CN103425051A - Unbalance identification and vibration suppression control system for magnetic suspension rotating machinery - Google Patents

Abstract

An unbalance identification and vibration suppression control system for a magnetic suspension rotating machinery comprises an unbalance identification module, an unbalanced force compensation module, a magnetic bearing power amplifier, an electromagnet rotor and a displacement sensor. Based on stable control of a magnetic suspension rotor, the unbalance of a magnetic bearing is identified in an online manner by a novel wave trap based on coordinate transformation, on one hand, the identification amount is used for compensating common-frequency current stiffness force, on the other hand, proper common-frequency current stiffness force is generated according to the identification amount to compensate common-frequency displacement stiffness force, and the influence of the low-pass characteristic of the power amplifier on compensation precision of the common-frequency displacement stiffness force is eliminated by leading a simplified inverse model of the magnetic bearing power amplifier into a feed-forward channel. When the magnetic suspension rotor rotates at a high speed, common-frequency bearing force is greatly reduced, and unbalanced vibration of the magnetic suspension rotor is remarkably suppressed. The unbalance identification and vibration suppression control system is simple, convenient, easy and particularly suitable for an actual high-speed magnetic suspension rotor system.

Description

A kind of amount of unbalance identification of magnetic levitation rotating machinery and vibration suppression control system

Technical field

The amount of unbalance identification of a kind of magnetic levitation rotating machinery of the present invention and vibration suppression control system, can, for the High Precision Automatic balance of magnetic suspension rotor, be particularly useful for the leptosomatic rotors such as magnetic suspension high speed motor.

Background technology

Magnetic suspension bearing is a kind of novel high-performance bearing, utilize controllable electric magnetic field that rotor stability is suspended in to given position, have contactless, without friction, without lubricated, high precision, long-life and the damping stiffness advantage such as controlled initiatively, be the desirable supporting way of rotor-support-foundation system, now be widely used in the fields such as industry, military affairs and Aero-Space.

Due to machining accuracy, can not accomplish that the mass distribution of rotor, geometrically definitely evenly, causes its geometrical axis and the axes of inertia not to overlap.And the size of centrifugal force is directly proportional to rotating speed square, therefore, even rotor very little deviation when High Rotation Speed also can produce very large centrifugal force, so reduce support accuracy, restriction rotating speed raising, cause vibration, so can shorten the serviceable life of magnetic bearing and produce noise.Therefore it is a core technology in the magnetic bearings control technology that the unbalance vibration of magnetic bearing rotor-support-foundation system suppresses.

The unbalance vibration method of magnetic suspension rotor system can roughly be divided into two classes: the principal axis of inertia, with frequency displacement removing method with frequency bearing removing method, is called for short displacement elimination and power and eliminates.

The displacement null method also claims the on-line dynamic balancing method usually, normally adopt armature spindle is increased the weight of or duplicate removal makes its geometrical axis overlap with the axes of inertia, the method implements complexity, need to drop into a large amount of man power and materials, and, when rotor has small variation, the work in early stage had just been lost efficacy, and obtained rebalancing, therefore the method is unfavorable for promoting the use of on a large scale, existing only use at the heavy-duty motor that requires armature spindle to have very high pointing accuracy or mechanical bearing to support etc.The power null method is the supporting gap that utilizes magnetic bearing and rotor, same frequency component in the filtering displacement signal, its displacement rigidity power is being compensated, make armature spindle around its principal axis of inertia rotation, and then the unbalance vibration of inhibition rotor, its core technology is exactly the identification to unbalanced component, and main method comprises: broad sense trapper, Neural Network Self-learning algorithm, LMS algorithm etc.Although with frequency electric current minimizing method, unbalance vibration is had to very strong inhibition ability, the method can't be eliminated the same frequency bearing that the magnetic bearing displacement rigidity causes.For eliminating fully with the frequency bearing, patent ZL200710176720 " a kind of magnetic levitation flywheel high precision initiative vibration control system " increases by one from the compensation path that is displaced to electric current (gain of compensation path is displacement rigidity and the ratio of current stiffness) on the basis with electric current minimization frequently, go to zero thereby make to turn the once per revolution vibration power at place frequently, realize that the magnetically levitated flywheel rotor rotates around the principal axis of inertia.This method can well suppress out-of-balance force for the low speed rotation period of the day from 11 p.m. to 1 a.m, but for high speed rotor, because the low-pass characteristic of magnetic bearing control system power amplifier link is remarkable, use the method to have very large high frequency attenuation and hysteresis to the displacement rigidity force compensating, compensation precision significantly descends.Above-mentioned impact enlarges markedly along with the rising of rotating speed, has reduced the inhibition of unbalance vibration, and therefore, the method can not meet the effective inhibition to the high speed magnetic suspension motor unbalance vibration.

Summary of the invention

The technical matters that the present invention solves is: the defect that overcomes existing high speed magnetic suspended rotor unbalance vibration inhibition method, provide a kind of novel trapper based on coordinate transform to the online real-time identification of same frequency component in displacement, introduce the simplification inversion model of magnetic bearing power amplifier in the feedforward compensation passage, eliminate the impact of the low-pass characteristic of magnetic bearing power amplifier on displacement rigidity force compensating precision, realize that the unbalance vibration of high speed magnetic suspended rotor suppresses.

Technical solution of the present invention is: a kind of amount of unbalance identification of magnetic levitation rotating machinery and vibration suppression control system comprise controller, power amplifier, electro-magnet rotor, displacement transducer, novel trapper, feedforward compensation.Wherein displacement transducer detects the position of electro-magnet rotor in real time, and the rotor displacement amount is converted into to voltage signal output; The output signal of novel trapper module received bit displacement sensor, pick out the once per revolution vibration signal in displacement signal, as the input signal of current stiffness power and displacement rigidity force compensating; The feedforward compensation module, by introducing the inversion model of power amplifier, in the control signal that it is suitable that the power amplifier input end superposes, makes power amplifier produce a current signal, and this current stiffness power just in time can compensate for displacement rigidity power; Controller receives reference-input signal, Displacement Feedback signal and same bit shift compensation signal frequently, and the operation magnetic bearing is stablized control algolithm, produces not containing the control signal with frequency component; Under the non-driving with the same frequency displacement signal that control signal and feedforward compensation produce frequently that the magnetic bearing power amplifier produces at controller, output current signal applies ACTIVE CONTROL to electro-magnet rotor, realizes the stable suspersion of rotor and makes unbalance vibration power be approximately zero.

Described amount of unbalance recognition module is to adopt the novel trapper based on coordinate transformation method that the same frequency component in displacement signal is extracted, be superimposed upon on the one hand the input end of controller with the frequency displacement signal, thereby the filtering controller is the same frequency component of output (1), power amplifier can not produced with the frequency electric current, and then can not produce with frequency current stiffness power; On the other hand by introduce the simplification inversion model of power amplifier in feedforward compensation, utilize identification same frequency component out to produce compensating signal, produce control signal after computing, power amplifier is produced with the frequency offset current, thereby, by current stiffness force compensating displacement rigidity power, make unbalance vibration power minimum or trend towards zero.

Described amount of unbalance recognition module is to adopt the novel trapper based on coordinate transformation method to realize, it realizes that principle is specially: the rotating speed that at first utilizes sensor Real-time Obtaining rotor, and then generation coordinate transform battle array T (Ω t), the position signalling of the rotor geometric center M that will measure under sensor coordinate system is transformed into that under rotor coordinate, (initial point of this coordinate system is on the axes of inertia of rotor, with rotor, be connected, sensor coordinate system is with the rotating speed rotation of rotor relatively), be shown below:

$\left[\begin{array}{c}{x}_r\\ y_r\end{array}\right]=T\left(\mathrm{\Ωt}\right)\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\Ωt}+\mathrm{\θ})& -\sin (\mathrm{\Ωt}+\mathrm{\θ})\\ \sin (\mathrm{\Ωt}+\mathrm{\θ})& \cos (\mathrm{\Ωt}+\mathrm{\θ})\end{array}\right]\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]$

[x wherein _sy _s] ^HFor the positional value that geometric center M measures under sensor coordinate system, [x _ry _r] ^HFor the coordinate figure of geometric center M under rotor system, the angular velocity that Ω is rotor, t is the time, the θ initial phase;

The movement locus of the geometric center M of rotor under sensor coordinate system is circle, and it is transformed into to its value under rotor coordinate will be a normal value, by low-pass filter, it be extracted, and the transport function G of low-pass filter (s) is:

$G\left(s\right)=\frac{1}{\frac{1}{2\mathrm{\π\Δf}}s+1}$

The bandwidth that wherein Δ f is low-pass filter, s is Laplace operator;

Direct current signal that filtering obtains, through the coordinate inverse transformation, has become again AC signal, and this AC signal namely with rotating speed with vibration signal frequently, realized being shown below with the identification of displacement signal frequently:

$\left[\begin{array}{c}{\hat{x}}_c\\ {\hat{y}}_c\end{array}\right]=\hat{T}\left(\mathrm{\Ωt}\right)\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\Ωt}+\mathrm{\θ})& \sin (\mathrm{\Ωt}+\mathrm{\θ})\\ -\sin (\mathrm{\Ωt}+\mathrm{\θ})& \cos (\mathrm{\Ωt}+\mathrm{\θ})\end{array}\right]\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]$

Wherein ${\left[\begin{array}{cc}{\hat{x}}_r& {\hat{y}}_r\end{array}\right]}^H$ For [x _ry _r] ^HThe direct current signal value extracted after low-pass filter, ${\left[\begin{array}{cc}{\hat{x}}_c& {\hat{y}}_c\end{array}\right]}^H$ For same displacement signal frequently,

Contrary for matrix T (Ω t), the angular velocity that Ω is rotor, t is the time, the θ initial phase.

The inversion model of described power amplifier is that matching obtains its transport function by off-line measurement power amplifier model, then gets its contrary acquisition; The method of testing of power amplifier model is for when the magnetic suspension rotor stable suspersion, fixed amplitude superposes in the output signal of controller, a series of sinusoidal signals that frequency is 1～2000Hz are as swept-frequency signal, obtain the output signal of power amplifier by current sensor measurement, it is current signal, then extract the current signal for excitation swept-frequency signal same frequency by bandpass filter, and then calculate when phase place poor of the amplitude of current signal and swept-frequency signal by the LMS algorithm, draw the Bode figure of power amplifier model, matching obtains the transport function of power amplifier model.

Described stability controller can be analog or digital, and centralized control or decentralised control can be to adopt PID to control, the control method of Sliding mode variable structure control or other various applicable and magnetic bearing stable suspersions.

Ultimate principle of the present invention is: during the magnetic suspension rotor rotation, the rotor inertia main shaft causes that with not overlapping of geometrical axis radial direction magnetic bearing produces with frequency bearing (consisting of current stiffness power and displacement rigidity power two parts).In order to reach, the unbalance vibration of high speed magnetic suspended rotor is suppressed, propose to suppress and same frequency current stiffness power that backoff algorithm makes radial direction magnetic bearing with displacement rigidity power frequently to make a concerted effort be zero, that is be zero with bearing frequently.Because high-speed rotating machine has self alignment effect, when once per revolution vibration power is zero, the rotor of High Rotation Speed just will trend towards its principal axis of inertia rotation, thereby reach the inhibition to unbalance vibration.Utilize novel trapper to carry out identification to the rotor radial displacement signal, extract the same frequency component of rotating speed, be used on the one hand the same frequency component of compensate for displacement feedback channel, eliminate current stiffness power, to be input in the feedforward compensation device with the frequency displacement signal on the other hand, compensate with frequency displacement rigidity power by producing with frequency current stiffness power, thereby once per revolution vibration power is gone to zero, reach the purpose that weakens unbalance vibration.

Usually, radial direction magnetic bearing is designed to difference structure, adopts eddy current displacement sensor to measure the displacement of rotor, and controller receives displacement signal, makes the magnetic bearing center that is suspended in of rotor stability through computing output current control signal.Current signal consists of two parts, and a part is bias current i ₀, being used to provide current stiffness, another part is for controlling current i _c, be used for stablizing the magnetic bearing rotor, control electric current opposite direction in two relative coils, form Differential Control.Be that electromagnetic force can be expressed as at this:

$F=\frac{1}{4}{\mathrm{\&mu;}}_0{n}^{2}A[\frac{{(i_0+i_x)}^2}{{(s_0-x)}^2}-\frac{{(i_0-i_x)}^2}{{({\mathrm{<figure-callout\; id="0"\; label="s"\; filenames="HDA0000370113750000013.png,HDA0000370113750000032.png"\; state="\{\{state\}\}">s</figure-callout>}}_0+x)}^2}]$

The number of turn that wherein n is solenoid, A is that magnet surface is long-pending, μ ₀For permeability of vacuum, s ₀Displacement for rotor when the equilibrium position, x is the displacement that rotor departs from equilibrium position.

When armature spindle is made little displacement movement in equilibrium position, can launch electromagnetic force at the Taylor of place, equilibrium position, ignore the high-order term linearity and turn to:

F _m＝k _hx+k _ii _x

Wherein $\left\{\begin{array}{c}{k}_h\stackrel{\mathrm{\&Delta;}}{=}\frac{\&PartialD;F_m}{\&PartialD;x}{|}_{i_c=0,x=0}=\frac{{\mathrm{\&mu;}}_0{n}^{2}A{i}_0^2}{{\mathrm{<figure-callout\; id="0"\; label="s"\; filenames="HDA0000370113750000013.png,HDA0000370113750000032.png"\; state="\{\{state\}\}">s</figure-callout>}}_0^3}\\ k_i\stackrel{\mathrm{\&Delta;}}{=}\frac{\&PartialD;F_m}{\&PartialD;i_c}{|}_{i_c=0,x=0}=\frac{{\mathrm{\&mu;}}_0{n}^{2}A{i}_0}{{\mathrm{<figure-callout\; id="0"\; label="s"\; filenames="HDA0000370113750000013.png,HDA0000370113750000032.png"\; state="\{\{state\}\}">s</figure-callout>}}_0^2}\end{array}\right.$

The same frequency component that identification obtains is on the one hand with removing the same frequency component in the Displacement Feedback link, and the control signal that controller is produced does not comprise with the frequency control signal, therefore the current signal that power amplifier produces does not comprise current stiffness power; On the other hand, according to the size of same frequency component, by producing with current stiffness power frequently, come feedforward compensation with frequency displacement rigidity power, and insertion power amplifier inversion model overcome the impact of the low-pass characteristic of power amplifier model on displacement rigidity force compensating precision in feedforward path.Gained to sum up, the electric current that power amplifier produces is:

$i\left(s\right)=G_w\left(s\right)\left\{G_c\left(s\right)\right[\mathrm{ref}-x_s+{\hat{x}}_c]-G_w^{-1}\left(s\right)k_i^{-1}{k}_s^{-1}{k}_h{\hat{x}}_c\}$

Wherein s is Laplace operator, G _c(s) be the transport function of controller, G _w(s) be the transport function of power amplifier, k _iFor current stiffness, k _hFor displacement rigidity, k _sFor the gain of displacement transducer, the desirable levitation position that ref is magnetic suspension rotor, x _sThe actual levitation position of the magnetic suspension rotor obtained for displacement sensor,

The same frequency component picked out for novel trapper.

Therefore the electromagnetic force of magnetic bearing output is:

$F=k_i{G}_w\left(s\right)G_c\left(s\right)[\mathrm{ref}-x_s+{\hat{x}}_c]+k_s^{-1}{k}_h(x_s-{\hat{x}}_c)$

For x _sIn same frequency component, therefore adopt as can be seen from the above equation the method can eliminate theoretically the unbalance vibration power in magnetic bearing fully, and then overcome the unbalance vibration of magnetic suspension rotor fully.

The present invention's advantage compared with prior art is:

(1) the present invention by the novel trapper based on the changes in coordinates method in displacement signal, with rotating speed, with vibration signal frequently, carrying out on-line identification, novel trapper is compared with traditional broad sense trapper, novel trapper not only has the function with the rotating speed adaptive change, and limit can freely configure the stability of assurance system.In the feedforward compensation passage of displacement rigidity power, introduce the simplification inversion model of magnetic bearing power amplifier, overcome the compensate for attenuation problem of the low-pass characteristic of power amplifier to high frequency displacement rigidity power, improved the compensation precision of displacement rigidity power.

(2) also to have an algorithm simple in the present invention, and operand is little, has overcome the existing calculation of complex to unbalance vibration inhibition method, is not suitable for the defects such as high speed magnetic suspended rotor system, and the inhibition of magnetic bearing unbalance vibration is significantly improved.

The accompanying drawing explanation

The amount of unbalance identification that Fig. 1 is a kind of magnetic levitation rotating machinery of the present invention and the structural representation of vibration suppression control system;

The principle schematic that Fig. 2 is the novel trapper based on coordinate transform of the present invention;

The block diagram of realizing that Fig. 3 is the novel trapper based on changes in coordinates of the present invention;

Fig. 4 is principle of work block diagram of the present invention;

Fig. 5 is magnetic bearing principle of work schematic diagram of the present invention;

Fig. 6 is power amplifier model measurement theory diagram of the present invention;

Fig. 7 is unbalance vibration inhibitory control algorithm flow chart of the present invention;

Fig. 8 is the vibration acceleration spectrogram that does not carry out the magnetic suspension rotor of unbalance vibration inhibition of the present invention;

Fig. 9 is the vibration acceleration spectrogram that adds the magnetic suspension rotor of unbalance vibration inhibition method of the present invention.

Embodiment

As shown in Figure 1, the present invention mainly comprises stability controller 1, power amplifier 2, electro-magnet rotor 3, displacement transducer 4, amount of unbalance identification 5, feedforward compensation 6.Wherein displacement transducer 4 detects the position of electro-magnet rotor 3 in real time, and the rotor displacement amount is converted into to voltage signal output; The displacement signal of amount of unbalance recognition module 5 received bit displacement sensor 4 outputs, pick out the once per revolution vibration signal in displacement signal by the novel trapper based on coordinate transform, as current stiffness power and displacement rigidity force compensating input signal; Feedforward compensation 6, by introducing the inversion model of power amplifier, in the control signal that it is suitable that the power amplifier input end superposes, makes power amplifier 2 produce a current signal, and this current stiffness power just in time can compensate for displacement rigidity power; (1) controller receives reference-input signal, and Displacement Feedback signal and same bit shift compensation signal frequently, implement magnetic bearing and stablize control algolithm, produces non-with the frequency control signal; Magnetic bearing power amplifier 2 receives the control signal that does not comprise same frequency component of controller generation and the same frequency displacement rigidity force compensating signal output current signal of feedforward compensation 6 generations applies ACTIVE CONTROL to electro-magnet rotor 3, realizes the stable suspersion of rotor and makes unbalance vibration power be approximately zero.

As described in Figure 4, as described in Figure 7, the specific implementation step is as described below for the algorithm realization flow for the implementation structure block diagram that the present invention proposes.

Step 1: the off-line measurement matching obtains the power amplifier model

In order to measure accurately the power amplifier model of magnetic bearing, at first guarantee that rotor stability suspends, add fixed amplitude in control signal, the swept frequency excitation signal that frequency is 1～2000Hz, obtain current signal by current sensor measurement, bandpass filter extracts the component with the pumping signal same frequency, utilizes the LMS algorithm to ask for the poor of the ratio of current signal and the amplitude of pumping signal and phase place, and its theory diagram as shown in Figure 6.The suspending reference position that wherein Ref_S is magnetic suspension rotor, k _sFor the Displacement Feedback coefficient, the forward direction amplification coefficient that AMP_KP is the power amplifier link, I_co is the current feedback coefficient, PWM_R is for producing the required dutycycle of fixed bias current (in permanent magnet biased magnetic bearing, this value can think 0), PWM_C is the control signal that controller produces, and makes the magnetic bearing rotor position of tending to balance, and i_PWM_R is bias current value.

The data of finally utilizing matching to obtain, can draw the Bode figure of power amplifier model, and then can obtain the mathematical model of power amplifier model.

Step 2: the identification of unbalanced component

Set the displacement transducer that is mutually 90 ° for stationary coordinate, rotational coordinates and rotor are connected, and with rotating speed Ω, rotate.As shown in Figure 3, coordinate system CX _sY _sFor sensor coordinate system, coordinate system CX _rY _rFor the rotating coordinate system of rotating speed Ω, rotor is, the barycenter that C is rotor is the axes of inertia, the geometric center that M is rotor, and the purpose of imbalance compensation is exactly to make the geometric center M of rotor around its axes of inertia C rotation, reaches transient equilibrium.Because the rigidity of rotor is almost equal at X and Y-direction, thereby the movement locus of geometric center M is just approaching round, as shown in Fig. 2 dotted line.If the coordinate of geometric center M in static coordinate system is: (x _s, y _s), transforming in rotating coordinate system of correspondence is: (x _r, y _r), be easy to draw both relations from figure:

$\left[\begin{array}{c}{x}_r\\ y_r\end{array}\right]=T\left(\mathrm{\&Omega;t}\right)\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})& -\sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})\\ \sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})\end{array}\right]\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]$

The angular velocity that wherein Ω is rotor, t is the time, the θ initial phase.

As can also be seen from Figure, if the rotating speed of rotor is Ω, rotational coordinates and geometric center M rotate around barycenter C with rotating speed Ω all the time, and the coordinate figure of geometric center M in rotational coordinates will be normal value, and the signal that enters wave filter also is approximately D. C. value.

For the wave filter of native system, adopted the first order inertial loop with low-pass characteristic, its transport function is:

$G\left(s\right)=\frac{1}{\frac{1}{2\mathrm{\&pi;\&Delta;f}}s+1}$

The bandwidth that wherein Δ f is low-pass filter, s is Laplace operator.

If the rotational frequency that f is rotor, the unbalance vibration frequency, can measure acquisition in real time by Hall element uniform velocity sensor.If f _iFor the frequency of actual displacement signal, by being permitted, the stack of multifrequency signal forms this value.Due to the existence of coordinate transform battle array T (Ω t), after coordinate transform, the frequency of signal is f _i-f, angular frequency is 2 π (f _i-f).So, the frequency characteristic of first order inertial loop is:

$G\left(\mathrm{j\&omega;}\right)=\frac{1}{1+\frac{1}{2\mathrm{\&pi;\&Delta;f}}\mathrm{j\&omega;}}$

In formula, ω=2 π (f _i-f), j is complex operator, and π is the circular constant constant.Above formula turns to:

$G\left(\mathrm{j\&omega;}\right)=\frac{1}{1+\frac{1}{2\mathrm{\&pi;\&Delta;<figure-callout\; id="2"\; label="f"\; filenames="HDA0000370113750000011.png"\; state="\{\{state\}\}">f</figure-callout>}}2\mathrm{\&pi;j}(f_i-f)}=\frac{1}{1+j\frac{f_i-f}{\mathrm{\&Delta;f}}}$

For the vibration signal with the same frequency of rotating speed, its frequency f _i≈ f, | f _i-f| ≈ 0:

$G\left(\mathrm{j\&omega;}\right)=\frac{1}{1+j\frac{f_i-f}{\mathrm{\&Delta;f}}}\&ap;1$

And for frequency higher harmonic wave and noise signal, its frequency f _iF, | f _i-f|>>Δ f:

$G\left(\mathrm{j\&omega;}\right)=\frac{1}{1+j\frac{f_i-f}{\mathrm{\&Delta;f}}}\&ap;\frac{1}{j\frac{f_i-f}{\mathrm{\&Delta;f}}}=\frac{\mathrm{\&Delta;f}}{j(f_i-f)}\&ap;0$

In sum, can find out with rotating speed can be decayed not with vibration signal frequently and pass through wave filter, and the higher noise signal of frequency will have very large decay while passing through wave filter, the degree of its decay is to be directly proportional to the height of frequency.When | f _i-f|>>during 10 Δ f, such high-frequency signal just has been difficult to by wave filter.

Being transformed into stationary coordinate from rotational coordinates is:

$\left[\begin{array}{c}{\hat{x}}_c\\ {\hat{y}}_c\end{array}\right]=\hat{T}\left(\mathrm{\&Omega;t}\right)\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})\\ -\sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})\end{array}\right]\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]$

Wherein ${\left[\begin{array}{cc}{\hat{x}}_r& {\hat{y}}_r\end{array}\right]}^H$ For [x _ry _r] ^HThe direct current signal value extracted after low-pass filter, ${\left[\begin{array}{cc}{\hat{x}}_c& {\hat{y}}_c\end{array}\right]}^H$ For same displacement signal frequently,

Contrary for matrix T (Ω t), the angular velocity that Ω is rotor, t is the time, the θ initial phase.

The direct current signal that filtering obtains, through the coordinate inverse transformation, has become again AC signal.And this AC signal namely with rotating speed with vibration signal frequently, realized with the identification of displacement signal frequently.

Step 3: according to the principle of step 1 and step 2, write control algolithm, the operation magnetic bearing is stablized control algolithm

Vibration suppression control algolithm flow process as shown in Figure 7.At first the AD by controller collects the displacement signal that the magnetic bearing rotor departs from equilibrium position, then the ECAP by controller catches or the outside rotating speed of inputting the magnetic bearing rotor, be input to uneven recognition module 5, displacement signal carried out to online real-time identification and obtain the once per revolution vibration signal.The once per revolution vibration signal that utilization picks out, the same frequency component be used on the one hand in filtering controller 1 input signal, stablize control algolithm for its operation magnetic bearing, such as existing ripe Decentralized PID control algolithm, produce the control signal that does not contain homogenous frequency signal, thereby suppressed with the generation of current stiffness power frequently; Be entered on the other hand feedforward compensation 6 passages, utilize the power amplifier model of determined off-line, get that it is contrary, carry out anticipatory control, produce with compensation rate frequently, with current stiffness force compensating frequently with displacement rigidity power is brought frequently unbalance vibration.Finally by non-, with the frequency controlled quentity controlled variable and with compensation rate stack frequently, synthesize the overhead control amount that magnetic bearing controller is exported, control power amplifier module 2 output current signals, make the unbalance vibration power of electro-magnet rotor 3 systems outputs as much as possible little.

For the effect that illustrates that unbalance vibration suppresses, Fig. 9, Figure 10 have provided magnetic suspension turbine blade rotor that rotating speed is 30000r/min (being that rotational frequency is 500Hz) and have been contrasted carrying out the vibration acceleration frequency spectrum of unbalance vibration before and after suppressing.Wherein, Fig. 9 is rotating speed while being 30000r/min, does not carry out the vibration acceleration frequency spectrum of the magnetic suspension turbine blade rotor of unbalance vibration inhibition, and its 500Hz component is larger, reaches 0.0279g (g is acceleration of gravity); Figure 10 is that rotating speed is while being 30000r/min, after the unbalance vibration Restrainable algorithms that adds the present invention to propose, the vibration acceleration frequency spectrum of magnetic suspension turbine blade rotor, its 500Hz component decrease is to 0.00276g, be only in Fig. 9 the 500Hz component 9.88%, the component of all the other frequencies is without significant change, and the once per revolution vibration that the magnetic suspension turbine blade rotor is described is that unbalance vibration is significantly suppressed, and has reached the purpose of experiment.

Amount of unbalance recognition module 5 of the present invention and feedforward compensation module 6 adopt DSP by the software programming Digital Implementation in the present embodiment, can also adopt the realizations such as CPLD, FPGA in actual applications.Stability controller 1 can be analog or digital, and centralized control or decentralised control can adopt PID control, Sliding mode variable structure control or other various applicable control methods.

In a word, the present invention is on the stable basis of controlling of magnetic suspension rotor, and by the novel trapper based on coordinate transform, to the on-line identification of magnetic bearing amount of unbalance, the identification amount is used for compensating with frequency current stiffness power on the one hand; Give birth to suitable same frequency current stiffness force compensating with frequency displacement rigidity power according to the identification volume production on the other hand, and pass through to introduce the simplification inversion model of magnetic bearing power amplifier in feedforward path, eliminate the low-pass characteristic of power amplifier to the impact with frequency displacement rigidity force compensating precision.When the magnetic suspension rotor High Rotation Speed, significantly reduced with the frequency bearing, the unbalance vibration of magnetic suspension rotor has obtained significant inhibition.The present invention is simple and easy to do, is particularly suitable for actual high speed magnetic suspended rotor system.

Non-elaborated part of the present invention belongs to the known prior art of those skilled in the art.

Claims (4)

1. the amount of unbalance identification of a magnetic levitation rotating machinery and vibration suppression control system, is characterized in that: comprise controller (1), power amplifier (2), electro-magnet rotor (3), displacement transducer (4), amount of unbalance identification (5) and feedforward compensation (6); Wherein displacement transducer (4) detects the position of electro-magnet rotor (3) in real time, and the displacement of rotor is converted into to voltage signal output; The displacement signal of amount of unbalance recognition module (5) received bit displacement sensor (4) output, pick out the once per revolution vibration signal in displacement signal by the trapper based on coordinate transformation method, as the input signal that produces offset current rigidity power and displacement rigidity power module; In feedforward compensation passage (6), by introducing the inversion model of power amplifier, in the control signal that it is suitable that the power amplifier input end superposes, make power amplifier (2) produce a current signal, this current stiffness power just in time can compensate for displacement rigidity power; Controller (1) receives reference-input signal, Displacement Feedback signal and same bit shift compensation signal frequently, and the operation magnetic bearing is stablized control algolithm, produces control signal; Magnetic bearing power amplifier (2) receives the non-of controller generation, with the same frequency displacement rigidity force compensating control signal output current signal that control signal and feedforward compensation (6) produce frequently, electro-magnet rotor (3) is applied to ACTIVE CONTROL, realizes the stable suspersion of rotor and makes unbalance vibration power be approximately zero.

2. the amount of unbalance identification of a kind of magnetic levitation rotating machinery according to claim 1 and vibration suppression control system, it is characterized in that: the identification of amount of unbalance is based on the novel trapper of coordinate transformation method and realizes, be specially: the rotating speed that at first utilizes sensor Real-time Obtaining rotor, and then generation coordinate transform battle array T (Ω t), the position signalling of the rotor geometric center M that will measure under sensor coordinate system is transformed under rotor coordinate, is shown below:

$\left[\begin{array}{c}{x}_r\\ y_r\end{array}\right]=T\left(\mathrm{\&Omega;t}\right)\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})& -\sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})\\ \sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})\end{array}\right]\left[\begin{array}{c}{x}_s\\ y_s\end{array}\right]$

[x wherein _sy _s] ^HFor the positional value that geometric center M measures under sensor coordinate system, [x _ry _r] ^HFor the coordinate figure of geometric center M under rotor system, the angular velocity that Ω is rotor, t is the time, the θ initial phase;

The movement locus of the geometric center M of rotor under sensor coordinate system is circle, the geometric center M of rotor is transformed under rotor coordinate at the position signalling under sensor coordinate system, the geometric center M of rotor will be a normal value at rotor coordinate upper/lower positions signal, extract the positional information of geometric center M under rotor system of rotor by low-pass filter, the transport function G of low-pass filter (s) is:

$G\left(s\right)=\frac{1}{\frac{1}{2\mathrm{\&pi;\&Delta;f}}s+1}$

The bandwidth that wherein Δ f is low-pass filter, s is Laplace operator;

Direct current signal that filtering obtains, through the coordinate inverse transformation, has become again AC signal, and this AC signal namely with rotating speed with vibration signal frequently, realized being shown below with the identification of displacement signal frequently:

$\left[\begin{array}{c}{\hat{x}}_c\\ {\hat{y}}_c\end{array}\right]=\hat{T}\left(\mathrm{\&Omega;t}\right)\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]=\left[\begin{array}{cc}\cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})\\ -\sin (\mathrm{\&Omega;t}+\mathrm{\&theta;})& \cos (\mathrm{\&Omega;t}+\mathrm{\&theta;})\end{array}\right]\left[\begin{array}{c}{\hat{x}}_r\\ {\hat{y}}_r\end{array}\right]$

Wherein ${\left[\begin{array}{cc}{\hat{x}}_r& {\hat{y}}_r\end{array}\right]}^H$ For [x _ry _r] ^HThe direct current signal value extracted after low-pass filter, ${\left[\begin{array}{cc}{\hat{x}}_c& {\hat{y}}_c\end{array}\right]}^H$ For same displacement signal frequently,

Contrary for matrix T (Ω t), the angular velocity that Ω is rotor, t is the time, the θ initial phase.

3. the amount of unbalance identification of a kind of magnetic levitation rotating machinery according to claim 1 and vibration suppression control system, it is characterized in that: the inversion model of described power amplifier (2) is by off-line measurement power amplifier model, matching obtains its transport function, then gets its contrary acquisition, the method of testing of power amplifier model is for when the magnetic suspension rotor stable suspersion, fixed amplitude superposes in the output signal of controller (1), a series of sinusoidal signals that frequency is 1～2000Hz are as swept-frequency signal, obtain the output signal of power amplifier (2) by current sensor measurement, it is current signal, then extract the current signal for excitation swept-frequency signal same frequency by bandpass filter, and then calculate when phase place poor of the amplitude of current signal and swept-frequency signal by the LMS algorithm, draw the Bode figure of power amplifier model, matching obtains the transport function of power amplifier model.

4. according to amount of unbalance identification and the vibration suppression control system of the described a kind of magnetic levitation rotating machinery of right 1, it is characterized in that: described stability controller (1) is analog or digital, centralized control or decentralised control, or adopt PID to control, or Sliding mode variable structure control.

Images