JP4466439B2 - Magnetic bearing device - Google Patents

Description

  The present invention relates to a magnetic bearing device used for a turbo molecular pump, a machine tool, or the like.

  Patent Document 1 discloses a 5-axis control type magnetic bearing device. In the magnetic bearing device, the displacement of the floating position (support position) of the rotor is detected by a displacement sensor, and the rotor is magnetically levitated to a desired position based on the detected displacement. The displacement sensor modulates an applied carrier wave in accordance with the displacement of the rotor, and performs demodulation processing after an analog modulation signal is changed to a digital value by an A / D converter.

JP 2004-144291 A

  However, since the above-described conventional apparatus converts the sensor signals corresponding to the five axes to digital signals, five channels are required as an A / D converter input.

The invention of claim 1 is applied to a multi-axis control type magnetic bearing device that supports a supported body in a non-contact manner in at least two biaxial directions, and the displacement of the support position in the first axial direction of the multiaxial directions is changed. First detection means for detecting, second detection means for detecting displacement of the support position in the second axial direction among the plurality of axial directions, a detection signal from the first detection means, and a second detection means A synthesizing unit that synthesizes the detection signal into one synthesized signal, an A / D converting unit that converts the synthesized signal into a digital signal, and a signal corresponding to the displacement in the first axial direction are extracted by demodulating the digital signal. The first demodulating means, the second demodulating means for demodulating the digital signal to extract a signal corresponding to the displacement in the second axial direction, and the signals extracted by the first and second demodulating means Based on the first and second axial support positions. Characterized in that a control means for.
According to a second aspect of the present invention, in the magnetic bearing device according to the first aspect, the first and second detecting means are displacement sensors that apply a carrier wave signal and modulate the carrier wave signal in accordance with the displacement of the support position. A carrier wave generating means for applying a sine wave signal to the first detecting means as a carrier wave signal and applying a cosine wave signal obtained by phase shifting the sine wave signal by 90 degrees to the second detecting means. is there.
According to a third aspect of the present invention, in the magnetic bearing device according to the second aspect, the first demodulating means multiplies the digital signal by a sine wave signal to extract a signal corresponding to the first axial displacement, The second demodulating means multiplies the digital signal by the cosine wave signal to extract a signal corresponding to the displacement in the second axial direction.

  According to the present invention, since the detection signal from the first detection means and the detection signal from the second detection means are combined into one composite signal and then converted into a digital signal, A / D conversion processing is performed. The load can be reduced, and the cost of the apparatus can be reduced.

  Hereinafter, the best mode for carrying out the present invention will be described with reference to the drawings. FIG. 1 is a sectional view showing a schematic configuration of a pump body 1 of a magnetic bearing turbomolecular pump. Hereinafter, a case where the present invention is applied to a magnetic bearing device of a magnetic bearing type turbo molecular pump will be described as an example. The magnetic bearing type turbo molecular pump shown in FIG. 1 includes a 5-axis control type magnetic bearing, and a rotor 4 formed with a plurality of stages of rotor blades 21 includes electromagnets 51 and 52 constituting a radial magnetic bearing, and an axial. Non-contact support is provided by the electromagnet 53 constituting the magnetic bearing.

  Inside the casing 20, a plurality of rotor blades 23 are provided in the axial direction (the vertical direction in the figure). The magnetic bearing is provided with radial displacement sensors 71 and 72 and an axial displacement sensor 73 corresponding to the radial electromagnets 51 and 52 and the axial electromagnet 53. The receptacle 25 is connected to a cable connecting the pump main body 1 and a controller (not shown), and the pump main body 1 is driven and controlled by the controller. The rotor 4 is rotationally driven by the motor 6 while being supported in a non-contact manner by the electromagnets 51, 52, 53, thereby generating a pump action.

  27 and 28 are emergency mechanical bearings. When the rotor 4 is not magnetically levitated, the rotor 4 is supported by these bearings 27 and 28. The bearing 27 constrains the movement of the rotor 4 in the radial direction (x-axis and y-axis) in an emergency, and the bearing 28 includes two radial directions (x-axis and y-axis) and one axial direction (z-axis). ).

  FIG. 2 is a conceptual diagram of a 5-axis control type magnetic bearing, in which the rotation axis J of the rotor 4 is shown to coincide with the z-axis. The radial electromagnet 51 shown in FIG. 1 includes a pair of electromagnets 51x related to the x axis and a pair of electromagnets 51y related to the y axis. Similarly, the radial electromagnet 52 also includes a pair of electromagnets 52x related to the x axis and a pair of electromagnets 52y related to the y axis. In addition, the axial electromagnet 53 includes a pair of electromagnets 53z disposed to face each other so as to sandwich the disk 41 provided at the lower end of the rotor 4 along the z axis. The displacement sensors 71 and 72 in FIG. 1 are also constituted by a pair of radial displacement sensors corresponding to the electromagnets 51x, 51y, 52x, and 52y. These five sets of electromagnets 51x, 51y, 52x, 52y, 53 and displacement sensors 71-73 constitute a 5-axis control type magnetic bearing.

  FIG. 3 is a block diagram showing the basic configuration of the magnetic levitation control system of the magnetic bearing device, and only one axis in the radial direction is shown for the magnetic bearing portion. The electromagnet 5 is a pair of radial electromagnets 51x, and the displacement sensor 7 is a radial displacement sensor 71x corresponding to the electromagnet 51x. The displacement sensor 7 is an inductance type sensor, and converts the gap displacement into an electrical signal by using a change in sensor unit impedance caused by a change in the support position. The sensor facing surface of the rotor 4 is made of a ferromagnetic material or a conductor.

  A controller that drives the pump body 1 is provided with a sensor circuit 2, a control circuit 3, and an excitation amplifier 8. The control circuit 3 includes an A / D converter 301, a D / A converter 302, a DSP (digital signal processor) 307 that is a calculation unit, a storage unit 306 having a ROM 304, a RAM 305, and the like. A carrier wave signal having a frequency fc is applied to the displacement sensor 7 by the sensor circuit 2, and the carrier wave is amplitude-modulated in accordance with a change in impedance of the sensor portion caused by the gap displacement.

  This amplitude modulated wave (AM wave) is input to the control circuit 3 via the sensor circuit 2 as a sensor signal. In the case of the radial displacement sensor 71x, the sensor circuit 2 calculates the difference between the sensor signals from each displacement sensor 71x, and the difference component is input to the control circuit 3 as a sensor signal.

  The analog sensor signal input to the control circuit 3 is converted into a digital value by the A / D converter 301 and input to the DSP 307. A magnetic levitation control constant is input to the storage unit 306 in advance, and the DSP 307 calculates an excitation current to flow through the electromagnet 5 based on the output of the displacement sensor 7 and the control constant. The DSP 307 outputs a control signal corresponding to the excitation current to be supplied. The control signal is converted into an analog value by the D / A converter 302 and then input to the excitation amplifier 8.

  Generally, when an analog signal is converted into a digital signal by an A / D converter and arithmetic processing is performed, in order to prevent aliasing, the sampling frequency is determined based on the sampling theorem. Is set to be at least twice the maximum frequency of the signal input to. This concept is also applied to the conventional magnetic bearing device, and the sampling frequency fs is set so that fs> 2 (fc + fr). Here, fc + fr is the maximum frequency of the sensor signal input to the A / D converter, fc is the carrier frequency, and fr is the modulation frequency due to the rotor displacement.

However, in the case of a turbo molecular pump, the sampling frequency fs can be set to a lower frequency as in the following equations (1) and (2). In the case of a turbo molecular pump, the frequency fr of the rotor displacement signal is limited to a relatively narrow range. If the sampling frequency fs is set so as to satisfy the following expressions (1) and (2), aliasing is performed. It is possible to reduce the arithmetic processing without generating.
(Case 1): (1/2) fs <fc−fr <fc + fr <fs (1)
(Case 2): fs <fc−fr <fc + fr <(3/2) fs (2)

In the demodulation process, the discretized signal is multiplied by a sine wave value of frequency (fs−fc) every sampling period Ts. Therefore, when the sampling frequency fs is set as shown in the following equation (3), the demodulating sine wave discrete value signal may be generated so as to have the same value every time nTs, and the arithmetic processing is performed more easily. be able to. However, n is an integer.
| Fs−fc | = fs / n (3)

  In the case 1 described above, fc = (2/3) fs, fc = (3/4) fs, fc = (4/5) fs, fc = (5 / 6) fs..., But fs / n is the largest when fc = (3/4) fs where the frequency range is 0 to (1/2) fs in the middle (1/4) fs. Can be taken. In case 2, fc = (4/3) fs, fc = (5/4) fs, fc = (6/5) fs, fc = (7/6) fs... ), But fs / n is the middle (1/4) fs in the frequency range 0 to (1/2) fs. When fc = (5/4) fs, the maximum frequency fr is taken. Can do.

  That is, it is preferable to set the sampling frequency fs such that fc = (3/4) fs or fc = (5/4) fs. Hereinafter, a case where the sampling frequency is set to fs = (4/3) fc will be described. However, the present invention can be applied regardless of the set value of the sampling frequency fs.

  4 and 5 are block diagrams showing the configuration of the DSP 307 provided in the sensor circuit 2 and the control circuit 3 shown in FIG. FIG. 4 is a diagram showing the displacement sensors 71x, 71y, 72x, 72y, 73 and the sensor circuit 2, and FIG. 5 shows the control circuit 3. Radial displacement sensors 71x, 71y, 72x, 72y are displacement sensors provided corresponding to the radial electromagnets 51x, 51y, 52x, 52y shown in FIG. FIGS. 6 to 8 show the signal waveforms of the portions indicated by reference numerals (a) in FIGS. 4 and 5. FIG. 8 shows a part of FIG. 7 for easy comparison of the signal waveforms. Shown twice.

  The sine wave discrete value generation unit 320 generates a sine wave discrete value signal (a) as shown in FIG. 6A, and the generated sine wave discrete value signal (a) is converted into a D / A converter 313 and a phase shift unit. 401 is input. The signal (a) is obtained by sampling the sine wave signal asin (2πfc · t) at a time interval Ts / 4, and is expressed as asin (2πfc · nTs / 4). Here, Ts is a period with respect to the sampling frequency fs of the A / D converter 301.

  When the signal (a) is input to the D / A converter 313, the signal (b) as shown in FIG. 6B is output from the D / A converter 313. The waveform of the signal (b) is stepped and contains harmonics. By filtering this signal (b) with a filter circuit 205 such as a low-pass filter or a band-pass filter, the waveform shown in FIG. A smooth carrier signal (c) = csin (2πfct) as shown in FIG. Here, the phase delay due to filtering is ignored for convenience.

  The signal (c) is input as a sine wave carrier signal to three of the five axes of the displacement sensors 71x, 72x, and 73 and also to the phase shift unit 206. The phase shift unit 206 shifts the phase of the input signal by 90 degrees. When the signal (c) is phase-shifted by 90 degrees, the signal (c ′) = c′cos (2πfct), which is a carrier signal of a cosine wave. ). This carrier wave signal (c ′) is input to the other two-axis displacement sensors 71y and 72y, respectively.

  Each radial displacement sensor 71x, 71y, 72x, 72y is provided in pairs so as to sandwich the rotor 4, and the carrier wave signal input to each displacement sensor is amplitude-modulated by an impedance change that changes according to the position of the rotor 4. Thus, an amplitude-modulated wave signal is obtained. The amplitude-modulated wave signal output from each of the pair of displacement sensors 71x is input to the differential amplifier 203, where a differential signal (D) is generated. Difference signals (E), (d), and (e) are also generated in the differential amplifiers 203 for the other displacement sensors 71y, 72x, and 72y.

  In the case of the axial displacement sensor 73, the difference between the amplitude-modulated wave signal from the displacement sensor 73 and the signal (carrier reference signal) obtained by gain correction and phase adjustment of the carrier wave signal (c) by the gain correction unit 202 and the phase shift 204 is the difference. Input to the dynamic amplifier 203. Then, a difference signal between the carrier wave reference signal and the amplitude modulation wave signal is generated.

  Here, when the displacement of the rotor 4 is constant, that is, the differential signals (d), (e), (D), and (E) output from each differential amplifier 203 are respectively (d) = dsin (2πfct + φ ), (E) = ecos (2πfct + φ), (D) = Dsin (2πfct + Φ), and (E) = Ecos (2πfct + Φ). In this case, as described later, when demodulated, a 0 Hz signal (DC signal) is obtained. FIGS. 6D and 6E show the differential signals (d) and (e).

The difference signal (d) output from the differential amplifier 203 related to the displacement sensor 72x and the difference signal (e) output from the differential amplifier 203 related to the displacement sensor 72y are added in the addition processing unit 207 to be a combined signal ( f). The synthesized signal (f) is input to the filter circuit 400, where bandpass filter processing with the carrier frequency fc as the center frequency is performed. In the description of the signal waveform, the processing of the filter circuit 400 is ignored, and the sum of the signal (d) and the signal (e) is output as the synthesized signal (f ′). The synthesized signal (f ′) is expressed as the following equation (4) and has a waveform as shown in (f ′) of FIG. Here, f ′ = √ (d ² + e ² ), φ ′ = φ + arctan (e / d).
(f ′) = dsin (2πfct + φ) + ecos (2πfct + φ)
= F'sin (2πfct + φ ') (4)

Similarly, the difference signal (D) output from the differential amplifier 203 related to the displacement sensor 71x and the difference signal (E) output from the differential amplifier 203 related to the displacement sensor 71y are added together in the addition processing unit 207 to obtain a signal. (F), and band-pass processing with the carrier frequency fc as the center frequency is performed by the filter circuit 400 to obtain a signal (F ′). Here, F ′ = √ (D ² + E ² ) and Φ ′ = Φ + arctan (E / D).
(F ′) = Dsin (2πfc · t + Φ) + Ecos (2πfc · t + Φ)
= F'sin (2πfc · t + Φ ')

  When a sine wave carrier signal is applied to one of the displacement sensors 72x and 72y and a cosine wave is applied to the other displacement sensor 72y, the phase of the complex impedance of the displacement sensors 72x and 72y (inductance L and If the arctan (2πfcL / R)) in the series connection value of the resistor R is substantially the same, the phase difference of the carrier frequency component (in this case, 90 degrees advance or delay) is also generated between the sensor output signals modulated by the displacement equivalent signal. Almost the same phase relationship is maintained.

  Normally, the displacement sensors 71x, 72x, 72x, 72y use sensors having the same structure, and therefore the impedance values of these displacement sensors are substantially the same. As a result, a signal obtained by adding the sensor output signals of the respective displacement sensors selected as a pair becomes an addition value of a sine wave and a cosine wave having the same center frequency, that is, a composite value of two signals having an orthogonal relationship. Since this composite value is composed of two signals having an orthogonal relationship, they can be separated and extracted by performing appropriate arithmetic processing.

The synthesized signal (f ′) output from the filter circuit 400 is converted into a digital value by each A / D converter 301 shown in FIG. The digital signal output from the A / D converter 301 is filtered by the filter processing unit 341 to become a signal (g). Here, the signal waveform will be described ignoring the filter processing unit 341. The discrete value signal (g) obtained as a result of sampling is expressed as the following equation (5). As shown in FIG. 7, n takes values of 0, 1, 2,..., and a circle signal is sampled. Further, g = f ′ and φ ₁ = π−φ ′.
(g) = dsin (2πfcnTs + φ) + ecos (2πfcnTs + φ)
= Dsin {2π (fs−fc) nTs + (π−φ)} + ecos {2π (fs−fc) nTs−φ)}
= Gsin {2π (fs−fc) nTs + φ ₁ } (5)
Here, when the sampling frequency fs of the A / D converter 301 is set as fs = (4/3) fc as described above, the signal (g) has a cycle of 4Ts as shown in FIG. Will have.

On the other hand, the discrete value signal (G) related to the displacement sensors 71x and 71y is expressed as the following equation (6), assuming that the sampling timing is shifted by Δt compared to the case where the discrete value signal (g) is acquired. Is done. Here, G = F ′ and Φ ₁ = π−Φ′−2πfs · Δt.
(G) = Dsin {2π (fs−fc) (nTs + Δt) + (π−Φ−2πfs · Δt)}
+ Ecos {2π (fs−fc) (nTs + Δt) + (− Φ−2πfs · Δt)}
= Gsin {2π (fs−fc) (nTs + Δt) + Φ ₁ } (6)

The signal (g) is obtained by discretizing the synthesized signal (f ′) obtained by adding the signals (d) and (e) by the addition processing unit 207. Then, when the multiplication unit 314 multiplies the signal (g) by the sine wave signal (h) = sin {2π (fs−fc) nTs + (π−φ)} that is in phase with the sine wave signal before addition synthesis, Signal (i) is obtained as shown in the equation.
(i) = dsin ² {2π (fs−fc) nTs + (π−φ)}
+ Ecos {2π (fs−fc) nTs−φ} · sin {2π (fs−fc) nTs + (π−φ)}

  The sine wave signal (h) is obtained by phase shifting the sine wave discrete value signal (a) generated by the sine wave discrete value generation unit 320 by the phase shift unit 401. The displacement equivalent signal (j) = d / 2 relating to the displacement sensor 72x can be extracted by removing the frequency (image frequency) twice the sine wave signal included in the signal (i) by the low-pass filter 315. it can.

Further, in the multiplication unit 314, a cosine wave signal (h ′) = cos {2π (fs−fc) nTs + () obtained by phase-shifting the signal (g) by the phase shift unit 401 with the sine wave discrete value signal (a). Multiplying −φ)}, a signal (i ′) as shown below is obtained. Then, by removing a frequency (image frequency) twice the cosine wave signal included in the signal (i ′) by the low-pass filter 315, a displacement equivalent signal (j ′) = e / 2 regarding the displacement sensor 72y is extracted. be able to.
(i ′) = ecos ² {2π (fs−fc) nTs + (− φ)}
+ Dsin {2π (fs−fc) nTs + (π−φ)} · cos {2π (fs−fc) nTs−φ}

On the other hand, the same processing is performed for the discrete value signal (G) related to the displacement sensors 71x and 71y. That is, in the multiplication unit 314, the sine wave signal (H) = sin {2π (fs−fc) (nTs + Δt) obtained by phase-shifting the signal (G) by the sine wave discrete value signal (a) by the phase shift unit 401. When multiplied by + (π−Φ−2πfs · Δt)}, a signal (I) represented by the following equation is obtained. Then, by processing the signal (I) by the low-pass filter 315, the displacement equivalent signal (J) = D / 2 regarding the displacement sensor 71x can be extracted.
(I) = Dsin ² {2π (fs−fc) (nTs + Δt) + (π−Φ−2πfs · Δt)}
+ Ecos {2π (fs−fc) (nTs + Δt) + (− Φ−2πfs · Δt)}
× sin {2π (fs−fc) (nTs + Δt) + (π−Φ−2πfs · Δt)}

Further, in the multiplication unit 314, a cosine wave signal (H ′) = cos {2π (fs−fc) (nTs + Δt) obtained by phase-shifting the signal (G) by the sine wave discrete value signal (a) by the phase shift unit 401. ) + (− Φ−2πfs · Δt)}, a signal (I ′) like the following equation is obtained.
(I ′) = Ecos ² {2π (fs−fc) (nTs + Δt) + (− Φ−2πfs · Δt)}
+ Dsin {2π (fs−fc) (nTs + Δt) + (π−Φ−2πfs · Δt)}
× cos {2π (fs−fc) (nTs + Δt) + (− Φ−2πfs · Δt)}

  Then, by processing the signal (I ′) by the low-pass filter 315, the displacement equivalent signal (J) = E / 2 regarding the displacement sensor 71y can be extracted. The gain / offset calculation unit 316 performs gain correction and offset correction processing on the signal output from each low-pass filter 315, and the control calculation unit 311 performs control calculation based on the corrected signal.

  By the way, in the prior art, each pair of displacement sensors 71x, 71y, 72x, 72y, that is, for each axis, is sampled by the A / D converter 301 to be converted into discrete values, and sine wave discrete values are generated in these discrete value signals. Demodulation is performed by multiplying the sine wave discrete value signal from the unit 320 to obtain a displacement equivalent signal for each axis.

  However, in the above-described embodiment, the sensor signals for two axes are synthesized, the synthesized signal is A / D converted, and the signal after A / D conversion is again separated into signals for each axis. Therefore, the number of channels used by the A / D converter 301 and the number of band-pass filters 400 provided in each sensor signal line before A / D conversion can be reduced. For example, in the case of the above-described 5-axis control configuration, the number of channels of the A / D converter 301 is conventionally required to be 5, but in this embodiment, since the two axes of the radial displacement sensor are combined into one, there are 3 channels. The number of bandpass filters 400 can be reduced to three.

  As a result, the circuit scale can be reduced and the cost can be reduced. Further, since the number of conversion commands within one sampling period from the arithmetic unit (DSP 37) to the A / D converter 301 is reduced, the waiting time for A / D conversion can be reduced, and the arithmetic processing configuration can be simplified. it can.

  In the above-described embodiment, the sensor signals of the upper radial displacement sensors 71x and 71y and the lower radial displacement sensors 72x and 72y are added, but any of the four axes in the radial direction is added. You may choose and combine the two. Further, a combination of the radial displacement sensors 71x to 72y and the axial displacement sensor 73 may be used. In that case, it is preferable to configure the displacement sensor so that the phases of the complex impedances of the radial displacement sensors 71x to 72y and the axial displacement sensor 73 are substantially the same. In the present embodiment, the magnetic bearing device used in the turbo molecular pump has been described as an example. However, the present invention can be applied to any multi-control type magnetic bearing device.

  In the correspondence between the embodiment described above and the elements of the claims, the displacement sensors 71x and 72x are first detection means, the displacement sensors 71y and 72y are second detection means, and the A / D converter 301 is A / D conversion means, addition processing section 207 as synthesis means, sine wave discrete value generation section 320, phase shift section 401 and multiplication section 314 as first and second demodulation means, and control calculation section 311 as control means. The sine wave discrete value generation unit 320 and the phase shift unit 206 constitute carrier wave generation means, respectively. The above description is merely an example, and when interpreting the invention, there is no limitation or restriction on the correspondence between the items described in the above embodiment and the items described in the claims.

It is sectional drawing which shows schematic structure of the pump main body 1 of a magnetic bearing type turbo molecular pump. It is a conceptual diagram of a 5-axis control type magnetic bearing. It is a block diagram which shows the basic composition of the magnetic levitation control system of a magnetic bearing apparatus. 2 is a block diagram showing a configuration of a sensor circuit 2. FIG. 3 is a block diagram showing a configuration of a DSP 307 provided in the control circuit 3. FIG. It is a figure which shows the signal waveform of signal (a)-(e) and signal (c '), (f'). It is a figure which shows the signal waveform of signal (g)-(j) and signal (h ')-(j'). It is a figure which shows the signal waveform of signal (g), (h), (h '), (G)-(J), (H')-(J ').

Explanation of symbols

4 Rotors 51-53, 51x, 52x, 52x, 52y, 53z Electromagnets 71-73, 71x, 71y, 72x, 72y Displacement sensor 206, 401 Phase shift unit 207 Addition processing unit 301 A / D converter 311 Control operation unit 314 Multiplication Unit 320 sine wave discrete value generation unit 400 filter circuit

Claims (3)

In a multi-axis control type magnetic bearing device for supporting a supported body in a non-contact manner in at least two axial directions,
First detection means for detecting displacement of a support position in the first axial direction of the plurality of axial directions;
Second detection means for detecting a displacement of a support position in a second axial direction of the plurality of axial directions;
Combining means for combining the detection signal from the first detection means and the detection signal from the second detection means into one combined signal;
A / D conversion means for converting the synthesized signal into a digital signal;
First demodulation means for demodulating the digital signal to extract a signal corresponding to the displacement in the first axial direction;
Second demodulating means for demodulating the digital signal to extract a signal corresponding to the displacement in the second axial direction;
A multi-axis control type magnetism comprising control means for controlling the support positions in the first and second axial directions based on the signals extracted by the first and second demodulation means; Bearing device. The magnetic bearing device according to claim 1,
The first and second detection means are displacement sensors to which a carrier wave signal is applied and modulate the carrier wave signal according to the displacement of the support position,
Carrier wave generating means for applying a cosine wave signal obtained by applying a sine wave signal to the first detecting means as the carrier wave signal, and applying a cosine wave signal obtained by phase shifting the sine wave signal by 90 degrees to the second detecting means is provided. A multi-axis control type magnetic bearing device. The magnetic bearing device according to claim 2,
The first demodulating means multiplies the digital signal by a sine wave signal to extract a signal corresponding to the displacement in the first axial direction, and the second demodulating means multiplies the digital signal by a cosine wave signal. Then, a signal corresponding to the displacement in the second axial direction is extracted, and a multi-axis control type magnetic bearing device.

Images