also be used as a device for diagnosing damage to a gear (moving member) 2. In this case, the processor 14 may determine the angular position of the gear at each time based on the measurement results of an angle sensor that measures the angular position of the rotating shaft 1A, and identify a damaged tooth of the gear based on the acceleration and angular position at each time. Such a gear diagnosis algorithm may be the Per-tooth method, which calculates an acceleration index for each tooth based on the acceleration at each time.

[0061] 11 shows a part of a movement mechanism according to a modified embodiment of the present invention, in which the inverter noise removal device 10 is omitted.

[0062] The movement mechanism shown in Fig. 11 is a ball screw mechanism. The ball screw mechanism includes a rotary motor 21, a screw shaft 22, a linear motion nut 23, a bearing 24, a housing 25, a ring 26, an acceleration sensor 7, a PWM inverter 28, and an inverter noise eliminator (not shown).

[0063] A current generated by a PWM inverter 28 is supplied to the rotary motor 21. The screw shaft (moving member) 22 is connected to a rotary shaft 21A of the rotary motor 21, and rotates together with the rotary shaft 21A.

[0064] The screw shaft 22 is rotatably supported by a bearing (support member) 24 and another bearing (not shown). The bearing 24 has an inner ring 24A and an outer ring 24B, with the inner ring 24A fixed to the end of the screw shaft 22 and the outer ring 24B fixed to a housing 25 of the bearing 24. The illustrated bearing 24 is a ball bearing, but may be another type of bearing.

[0065] A plurality of balls 23A are arranged inside the linear motion nut (moving member) 23, and the balls 23A are arranged in the screw grooves of the linear motion nut 23 and the screw grooves of the screw shaft 22, so that the linear motion nut 23 meshes with the screw shaft 22. Therefore, when the screw shaft 22 rotates, the linear motion nut 23 moves linearly along the longitudinal direction of the screw shaft 22.

[0066] The ring 26 is fixed to the outer ring 24B of the bearing 24. An acceleration sensor 7 that measures the acceleration of the screw shaft 22 is attached to the ring 26. The acceleration sensor 7 may be the same as the acceleration sensor 7 described above.

[0067] As the screw shaft 22 rotates, the inner ring 24A of the bearing 24 is subjected to vibrations at the rotational frequency of the screw shaft 22. The acceleration sensor 7, which is attached to the ring 26 fixed to the outer ring 24B of the bearing 24, also vibrates at the rotational frequency of the screw shaft 22. The acceleration sensor 7 measures the acceleration caused by the vibration of the screw shaft 22. If the screw shaft 22, the bearing 24, or the linear motion nut 23 is damaged, the acceleration sensor 7 will be subjected to periodic impacts that depend on the rotational frequency of the rotary motor 21.

[0068] Rotary motor 21 is driven by PWM inverter 28, and therefore vibrates at the PWM carrier frequency, harmonic frequencies of the carrier frequency, and their sideband frequencies. The vibration of rotary motor 21 is transmitted to screw shaft 22, which is rotated by rotary motor 21, and ultimately to inner ring 24A of bearing 24. Acceleration sensor 7, which is attached to ring 6 fixed to outer ring 24B of bearing 24, is also affected by rotary motor 21 and vibrates.

[0069] The acceleration sensor 7 supplies the acceleration data to an inverter noise elimination device (not shown). The inverter noise elimination device may be the same as the inverter noise elimination device 10. The inverter noise elimination device is used as a device for diagnosing damage to the screw shaft 22, the bearing 24, or the linear motion nut 23, and is capable of removing inverter noise from the acceleration data measured by the acceleration sensor 7. The inverter noise can be removed, for example, by performing the operations shown in FIG. 2, FIG. 7, or FIG. 8.

[0070] In the example of FIG. 11, the acceleration sensor 7 is disposed on the ring 26, but the acceleration sensor 7 may also be disposed on the linear motion nut 23.

[0071] 12 shows a part of a movement mechanism according to a modified example of the embodiment of the present invention, in which the inverter noise removing device 10 is omitted.

[0072] 12 includes a linear guide 30. The linear guide 30 includes a rail 31, a carriage 32, caps 33 and 34, seals 35 and 36, and a table 37.

[0073] The rail 31 is linear, and grooves 31A are formed on both sides of the rail 31. The carriage 32 is movable along the longitudinal direction of the rail 31. Inside the carriage 32, a plurality of balls (not shown) are arranged in two rows, and the balls can circulate in each row. The balls roll within the grooves 31A of the rail 31.

[0074] Caps 33 and 34 are disposed on both ends of the carriage 32, and are further sealed with seals 35 and 36. A table 37 is fixed to the upper surface of the carriage 32. The table 37 is connected to, for example, the linear motion nut 23 of the ball screw mechanism shown in FIG. 11. Therefore, the carriage 32 moves linearly along the longitudinal direction of the rail 31 in accordance with the linear motion of the linear motion nut 23.

[0075] Therefore, the linear guide 30 itself is not provided with a drive source for moving the carriage 32, but the carriage 32 may be a moving member that is moved by a rotary motor (for example, the rotary motor 21 shown in FIG. 11). In this case, the rail 31 is a support member that movably supports the carriage 32.

[0076] An acceleration sensor 7 that measures the acceleration of the carriage 32 is attached to the carriage 32 or a member fixed thereto. The acceleration sensor 7 may be the same as the acceleration sensor 7 described above.

[0077] As the rotary motor rotates, the carriage 32 is subjected to vibrations at the rotational frequency of the rotary motor. For example, if the table 37 is connected to the linear nut 23 of the ball screw mechanism shown in FIG. 11, when the rotary motor 21 shown in FIG. 11 rotates, vibrations are transmitted from the screw shaft 22 via the linear nut 23 to the carriage 32, causing the carriage 32 to vibrate at the rotational frequency of the screw shaft 22. The acceleration sensor 7 measures the acceleration caused by the vibration of the carriage 32. If there is damage to the carriage 32 or the rail 31, periodic impacts that depend on the rotational frequency of the rotary motor 21 are applied to the acceleration sensor 7.

[0078] Rotary motor 21 is driven by PWM inverter 28, and therefore vibrates at the PWM carrier frequency, harmonic frequencies of the carrier frequency, and their sideband frequencies. The vibration of rotary motor 21 is transmitted to screw shaft 22, which is rotated by rotary motor 21, and ultimately to carriage 32. Acceleration sensor 7 provided on carriage 32 is also affected by rotary motor 21 and vibrates.

[0079] The acceleration sensor 7 supplies the acceleration data to an inverter noise elimination device (not shown). The inverter noise elimination device may be the same as the inverter noise elimination device 10. The inverter noise elimination device is used as a device for diagnosing damage to the carriage 32 or the rail 31, and is capable of removing inverter noise from the acceleration data measured by the acceleration sensor 7. The inverter noise can be removed, for example, by performing the operations shown in FIG. 2, FIG. 7, or FIG. 8.

[0080] In the example of FIG. 12, the acceleration sensor 7 is disposed on the carriage 32 or a member fixed thereto, but the acceleration sensor 7 may also be disposed on the rail 31.

[0081] The above describes a technique for attenuating inverter noise from acceleration data of the moving members (gear 2, screw shaft 22, linear motion nut 23, carriage 32) moved by a rotary motor or the support members (bearings 4, 24, rail 31) that movably support these moving members. However, the present invention also encompasses attenuating inverter noise from acceleration data of the rotating shaft of a rotary motor or the bearings that are support members that support this rotating shaft, as will be described later with reference to Figure 13.

[0082] 13 shows a motor device according to a modified embodiment of the present invention. This motor device includes a rotary motor 40, the inverter noise removal device 10 described above, and an acceleration sensor 7.

[0083] Rotary motor 40 has a housing 41, a rotating shaft (moving member) 42, and bearings (support members) 43, 44. Rotary motor 40 has a known core and stator, but their description will be omitted. Rotary shaft 42, on which a stator is provided, is disposed inside housing 41, and both ends of rotating shaft 42 protrude from housing 41. Bearings 43, 44 are fixed in holes at both ends of housing 41, and bearings 43, 44 rotatably support rotating shaft 42.

[0084] As shown, acceleration sensor 7 is attached to housing 41 and measures the acceleration of rotating shaft 42 and / or bearings 43, 44. Acceleration sensor 7 may be attached to a ring (not shown) that is fixed to the outer ring of bearing 43 or 44 and does not contact rotating shaft 42, in which case it measures the acceleration of bearing 43 or 44. Acceleration sensor 7 may be attached to a ring (not shown) that is located away from bearings 43, 44 and that slidably contacts rotating shaft 42, in which case it measures the acceleration of rotating shaft 42.

[0085] Since the rotary motor 40 is driven by the PWM inverter 8, the rotary shaft 42 and bearings 43, 44 vibrate at the PWM carrier frequency, harmonic frequencies of the carrier frequency, and their sideband frequencies.

[0086] The inverter noise removal device 10 is used as a device for diagnosing damage to the rotating shaft 42 and / or the bearings 43, 44, and is capable of removing inverter noise from acceleration data measured by the acceleration sensor 7. The inverter noise can be removed, for example, by performing the operations shown in FIG. 2, FIG. 7, or FIG. 8.

[0087] Although the present invention has been shown and described with reference to preferred embodiments thereof, it will be understood by those skilled in the art that changes may be made in form and detail without departing from the scope of the invention as set forth in the appended claims. Such changes, modifications and alterations are intended to be included within the scope of the invention.

[0088] For example, in the above-described embodiment and modified examples, the moving member moved by the rotary motor is a gear, a screw shaft of a ball screw, or a carriage of a linear guide, but the moving member to which the present invention is applied may be another moving member.

[0089] In the inverter noise removal device 10, each function executed by the processor 14 may be executed by hardware instead of a processor, or may be executed by a programmable logic device such as an FPGA (Field Programmable Gate Array) or a DSP (Digital Signal Processor). [Explanation of symbols]

[0090] 1,21 Rotary motor 2 Gears (moving parts) 4,24 Bearing (support member) 7 Acceleration Sensor 8,28 PWM inverter 10 Inverter noise eliminator 14 processors (receiving section, generating section, Fourier transform section, equalizing section, inverse Fourier transform section, envelope processing section) 22 Screw shaft (moving member) 23 Linear motion nut (moving member) 30 Linear guide 31 Rail (support member) 32 Carriage (moving member) 40 RPM motor 42 Rotating shaft (moving member) 43, 44 Bearing (support member)

Claims

1. a receiving unit that sequentially receives acceleration measurement values ​​from an acceleration sensor that measures the acceleration of a rotary shaft of a rotary motor driven by an inverter, a moving member moved by the rotary motor, or a support member that movably supports the rotary shaft or the moving member; a generating unit that generates acceleration data in a first time domain from the acceleration measurement values ​​received by the receiving unit; a Fourier transform unit that performs a Fourier transform on the acceleration data in the first time domain generated by the generation unit to generate acceleration data in a first frequency domain; an equalization unit that performs an equalization process including dividing each acceleration value of the acceleration data in the first frequency domain generated by the Fourier transform unit by the acceleration value, thereby generating acceleration data in the second frequency domain having uniform acceleration values ​​across all frequencies in the data; an inverse Fourier transform unit that performs an inverse Fourier transform on the acceleration data in the second frequency domain generated by the equalization unit to generate acceleration data in a second time domain; An inverter noise removal device comprising:

2. an envelope processing unit that performs envelope processing on the acceleration data in the second time domain generated by the inverse Fourier transform unit to generate acceleration data in a third time domain; The Fourier transform unit performs a Fourier transform on the acceleration data in the third time domain generated by the envelope processing unit to generate acceleration data in the third frequency domain.

2. The inverter noise removal device according to claim 1.

3. The inverter noise removal device according to claim 1 or 2; the rotary motor; the moving member that is moved by the rotary motor; the support member that movably supports the moving member; the acceleration sensor arranged to measure the acceleration of the moving member or the support member supporting the moving member; A moving mechanism comprising:

4. The inverter noise removal device according to claim 1 or 2; the rotary motor; the acceleration sensor is arranged to measure the acceleration of a rotary shaft of the rotary motor or a support member that supports the rotary shaft; A motor device comprising:

5. Sequentially receiving acceleration measurement values ​​from an acceleration sensor that measures the acceleration of a rotary shaft of a rotary motor driven by an inverter, a moving member moved by the rotary motor, or a support member that movably supports the rotary shaft or the moving member; generating first time-domain acceleration data from the received acceleration measurements; Fourier transforming the first time domain acceleration data to generate first frequency domain acceleration data; performing an equalization process that includes dividing each acceleration value of the first frequency domain acceleration data by the acceleration value to generate second frequency domain acceleration data having uniform acceleration values ​​across all frequencies in the data; inverse Fourier transforming the second frequency domain acceleration data to generate second time domain acceleration data; An inverter noise removal method comprising:

6. enveloping the second time-domain acceleration data to generate third time-domain acceleration data; Fourier transforming the third time domain acceleration data to generate third frequency domain acceleration data.

6. The inverter noise removal method according to claim 5, further comprising:

7. A program for causing a computer to function as the receiving unit, the generating unit, the Fourier transform unit, the equalizing unit, and the inverse Fourier transform unit according to claim 1.

Images