JP6450575B2 - Inverter noise elimination method and diagnostic method for equipment including inverter - Google Patents

Description

  The present invention relates to an inverter noise removal method and a facility diagnosis method including an inverter and the like.

  Recently, a motor driven by a PWM inverter system is often used particularly at a production site of a processing assembly plant. The motor driven by the PWM inverter system has an advantage that it can be easily operated by changing the number of revolutions only by changing the set value (modulation signal frequency).

  On the other hand, if any abnormality occurs in the rolling bearing or sliding bearing that supports the rotating shaft in a rotating device such as a motor, unusual vibrations and abnormal noise occur. The plain bearing may be damaged, resulting in equipment stoppage.

  Therefore, conventionally, in order to diagnose whether or not such an abnormality has occurred, for example, an acoustic signal is detected by detecting sound generated by contact between a rolling element and a rolling element rotating on a rotating shaft and a rolling element. The state of the equipment is monitored by taking out the sound signal and making a diagnosis using the acoustic signal (equipment diagnosis using vibration diagnosis) (see, for example, Patent Document 1).

Japanese Patent Laid-Open No. 9-136776

  However, if the motor is of the PWM inverter type drive as described above, a noise signal due to the carrier frequency is generated, so that it is difficult to monitor the state in equipment diagnosis such as vibration diagnosis. Specifically, it is as follows.

(1) Problems at the time of absolute value determination Absolute value determination (method) is a method of determining the current state by comparing measured values with reference values as shown in Tables 1 and 2 below. Table 1 is an example of criteria (AMD (Asahikasei Machine Diagnosis) speed judgment criteria) for judging swinging phenomena such as unbalance of rotating bodies, misalignment of couplings, loosening of joints and looseness. The frequency range of the vibration speed value is, for example, 10 Hz to 1 kHz.

Each class and zone in Table 2 will be described as follows.
Class I: Engines and machines incorporated as part of the overall finished machine under normal operating conditions [typical motors with a typical output of 15 kW or less]
Class II: Medium-sized machine without special foundation (typical 15kW to 75kW electric motor) and engine or machine firmly installed on special foundation (300kW or less)
Class III: Large motors and large rotating machines installed on rigid foundations or heavy foundations with relatively high rigidity in the direction of vibration Class IV: Vibration measurements on large motors and large rotating machines Installed on a foundation with relatively soft rigidity in the direction (typical example: turbo generator set and gas turbine with an output of 10 MW or more)

Zone A: A normal range of vibration values of a newly installed machine, and generally a vibration management value at the time of delivery of a newly installed machine.
Zone B: In general, a range in which long-term operation is possible without any limitation. Zone C: In a range where long-term continuous operation cannot be expected, and within a range that can be operated under this vibration condition for a limited period for improvement treatment. A driving alarm value is set in this zone. The alarm value should be set individually according to the characteristics of the machine, but if there is no experience / information, it is recommended to set it to 1.25 times the boundary value of zone B / C.
Zone D: Machines with vibration values in this zone are typically in a range that is considered severe enough to cause damage, and shutdown values are typically in this range.

  In this case, the absolute value determination of the rolling bearing is determined based on the reference as shown in FIG. The dN value on the horizontal axis is the product of the inner diameter of the bearing and the rotational speed, and indicates the peripheral speed of the rolling element of the rolling bearing. For example, since the bearing model 6312 has an inner diameter of 60 mmφ, the dN value is 9.0 × 10E4 at a rotation speed of 1500 rpm. At this time, if the measured value is 1.0 G, it is determined as “caution”. Due to the noise component, it is generated with a high normal value, so it is difficult to distinguish even if the abnormal component rises.

  Further, the method of determining by the magnification with the normal value is referred to as “relative determination method” (see FIGS. 24 and 25). 24 and 25 show the difference between the case where the same motor is driven by an inverter and the case where it is driven by a commercial power source. FIG. 24 shows that noise components not generated in FIG. 25 are generated at a high level. These components are all sidebands around the carrier frequency. Since the acceleration level increases due to the occurrence of this component, it may be determined as “careful though it is normal” based on the previous reference value. For this reason, it becomes difficult to use the absolute value determination criterion determined by the acceleration value. Moreover, it may be a caution area even in normal times. Due to the high amplitude of the noise component, the abnormality cannot be accurately detected when the amplitude of the abnormal signal is small at the time of a mild abnormality, so that it is difficult to detect the abnormality early.

(2) Problems at the time of precision diagnosis The term "precision diagnosis" here refers to the development of countermeasures by clarifying the type and mechanism of occurrence of abnormalities using signal processing such as frequency analysis and envelope analysis. Vibration diagnosis can be broadly divided into “simple diagnosis” and “precision diagnosis”. “Simple diagnosis” measures vibration values at intervals determined in daily life and manages vibration value trends (trend management). ) To perform state monitoring.

The purpose of the simple diagnostic technique is as follows.
(1) Early detection of equipment abnormalities
(2) Quantitatively trace the state of equipment through deterioration tendency management
(3) Monitoring and protection of equipment status
(4) Extraction of equipment for precision diagnosis

“Precise diagnosis” is performed when the “simple diagnosis” shows a tendency of being not normal (indicating an upward trend) or having entered the “attention range” with respect to the reference value. The purpose of precision diagnosis technology is as follows.
(1) Clarify the cause of abnormality
(2) Standardize the type and location of abnormalities and deterioration
(3) Knowing the degree of abnormality and deterioration (severity) and predicting the progress
(4) Determining the optimal repair method and time

Two combined state monitoring, “simple diagnosis” and “precision diagnosis” described above, are performed. However, when the vibration acceleration value in the motor unit when the inverter is driven is obtained, the level increases due to noise, which may cause a false diagnosis in the absolute determination method. Further, if the vibration acceleration waveform of the motor unit when the inverter is driven is enveloped, 2f _s (twice the modulation signal frequency) and its harmonics are generated, making it difficult to determine motor abnormality. In addition, there may be cases where 2f _s and its harmonics are close with rolling scratches frequency (frequency caused by the fact that there are scratches on rolling bearings), there can be a misdiagnosis.

  The present invention relates to an inverter noise removal method capable of accurately detecting an abnormal signal in vibration diagnosis by avoiding the influence of a noise signal caused by a carrier frequency when a PWM inverter type drive motor is used. And a diagnostic method for equipment including an inverter and the like.

In order to solve such a problem, the present invention is a method for removing noise generated during operation of an inverter, an induction motor controlled by the inverter, and a rotating device driven by the induction motor.
Based on the carrier frequency f _b when modulating the input signal to the induction motor by the inverter and the modulation signal frequency f _s , main frequency components other than the basic component are (2n−1) * f _b ± 2 mf _s
2n * f _b ± (2p−1) f _s
(However, n, m: 1,2,3, ..., p: 2,3,4)
And the frequency component of each sideband of the carrier frequency f _b of the current as f _H (n), the frequency of the torque generated by the induction motor | f _s ± f _H (n) |
To calculate
Based on the relationship that the calculated frequency of the torque matches or corresponds to the vibration frequency,
Noise generated in the vibration spectrum by replacing the amplitude value of the vibration spectrum corresponding to the calculated torque frequency with the average value of the amplitude values of the high and low frequencies before and after the vibration frequency. Remove the ingredients,
The spectrum from which the noise component is removed is subjected to inverse Fourier transform, and a vibration acceleration waveform collected by the induction motor is obtained in a form in which the noise component is removed.

  In general, the vibration acceleration of a motor unit collected from an acceleration sensor (piezoelectric type) includes a noise component in the case of inverter driving. The noise component (here, this is expressed as noise) is an acceleration component due to the carrier frequency of the inverter (a sideband due to the modulation frequency around the carrier frequency and its higher order components), and this component It is considered that it is difficult to evaluate the soundness of the motor (detect the abnormality) because it interferes. Moreover, since the vibration acceleration amplitude value includes noise, the vibration acceleration amplitude value is increased, and it is difficult to detect abnormal signs. When discriminating the type of abnormality, it was found that a higher-order component twice the modulation component was generated in the envelope spectrum, making it difficult to diagnose the abnormality.

The present invention finds the regularity of the generated frequency from these noise generation mechanisms, obtains the generated noise frequency, cuts the frequency on the spectrum, and replaces the spectrum level of the frequency with the average value of the frequencies before and after the noise frequency. To obtain a noise-cut spectrum. The noise-cut spectrum is subjected to inverse Fourier transform to return to an acceleration waveform from which noise has been removed, and using this, a vibration acceleration value indicating the state of the motor equipment can be obtained. In other words, the noise generation frequency can be calculated from the carrier frequency and modulation signal frequency values obtained before the vibration measurement, so that the spectrum without the noise is obtained by cutting the frequency from the spectrum, and the inverse Fourier is obtained from the spectrum. By converting, an acceleration waveform can be obtained, and a vibration acceleration value level not including noise can be obtained. The realized abnormality diagnosis with high accuracy using an envelope spectrum the acceleration waveform the 2f _s and the spectrum envelope converts its harmonics are not generated.

  In short, the present invention focuses on “current” and “vibration” that have different frequencies in the first place, finds the commonality between the current spectrum and the torque spectrum, and performs preprocessing to remove noise from the spectrum. Has one feature. In addition, since the noise component can be removed even in the case of discrimination of the type of abnormality, the type of abnormality can be clearly identified and appropriate maintenance can be handled. This leads to the prevention of a fatal failure, and it is not necessary to stop production, which can contribute to the improvement of productivity.

  The diagnosis method according to the present invention is a method for diagnosing an inverter, an induction motor controlled by the inverter, a rotating device driven by the induction motor, and a facility including the rotating device, and includes the inverter noise described above. Including a removal method.

  According to the present invention, when a PWM inverter drive motor is used, it is possible to accurately detect an abnormal signal in vibration diagnosis while avoiding the influence of a noise signal due to the carrier frequency.

It is a figure which shows the generation principle of a PWM pulse. It is a figure which shows the PWM pulse waveform in the case of a single phase. It is a figure which shows the analysis by a double Fourier series. It is a figure shown about the pulse generation of a three phase sine wave PWM inverter. It is a figure which shows the output frequency spectrum of a three-phase PWM inverter. It is a graph which shows the comparison (around 2kHz) of the vibration and current spectrum in an inverter noise frequency confirmation test. It is a graph which shows the comparison (around 4kHz) of a vibration and a current spectrum in an inverter noise frequency confirmation test. It is a figure which shows the sideband of the carrier frequency (2000Hz) in a current spectrum. It is a figure which shows the sideband of the carrier frequency (4000Hz) in a current spectrum. It is a figure which shows the sideband of the carrier frequency (2000Hz) in a vibration spectrum. It is a figure which shows the sideband of the carrier frequency (4000Hz) in a vibration spectrum. It is a figure which shows the vibration acceleration spectrum (commercial power supply 60Hz) of the motor bearing part of a test apparatus. It is a figure which shows the vibration acceleration spectrum (normal bearing) at the time of the inverter drive of the motor bearing part of a test apparatus. It is a figure which shows the acceleration spectrum (normal bearing) after inverter noise cut. It is a figure which shows the spectrum comparison after the inverter noise cut at the time of a bearing outer ring | flaw, and the time of a commercial power supply. It is a figure which shows the spectrum comparison at the time of the inverter noise generation | occurrence | production at the time of a bearing outer ring | flaw, and the time of a commercial power source. It is a figure which shows the envelope spectrum (outer ring wound bearing) after noise removal. It is a figure which shows the envelope spectrum (outer ring wound bearing) in a commercial power source. It is a figure which shows the envelope spectrum (normal bearing) after noise removal. It is a figure which shows the envelope spectrum (normal bearing) at the time of displaying a commercial power supply as a case where it is not an inverter drive. It is a block diagram which shows the outline | summary of the diagnostic system of an installation. It is a flowchart which shows the outline | summary of the flow of the noise removal in the diagnostic system of an installation. It is a figure which shows the carrier frequency, sideband, etc. in a current spectrum. It is a figure explaining replacing the carrier frequency in a current spectrum with the average value of each amplitude value of the spectrum located in the both sides (front and back). It is a graph which shows the AMD rolling bearing criteria. It is a figure which shows the vibration acceleration spectrum (carrier frequency 2kHz, modulation frequency 60Hz) in an inverter motor. It is a figure which shows the vibration acceleration spectrum (carrier frequency 2kHz, modulation frequency 60Hz) in commercial power supply (when it is not an inverter). It is a figure which shows the outline | summary of the conventional abnormality diagnosis system as a reference.

  Hereinafter, the configuration of the present invention will be described in detail based on an example of an embodiment shown in the drawings. The embodiment described below is an example of a preferred embodiment of the present invention, but is not limited thereto, and various modifications can be made without departing from the scope of the present invention.

<Current Spectrum Generation Principle in PWM Inverter (Pulse Width Modulation)>("PWM Power Conversion System", see Kyoritsu Publishing, Katsunori Taniguchi)
Currently, a PWM inverter is the most used inverter system in the production site of a processing and assembly plant. In this method, sideband waves generated around the carrier frequency and its higher order components are generated.

  FIG. 1 shows the principle of PWM pulse generation. Convert the alternating current, which is the frequency of the commercial power supply, into direct current, cut the input direct current voltage as shown in Fig. 1 to make the output voltage into pulses, and control the number, interval, width, etc. of the pulses to change the alternating current of the target frequency To get.

A case where a sine wave e _s is used as a modulation signal and a triangular wave e _b is used as a carrier signal is shown. A voltage waveform is obtained by comparing e _b and e _s and turning on and off the switching element. In other words, to obtain a pulse waveform (1) outputs Ed / 2 when _{e s> e b, (2} ) e s <e Figure by outputting -Ed / 2 when _b 1 (b).

  FIG. 2 shows the PWM pulse waveform obtained in the case of a single phase. This waveform is a waveform in which the pulse width is periodically modulated (Pulse width modulation waveform), and is represented by a double Fourier series, not a Fourier series.

  That is, when a plane OA perpendicular to the plane XOY is considered on a solid waveform in which walls of the same waveform having f (y) at the wall height H and arranged in a 2π cycle are arranged on the plane XOY as shown in FIG. FIG. 3B shows a pulse modulation waveform obtained by projecting a cross section of f (y) on OX.

Here, when _{x = ω b t, y =} ω s t, the double Fourier series of the waveform at this time is shown by Equation (1).
here,

It is.

Generally, it is represented in the form of a complex double Fourier series shown in Equation (2).
here,

It is.

Next, consider a single-phase half-bridge inverter (see FIGS. 1 and 2). The modulation signal e _s
In terms of phase angles θ ₁ and θ _{2 at which} pulse edges occur, the following equation is obtained from the relationship shown in FIG.

Where E _{s is} the amplitude value of the modulation signal, E _{b is} the amplitude value of the carrier triangular signal
M: Degree of modulation (= E _s / E _b )

The complex double Fourier coefficient in this case is expressed by Equation (3).
However,

  Further, it can be said that the pulse of the three-phase sine wave PWM inverter to be analyzed this time has a waveform having a phase difference of 120 ° in FIG. 2, which is a waveform as shown in FIG.

Here, the equation (3) is multiplied by the conversion coefficient λ _{b, n} ,
(1) DC component (m = 0, n = 0),
(2) Fundamental frequency component (m = 0, n = 1)
(3) Harmonic component of the fundamental wave (m = 0, n> 1)
(4) Harmonic component of carrier frequency (m ≧ 1, n = 0)
(5) Sideband component of carrier frequency (m ≧ 1, n ≠ 0)
When the respective complex double Fourier coefficients are obtained, the amplitude is zero except for only the fundamental frequency component and the carrier frequency sideband.

As a result, the output waveform of the three-phase sine wave PWM inverter becomes as shown in Equation (4). In other words, the frequency of the sideband is
When m = 1, n = ± 2, mω _b ± nω _s = ω _b ± 2ω _s
When m = 1, n = ± 4 …… mω _b ± nω _s = ω _b ± 4ω _s
When m = 1, n = ± 6 …… mω _b ± nω _s = ω _b ± 6ω _s
When m = 2 and n = ± 1, mω _b ± nω _s = 2ω _b ± ω _s
When m = 2 and n = ± 5 …… mω _b ± nω _s = 2ω _b ± 5ω _s
In addition, when n = ± 3, in the case of three phases, each phase cancels out and does not occur.
... (4)
here,
It is.

Therefore, the main frequency components other than the basic component are
f _b ± 2f _s , f _b ± 4f _s , f _b ± 6f _s ...
2f _b ± f _s , 2f _b ± 5f _s , 2f _b ± 7f _s ...
And all are sideband components of the carrier frequency. In addition,
f _b : carrier frequency, f _s : modulation signal frequency.

<Frequency of torque generated by induction motor> (Refer to “Forced torsional vibration of the compressor caused by the inverter”, p. 74, equation (2): Kenji Tanaka et al.)
The frequency of torque generated by the PWM inverter induction motor is obtained by the following equation (Equation (5)). That is, if the frequency component of each sideband of the carrier frequency of the current is f _H (n),

This torque is an electromagnetic force between the stator and the rotor, and vibration is generated by this electromagnetic force. Therefore, the noise vibration generated by the inverter matches the frequency of this torque (see FIG. 5).

<Inverter noise frequency confirmation test>
FIG. 6 and FIG. 7 show the results of vibration spectrum and current spectrum measured for a laboratory motor of 15 kW at an inverter carrier frequency of 2000 Hz and a modulation signal frequency of 40 Hz. It can be seen that the same frequency as the simulation result is generated for both current and vibration.

That is, a sideband of a carrier frequency is generated in the current spectrum based on the following regularity (see Table 3) (see FIGS. 8 and 9).

In addition, sidebands of carrier frequencies are generated in the vibration spectrum based on the following regularity (see Tables 4 and 5) (see FIGS. 10 and 11).

<Inverter noise cut test>
FIG. 12 shows the vibration acceleration spectrum of the test apparatus motor bearing portion driven at a commercial power supply of 60 Hz (when not driven by an inverter) (FFT12800 line). FIG. 13 shows an acceleration spectrum when the same motor is used and the carrier frequency is 2 kHz and the modulation signal frequency is 60 Hz. It can be seen that the spectrum is completely different, and almost all inverter noise is occupied.

  FIG. 14 shows a comparison between the spectrum obtained by calculating the noise frequency by the above-described method and the spectrum at the commercial power source shown in FIG. That is, it is a spectrum in which the amplitude of the noise frequency calculated from the carrier frequency 2000 Hz and the modulation signal 60 Hz using the formulas (4) and (5) is zero. It can be seen that most of the noise frequency can be cut although it remains in the high frequency region.

  FIG. 15 shows a comparison between the spectrum after the inverter noise cut when a slit flaw having a width of 0.3 mm is added to the outer ring of the motor bearing and the spectrum at the commercial power source. Inverter noise is cut even when there is a bearing flaw, and a spectrum that is the same as that at the time of commercial power is obtained. It can be seen that the inverter noise is removed as shown in FIG.

Next, FIG. 17 shows an envelope spectrum obtained by performing inverse Fourier transform on the acceleration spectrum from which noise is obtained in FIG. 15 to return to an acceleration waveform, performing envelope processing, and Fourier transform again. It can be seen that the outer ring flaw frequency f _out and its harmonics shown in Equation (6) are clearly generated. That is, the same result as in the commercial power supply is obtained (see FIG. 18).

  An envelope spectrum when noise is cut from a normal bearing is shown in FIG. It can be seen that a high-order component of 120 Hz (twice the modulation signal frequency of 60 Hz) is generated at a low level due to the effect of the inverter noise component that remained a little. FIG. 20 is a diagram showing an envelope spectrum (in the case of a normal bearing) when a commercial power source is displayed as a case where the inverter is not driven.

<Overview of diagnosis system and noise removal flow in diagnosis system>
Here, an example of the diagnostic system 100 that realizes the above-described diagnostic processing or function is shown in FIG. 21, and a flow of noise removal in the diagnostic system 100 is also described with reference to FIG. 22A. The diagnostic system 100 includes a vibration sensor 10, a preamplifier 20, and a computer 30.

  A vibration sensor (for example, a piezoelectric acceleration sensor) 10 detects vibration in a motor bearing or the like (hereinafter sometimes simply referred to as a bearing) and outputs a detection signal (an acceleration waveform including noise) as vibration data. The preamplifier 20 amplifies the detection signal of the vibration sensor 10 and transmits it to the computer 30.

  The computer 30 has equipment diagnosis software, and performs various processes and determinations of vibration data detected and transmitted by the vibration sensor 10 and outputs the results. The computer 30 of this embodiment has functions as each unit such as a filter unit 32, an amplifier unit 34, and an FFT unit 36, and executes the following processes (see FIG. 21).

  The filter unit 32 filters the vibration data amplified by the preamplifier 20. The amplifier unit 34 amplifies the filtered data. The FFT unit 36 performs fast Fourier transform on the amplified data.

Further, the computer 30 inputs the carrier frequency fb and the modulation signal frequency f _s (step SP1 in FIG. 22A), and automatically calculates the noise frequency. That is,
(2n-1) * f _b ± 2 mf _s
2n * f _b ± (2p−1) f _s
(Where n, m: 1, 2, 3,..., P: 1, 3, 4,...) Are calculated (step SP2), and then each side of the current carrier frequency fb The frequency component of the wave band is f _H (n), and the frequency of the torque generated by the induction motor | f _s ± f _H (n) |
Is calculated (step SP3).

  By the way, when n = 1, the carrier frequency is 1 time component, and m = 1 is the first sideband. Therefore, in the upper side (upper formula) of the above two formulas including n and m, a sideband wave of an even multiple of the modulation signal frequency is generated when the carrier frequency is an odd multiple, and in the lower formula, when the carrier frequency is an even multiple of the carrier frequency. A sideband wave having an odd multiple of the modulation signal frequency is generated. However, triple sidebands are not generated (because they cancel each other out because they are three phases). Therefore, the value of p is 1, 3, 4,.

Further, the computer 30 performs a noise frequency cut process from the calculation result and the data subjected to the fast Fourier transform. Specifically, by setting the calculated level (amplitude value) of | f _s ± f _H (n) | to 0 (step SP4), noise components generated in the current spectrum are removed (noise frequency). Cut processing). This is because the level (amplitude value) of the calculated | f _s ± f _H (n) | is calculated based on the high frequency spectrum and the low frequency spectrum located on both sides (front and rear) of the current carrier frequency f _b . This corresponds to replacement with the average value of each amplitude value (see FIGS. 22B and 22C). After the noise frequency cut processing, the computer 30 outputs data after the processing (acceleration spectrum after noise removal).

  At the same time, the computer 30 performs inverse Fourier transform on the processed data (acceleration spectrum after noise removal), and in the case of simple diagnosis, the O / A value of the converted data (acceleration waveform after noise removal) ( Overall value: product sum of power spectrum (Y-axis value) of each frequency subjected to FFT analysis is calculated (steps SP6 and SP7), and the calculated data (acceleration value after noise removal) is output (step SP10). . According to the diagnostic system 100, state monitoring can be performed by performing trend management of the acceleration value (trend management). In place of the O / A value, an RMS value (effective value) and an average value can also be calculated.

  Further, in the case of precise diagnosis, the computer 30 performs envelope processing (envelope processing) on the data after inverse Fourier transform (acceleration waveform after noise removal) (step SP6, SP8), and further performs fast Fourier transform (step SP9). ). Further, the computer 30 automatically calculates a flaw frequency of the bearing (a frequency generated due to a flaw in the rolling bearing or the like) from the rotation frequency of the bearing. The computer 30 automatically determines and discriminates the bearing abnormality type from the data after the automatic calculation and the data after the fast Fourier transform (acceleration envelope spectrum), and outputs the data after the determination (abnormality type and countermeasure). (SP10). Abnormal types include bearing flaws (outer rings, inner rings, rolling elements, cages), poor bearing lubrication (oil leaks and foreign matter), and bearing fit backlash. Moreover, as a countermeasure, there is a display of the contents of determination corresponding to various states of the bearing (see Table 1).

As described above, in the abnormality diagnosis method of the present embodiment described so far, the noise generation frequency can be calculated from the values of the carrier frequency and the modulation signal frequency grasped before the vibration measurement, so that the frequency is cut from the spectrum (or Obtain a spectrum that does not contain noise (by replacing the spectrum level average value on both sides of the cut frequency), and obtain an acceleration waveform by performing inverse Fourier transform from that spectrum, and obtain a vibration acceleration value level that does not contain noise. Can do. In addition, it is possible to perform highly accurate abnormality diagnosis using 2 f _s (twice the modulation signal frequency) and an envelope spectrum that does not generate harmonics from the spectrum obtained by envelope conversion of the acceleration waveform.

<Comparative example>
In conventional abnormality diagnosis, when obtaining vibration displacement and vibration speed values, the acceleration waveform (filtered and amplified by an amplifier) is passed through an integration circuit to obtain the level, and trend management based on that level (vibration acceleration value) is performed. (See FIG. 26). There, the relationship between vibration and sound torque frequencies and acceleration diagnostic techniques has not been found.

  The present invention is suitable for application when diagnosing an inverter, an induction motor controlled by the inverter, a rotating device driven by the induction motor, and equipment including the rotating device.

DESCRIPTION OF SYMBOLS 10 ... Vibration sensor 20 ... Preamplifier 30 ... Computer 100 ... Diagnostic system

Claims (2)

An inverter, an induction motor controlled by the inverter, and a method for removing noise generated during operation of equipment including a rotating device driven by the induction motor,
Based on the carrier frequency f _b when modulating the input signal to the induction motor by the inverter and the modulation signal frequency f _s , main frequency components other than the basic component are (2n−1) * f _b ± 2 mf _s
2n * f _b ± (2p−1) f _s
(However, n, m: 1,2,3, ..., p: 2,3,4)
After having expressed in the frequency components of each sideband before crisis Yaria frequency f _b as f _H (n), the frequency of the torque induction motor occurs _{| f s ± f H (n} ) |
To calculate
Based on the relationship that the calculated frequency of the torque matches or corresponds to the vibration frequency,
Noise generated in the vibration spectrum by replacing the amplitude value of the vibration spectrum corresponding to the calculated torque frequency with the average value of the amplitude values of the high and low frequencies before and after the vibration frequency. Remove the ingredients,
An inverter noise removal method, wherein the spectrum obtained by removing the noise component is subjected to inverse Fourier transform, and a vibration acceleration waveform collected by the induction motor is obtained in a form from which the noise component is removed. An inverter, an induction motor controlled by the inverter, and a method for diagnosing equipment including a rotating device driven by the induction motor,
An inverter noise removing method according to claim 1, the diagnostic method of the equipment including the inverter.