JP5773932B2 - Electric motor rotation information detection method, electric motor rotation information detection device, electric motor control device - Google Patents

Description

  The present invention relates to a method and apparatus for detecting rotation information such as the rotation speed and rotation angle of an electric motor based on a change in current flowing in the electric motor, and an apparatus for controlling the rotation of the electric motor based on the detected rotation information. .

  2. Description of the Related Art Conventionally, there has been proposed a method and apparatus for detecting rotation information such as the number of rotations of an electric motor based on a change in a current flowing through the electric motor being driven without using an external detection means such as an encoder.

  For example, in Patent Document 1, a current flowing through an electric motor equipped with a brush is detected, a frequency analysis of the current waveform is performed, and a ripple (ripple current) frequency is extracted. Every time the commutator that is in sliding contact with the brush of the electric motor is switched, a ripple occurs in the current flowing through the electric motor. Therefore, the number of rotations of the electric motor is obtained from the ripple frequency and the number of commutators.

  In Patent Document 2, the input current waveform of the electric motor is sampled at a predetermined interval, sampling data is collected, and the data is sequentially shifted at a constant cycle in the time axis direction to generate a plurality of shift sampling data. . Then, the sum of the absolute values of the differences between the sampling data and each shift sampling data is plotted at a constant period to obtain a normalized current waveform, and based on the fundamental frequency of the harmonic obtained by frequency analysis of the waveform Thus, the rotational speed of the electric motor is calculated.

  Patent Document 3 discloses that time-series waveform data of fluctuation values such as current is subjected to frequency analysis by an algorithm such as Fourier transform, converted to frequency waveform data, and a frequency having a high amplitude is extracted. As is done in general. In Patent Document 3, a failure of a machine coupled to an electric motor is diagnosed based on a gain level fluctuation of a frequency having a high peak.

  Incidentally, in the current flowing through the electric motor being driven, noise due to external factors and mechanical variations of the electric motor may occur in addition to ripple. For this reason, in the waveform obtained by frequency analysis of the current waveform, the amplitude of the noise frequency may protrude more greatly than the ripple frequency, and the noise frequency may be erroneously detected as the ripple frequency. There is. Thus, the rotation information of the electric motor cannot be accurately detected at the erroneously detected frequency.

JP 2004-80921 A JP-A-2005-341744 JP 2010-166686 A

  An object of the present invention is to reduce the influence of noise and accurately detect rotation information of an electric motor.

According to the present invention, in a rotation information detection method for an electric motor that detects current flowing in an electric motor equipped with a brush and detects rotation information of the electric motor based on a frequency of a ripple generated in the current, the electric current flows in the electric motor. Frequency analysis of current time-series waveform data, the frequency of the fundamental wave of the motor current having a maximum value in a predetermined low frequency of the obtained waveform data in frequency units is set as the fundamental frequency candidate, and the commutator of the electric motor The ripple frequency candidate is calculated by multiplying the fundamental frequency candidate by a coefficient set in advance based on the number and the change characteristic of the current flowing through the electric motor, and the maximum value that is adjacent to the ripple frequency candidate is calculated using waveform data in frequency units. Is added as a ripple frequency candidate, and when the maximum amplitude of the ripple frequency candidate is equal to or greater than a predetermined threshold value, the ripple frequency having the maximum amplitude is The number candidates to the ripple frequency.

Further, in the present invention, there is provided a current detection means for detecting a current flowing in the electric motor provided with the brush, and the electric motor detects the rotation information of the electric motor based on the frequency of the ripple generated in the current flowing in the electric motor. The rotation information detection apparatus further includes an analysis unit that performs frequency analysis on the time-series waveform data of the current detected by the current detection unit, and an estimation unit that estimates the ripple frequency. The estimation means uses the frequency of the fundamental wave of the motor current having a local maximum value in a predetermined low range of the waveform data in frequency units obtained by the frequency analysis by the analysis means as a fundamental frequency candidate, and the number of commutators of the electric motor The ripple frequency candidate is calculated by multiplying the fundamental frequency candidate by a coefficient set in advance based on the change characteristic of the current flowing through the electric motor, and has a maximum value on both sides of the ripple frequency candidate in the frequency unit waveform data. A frequency is added as a ripple frequency candidate, and when the maximum amplitude of the ripple frequency candidate is greater than or equal to a predetermined threshold, the ripple frequency candidate having the maximum amplitude is estimated as the ripple frequency.

  The present invention also includes current detection means for detecting the current flowing through the electric motor provided with the brush, and detects rotation information of the electric motor based on the frequency of the ripple generated in the current flowing through the electric motor. The electric motor control apparatus that controls the driving of the electric motor based on the information further includes the analysis means and the estimation means.

  According to the above, in the frequency unit waveform data obtained by frequency analysis of the time series waveform data of the electric motor current, it is possible to reliably detect the ripple frequency that is substantially proportional to the fundamental frequency in the low band. . For this reason, the influence of noise can be reduced and the rotation information of the electric motor can be accurately detected based on the ripple frequency. Further, the rotation of the electric motor can be appropriately controlled based on the rotation information of the electric motor detected with high accuracy.

  Further, in the present invention, in the rotation information detection method for the electric motor, before performing frequency analysis on the time series waveform data of the current flowing through the electric motor, an approximate function of the time series waveform data of the current is calculated, and the approximation function The trend of the current time-series waveform data may be corrected based on the above.

  In the present invention, the rotation information detecting device for the electric motor further calculates an approximate function of the current time-series waveform data before performing frequency analysis of the time-series waveform data of the current flowing through the electric motor by the analyzing means. Then, a correction means for correcting the tendency of the current time-series waveform data based on the approximate function may be provided.

  According to the present invention, in the method for detecting rotation information of the electric motor, the filter coefficient is calculated based on the ripple frequency, the time-series waveform data of the current flowing through the electric motor is filtered by the filter coefficient, and the ripple current is detected. The series waveform data may be extracted, and the ripple pulse signal may be generated from the time series waveform data of the ripple current.

  According to the present invention, in the rotation information detection apparatus for the electric motor, the calculation means for calculating the filter coefficient based on the ripple frequency estimated by the estimation means, and the time series waveform data of the current flowing through the electric motor are filtered. Filtering means for extracting ripple current time-series waveform data and generating means for generating ripple pulse signals from the ripple current time-series waveform data.

  According to the present invention, it is possible to reliably detect the ripple frequency and accurately detect the rotation information of the electric motor without being affected by noise. Moreover, the rotation of the electric motor can be appropriately controlled based on the rotation information detected with high accuracy.

It is a block diagram of the electric motor control apparatus by one Embodiment of this invention. It is a control block diagram for electric motor rotation information detection of the same device. FIG. 3 is a detailed diagram of the trend correction block of FIG. 2. FIG. 3 is a detailed diagram of the FFT block of FIG. 2. It is the flowchart which showed the procedure of the estimation process of a ripple frequency. It is the figure which showed an example of the spectrum of the frequency obtained by carrying out the fast Fourier transform of the motor current. It is the figure which showed an example of the time series waveform data of a motor current. It is the figure which showed an example of the spectrum of the frequency of FIG. FIG. 3 is a detailed diagram of the bandpass filter design block of FIG. 2. FIG. 3 is a detailed diagram of a filtering processing block in FIG. 2. FIG. 3 is a detailed diagram of the ripple pulse processing block of FIG. 2. It is the flowchart which showed the procedure of the binarization process of a ripple current. It is the figure which showed an example of a ripple current and a ripple pulse signal.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the drawings, the same or corresponding parts are denoted by the same reference numerals.

  First, the configuration of an electric motor control device 100 according to an embodiment of the present invention will be described with reference to FIG.

  The control unit 1 of the electric motor control device 100 is composed of a microcomputer. The motor drive unit 2 includes a switching circuit including a pair of relays or four switching elements. The control unit 1 operates the motor driving unit 2 to cause a current to flow from the battery 5 to the electric motor 4 to rotate the electric motor 4.

  The electric motor 4 includes a DC motor provided with a brush and a commutator. The control unit 1 operates the drive target 30 with the power of the electric motor 4. For example, the drive object 30 is a power window of a vehicle, and the electric motor control device 100 is an ECU (electronic control device) for the power window.

  The current detection unit 3 detects the motor current flowing through the electric motor 4 at a predetermined period and outputs it to the control unit 1. While the electric motor 4 is being driven, a ripple (ripple) is generated in the motor current each time the commutator that is in sliding contact with the brush of the electric motor 4 is switched. The rotation information detection unit 10 of the control unit 1 detects rotation information such as the rotation speed and rotation angle of the electric motor 4 based on the ripple frequency of the motor current, as will be described later.

  The electric motor control device 100 is an example of the “electric motor rotation information detection device” in the present invention. The current detection unit 3 is an example of the “current detection unit” in the present invention.

  The position detection unit 20 of the control unit 1 detects the operation position of the drive object 30 based on the rotation angle of the electric motor 4 detected by the rotation information detection unit 10. The control unit 1 controls the rotational speed of the electric motor 4 by operating the motor driving unit 2 based on the rotation speed detected by the rotation information detection unit 10. Further, the control unit 1 operates the motor drive unit 2 based on the rotation angle detected by the rotation information detection unit 10 and the operation position detected by the position detection unit 20, and the rotation amount of the electric motor 4 The operation amount of the driving object 30 is controlled.

  Next, a method for detecting rotation information of the electric motor 4 according to an embodiment of the present invention will be described with reference to FIGS. It is the control unit 1 of the electric motor control device 100 that executes this method.

  The control unit 1 is provided with control blocks 11 to 16 for detecting rotation information of the electric motor 4 as shown in FIG. Each function of these blocks is actually realized by software, but may be realized by hardware. The motor current value of the electric motor 4 detected by the current detection unit 3 in FIG. 1 is input to the control unit 1 at any time and stored in a predetermined number in the built-in memory of the control unit 1. Then, the motor current value of a predetermined number of samples is input to the trend correction block 11 of FIG. 2 as time series waveform data. The tendency correction block 11 is an example of the “correction unit” in the present invention.

  The trend correction block 11 approximates the time series waveform data of the motor current as shown in FIG. 3B by a quadratic function by the least square method in the quadratic function approximation block 11a shown in FIG. Approximate data as shown in FIG. 3C is calculated. Then, each value of the approximate data is subtracted from each value of the time series waveform data of the motor current to correct the tendency of the time series waveform data of the motor current.

  As a result, for example, the time series waveform data of the motor current having a large degree of decrease and increase over the entire waveform as shown in FIG. 3 (b) suppresses the degree of decrease and increase as shown in FIG. 3 (d). Waveform data. The time series waveform data of the motor current after the trend correction is output from the trend correction block 11 to the FFT (Fast Fourier Transform) block 12 in FIG. The FFT block 12 is an example of the “analyzing unit” in the present invention.

  The FFT block 12 performs frequency analysis on the time series waveform data of the motor current after the trend correction by FFT. As a result, the time-series waveform data of the motor current after the trend correction as shown in FIG. 4B is converted into a frequency spectrum (waveform unit waveform data) as shown in FIG. 4C, for example.

  The spectrum obtained by the frequency analysis is output from the FFT block 12 to the ripple frequency estimation block 13 in FIG. The ripple frequency estimation block 13 is an example of the “estimator” of the present invention.

  In the ripple frequency estimation block 13, first, as shown in FIG. 6, all frequencies having a maximum value in a predetermined low band of the spectrum are set as fundamental frequency candidates Fb (step S1 in FIG. 5). Next, the frequency Fbm having the maximum amplitude is detected from the fundamental frequency candidates Fb (step S2 in FIG. 5). Then, the ripple frequency candidate Frb is calculated by multiplying the frequency Fbm by a predetermined coefficient K (step S3).

  The coefficient K is set in advance based on the number R of commutators provided in the electric motor 4 and the change characteristics of the current flowing through the electric motor 4. Specifically, a value obtained by dividing the number of commutators R by the number of fundamental waves B of the time-series waveform data of the current flowing in the electric motor 4 when the rotor makes one rotation is set as the coefficient K. (Coefficient K = commutator number R / fundamental wave number B).

  For example, if the electric motor 4 has eight commutators and the time series waveform data of the motor current when the rotor of the electric motor 4 makes one revolution changes as shown by the solid line in FIG. Eight have occurred. This is because when the rotor of the electric motor 4 makes one rotation, the commutator that is in sliding contact with the brush is switched eight times, and the current flow (large or small) changes between the brush and the commutator each time. is there.

  Further, in FIG. 7, while the eight ripples are generated, two large fundamental waves of the motor current on which the ripples are superimposed are generated (two cycles) as indicated by a broken line. In this case, the number of commutators R = 8 is divided by the fundamental wave number B = 2 to obtain a coefficient K = 4 (coefficient K = commutator number R / fundamental wave number B = 8/2 = 4). That is, the coefficient K is the ratio of the commutator number R (= ripple number per rotation) of the electric motor 4 to the fundamental wave number B.

  When frequency analysis is performed on the time series waveform data of the motor current as shown in FIG. 7 by FFT, a spectrum as shown in FIG. 8 is obtained. In FIG. 8, the frequency having the maximum value in the low band is the frequency of the fundamental wave of the motor current, that is, the fundamental frequency. And the frequency which has the maximum value of the maximum amplitude in a high region is a ripple frequency. The ripple frequency is about four times the fundamental frequency. That is, the ripple frequency is approximately K times the fundamental frequency, and the ripple frequency and the fundamental frequency are approximately proportional.

  The number of commutators R of the electric motor 4 is disclosed in, for example, the specification of the electric motor 4. Since the fundamental wave number B per rotation of the electric motor 4 is different for each electric motor 4, for example, it is measured in advance by an experiment.

  Focusing on the data characteristics as described above, in step S3 of FIG. 5, the fundamental frequency candidate Fbm is multiplied by a coefficient K to calculate the ripple frequency candidate Frb.

  In addition, in the step S1 of FIG. 5, the low range of the spectrum for detecting the fundamental frequency candidate Fb (step S1, FIG. 6 and FIG. 8 of FIG. 5) Is set to a range equal to or less than the value multiplied by the fundamental wave number B. For example, when the use area of the electric motor 4 is 50 Hz and the fundamental wave number B is 2, since 50 Hz × 2 = 100 Hz, the low frequency is 0 to 100 Hz.

  When the ripple frequency candidate Frb is calculated in step S3 of FIG. 5, as shown in FIG. 6, the frequencies having the maximum values adjacent to the ripple frequency candidate Frb in the spectrum are added as ripple frequency candidates Frr and Frl (FIG. 5). Step S4). Then, the maximum amplitude Am is detected from the amplitudes of these ripple frequency candidates Frb, Frr, and Frl (step S5), and it is determined whether or not the maximum amplitude Am is equal to or greater than a predetermined threshold value (step S6).

  If the maximum amplitude Am is greater than or equal to a predetermined threshold (step S6: YES), the ripple frequency candidates Frb, Frr, Frl having the maximum amplitude Am are estimated as ripple frequencies (step S7). In the example of FIG. 6, since the amplitude of the ripple frequency candidate Frl among the ripple frequency candidates Frb, Frr, and Frl is the maximum and equal to or greater than the threshold value, the ripple frequency candidate Frl is estimated as the ripple frequency.

  On the other hand, if the maximum amplitude Am is less than the predetermined threshold (step S6 in FIG. 5: NO), the frequency Fbm is excluded from the fundamental frequency candidates Fb (step S8). And the process after step S2 is performed again.

  The ripple frequency estimated by the ripple frequency estimation block 13 is output from the ripple frequency estimation block 13 to the bandpass filter design block 14 of FIG. The bandpass filter design block 14 is an example of the “calculation unit” in the present invention.

In the bandpass filter design block 14, as shown in FIG. 9, the low frequency side edge frequency is set in the block 14a and the high frequency side edge frequency is set in the block 14b based on the input ripple frequency. Specifically, values obtained by increasing / decreasing the ripple frequency by a predetermined ratio C% are set as edge frequencies on the high frequency side and the low frequency side, respectively.
Ripple frequency + (C% of ripple frequency) = High frequency side edge frequency Ripple frequency-(Cripple frequency C%) = Low frequency side edge frequency

  Then, a band pass filter coefficient is calculated in block 14c based on the low-frequency and high-frequency edge frequencies. Specifically, a general FIR (finite impulse response) filter based on the window function method is designed, and each filter coefficient is calculated. The procedure will be described below.

After setting the edge frequency Fl on the low frequency side and the edge frequency Fh on the high frequency side, the edge angular frequencies ωl and ωh normalized by the sampling frequency Fs are determined.
ωl = 2πFl / Fs, ωh = 2πFh / Fs
Next, a low-pass filter Hl (n) having a cutoff angular frequency of ωc = (ωh−ωl) / 2 is designed.
Hl (n) = sin (n × ωc) / πn (n = 0, ± 1,..., ± M)
Next, the window function W (n) is multiplied.
Hlw (n) = Hl (n) × W (n)
Then, the filter coefficient Hbp (n) of the bandpass filter is calculated using the following relational expression.
ω0 = (ωl + ωh) / 2
Hbp (n) = 2 cos (n × ω0) × Hlw (n)

  The calculated filter coefficients are output from the bandpass filter design block 14 to the filtering processing block 15 in FIG. The filtering processing block 15 is an example of the “extraction means” in the present invention.

  The filtering processing block 15 filters the time-series waveform data of the motor current with the input filter coefficient to extract the time-series waveform data of the ripple current. Specifically, by performing a general FIR filter calculation using the input filter coefficient in the filter calculation block 15a shown in FIG. 10A, a time-series waveform of the motor current as shown in FIG. 10B. From the data, time series waveform data of only ripple current as shown in FIG. 10C is extracted.

  The extracted time-series waveform data of the ripple current is output from the filtering processing block 15 to the ripple pulse processing block 16 of FIG. The ripple pulse processing block 16 is an example of the “generating unit” in the present invention.

  The ripple pulse processing block 16 binarizes the time-series waveform data of the ripple current as shown in FIG. 11 (b) to generate a rectangular wave ripple pulse signal as shown in FIG. 11 (c).

  Specifically, as shown in FIGS. 12 and 13, a minimum value is detected following the time series of time series waveform data of the ripple current, and the minimum value is a predetermined negative band threshold value on the minus side. If the previous ripple pulse is “High” (high level) (step S11: YES), the ripple pulse is set to “Low” (low level) (step S12). Further, if the local maximum value is detected, the local maximum value is equal to or greater than the predetermined positive (+) dead zone threshold, and the previous ripple pulse is “Low” (step S13: YES), the ripple pulse is detected. Is set to “High” (step S14).

  For the first time, for example, when a minimum value is detected from time-series waveform data of a ripple current and the minimum value is equal to or smaller than a predetermined negative-side dead band threshold, the ripple pulse is set to “Low”. If the maximum value is detected and the maximum value is equal to or greater than a predetermined positive-side dead band threshold, the ripple pulse is set to “High”.

  The ripple pulse signal generated as described above is output from the ripple pulse processing block 16 to the rotation information detection unit 10 of FIG.

  The rotation information detection unit 10 detects rotation information such as the rotation speed and rotation angle of the electric motor 4 based on the input ripple pulse signal. For example, the rotation speed is calculated based on the number of commutators of the electric motor 4 and the number of pulses per unit time, or the reciprocal of the width of one pulse is calculated. The rotation angle is calculated based on the number of pulses from the base point. The rotation information detection unit 10 continuously detects rotation information at a plurality of points in time from the ripple pulse signal.

  The number of rotations of the electric motor 4 can also be calculated from the ripple frequency estimated by the ripple frequency estimation block 13 of FIG. 2 (number of rotations = ripple frequency / number of poles of the electric motor 4 × 60). In this case, the rotational speed of the temporary point is detected.

  From the time taken for the ripple frequency estimation block 13 to estimate the ripple frequency after passing through the trend correction block 11 and the FFT block 12, the rotation information detection unit 10 detects the rotation information of the electric motor 4 at each time point from the ripple pulse signal. The time it takes to do is shorter.

  According to the above embodiment, in the frequency spectrum obtained by frequency analysis of the time-series waveform data of the current of the electric motor 4 by the FFT, the ripple frequency that is substantially proportional to the fundamental frequency in the low band is reliably detected. be able to. For this reason, as shown by a broken line in FIG. 6, even when noise having a large amplitude that is not a ripple frequency candidate is generated in the motor current, it is possible to prevent the noise frequency from being erroneously detected as the ripple frequency.

  Therefore, the rotation information of the electric motor 4 can be accurately detected based on the ripple frequency without being affected by noise. Further, the rotation of the electric motor 4 can be appropriately controlled based on the rotation information of the electric motor 4 detected with high accuracy.

  By the way, just before the electric motor 4 starts to move or immediately before it stops, the current of the electric motor 4 tends to increase or decrease. For this reason, if the time series waveform data of the motor current at that time is subjected to frequency analysis by FFT, the maximum value does not appear remarkably in the low region of the spectrum of the obtained frequency, and the fundamental frequency candidate Fb may not be detected.

  However, in the above embodiment, the trend correction block 11 corrects the tendency of the time series waveform data of the motor current with an approximate function (secondary function), and then the FFT block 12 performs frequency analysis. For this reason, the fundamental frequency candidate Fb can be reliably detected by prominently producing a maximum value in the low frequency spectrum obtained by FFT.

  Further, in the above embodiment, time series waveform data of the ripple current is extracted based on the ripple frequency, and a ripple pulse signal is generated from the waveform data. For this reason, the rotation speed and rotation angle of the electric motor 4 can be continuously detected from a ripple pulse signal at a plurality of time points. Further, the detection accuracy and resolution can be improved by detecting the rotation speed of the electric motor 4 from the ripple pulse signal rather than directly calculating the rotation speed of the electric motor 4 from the ripple frequency.

  The present invention can employ various embodiments other than those described above. For example, in the above-described embodiment, an example in which the frequency analysis of the time series waveform data of the motor current is performed by FFT is shown. However, the present invention is not limited to this, and the frequency is calculated by other algorithms such as DFT (Discrete Fourier Transform). You may make it analyze.

  Further, in the above embodiment, an FIR filter based on the window function method is designed based on the ripple frequency, the band pass filter coefficient is calculated, and the time series data of the ripple pulse current is extracted from the time series data of the motor current by the FIR filter calculation. However, the present invention is not limited to this example. The time series data of the ripple pulse current may be extracted from the time series data of the motor current by a signal processing method based on a ripple frequency other than this.

  Further, in the above-described embodiment, the example in which the present invention is applied to the electric motor control device 100 including the ECU for the power window has been described. However, the present invention is applied to various devices that require rotation information of the electric motor including the brush. The present invention can be applied.

DESCRIPTION OF SYMBOLS 3 Current detection part 4 Electric motor 11 Trend correction block 12 FFT block 13 Ripple frequency estimation block 14 Band pass filter design block 15 Filtering processing block 16 Ripple pulse processing block 100 Electric motor controller Am Maximum amplitude of ripple frequency candidate B Number of commutators Fb Candidate fundamental frequency Frb, Frl, Frr Ripple frequency candidate K coefficient

Claims (7)

In the rotation information detection method of the electric motor, the current flowing through the electric motor including the brush is detected, and the rotation information of the electric motor is detected based on the frequency of the ripple generated in the current.
Analyzing the time-series waveform data of the current flowing through the electric motor, and having the maximum value in a predetermined low frequency of the obtained waveform data in frequency units, the fundamental frequency of the motor current as a fundamental frequency candidate,
Multiplying the fundamental frequency candidate by a coefficient set in advance based on the number of commutators of the electric motor and a change characteristic of the current flowing through the electric motor, the ripple frequency candidate is calculated,
Add the frequency having the maximum value on both sides of the ripple frequency candidate in the waveform data of the frequency unit as a ripple frequency candidate,
A method for detecting rotation information of an electric motor, wherein when the maximum amplitude of the ripple frequency candidate is equal to or greater than a predetermined threshold, the ripple frequency candidate having the maximum amplitude is set as a ripple frequency. In the rotation information detection method of the electric motor according to claim 1,
Before frequency analysis of the time series waveform data of the current flowing through the electric motor,
A method for detecting rotation information of an electric motor, comprising: calculating an approximate function of the time-series waveform data of the current and correcting a tendency of the time-series waveform data of the current based on the approximate function. In the rotation information detection method of the electric motor according to claim 1 or claim 2,
Calculating a filter coefficient based on the ripple frequency;
Filtering the time series waveform data of the current flowing through the electric motor with the filter coefficient to extract the time series waveform data of the ripple current,
A method for detecting rotation information of an electric motor, wherein a ripple pulse signal is generated from time-series waveform data of the ripple current. A current detecting means for detecting a current flowing in an electric motor having a brush;
In the rotation information detecting device for the electric motor that detects the rotation information of the electric motor based on the frequency of the ripple generated in the current flowing through the electric motor,
Analyzing means for frequency analysis of time-series waveform data of current detected by the current detecting means;
An estimation means for estimating the ripple frequency,
The estimation means includes
The frequency of the fundamental wave of the motor current having a maximum value in a predetermined low frequency of the waveform data in frequency units obtained by the frequency analysis by the analyzing means is set as a fundamental frequency candidate,
Multiplying the fundamental frequency candidate by a coefficient set in advance based on the number of commutators of the electric motor and a change characteristic of the current flowing through the electric motor, the ripple frequency candidate is calculated,
Add the frequency having the maximum value on both sides of the ripple frequency candidate in the waveform data of the frequency unit as a ripple frequency candidate,
An apparatus for detecting rotation information of an electric motor, wherein the ripple frequency candidate having the maximum amplitude is estimated as a ripple frequency when the maximum amplitude of the ripple frequency candidate is equal to or greater than a predetermined threshold value. In the rotation information detection device of the electric motor according to claim 4,
Before performing frequency analysis on the time series waveform data of the current flowing through the electric motor by the analyzing means, an approximate function of the time series waveform data of the current is calculated, and the time series waveform data of the current is calculated based on the approximation function. An apparatus for detecting rotation information of an electric motor, comprising correction means for correcting the tendency. In the rotation information detection device for the electric motor according to claim 4 or 5,
Calculating means for calculating a filter coefficient based on the ripple frequency estimated by the estimating means;
Extraction means for filtering time-series waveform data of current flowing in the electric motor with the filter coefficient and extracting time-series waveform data of ripple current;
An apparatus for detecting rotation information of an electric motor, comprising: generating means for generating a ripple pulse signal from time-series waveform data of the ripple current. A current detecting means for detecting a current flowing in an electric motor having a brush;
In the electric motor control device that detects rotation information of the electric motor based on the frequency of the ripple generated in the current flowing in the electric motor, and controls the driving of the electric motor based on the rotation information.
Analyzing means for frequency analysis of time-series waveform data of current detected by the current detecting means;
An estimation means for estimating the ripple frequency,
The estimation means includes
The frequency of the fundamental wave of the motor current having a maximum value in a predetermined low frequency of the waveform data in frequency units obtained by the frequency analysis by the analyzing means is set as a fundamental frequency candidate,
Multiplying the fundamental frequency candidate by a coefficient set in advance based on the number of commutators of the electric motor and a change characteristic of the current flowing through the electric motor, the ripple frequency candidate is calculated,
Add the frequency having the maximum value on both sides of the ripple frequency candidate in the waveform data of the frequency unit as a ripple frequency candidate,
An electric motor control device, wherein when the maximum amplitude of the ripple frequency candidate is greater than or equal to a predetermined threshold, the ripple frequency candidate having the maximum amplitude is estimated as a ripple frequency.

Images