JPH0610485B2 - Magnetic bearing control device - Google Patents

Description

TECHNICAL FIELD The present invention relates to a magnetic bearing control device suitable for use in space equipment, high-speed rotating machines, and the like.

[Prior Art] As its name suggests, a magnetic bearing supports an object in a non-contact manner by utilizing magnetic attraction and repulsion, and has features such as low friction, low vibration, and low noise. is there. Further, since it can be used for ultra-high speed rotation and rotation in vacuum, it has begun to be used in attitude control devices for artificial satellites, high-speed rotating machines, and the like.

The magnetic bearings are broadly classified into one-axis control type and multi-axis control type according to the number of control axes.

The uniaxial control magnetic bearing is applied to a thrust bearing or the like. As the control device, a PID (proportional / integral / derivative) adjuster is used, and it is common to compensate for unbalanced stiffness and coil current delay together with current minor feedback compensation. Here, the unbalanced stiffness is
In the case of magnetic attraction, when the rotating shaft deviates from the equilibrium state, it acts in a direction that promotes the deviation.Therefore, the value obtained by dividing the unbalanced portion of the magnetic attraction by the amount of displacement, that is, the destabilizing force It means the magnetic spring constant. This corresponds to the coefficient of δj in the equation (14) described later.

On the other hand, a multi-axis control magnetic bearing requires a multi-input / multi-output control device, and the following control system is adopted.

System in which each bearing is independently controlled by PID This is a system in which the above-described uniaxial control system is independently applied to each bearing, and each bearing is independently controlled by PID.

Motion of each axis of the rotor center of gravity is independently PI
D control method This is a method in which PID control is independently performed on the center of gravity of the rotor for movements in three translation directions and two rotation directions.

FIG. 9 is a diagram for explaining this type of system. The center of gravity of the rotor of the control target (magnetic bearing) 120 is the axial direction,
It performs translational motion in the radial direction (two directions) and rotational motion in two directions. These center-of-gravity displacements are detected by the displacement detector 12
1 and supplied to the displacement estimator 122.
However, since the movement in the axial direction can be controlled completely independently of the movement in the other directions, it is omitted in FIG.

The displacement estimator 122 uses the detected amount to estimate the displacement of the center of gravity of the rotor. Is calculated and fed back to the adder 123. The adder 123 subtracts the feedback amount from the center-of-gravity displacement command value and supplies the deviation to the PID controller 124. PID
The output of the controller 124 relates to the command value of the force acting on the center of gravity, which is converted by the load command converter 125 into the command value of the magnetic force acting on the bearing portion and sent to the current command device 126. . The current command device 126 creates a current command value to be supplied to the electromagnet of the bearing unit from the command value, and thereby controls the control target 120. In this way, the rotor is supported by the bearing portion without contact.

Meanwhile, the estimated center of gravity displacement Are respectively supplied to the multiplier 128 through the differentiator 127, are multiplied by the rotation speed sent from the rotation speed detector 129, and are added to the subsequent stage of the PID controller 124.
It is fed back to 30 in a crossed form to compensate for the gyro moment.

Fumio Matsumura et al. "Fundamental equation and control system design of horizontal shaft type magnetic bearing", IEEJ Transactions C, 101, 137 (1)
981). With this control method,
The feedback coefficient is obtained by numerical calculation so that the predetermined evaluation function is minimized.

[Problems to be Solved by the Invention] By the way, the above-described conventional control method has the following drawbacks.

(1) The method of controlling each bearing independently cannot compensate the mutual interference between the left and right bearings and the influence of the gyro moment accompanying the rotation of the rotor.

(2) In the control method in which the motion of the center of gravity of the rotor is independently controlled by PID, when the structure has a bilateral symmetry, the motion of the center of gravity in each axial direction can be controlled. However, in the case of left-right asymmetry, the influence of the unbalanced rigidity of the left and right bearings affects the bearings on the opposite side, causing mutual interference, and this control method cannot compensate for this.

Also, when the rotor starts rotating, a gyro moment is generated, and the horizontal and vertical rotational movements cause mutual interference. In order to compensate for this, in the conventional method shown in FIG. Is being fed back to compensate. That is, the angular displacement The gyro moment was compensated by differentiating from to obtain the angular velocity and feeding it back.

However, with this gyro-moment compensation, the angular displacement is Since the gyro moment compensation was performed only by the value obtained by differentiating, the delay of the coil of the electromagnet could not be compensated.
As a result, there has been a problem that the response of the coil is delayed and the response is insufficient.

The above-mentioned mutual interference between the left and right bearings and the mutual interference around each axis due to the gyro moment corresponds to the appearance of an off-diagonal term in the coefficient matrix of the feedback loop of the displacement of the center of gravity. It will be described later.

(3) With the control method based on the modern control theory, the disadvantages of the above (1) and (2) are solved, but the optimum value of the feedback coefficient also changes when the rotational speed changes, so that optimum control cannot be performed. Will end up. Further, since it is necessary to take feedback of all the state variables, a large number of feedback loops are required, and it is difficult to set the feedback coefficient obtained by numerical calculation and to make fine adjustments.

The present invention has been made under such a background, and an object thereof is to provide a magnetic bearing control device capable of compensating for mutual interference due to a gyro moment and always performing optimum control.

[Means for Solving the Problems] In order to solve the above problems, according to the present invention, a rotor is supported in a contactless manner by controlling an electric current flowing through an electromagnet arranged in a bearing portion. In the magnetic bearing described above, a displacement detector that detects the radial displacement of the rotary shaft of the rotor at at least two points and outputs the detection results of the displacement detector, the output of the displacement detector, the displacement detector, and the A displacement estimator that performs a calculation based on the mutual positional relationship of the center of gravity of the rotor to estimate the center of gravity displacement of the rotor, and the estimated center of gravity displacement output from the displacement estimator is I-PD.
An I-PD type main controller that controls the translation and rotational movement of the center of gravity by adding an operation and feeding back the operation, a rotation speed detector that detects the rotation speed of the rotor, and a rotation of the estimated center of gravity displacement. A gyro moment compensator for compensating for a gyro moment generated by the rotation of the rotor by feeding back a value obtained by multiplying the component by I-PD operation and the output of the rotation speed detector. It is characterized by

[Operation] According to the above means, the translation and rotation of the center of gravity of the rotor are
The gyro that is I-PD controlled based on the estimated center-of-gravity displacement instead of the detection result of the displacement detector itself and that is generated by the rotation of the rotor according to the rotation component of the estimated center-of-gravity displacement and the rotation speed of the rotor. The moment is compensated.

It should be noted that each gain of the I-PD may be determined in consideration of the current characteristics of the electromagnet coil of the bearing portion, in which case gyro moment compensation with good responsiveness that allows for delay due to the coil is possible. is there.

Embodiments Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a configuration of a magnetic bearing control device according to an embodiment of the present invention. Before entering the description of this control device, first, referring to FIG. The magnetic bearing to be controlled will be described.

Structure of Magnetic Bearing (Control Target) and Relevant Amounts FIG. 2 is a perspective view showing the structure of the magnetic bearing.

In the figure, 1 is a rotor, and 2 is a rotation axis of the rotor 1. The rotary shaft 2 is supported by the left and right bearings 3l, 3r without contact. The bearings 3l and 3r each have four electromagnets (3l _{1 to} 3l ₄ and 3r _{1 to} 3r in FIG. 1).
₄ ), and try to hold the rotary shaft 2 in the equilibrium position by sucking it. Left and right detectors 4l and 4r are arranged outside the bearings 3l and 3r. Detection unit 4
l and 4r are, for example, eddy current type non-contact displacement gauges, and detect the radial displacement of the rotary shaft 2.

The following stationary coordinate system is defined for the rotor 1.
First, the origin O is the position of the center of gravity G of the rotor 1 in the equilibrium state. In addition, the axis of the rotating shaft 2 in the equilibrium state is x
The axis and the vertically downward direction are the z-axis, and the y-axis is defined so that each of these axes forms a right-handed system. The center of gravity G and the center of the rotor 1 do not always coincide with each other (see FIG. 3 (a)).

FIG. 3 is a diagram showing the positional relationship of the above-mentioned parts.
(a) and (c) are front views and (b) is a plan view. As shown in FIG. 3A, the distance between the center of gravity G and the center of the rotor 1 is l ₀ , and the distance between the center of the rotor 1 and the left and right bearings 3l, 3r are respectively.
Let l ₁ and l ₂ be the distances between the center of the rotor 1 and the detectors 4l and 4r, and are l ₁ ′ and l ₂ ′, respectively. Further, as shown in FIG. 6B, the displacement amount of the rotary shaft 2 in the y-axis direction will be described. That is, the displacement amount in the y-axis direction (displacement amount from the equilibrium position) of the rotary shaft 2 in the left and right bearing parts 3l, 3r is y _l , yr, and the displacement amount in the y-axis direction of the rotary shaft 2 in the left and right detection parts 4l, 4r. Be y _l ′, y
Let r ′. Further, the displacement amount in the y-axis direction of the rotating shaft 2 at the position of the center of gravity G and y _G. Furthermore, as shown in FIG.
The displacements in the axial direction are z _l , zr, z _l ′, zr ′, and z _G , respectively.

Next, in the left bearing portion 3l, the suction force acting on the rotary shaft 2 is f _l1 vertically upward, f _l2 vertically downward, and horizontally leftward.
f _l3, and a f _l4 to the horizontal right side. Also, the right bearing part 3r
, The suction force acting on the rotary shaft 2 is fr ₁ ,
Let fr ₂ , fr ₃ , fr ₄ . By doing this, in the left and right bearings 3l, 3r, the forces fy _l ,
fyr, fz _l , fzr are as follows.

_{fy l = f l3 -f l4,} fyr = fr 3 -fr 4, fz l = f l2 -f l1, fzr = fr 2 -fr 1 ... (1) In addition, the rotor 1 x, y, z axes The amount of rotation around θ _x , θ _y ,
Let θ _{z be} the rotational angular velocity of the rotor 1 around the x-axis, and ω _x .

Transfer Characteristics of Magnetic Bearing Next, the transfer characteristics of the magnetic bearing will be obtained starting from the equation of motion of such a magnetic bearing, and the block diagram thereof will be determined.

(1) Equation of motion Consider the case where the rotor 1 is regarded as a rigid body having an axisymmetric structure and 5 degrees of freedom are controlled except rotation about the x axis.
Here, in the direction of the rotary shaft 2, that is, in the thrust direction (x-axis direction), the amount of displacement is detected, the force is applied,
Since it can be performed independently of the movement in the other direction, it can be controlled independently.

However, the four axes in the radial direction, that is, the radial directions (y-axis and z-axis directions) affect each other. For example, when a suction force acts on the left bearing portion 3l, the right bearing portion 3r also moves.

Further, in general, the positions of the detection parts 4l and 4r and the bearing parts 3l and 4r
Since it is different from the position of 3r, the displacement of the rotary shaft 2 in the left bearing 3l must be estimated from the displacements of the left and right detectors 4l, 4r.

In this way, the left and right movements, even when the suction force is generated,
It is also relevant in detecting displacement. When the rotor 1 starts to rotate, a gyro moment is generated,
The y-axis and z-axis motions also interfere with each other. That is,
The four radial axes are all related.

Now, in FIGS. 2 and 3, the equation of motion of the translational motion of the center of gravity G is as follows.

m _G = Fx, m _G = Fy, m _G = Fz (2) Further, the equation of motion of the rotational motion of the center of gravity G is as follows.

Jy _z -ω _x Jx _y = Mz Jy _y -ω _x Jx _z = My (3) where m is the mass of the rotor 1, Jx and Jy are the x-axis and the moment of inertia about the y-axis (however, z Moment of inertia about axis Jz
= Jy), x _G , y _G , z _G is the displacement amount of the center of gravity G from the equilibrium position in each axial direction, θ _y and θ _z are the displacement amounts around the center of gravity G, and ω _x
Are rotational angular velocities of the rotor 1, F _x , F _y and F _z are forces acting on the center of gravity G, and M _{y and} M _z are moments acting on the center of gravity G.

The equation of motion of the radial 4-axis is expressed as a matrix as follows.

If you write this in direct notation, Then, when Laplace conversion is performed, Becomes This is the equation of motion that describes the motion of the rotor.
In this specification, for convenience, the same symbol is used for both the function in the time domain and the function in the complex frequency domain after Laplace transform.

(2) Bearings 3l, 3r, detectors 4l, 4r, and center of gravity G
The relationship of the displacement amount of the rotary shaft 2 will be described.

First, the relationship between the displacement of the center of gravity and the displacement of the bearing portion is as follows.

Or in direct notation:

Further, the relationship between the displacement of the center of gravity and the displacement of the detection unit is as follows.

In direct notation, Becomes

If you solve equation (6) above, the displacement of the detector Center of gravity displacement estimated from Can be estimated. That is, Becomes In direct notation Becomes

Substituting equation (7) into equation (5), the estimated bearing displacement Can be estimated. That is, Becomes When expressed in direct notation, Becomes

Thus, the detector displacement Estimated values of bearing displacement and center of gravity displacement Can be asked.

On the other hand, the suction force generated in the bearings 3l, 3r (See the formula (1) above) and the generalized force acting on the center of gravity G of the rotor 1. The relationship with and is as follows.

When expressed in direct notation, Becomes

In this way, the relationship between the displacement amount and the force in the detection unit 4, the center of gravity G, and the bearing unit 3 was obtained.

(3) Obtain the suction force generated in the bearings 3l, 3r. The attraction force f of the electromagnet (horse-shoe shape) can be generally expressed as follows.

Here, f is the attractive force, i is the coil current, δ is the size of the gap between the electromagnet and the rotary shaft 2, φg is the magnetic flux, Pg is the permeance of the gap, Ag is the cross-sectional area of the gap, and μ ₀
Is the magnetic permeability in vacuum, and N is the number of turns of the coil.

In the above formula (10), the following formulas hold when the attraction force, the current, and the size of the gap are divided into a steady component and a variable component, respectively.

Here, the stationary component is a value at equilibrium and is shown in capital letters. Further, the variation corresponds to the deviation from the equilibrium state and is shown in lowercase.

Assuming that the variation is sufficiently smaller than the steady component, minute terms of the second order and above can be omitted, and the following equation holds.

Therefore, The following two equations hold.

Here, Fj, Ij, and W are the steady components of the attraction force, the current, and the size of the gap, and fj, i _j , and δj are the variations thereof.
Further, the suffix j takes the _{l1, l2, l3, l4,} r 1, r 2, r 3, r 4. That is, the steady force and the variation of the attraction force, the current, and the size of the gap are determined for each direction of the left and right bearing portions 3l, 3r.

Next, the variation i _j of the coil current, the bearing displacement y _l , yr, z _l ,
zr, bearing attraction forces fy _l , fyr, fz _l , and fzr are calculated.
First, when the directions of the currents that generate the positive attraction forces fy _l , fyr, fz _l , and fzr are taken as the positive directions of the currents iy _l , iyr, iz _l , and izr, the following equation is established.

_{y l = -δ l3 = δ l4} yr = -δr 3 = δr 4 z l = -δ l2 = δ l1 zr = -δr 2 = δr 1 ... (15) iy l = i l3 = -i l4 iyr = ir ₃ = -ir ₄ iz _l = i _l2 = -i _l1 izr = ir ₂ = -ir ₁ (16) From these expressions and expression (14), the following expression is obtained.

If the formula is expressed as a matrix, Becomes Also, when expressed in direct notation, Becomes

Thus, the current fluctuation And suction force fluctuation And the amount of fluctuation in the bearing I got a relationship with.

(4) Current characteristics of electromagnet Current command given to electromagnet And response (actual coil current) The relationship with is expressed by the following formula.

Here, assuming that all four sets of radial bearings have the same characteristics, the following is obtained.

However, Is a unit matrix Note that the following is a specific example of the transfer characteristic G (s).

When there is no current minor feedback As shown in Fig. 8 (a), when there is no current minor feedback, 1 / G (s) = 1 + as. In addition, a represents the time constant of the coil.

When there is proportional current minor feedback In this case, G (s) = K / (1 + as + K) 1 / G (s) = (1 + K) / K + as / K Becomes Note that K is a proportional gain.

In the case of PI current minor feedback In this case, the configuration is as shown in FIG. 8 (c). Here, in order to stabilize the system, if the proportional gain Kp and the time constant T are set to Kp = 2a / T, 1 / G
(S) is as follows.

1 / G (s) = 1 + T 2 s / 2a + T 2 (aT) s 2 / 2a -T 3 s 3 (aT) / 2a + T 4 (aT) s 4 / 2a + ...... PI current minor feedback and first-order lag filter (time Combination with constant T) In this case, FIG. 8 (d) is obtained, and 1 / G (s) is given by the following equation.

1 / G (s) = 1 + T (1 + T / 2a) s + T 2 s 2/2 (5) and synthesizing the above transfer characteristic of the controlled object, the magnetic bearing, i.e. determine the transfer characteristic of the controlled object.

Rewriting the above equations of motion, Becomes

The above formulas (4b), (5a), (6a) and (9a), (17b),
The following equation is obtained from the equation (18a).

here, By substituting these expressions into the expression (19), the following expression is established.

Therefore, Becomes This is the command value of the generalized force that acts on the center of gravity. And center of gravity displacement Is an expression showing a relationship with.

Comparing this equation (23) with the equation of motion (4b), It can be seen that the item of is newly added. this (Hereafter, the unbalanced stiffness matrix Is called bearing displacement It is a term that represents the influence of unbalanced stiffness, and has a component represented by the following equation.

When the mechanical structure of the magnetic bearing is symmetrical, l ₀ = 0, l ₁ = l ₂ , F _l3 + F _l4 = Fr ₃ + Fr ₄ , F _l2 + F _l1 = Fr ₂ + Fr ₁ … (25) Therefore, the unbalanced stiffness matrix Is a diagonal matrix. So this matrix Center of gravity displacement Is fed back, mutual interference of each axis does not occur. However, in the case of left-right asymmetry, the unbalanced stiffness matrix Is non-diagonal, the y-axis and θ _z- axis, and the z-axis and θ _y- axis motions interfere with each other.

Thus, the motion of the center of gravity G expressed by the above equation (23) is The effect of unbalance stiffness expressed by and the gyro moment matrix Since there is an influence of the gyro moment expressed by
The movements of each axis are interrelated.

FIG. 4 is a block diagram showing the transfer characteristic represented by the above equation (23).

In this figure, the part surrounded by the alternate long and short dash line is the control target. In the previous stage of this controlled object, the command value of the force acting on the center of gravity From suction force command value Load command converter (the matrix of equation (21) above) for calculating Equivalent to) and suction force command value To current command value Current command device for calculating (the matrix of the above equation (20) Equivalent to) is provided.

The control target is the current command value According to the coil current (Variation) is made to flow through the electromagnet to attract the bearing. (Variation) is generated. This suction power Forces the center of gravity G of the rotor 1 And the rotor 1 performs the motion of the formula (4b). This allows the center of gravity displacement Is the matrix Positive feedback is provided to the output side of. This is the matrix of unbalanced stiffness Equivalent to.

The system shown in the block diagram of FIG. 4 is the command value of the generalized force that acts on the center of gravity. Center of gravity displacement using the four components of It is a multi-input / multi-output system that outputs four components of.

In this case, the unbalanced stiffness matrix And the gyro moment matrix Are non-diagonal, so each axis interferes with each other. Also the matrix Is a positive system, so it is an unstable system.

Configuration Method of Control Device Next, a control device for controlling the controlled object will be described. As a characteristic of a control system in which a control device is added to a controlled object, it is desirable that a certain one output variable be made non-interfering so that it can be controlled by one input. In addition, it is preferable that the decoupling subsystems have sufficiently good performance in terms of controllability, stability, and quick response.

It is considered to realize these control characteristics by using an I-PD type control device. Hereinafter, when referred to as the I-PD system,
In addition to I-PD, I-PDD ² , I-PDD ² D ³ ...
Each method is included.

The configuration of the control system of the I-PD system is as shown in FIG.
Transfer function Control object represented by (This corresponds to the control object in FIG. 4 with the load command converter and the current command device added)
, The transfer function P (proportional), D (derivative), D ² (2nd derivative)
...... The local feedback is applied to adjust the pole placement to improve the characteristics while controlling the control amount. The target value The feedback is directly connected to the side, and the steady position deviation is controlled to zero by the I (integration) operation. The above transfer function The underline of represents an ascending polynomial of complex frequency s. For example, Becomes

Integral gain included in I-PD controller Proportional gain Toshiyuki Kitamori “I-PD based on partial knowledge of controlled object
System non-interference control system design "Proceedings of measurement and automatic control, 16-1, 1
Design based on the partial model matching method described in 12/117 (1980).

(1) Partial model matching method Consider a reference model having a transfer characteristic represented by the following equation as a control system that sufficiently satisfies design specifications in terms of incoherence, controllability, stability, and quick response.

Is.

Since the inside of the {} in the equation (26) is a diagonal matrix, the input and output are decoupled and divided into subsystems.
Further, by selecting an appropriate coefficient sequence, it is possible to provide each subsystem with appropriate damping and responsiveness.

Here, the relationship between the input and output of the control system when the magnetic bearing to be controlled is controlled by the I-PD type controller is (2
The operation of matching with the reference model of equation 6) is performed. It is called partial model matching method to make the coefficient of each term of the transfer characteristic of this control system equal to the coefficient of the above reference model from the lower order of s to the order according to the compensation element. The design formula for each gain is derived. That is, the proportional element as the compensation element in the feedback loop. When using and, each gain is given by

Here, [] _diag represents a diagonal element of the matrix. (2
If each gain given by equations (7) to (30) is used as a compensation element, a non-interfering and stable control system having the transfer characteristics of the reference model represented by equation (26) can be obtained.

Even when the compensating element in the feedback loop includes a higher order differentiating element, that is, a second-order differentiating element, a third-order differentiating element, etc., each gain is the same as in the equations (27) to (30). Given.

(2) Next, the transfer characteristic of this embodiment shown in FIG.
Matrix of the above I-PD control method By equalizing and, the integral gain To calculate.

First, the coil current characteristic G (s) of the device of this embodiment is expressed by the following equation.

1 / G (s) = g ₀ + g ₁ s (31) This formula strictly holds when there is no current minor feedback or when proportional current minor feedback is used. In addition, in the case of other current minor feedback systems, it holds approximately. This equation (31) corresponds to the block diagram of FIG. 4 (2
Substituting into equation 3), Becomes To the expression Is the matrix of Since they correspond to And the following equations hold.

Therefore, the rise time, the integral gain, the proportional gain, and the derivative gain are defined by the following equations.

Rise time Integral gain Proportional gain Derivative gain Thus, the integral gain, the proportional gain, and the derivative gain were obtained. Here, the first term on the right side of each gain is
Mass matrix The component related to (diagonal matrix), the second term is the gyro moment matrix It is a component related to (non-diagonal matrix). The third term of the proportional gain and the derivative gain is the unbalanced stiffness matrix. It is a component related to (non-diagonal matrix) and compensates for unbalanced stiffness. When this I-PD control device is added to the controlled object shown in FIG. 4, it becomes as shown in FIG.

In FIG. 6, the adder 7 corresponds to the adder on the input side of the integrator in FIG. 5, and the adder 8 corresponds to the adder on the output side of the integrator in FIG. In addition, from the load command converter of FIG. 6 to the displacement estimator, Corresponding to In Fig. 5 Corresponding to.

In FIG. 6, the detector displacements output from the detectors 4l and 4r Center of gravity estimated by Is converted to. This estimated centroid displacement Is directly fed back to the adder 7 on the target value side, and the PD element Is fed back to the adder 8 on the output side of the integrating element. The command value of the generalized force on which the output of the adder 8 acts on the center of gravity Becomes

In this figure, the feedback loop in the controlled object is the unbalanced stiffness matrix. It is represented by. This is equivalent conversion of the feedback loop of FIG.

Where the mass matrix of integral, proportional, and derivative gains Although each element related to is only a diagonal element, the gyro moment matrix And the unbalanced stiffness matrix Each component related to has a non-diagonal component. That is, the gyro moment matrix The effect of the gyro moment expressed by In order to control the axes independently of each other by decoupling the movements of the axes around the center of gravity G that are related to each other by the influence of the unbalanced rigidity expressed by It is necessary to cross it.

Of these, the unbalanced stiffness compensation portion can be a diagonal matrix. That is, it is the unbalanced stiffness compensating portion of both equations (37) and (38). This part can be diagonalized by separating the component with respect to the other part.

For this reason, the PD element represented by the above equations (37) and (38) Divided into sections. Here, the above equations (36) to (38)
At the matrix The term related to the translation of the center of gravity of the rotor 1 and the compensation part of the rotational movement (hereinafter referred to as the main controller) The term relating to the gyro moment compensation part of the rotor 1 (hereinafter referred to as gyro moment compensator) The term concerning the compensation part of the unbalanced rigidity of the rotor 1 (hereinafter,
Unbalanced stiffness compensator), respectively.

Then, the input of the unbalanced stiffness compensator is estimated Bearing displacement estimated from To the load command converter By changing from the front to the back of the As shown in the figure, the configuration is simplified and the adjustment is easy.

That is, as shown in FIG. The coefficient matrix Depending on the bearing displacement And convert this to a PD element Feedback is made to the adder 9 via.
Even with such a configuration, it is completely equivalent to FIG. This allows for a diagonal matrix during the compensation feedback loop. Thus, the unbalanced stiffness compensation portion can be diagonalized.

If it can be diagonalized, the compensating feedback loop of unbalanced stiffness will not cross, the control device will be simplified, and the fine adjustment of the gain in the field will be easy.

In addition, in FIG. 7, the mass matrix of each gain of integral, proportional, and derivative And gyro moment matrix Is the estimated center of gravity displacement Form a feedback loop with

Configuration of the Embodiment By embodying the configuration of FIG. 7, the configuration of FIG. 1 is obtained.

In FIG. 1, 30 is a displacement estimator. The displacement estimator 30 is connected to the detection units 4l and 4r. And
Detector displacement Eq. (7a) conversion is applied to the estimated center of gravity displacement And the displacement of the detection part by converting equation (8a). Bearing displacement estimated from Ask for. That is, the seventh component This is the part corresponding to and. Estimated center-of-gravity displacement found here Is a main controller 40, 50, a gyro moment compensator 6
0, the estimated bearing displacement Are supplied to the unbalanced stiffness compensator 80.

These components 40, 50, 60 are proportional to those in FIG.
The control of the I-PD system is executed corresponding to the differential element and the integral element.

First, the main controllers 40, 50 are And center of gravity displacement command value From, the command value of the generalized force that acts on the center of gravity Is to seek. That is, the main controller 40 is
The adder 7 shown in the figure and the mass matrix in the integral gain I elements 41a and 41b corresponding to components related to and mass matrix in proportional gain and derivative gain PD conversions 42a and 42b corresponding to the components related to
Equipped with an adder 8 and the estimated center of gravity displacement Command value of generalized force that acts on the center of gravity from Is calculated.

The I elements 41a and 41b are the center of gravity displacement command values. Center of gravity displacement estimated from For the deviation minus The PD element 42 is integrated by the gain of (see Expression (36)).
a and 42b are estimated center-of-gravity displacements Is calculated (Equations (37) and (38)). Further, the main controller 50 is also configured in the same manner, and includes the adders 7 and 8 and the I element 51.
a, 51b and PD elements 52a, 52b.

Next, the gyro moment compensator 60 uses the integral gain, the proportional gain, and the differential gain ((36) to (3) described above.
Gyro-moment matrix of equation 8)) The calculation of each component is performed. That is, each speed ω _{x of the} rotor 1 supplied from the rotation speed detector 6,
Matrix by taking the product of the moment of inertia Jx about the x-axis The component ω _x Jx of (see the equations (4) to (4b)),
This is multiplied by the I-PD output to calculate the compensation value of the gyro moment of the equations (36) to (38).

This gyro moment compensator 60 has The rotation related component of, ie Is supplied and the following calculation is performed. First, the deviation between the command value and the estimated value Are respectively supplied from the adder 7 to the I elements 61a and 61b. Also, the estimated value Is the estimated value in the P element 62a and the D element 63a. Are supplied to the P element 62b and the D element 63b, respectively. The outputs of these elements are added and subtracted by the adder 8 and supplied to the multipliers 64a and 64b, respectively. The value ω _x Jx is supplied from the amplifier 65 to the multipliers 64a and 64b, and multiplication is performed.

The above-mentioned P elements 62a and 62b are (a ₁
Amplification is performed by g ₁ / a ₃ σ ² ) times (see the equation (37)). Further, the D elements 63a and 63b are (a ₂ g ₁ /
a ₃ σ-g ₀₎ performs computation of s (the (38) equation). Further, the I elements 61a and 61b are (g ₁ / a ₃
σ ³ ) (1 / s) is calculated (see the above equation (36)).

The compensation value of the gyro moment thus obtained is supplied to the adders 11 and 12, and the moment about the z axis and
The command value of the generalized force that acts on the center of gravity by compensating the moment around the y-axis Is calculated and supplied to the load / dwell command converter 70.

The load command converter 70 is a matrix element of FIG. And the command value of the generalized force acting on the center of gravity calculated by the equation (21). Is the command value of the bearing suction force Is to be converted to. Where the coefficient matrix Inverse matrix of Is given by the following equation.

In order to execute the above calculation, the load command converter 70
Amplifier 71 for multiplying the input by (l ₂ + l ₀ ) / (l ₁ + l ₂ ).
a, 73a, amplifiers 71b, 73b for multiplying (l ₁ −l ₀ ) / (l ₁ + l ₂ ) times, and amplifiers 72a, 72b, 74a, 74b for multiplying 1 / (l ₁ + l ₂ ) times, and addition Container 75a,
It has 75b and adders 76a and 76b. Then, from these adders 75a to 76b, the command value of the bearing portion attraction force is calculated. Is output and supplied to the unbalanced stiffness compensator 80.

The unbalanced stiffness compensator 80 was inserted in the feedback of FIG. Corresponding to the element and the adder 9, and the command value of the above-mentioned bearing portion attraction force Is supplied to the adder 9 and the estimated bearing displacement PD elements 81a, 81b for executing the proportional / derivative calculation of
It is composed of 82a and 82b.

The P element of the PD elements 81a to 82b is the estimated bearing displacement. And the D element is the estimated bearing displacement Is calculated. And the command value of the bearing part suction force Deviation value obtained by subtracting the outputs of PD elements 81a to 82b from Is supplied to the current command device 90.

The current command device 90 is shown in FIG. It corresponds to an element. That is, the amplifiers 91a, 91b, 92a, 92b that amplify the deviation value supplied from the adder 9 and output the command value corresponding to the coil current fluctuation.
And the adders 93a to 96b for adding and subtracting these fluctuation command values from the steady portion of the coil current.
Then, from the adders 93a to 96b, the respective coils 3l ₁ to
3l _4, and _3r command value of the current supplied to the 1 ~3r ₄ is output.

Since the magnetic bearing control device according to this embodiment is matched with the reference model of the I-PD system as described above, good control characteristics can be obtained. That is,
It has excellent characteristics in terms of controllability, stability, and quick response.

Also, the estimated bearing displacement Is fed back to each input side of the PD elements 81a to 82 of the unbalanced stiffness compensator 80 to obtain a diagonal matrix of unbalanced stiffness. Therefore, the compensation control of the unbalanced rigidity can be simplified. This gives the y-axis and θ _z
Mutual interference between the axis or the z axis and θ _y can be avoided.

Furthermore, since the influence of the gyro moment is eliminated by the gyro moment compensator 60, stable control can always be performed regardless of the rotation speed of the rotor 1. In particular, since the gyro moment compensator 60 is adapted to compensate the current delay due to the coil characteristics of the electromagnet, it is possible to perform compensation with good responsiveness.

In the embodiment described above, the main controllers 40, 50
In addition, although the unbalanced stiffness compensator 80 and the gyro moment compensator 60 are used together, the unbalanced stiffness compensator 80 is omitted and only the gyro moment compensator 60 is added to the main controllers 40 and 50. Even with the configuration, in the magnetic bearing whose left and right are asymmetrical, the control corresponding to the actual displacement at the center of gravity position can be performed, and the optimum control can always be performed even if the rotational speed changes.

EFFECTS OF THE INVENTION As described above, the present invention is based on the estimated center-of-gravity displacement, not the detection result of the displacement detector itself.
Since control and gyro moment compensation are performed,
In particular, even with an asymmetrical rotating body, control corresponding to the actual displacement at the position of the center of gravity becomes possible.

Further, since the feedback constant does not depend on the rotation speed, optimum control can always be performed even if the rotation speed changes.

Further, if each gain of the I-PD is determined in consideration of the current characteristics of the electromagnet coil of the bearing portion, it is possible to perform gyro moment compensation with good responsiveness in consideration of a delay due to the coil.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a magnetic bearing control device according to an embodiment of the present invention, FIG. 2 is a perspective view showing a configuration of a magnetic bearing, and FIG. 3 is a concept for explaining a coordinate system of the magnetic bearing. It is a figure, the same figure (a) and (c) is a front view and the same figure (b).
Is a plan view, FIG. 4 is a block diagram showing the transfer characteristics of the magnetic bearing to be controlled, FIG. 5 is a block diagram showing the configuration of the control system of the I-PD system, and FIG. 6 is the control of FIG. I for the target
-A block diagram showing the characteristics of a device configured by adding a PD control device and a detection unit, Fig. 7 is a block diagram showing the properties of the device of the first embodiment, and Fig. 8 is a bearing electromagnet. FIG. 9 is a block diagram showing a characteristic example, and FIG. 9 is a block diagram for explaining a conventional control method. 1 ... Rotor, 2 ... Rotation axis, 3 ... Bearing section, 4 ... Detection section, 30 ... Displacement estimator, 40, 50 ... Main controller,
60 ... Gyro-moment compensator, 70 ... Load command converter, 80 ... Unbalance stiffness compensator, 90 ... Current command device.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (56) Reference JP-A-58-81217 (JP, A) JP-A-59-89821 (JP, A) JP-A-62-35114 (JP, A) JP-A-62- 46016 (JP, A) Proceedings of the Society of Instrument and Control Engineers 16 [1] (Sho 55 ・ 29) 112-117

Claims (1)

[Claims] 1. A magnetic bearing in which a rotor is supported in a contactless manner by controlling an electric current flowing through an electromagnet disposed in a bearing portion, and at least radial displacement of a rotary shaft of the rotor is achieved. Displacement detectors that detect at two locations and output the respective detection results, calculations are performed based on the output of this displacement detector and the mutual positional relationship between the displacement detector and the center of gravity of the rotor, and The displacement estimator for estimating the center of gravity displacement of the rotor and the estimated center of gravity displacement output from the displacement estimator are I-PD.
An I-PD type main controller that controls the translation and rotational movement of the center of gravity by adding an operation and feeding back the operation, a rotation speed detector that detects the rotation speed of the rotor, and a rotation of the estimated center of gravity displacement. A gyro moment compensator for compensating for a gyro moment generated by the rotation of the rotor by feeding back a value obtained by multiplying the component by I-PD operation and the output of the rotation speed detector. A magnetic bearing control device characterized by the above.