CN115225000A

CN115225000A - Compensation device and method for radial run-out component of magnetic suspension rotor position

Info

Publication number: CN115225000A
Application number: CN202210826036.XA
Authority: CN
Inventors: 赵一凡; 赵远; 骆淳
Original assignee: Suzhou Mengnuoda Precision Machinery Technology Co ltd
Current assignee: Suzhou Mengnuoda Precision Machinery Technology Co ltd
Priority date: 2022-07-14
Filing date: 2022-07-14
Publication date: 2022-10-21

Abstract

The invention discloses a compensation device and a compensation method for radial run-out components of a magnetic suspension rotor position, which relate to the technical field of high-speed magnetic suspension motors and comprise a hardware-in-loop simulation system; the hardware-in-the-loop simulation system comprises: the system comprises a high-speed magnetic suspension motor and a control system thereof, wherein the high-speed magnetic suspension motor and the control system thereof form a hardware system physically existing in a hardware-in-loop simulation system, and the system also comprises a simulation system in which hardware formed by dynamic mathematical models of the high-speed magnetic suspension motor and the control system thereof exists in a computer digital space in the loop system. The high-speed magnetic suspension motor comprises: the magnetic suspension rotor assembly, the radial/axial electromagnetic bearing, the rotor assembly radial/axial displacement sensor, the three-phase motor stator and the magnetic suspension rotor assembly angular displacement sensor. The method determines the Fourier coefficient parameters of each subharmonic of the radial runout component in a virtual digital space through hardware-in-loop simulation and artificial intelligence computer optimization, and then reversely reconstructs the radial runout by the obtained Fourier coefficient parameters in an actual system to counteract the adverse effect of the actually existing radial runout component.

Description

Compensation device and method for radial run-out component of magnetic suspension rotor position

Technical Field

The invention relates to the technical field of high-speed magnetic suspension motors, in particular to a device and a method for compensating radial run-out components of a magnetic suspension rotor position.

Background

The rotor of the high-speed magnetic suspension motor rotates in a suspension state under the action of the electromagnetic force of the magnetic suspension bearing, so that the high-speed magnetic suspension motor has the advantages of no mechanical wear, no lubricating oil pollution, low energy consumption, low noise, long service life and the like, and the high-speed magnetic suspension motor replaces the traditional mechanical bearing motor more and more in application scenes with special requirements on high speed, high efficiency, vacuum, ultra-cleanliness and the like. One of the main application fields of the high-speed magnetic suspension motor is various high-speed turbomachines, such as a centrifugal air compressor, a centrifugal refrigerant compressor, and the like.

The high-speed centrifugal compressor equipment adopting the high-speed magnetic suspension motor driving technology is mainly characterized by high efficiency, and compared with the traditional compressors in other forms, the high-speed centrifugal compressor equipment has the advantage that the electric energy is saved by 20-40% generally. One of the key factors for realizing large-scale energy saving of the high-speed centrifugal compressor is high-precision position control of a magnetic suspension rotor assembly (a rotating body comprising a turbine wheel, a motor rotor, a magnetic suspension bearing rotor magnetic circuit and other components) in a suspension state, and one of the key conditions for realizing the high-precision position control is the position detection precision of a rotor assembly radial position sensor. In a practical magnetic suspension system, a radial run-out component often exists in a rotor position signal output by a radial displacement sensor of a magnetic suspension rotor assembly.

The factors that cause radial run-out are mainly: the manufacturing error causes the surface circumference of the electromagnetic target ring of the rotor assembly radial displacement sensor to be eccentric with the geometric rotation axis of the rotor assembly, the surface deformation of the target ring, the non-uniform electromagnetic property of the material for manufacturing the target ring and the like.

The radial run-out component is represented as a harmonic component having a frequency of one and several times the motor speed. Under ideal conditions, the rotor assembly of the magnetic suspension motor is suspended in a motion space in the magnetic suspension bearing in a non-contact way to rotate at a high speed under the position closed-loop control, and the rotating shaft of the rotor assembly is superposed with the geometric center line of the electromagnet of the magnetic suspension bearing. The presence of the radial run-out component will break this balance, causing radial oscillation of the motion of the rotor assembly, causing the actual axis of rotation of the rotor to deviate from the designed axis of rotation. In the case that the load device of the high-speed magnetic suspension motor is a high-speed turbomachine, such as a centrifugal air compressor or a centrifugal refrigerant compressor, in order to avoid the friction between the turbine wheel and the volute due to the deviation of the rotating shaft, the design needs to increase the gap between the turbine wheel and the volute, but the increase of the gap will reduce the efficiency of the compressor.

In addition, the radial oscillation of the rotor assembly can cause overhigh temperature of the electromagnet of the magnetic suspension bearing, and the material of the rotor assembly is subjected to fatigue damage even breakage due to internal periodic stress.

Therefore, reducing the run-out in the detection of the radial position of the rotor assembly is a problem that must be considered in the manufacturing process of the magnetic levitation motor.

The conventional method for reducing radial run-out not only ensures the machining and assembling precision of the rotor assembly, but also needs to obtain radial run-out parameters through a testing means in the manufacturing process of a product, or adds a radial run-out parameter identification or self-adaption function in a control algorithm to compensate residual radial run-out components.

However, the above-mentioned run-out compensation process often needs to be performed in a trial and error manner during the operation of the equipment, so that there is a risk that the equipment is damaged due to improper parameter selection.

Disclosure of Invention

The invention aims to provide a compensation device and a compensation method for radial runout components of magnetic suspension rotor positions, which are characterized in that Fourier coefficient parameters of each subharmonic of the radial runout components are determined in a virtual digital space through hardware-in-loop simulation and artificial intelligence computer optimization, and then the obtained Fourier coefficient parameters are used for carrying out reverse reconstruction on the radial runout in an actual system so as to counteract the adverse effect of the actually existing radial runout components.

In order to achieve the purpose, the invention provides the following technical scheme: a device for compensating for a radial runout component of a magnetically levitated rotor position, comprising: a hardware-in-the-loop simulation system;

the hardware-in-the-loop simulation system comprises: the system comprises a high-speed magnetic suspension motor and a control system thereof, wherein the high-speed magnetic suspension motor and the control system form a hardware system of a hardware-in-loop simulation system in physical existence; a simulation system of hardware in a computer digital space in a ring simulation system is formed by a high-speed magnetic suspension motor dynamic mathematical model, a magnetic suspension control simulation system and a radial run-out compensation parameter computer optimization algorithm;

the high-speed magnetic suspension motor comprises: the device comprises a magnetic suspension rotor assembly, a radial electromagnetic bearing, a rotor assembly radial displacement sensor, a three-phase motor stator, an axial electromagnetic bearing, an axial displacement sensor and a magnetic suspension rotor assembly angular displacement sensor.

Optionally, the rotor assembly radial displacement sensor is used for detecting the position of the magnetic suspension rotor assembly in the axial electromagnetic bearing air gap.

Optionally, the control system is a high-speed magnetic suspension motor controller;

the high-speed magnetic suspension motor controller comprises a rotor assembly radial/axial displacement sensor signal processing circuit, an electromagnetic bearing current PWM power amplifier, a magnetic suspension rotor assembly angular displacement signal processing circuit, a magnetic suspension position closed-loop controller and a motor PWM driving and controller.

Optionally, the magnetic levitation rotor assembly includes:

the device comprises a turbine impeller, a radial magnetic suspension bearing rotor magnetic circuit, a motor rotor, a rotor assembly radial displacement sensor electromagnetic target ring and an axial magnetic suspension bearing rotor;

the magnetic suspension rotor assembly is a rigid body rotator neglecting a flexible mode.

Optionally, the radial electromagnetic bearing, the axial electromagnetic bearing, the rotor assembly radial/axial displacement sensor, the magnetic suspension position closed-loop controller, and the attached sensor signal processing circuit and power amplifying circuit constitute a magnetic suspension control system of the magnetic suspension rotor assembly.

Optionally, the dynamic mathematical model of the high-speed magnetic levitation motor includes:

a position run-out component in the output signal of the parameter adjustable radial position sensor;

and the functional module simulates the dynamic characteristics of a current control ring of the magnetic suspension bearing and the dynamic characteristics of a radial position sensor of the rotor assembly.

Optionally, the hardware-in-the-loop simulation system further includes:

a reverse dynamics control algorithm, the reverse dynamics control algorithm being a radial position closed loop control algorithm of the magnetic levitation rotor assembly.

The use method of the device for compensating the radial runout component of the position of the magnetic suspension rotor comprises the following steps:

s1: determining characteristic parameters of position jumping components in output signals of a radial position sensor of a rotor of the magnetic suspension motor by operating the hardware-in-the-loop simulation system;

s2: then storing the obtained characteristic parameters of the position jumping components in a digital storage unit of the control system of the high-speed magnetic suspension motor;

s3: reconstructing the jumping component in the radial position sensor in real time by using the stored characteristic parameters of the position jumping component in the operation process of the control system of the high-speed magnetic suspension motor;

s4: and injects it back into the radial displacement feedback signal of the control system to achieve compensation for the jitter component in the feedback signal.

Compared with the prior art, the invention has the following beneficial effects:

1. the method determines the Fourier coefficient parameters of each subharmonic of the radial run-out component in a virtual digital space by hardware-in-loop simulation and artificial intelligence computer optimization, and then performs reverse reconstruction on the radial run-out component by using the obtained Fourier coefficient parameters in an actual system to offset the adverse effect of the actually existing radial run-out component, so that the method provided by the invention does not need trial and error in equipment operation, can greatly reduce the risk of equipment damage and improve the automation degree of a production process.

2. The invention utilizes the mechanical, electromagnetic and control system structure of the high-speed magnetic suspension motor, determines the radial position jumping harmonic component existing in the output signal of the radial position sensor of the rotor of the high-speed magnetic suspension motor through the methods of signal acquisition, system simulation and computer artificial intelligent parameter optimization, and eliminates the adverse effect of the radial position jumping on the control of the high-speed magnetic suspension motor through reverse compensation; the purposes of reducing the electromagnet loss of the magnetic suspension bearing, improving the control precision of the rotor position, improving the efficiency of the magnetic suspension motor driving equipment and prolonging the service life of the equipment are achieved.

Drawings

FIG. 1 is a front view of the structure of the present invention;

FIG. 2 is a schematic diagram of a high-speed magnetic levitation motor and a controller thereof according to the present invention;

FIG. 3 is a schematic view of a magnetic levitation rotor assembly of the present invention;

in fig. 4, a is a schematic diagram of a coordinate system involved in radial motion control of the magnetic levitation rotor assembly in the electromagnetic bearing;

b is a schematic diagram of the rotor assembly having six freedom of motion or motion state variables relative to an inertial coordinate system in a suspension state;

c is a coordinate vector diagram of the intersection point of the geometric central axis of the rotor assembly and the sensor coordinate system plane and a vector diagram of the bearing electromagnetic force in the bearing coordinate system plane;

in fig. 5, a is a block diagram of a control system formed by a mathematical model in the form of a state space of a magnetic suspension motor and a closed-loop control of the radial position of a motor rotor assembly by adopting a reverse dynamic algorithm, and does not include the influence of rotor radial position jumping and dynamic balance unbalance;

b is a control system block diagram formed by a mathematical model in a state space form of the magnetic suspension motor and the closed-loop control of the radial position of the motor rotor assembly by adopting a reverse dynamics algorithm, and comprises the influence of rotor radial position jumping and dynamic balance unbalance;

FIG. 6 is a structural diagram of the dynamic simulation of a magnetic levitation motor in a hardware-in-the-loop simulation system according to the present invention;

FIG. 7 is a block diagram of an embodiment of a hardware-in-the-loop simulation compensation device for radial runout of a rotor position of a magnetic suspension motor according to the present invention.

In the figure: 1. a high-speed magnetic suspension motor; 2. a high-speed magnetic suspension motor controller; 3. a dynamic mathematical model of the high-speed magnetic suspension motor; 4. a magnetic suspension control simulation system; 5. a radial run-out compensation parameter computer optimization algorithm;

the high-speed magnetic suspension motor 1 comprises: 1_1, a magnetic suspension rotor assembly; 1_2, radial electromagnetic bearing; 1_3, a rotor assembly radial displacement sensor; 1_4, a three-phase motor stator; 1_5, axial electromagnetic bearing; 1_6, magnetic levitation rotor assembly angular displacement sensor, as shown in fig. 2;

the high-speed magnetic suspension motor controller 2 comprises: 2-1, a signal processing circuit of the rotor assembly radial displacement sensor; 2_2, an electromagnetic bearing PWM power amplifier; 2_3, and an angular displacement signal processing circuit of the magnetic suspension rotor assembly; 2_4, a magnetic suspension position closed-loop controller; 2_5, a motor PWM drive and controller;

the magnetic suspension rotor assembly 1_1 comprises: 1_1, turbine impeller; 1_1 \u2, radial magnetic suspension bearing rotor magnetic circuit; 1_1_3, motor rotor; 1_1 \u4, and a rotor assembly radial displacement sensor electromagnetic target ring; 1_1 \u5, axial magnetic suspension bearing rotor.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

Referring to fig. 1 to 7, the present embodiment provides a technical solution: a device for compensating the radial run-out component of the position of a magnetic levitation rotor comprises: a hardware-in-the-loop simulation system;

the hardware-in-the-loop simulation system comprises: the high-speed magnetic suspension motor 1 and the control system 2 thereof, wherein the high-speed magnetic suspension motor 1 and the control system 2 form a hardware system of a hardware-in-loop simulation system in physical existence;

a high-speed magnetic suspension motor dynamic mathematical model 3, a magnetic suspension control simulation system 4 and a radial run-out compensation parameter computer optimization algorithm 5 form a simulation system of which hardware exists in a computer digital space in a ring simulation system.

More specifically, in the present embodiment: the device utilizes the self mechanical, electromagnetic and control system structures of a high-speed magnetic suspension motor 1 and a control system 2 thereof, determines the radial position jumping harmonic component existing in the output signal of a radial position sensor of a rotor of the high-speed magnetic suspension motor through a signal acquisition, system simulation and computer artificial intelligent parameter optimization method, and then eliminates the adverse effect of radial position jumping on the control of the high-speed magnetic suspension motor through reverse compensation, thereby achieving the purposes of reducing the electromagnet loss of a magnetic suspension bearing, improving the control precision of the rotor position, improving the efficiency of magnetic suspension motor driving equipment and prolonging the service life of the equipment;

specifically speaking: in a hardware system formed by a high-speed magnetic suspension motor 1 and a control system 2 thereof, wherein hardware physically exists in a ring simulation system, a magnetic suspension rotor assembly 1_1 is controlled by a magnetic suspension motor controller 2 to surround a geometric central shaft x of the magnetic suspension rotor assembly _r o _r Rotates at a low speed at a constant rotation speed omega, and the geometric central axis of the rotor and the geometric central line x of the electromagnet _i o _i Overlapping;

radial run-out compensation parameter computer optimization algorithm 5:

firstly, determining the frequency spectrum characteristics of a position jumping component in an output signal of a rotor assembly radial displacement sensor 1_3 through fast Fourier transform; determining a main harmonic component of the position beating component according to the frequency spectrum characteristic, namely a pilot frequency spectrum; carrying out position jitter component characteristic parameter computer parameter optimization in a targeted manner according to harmonic waves in the pilot frequency spectrum;

respectively extracting characteristic parameter vectors of disturbance components caused by the position jitter in output signals of rotor assembly radial displacement sensors 1_3 of an actual physical system and a computer simulation system by adopting a Fourier series expansion method, and forming a target function of a computer optimization algorithm by using a vector difference value of the characteristic parameter vectors and the characteristic parameter vectors;

adopting but not limited to Hu Keji Weiss Algorithm (Hooke-Jeeves Algorithm) as the search engine of computer optimization;

and performing parameter optimization on the characteristic parameter vectors of the harmonic waves in the pilot frequency spectrum one by utilizing a linear system superposition principle, and then calculating a compensation value of the total position jitter component according to the superposition principle.

Further, in the present embodiment: the high-speed magnetic levitation motor 1 includes:

magnetic levitation rotor assembly 1_1, radial electromagnetic bearing 1_2, rotor assembly radial displacement sensor 1_3, three-phase motor stator 1_4, axial electromagnetic bearing 1_5 and axial displacement sensor, and magnetic levitation rotor assembly angular displacement sensor 1_6;

rotor assembly radial displacement sensor 1_3 is used to detect the position of magnetically levitated rotor assembly 1_1 in the air gap of radial electromagnetic bearing 1_2.

Further, in the present embodiment: the control system is a high-speed magnetic suspension motor controller 2;

the high-speed magnetic levitation motor controller 2 comprises a rotor assembly radial displacement sensor signal processing circuit 2_1, an electromagnetic bearing current PWM power amplifier 2_2, a magnetic levitation rotor assembly angular displacement signal processing circuit 2_3, a magnetic levitation position closed-loop controller 2_4 and a motor PWM driving and controlling device 2_5. The rotor assembly axial position control system is also an indispensable part for controlling the high-speed magnetic suspension motor, but the invention is not described in detail because of low relevance.

Further, in this implementation: magnetic levitation rotor assembly 1_1 comprises:

1 \u1 of a turbine impeller, 1_1 _2of a radial magnetic suspension bearing rotor magnetic circuit, 1_1 _3of a motor rotor, 1_1 _4of an electromagnetic target ring of a rotor assembly radial displacement sensor and 1_1 _5of an axial magnetic suspension bearing rotor;

the magnetic suspension rotor assembly 1_1 is a rigid body rotator with flexible modes omitted.

Further, in this implementation: the radial electromagnetic bearing 1_2, the rotor assembly radial displacement sensor 1_3, the magnetic suspension position closed-loop controller 2_4, and the attached sensor signal processing circuit and power amplification circuit form a magnetic suspension radial position control system of the magnetic suspension rotor assembly 1_1.

More specifically, in the present embodiment: the magnetic suspension rotor assembly 1_1 is supported in a space in a magnetic suspension bearing in a non-contact mode, the geometric center line of the magnetic suspension rotor assembly 1_1 is coincident with the geometric center line of an electromagnet, and the magnetic suspension rotor assembly 1_1 performs constant-speed rotary motion in a suspension state under the control of a motor PWM driving and control device 2_5.

Further, in the present embodiment: the dynamic mathematical model 3 of the high-speed magnetic suspension motor comprises:

more specifically, the method can be used for performing computer simulation on the response of the high-speed magnetic suspension motor 1 and the control system 2 thereof to the jumping component;

the functional module simulates the dynamic characteristics of a current control ring of the magnetic suspension bearing and the dynamic characteristics of a radial position sensor of the rotor assembly;

more specifically, the functional module adopts, but is not limited to, an inertial filter structure, and the filtering time constant of the inertial filter structure is respectively set according to the bandwidth and the transmission delay of the magnetic bearing current control ring and the rotor assembly radial position sensor in the actual system, so that the dynamic characteristic of the simulation system is closer to the actual system.

Further, in this implementation: the hardware-in-the-loop simulation system further comprises:

and the reverse dynamics control algorithm is a radial position closed-loop control algorithm of the magnetic suspension rotor assembly 1_1.

More specifically, in the present embodiment: the control structure of 'proportional differential control + rotor gyroscopic effect compensation + magnetic suspension bearing negative stiffness characteristic compensation' designed according to a radial motion mathematical model of a rotor assembly is adopted;

in order to detect the position jumping component in the output signal of the rotor radial position sensor, the control setting of the radial position closed loop is to make the geometric central axis of the rotor assembly coincide with the geometric central line of the magnetic suspension bearing electromagnet and to make the rotor assembly rotate at a low speed with the centrifugal force generated by the dynamic balance unbalance of the rotor assembly negligible.

A method for compensating the radial runout component of the position of a magnetically levitated rotor includes the following steps:

s1: determining characteristic parameters of position jumping components in output signals of a radial position sensor of a rotor of a magnetic suspension motor by operating hardware in a ring simulation system;

s2: then storing the obtained characteristic parameters of the radial position jumping component in a digital storage unit of a control system of the high-speed magnetic suspension motor;

s3: the jumping component in the radial position sensor is reconstructed in real time by using the stored characteristic parameters of the jumping component of the radial position in the running process of a control system of the high-speed magnetic suspension motor;

Specific description and operation principle:

1) Radial motion control of the magnetically levitated rotor assembly 1_1 in electromagnetic bearings involves 4 coordinate systems, as shown at a in fig. 4, where o _i _x _i y _i z _i The coordinate system is an inertial coordinate system fixed with the magnetic suspension bearing electromagnet or a static coordinate system; o _r _x _r y _r z _r Moving coordinate system, o, fixed to the rotor assembly _r Ideally coinciding with the rotor assembly centre of gravity COG, o _r _x _r y _r z _r Also known as COG coordinate system; o _a _y _a z _a And o _b _y _b z _b Jointly form a bearing coordinate system of a plane where the electromagnetic force vector of the magnetic suspension bearing is located; o _as _y _as z _as And o _bs _y _bs z _bs Jointly forming a sensor coordinate system of the plane in which the rotor radial position sensor is located, the magnetic levitation rotor assembly 1_1 has 6 degrees of freedom of movement, or state of movement variables, in a levitated state, as shown at b in fig. 4, with respect to an inertial coordinate system, where [ x ] is _r y _r z _r ]And

respectively, relative inertial frame o of rotor assembly _i _x _i y _i z _i The translational and rotational motion state variables, among the above 6 motion state variables, represent the axial translational and rotational motion state variable x _r (t) and

in general, the control can be performed independently and the influence of the radial motion of the rotor is negligible, so that the realization of the axial rotation control and the displacement control is not deeply analyzed in the following description of the present invention. The remaining 4 motion state variables y of the rotor assembly _r ，z _r Phi, theta constitute the state variable q = [ y ] for radial motion in the COG coordinate system _r z _r ψ θ] ^T 。

2) Radial electromagnetic force vector of magnetic suspension bearing in bearing coordinate system o _a _y _a z _a And o _b _y _b z _b Can be expressed by a vector

U _F ＝[f _ay f _by f _az f _bz ] ^T

3) Radial electromagnetic force U _F Is a linearized mathematical model of

U _F ＝-K _g B ^T q+K _i I _MB

Wherein

Is a magnetic suspension bearing position negative rigidity matrix, K _ga 、K _gb Respectively are negative rigidity coefficients of the positions of the radial magnetic suspension bearings at the front end and the rear end,

transpose thereof B ^T Is a transformation matrix from the state variable of the COG coordinate system to the displacement of the bearing coordinate system,

is a current stiffness matrix of the magnetic suspension bearing, K _ia 、K _ib Respectively are the current rigidity coefficients of the radial magnetic suspension bearings at the front end and the rear end,

I _MB ＝[i _ay i _by i _az i _bz ] ^T the radial magnetic suspension bearing current vector is also the control output of the radial magnetic suspension control system.

4) The radial electromagnetic force vector U of the magnetic suspension bearing can be converted by the following linear transformation _F Conversion to rotor COG coordinate system

F _coc ＝BU _F

Wherein B is a bearing coordinate system o _a _y _a z _a And o _b _y _b z _b A force transformation matrix to COG coordinate system.

5) Radial motion state variable of COG coordinate system

q＝[y _r z _r ψ θ] ^T

Can be equivalently formed by the geometric central axis of the rotor assembly and the sensor coordinate system plane o _as _y _as z _as And o _bs _y _bs z _bs Point of intersection o _ras And o _rbs Coordinate vector of

q _s ＝[y _as z _as y _bs z _bs ] ^T

As shown in 4.c). Equivalent motion state variable q _s Is measured atSensor coordinate system o _as _y _as z _as And o _bs _y _bs z _bs In (1) implementation, measured state variable q _s Mapping to the rotor COG coordinate system can be done by the following linear transformation

q＝C ^-1 q _s

Wherein

Its inverse matrix C ^-1 As sensor coordinate system o _as _y _as z _as And o _bs _y _bs z _bs To the COG coordinate system.

6) The radial motion linearization dynamic mathematical model of the state space form of the magnetic suspension motor can be derived according to the coordinate system, the state variable and the magnetic suspension bearing electromagnetic force vector defined above

Wherein

Is a state space state variable of the radial motion,

in the form of a matrix of states,

is a rotor assembly generalized mass matrix, m is the mass of the rotor assembly, I _yy For rotor assembly y _r The rotational inertia of the shaft is reduced,

gyro effect matrix of which I _xx Is a rotor assembly x _r The rotational inertia of the shaft, omega is the rotational speed of the rotor assembly,

K _gCOG ＝BK _g B ^T from the bearing coordinate system o _a _y _a z _a And o _b _y _b z _b The negative stiffness matrix of the position of the magnetic suspension bearing is transformed to a COG coordinate system,

in order to input the matrix, the input matrix is,

is an output matrix.

7) In the invention, in order to realize that the magnetic suspension rotor assembly 1_1 surrounds the geometric central axis x thereof under the control of a magnetic suspension control system _r o _r Rotating and the geometric central shaft of the rotor and the geometric central line x of the electromagnet of the magnetic suspension bearing _i o _i The closed-loop controller for the magnetic suspension position adopts a reverse dynamics algorithm, namely a structure of adopting a COG coordinate system PD control + rotor gyroscopic effect compensation + magnetic suspension bearing negative stiffness characteristic compensation, and the magnetic suspension controller corresponding to the algorithm outputs I _MB Can be expressed by the following formula

K _p ＝(BK _i ) ^-1 MP _COG C ^-1 Proportional control gain for PD control, wherein

P _par ，P _con Proportional control of gain for translation and rotation, respectively，K _d ＝(BK _i ) ^-1 MD _COG C ^-1 A differential control gain for PD control, wherein

D _par ，D _con Differential control gain, C, for translation and rotation, respectively _S (BK _i ) ^-1 K _gCOG C ^-1 The compensation coefficient of the negative rigidity characteristic of the magnetic suspension bearing position,

C _G ＝(BK _i ) ^-1 GC ^-1 and the compensation coefficient is the rotor gyro effect.

8) Fig. 5 a shows a control system block diagram formed by the mathematical model in the form of the state space of the magnetic levitation motor and the closed-loop control of the radial position of the motor rotor assembly by using the inverse dynamics algorithm;

under the ideal condition, the magnetic suspension rotor assembly 1_1 rotates at high speed in a motion space which is suspended in a magnetic suspension bearing in a non-contact manner under the position closed-loop control, the rotating shaft of the magnetic suspension rotor assembly 1_1 coincides with the geometric center line of an electromagnet of the magnetic suspension bearing, the average value of a rotor position signal output by the rotor assembly radial displacement sensor 1_3 is zero and does not contain a periodic component with the same integral multiple frequency as the rotating speed omega or omega, but in an actual system, due to the existence of factors such as the fact that the circumference where the electromagnetic target ring 1 u 4 of the rotor assembly radial displacement sensor is located is not concentric with the geometric rotating shaft of the rotor assembly, the deformation of the surface of the target ring, the non-uniform electromagnetic characteristics of a material for manufacturing the target ring and the like, the rotor position signal output by the rotor assembly radial displacement sensor 1_3 often contains a periodic component with the same integral multiple frequency as the rotating speed or omega frequency, namely the radial component, and is represented by a vector R.

R＝[r _ay r _az r _by r _bz ] ^T

Wherein

o _as _y _as Position of coordinate axisA sensor bounce component;

α _r = Ω t being o _r _y _r And o _i _y _i The rotation angle between the two coordinate axes is measured by an angular displacement sensor 1_6 of the magnetic suspension rotor assembly;

due to o _as _z _as Position of coordinate axis position sensor lags behind in the direction of rotation of rotor assembly _as _y _as Coordinate axis position sensor 90 °, o _as _z _as The coordinate axis position sensor runout component may be expressed as

o _bs _y _bs A coordinate axis position sensor run-out component;

and r _az Same principle of r _bz Can be expressed as

Since the elements of the vector R are Fourier series containing each harmonic of the jitter component, the characteristic parameter of R can be represented by a 4-dimensional vector formed by Fourier coefficients of each harmonic

ξ _i ＝[A _ai B _ai A _bi B _bi I ^T

i =1,2, …, n being the harmonic order.

9) In xi _i R can be reduced by the following formula in the known case

Wherein

R _i ＝[r _ayi r _azi r _byi r _bzi ] ^T ＝Γ _i ξ _i

I.e. R _i Represents the ith harmonic in R after harmonic decomposition of R.

10 In addition to the above-mentioned factors causing the radial displacement runout component, the magnetic levitation rotor assembly 1_1 inevitably has a certain degree of dynamic balance unbalance during the manufacturing process, and when the magnetic levitation rotor assembly 1_1 has dynamic balance unbalance, it will generate radial centrifugal force to force the geometric center line of the rotor assembly to deviate from the geometric center line of the electromagnet of the magnetic levitation bearing under the rotation state, and also generate a periodic component with the same frequency as the rotation speed Ω in the output signal of the rotor assembly radial displacement sensor 1_3. The centrifugal forces caused by dynamic balance imbalances can be equivalently applied with the bearing coordinate system o _a _y _a z _a And o _b _y _b z _b Vector F in (1) _u To indicate.

11 Has consideration of the runout component R in the output of the rotor assembly radial displacement sensor 1_3 and the centrifugal force F caused by dynamic balance unbalance _u The state space form radial motion linearization dynamic mathematical model of the magnetic suspension motor becomes the following form

q _se ＝q _s +R

Wherein

As a centrifugal force F _u Input matrix of q _se For radial displacement sensor outputs containing a run-out component R, q _se ＝[y _ase z _ase y _bse z _bse ] ^T 。

12 5.b) are shown to take into account the component of the runout disturbance R in the radial displacement sensor output and the centrifugal force disturbance F created by the rotor assembly dynamic balance imbalance _u And later, a closed-loop control system block diagram of the radial position of the motor rotor assembly.

Comparing a and b in fig. 5, it can be seen that the system output q actually needs to be controlled in a _s Feedback variable q measurable in b due to disturbance of jitter component R becoming undetectable _se Not only contains the direct influence of R, but also contains R and F _u Closed loop formed by magnetic suspension position controller and magnetic suspension motor is in q _s The induced response. q. q.s _se Medium pulsation component R and centrifugal force F _u The displacement generated by the disturbance is also a periodic component of omega or an integral multiple frequency of omega, and can be expressed by the following formula

Q＝[s _ay s _az s _by s _bz ] ^T

Wherein

o _as _y _as Disturbance components of the coordinate axis position sensor;

o _as _z _as disturbance components of the coordinate axis position sensor;

o _bs _y _bs disturbance components of the coordinate axis position sensor;

o _bs _z _bs and (4) disturbance components of the coordinate axis position sensor.

13 Are) andvector xi formed by Fourier coefficients _i Similarly, the disturbance component Q can be represented by a vector formed by Fourier coefficients of its respective harmonics

γ _i ＝[C _ai D _ai C _bi D _bi l ^T

i =1,2, …, n being the harmonic order.

γ _i The functional relationship with Q is

Q _i ＝Γ _i γ _i

The runout component R in the output of the rotor assembly radial displacement sensor 1_3 will cause the motion of the magnetic suspension rotor assembly 1_1 to generate radial oscillation, so that the actual rotation axis of the rotor deviates from the designed rotation axis, when the load device of the high-speed magnetic suspension motor is a high-speed turbomachine, such as a centrifugal air compressor or a centrifugal refrigerant compressor, in order to avoid the friction between the turbine wheel and the volute due to the deviation of the rotation axis, the gap between the turbine wheel and the volute needs to be increased in the design, but the increase of the gap will cause the efficiency of the compressor to be reduced, in addition, the radial oscillation of the rotor assembly will also cause a series of problems such as the temperature of the electromagnet of the magnetic suspension bearing is too high, and the rotor assembly material is subjected to fatigue damage or even fracture due to internal periodic stress, so that the compensation of the disturbance component in the control algorithm to eliminate the adverse effects caused by the disturbance component is necessary to improve the control performance, efficiency, service life and reliability of the high-speed magnetic suspension motor. Due to R and q _s The attribute which can not be directly measured, the related parameter of R is usually obtained by means of indirect test in the manufacturing and using processes of the product, or the radial run-out parameter identification or the self-adapting function is added in the control algorithm to compensate the radial run-out component. The above-mentioned run-out compensation process often needs to be performed in a trial and error manner during the operation of the equipment, and therefore there is a risk that the equipment is damaged due to improper parameter selection. The compensation method provided by the invention is realized through hardwareRing simulation and artificial intelligence computer optimization method for obtaining relevant parameters of radial runout component R in virtual digital space according to b and R and q in figure 5 _se Analysis of the composition reveals that q _se The periodic disturbance component Q in (2) can be a vector gamma formed by Fourier coefficients _i Represented by a vector xi where R can be constructed by fourier coefficients _i Is represented by, and γ _i Is xi _i As a function of (c).

14 The hardware of the invention is therefore used in a loop simulation system with the gamma values obtained in the actual system and in the digital simulation, respectively _i And

(the crown is represented by ^ a. The difference between variables or models in a simulation system, the same applies below)

Is formed by

For the objective function of the variable, the objective function is converged to the minimum extreme value by artificial intelligence optimization of a computer, and the objective function variable corresponding to the extreme value

Reconstructing the generated hop component

I.e. the hopping component R in the actual system _i Can be used to compensate the jump component Ri existing in the actual system, i.e. the optimal approximation value of (b) can be used to compensate the jump component Ri existing in the actual system

Substituted q _se To achieve cancellation of q _se Radial runout component R of _i The purpose of (1). Due to the acquisition

Is performed in a virtual digital space, without the need forThe running state of the actual system is changed, so the invention can greatly reduce the risk of equipment damage in the radial runout compensation process and improve the automation degree of the production process.

15 It is noted here that, as shown at b in fig. 5, in addition to the runout component R in the radial displacement sensor output, a radial position feedback signal q for the rotor assembly is provided _se Causing disturbances, e.g. centrifugal force vectors F resulting from dynamic balance unbalance _u Will also be at q _se Causing a disturbance. Due to F _u Is synchronized with the motor speed omega, so q _se Middle corresponds to F _u Will be in phase with the fundamental component R of R ₁ Same frequency, thus obtaining R ₁ Forming interference, known as F _u The amplitude of the radial displacement sensor is in direct proportion to the square of the omega, and the parameter change of the R is influenced by the change of the omega very little according to the mechanism generated by the jumping component in the output of the radial displacement sensor, so that in the process of acquiring the relevant parameters of the radial jumping component R in a hardware-in-loop simulation and artificial intelligence computer optimization way, the rotating speed omega of the motor should be controlled in the range as low as possible to meet the stability requirement so as to reduce the dynamic balance unbalance centrifugal force F to the maximum extent _u The influence of (c).

16 Hardware-in-the-loop simulation system is shown in fig. 6. The magnetic suspension controller in the simulation system and the controller in the actual system adopt the same structure and control parameters. Module in system

And

the dynamic characteristics of the magnetic suspension bearing current control ring and the rotor assembly radial position sensor are simulated respectively.

And

in the invention, a structure without limitation of an inertial filter is adopted, and a filtering time constant of the inertial filter needs to be set according to the bandwidth and the transmission delay of a magnetic bearing current control ring and a rotor assembly radial position sensor in an actual system respectively so as to enable the dynamic characteristic of a simulation system to be closer to the actual system.

17 In the process of artificial intelligence computer optimization, the input of the magnetic suspension motor dynamic simulation system is the radial run-out given in the optimization path

Vector of fourier coefficients

By means of a formula

Reduction to form

Then the magnetic suspension position closed-loop controller is superposed on the magnetic suspension position closed-loop controller for simulation input; the output of the magnetic suspension motor dynamic simulation system relative to the computer optimizing system is

By pairs

And performing Fourier series expansion to obtain the optimal target function, and using the optimal target function as the input of the artificial intelligence computer optimization system to calculate the optimal target function.

In the process of optimizing the artificial intelligence computer, the rotating speed omega of the magnetic suspension motor is controlled to make the dynamic balance unbalance centrifugal force F _u Can be ignored, so in fig. 6) F can be set _u ＝0。

18 Fig. 7 shows an implementation block diagram of the hardware-in-the-loop simulation compensation device for radial runout of the rotor position of the magnetic levitation motor, and a specific implementation process is described as follows:

in the course of heightThe fast magnetic suspension motor 1 and the control system 2 thereof form a hardware system in which hardware physically exists in a ring simulation system, and a rotor assembly is controlled by a magnetic suspension motor controller to surround a geometric central axis x of the rotor assembly _r o _r Rotates at a low speed at a constant rotation speed omega, and the geometric central axis of the rotor and the geometric central line x of the electromagnet _i o _i And (4) overlapping. In this state, if a radial run-out component R exists in the rotor position sensor output at a frequency Ω or several times Ω, the feedback input variable q for closed-loop control of the rotor assembly position _se Will appear with periodic components of the same frequency.

19 First needs to determine q by fast fourier transform _se Pilot spectrum in (1), q _se The major harmonic component of (a). The necessity of determining the pilot spectrum is targeted for subsequent optimization of the computer parameters. The pilot spectrum is often q _se Of the frequency omega and some few harmonics in the sequence of harmonics. Without loss of generality, these harmonics can be assumed to be

i＝n ₁ ，n ₂ ，…，n _N

Where N is the number of harmonics in the pilot spectrum, N ₁ ，n ₂ And … is the order of the harmonics.

20 An alternative way of optimizing the elements of the characteristic parameter vector of R after the pilot spectrum has been determined is to simultaneously optimize all harmonics in the pilot spectrum, i.e. to construct an optimization variable parameter vector

And an objective function

Xi optimal value xi of optimizing search by computer ^opt Make J (xi) converge on the minimum extreme value J (xi) ^opt )。

The calculation amount of computer optimization and the dimension of variable parameter are increased in geometric series. Due to xi _i A 4-dimensional vector, and the dimension of vector xi is 4N. Therefore, the above method for simultaneously optimizing all harmonics in the pilot spectrum greatly increases the amount of calculation and the uncertainty of the optimization result.

21 On the other hand, the mathematical model of the magnetic suspension motor and the control system thereof near the working point is a linearized model, and the superposition principle of the linear system is satisfied. Therefore, the method for optimizing the characteristic parameter vector of R provided by the invention utilizes the characteristic that the system meets the superposition principle of a linear system, and firstly carries out optimization on the characteristic parameter vector of each harmonic in the pilot frequency spectrum

Optimizing parameters one by one, and calculating radial position runout corresponding to the pilot frequency spectrum according to the superposition principle

With i = n _k Illustrating computer optimization by way of example

The process of (1).

Is provided with

The corresponding computer optimizing objective function is

Wherein

I.e. the diameter of the actual systemOutputs q to the displacement sensor _se In (1) corresponds to n _k Harmonic parameter vector of subharmonic component, by pair q _se Performing Fourier series expansion to obtain;

obtained by simulating a real system in the digital space of a computer. The operation condition of the magnetic suspension motor and the control system thereof in the simulation system is kept consistent with that of the actual system, and the simulation output of the radial displacement sensor

Middle injection

22 The computer optimization process is

Four-dimensional space of the place

Searching out a 'best' path to approximate the target function

Harmonic parameter vector corresponding to the minimum value of

The computer optimization algorithm for realizing the process belongs to the category of artificial intelligence, and various choices for realizing the aim of the invention exist in the background of the high-speed development of the artificial intelligence at present, and the best is relative to the selected optimization algorithm. In the present invention, as one of the specific implementation schemes of the computer optimization Algorithm, hu Keji Weiss Algorithm (Hooke-Jeeves Algorithm) is selected but not limited to be used.

23 Hu Keji wies algorithm consists of two moving processes, probe movement and pattern movement, and is also called pattern search method. By detecting movement→ modal shift → exploring shift … are repeated to approximate the extremum of the objective function one after another. The j +1 th detection movement is from the current parameter variable vector

Are started respectively at

Perturbing the components one by one in positive and negative directions by set harmonic component amplitude increment step length h, and obtaining each time parameter variable perturbation through simulation

(l =1,2, …,8, as represented in

8 perturbations on 4 components) of the target function

To determine the optimal direction of movement of the perturbation components, i.e. to increase, or to decrease or to remain the same. In turn accomplish

Perturbation of 4 components to obtain the parameter variable vector

Of incremental vectors

And as a result of detecting the movement.

24 Mode shift is the result of a probe shift

Order under the condition of non-zero

Along the edge

Is continuously moved in the direction of (i.e. is

For new obtained after each mode shift

Performing simulation and calculating corresponding objective function

The condition for completion of the mode shift is

J(j，m)＜J(j，m+1)

I.e. obtained by the movement of the objective function along the probe

No further convergence in direction is achieved. The parameter variable vector obtained at this time will be used as the new current value

A new probing movement is performed.

24 If the result of the movement is detected

Equal to zero, i.e. the result of the detected movement is

Remain static, then explain

The optimum value has been approached. The probing moving step h can be reduced at this time to

And returning to the detection moving stage as the current parameter variable vector to continue optimizing so as to improve the accuracy of radial run-out compensation. If h is already less than the set minimum step size, the computer optimization process ends. At this time

I.e. for n in the pilot spectrum _k Optimal target value of parameter variable for computer optimization by subharmonic

26 By repeatedly applying Hu Keji Weiss algorithm to complete computer optimization of all harmonic components in pilot frequency spectrum

Position run-out component characteristic parameter obtained by running hardware in loop simulation system

k =1,2, …, N is stored in a digital storage unit of a high-speed magnetic suspension motor control system, and a position jumping component characteristic parameter in the digital storage unit is used during the operation of the actual high-speed magnetic suspension motor system

k =1,2, …, N reconstructs the jitter component in the radial position sensor in real time, i.e. in real time

And will be

Feedback input q of reverse injection radial displacement sensor _se Namely, the q pair can be realized _se Compensation of the jitter component R in (1).

Claims

1. A compensation device for radial run-out component of magnetic suspension rotor position is characterized in that: the method comprises the following steps: a hardware-in-the-loop simulation system;

the hardware-in-the-loop simulation system comprises: the system comprises a high-speed magnetic suspension motor (1) and a control system (2) thereof, wherein the high-speed magnetic suspension motor (1) and the control system (2) form a hardware system of a hardware-in-loop simulation system in physical existence;

a dynamic mathematical model (3) of the high-speed magnetic suspension motor, a magnetic suspension control simulation system (4) and a computer optimization algorithm (5) of radial run-out compensation parameters form a simulation system of which hardware exists in a computer digital space in a ring simulation system;

the high-speed magnetic suspension motor (1) comprises: magnetic levitation rotor assembly (1_1), radial electromagnetic bearing (1_2), rotor assembly radial displacement sensor (1_3), three-phase motor stator (1_4), axial electromagnetic bearing (1_5) and axial displacement sensor, and magnetic levitation rotor assembly angular displacement sensor (1_6).

2. The device for compensating for the radial run-out component of a magnetic levitation rotor position of claim 1, wherein: the rotor assembly radial displacement sensor (1_3) is used for detecting the position of a magnetic levitation rotor assembly (1_1) in the air gap of the axial electromagnetic bearing (1_5).

3. The device for compensating for the radial runout component of a magnetically suspended rotor position of claim 1, wherein: the control system is a high-speed magnetic suspension motor controller (2);

the high-speed magnetic suspension motor controller (2) comprises a rotor assembly radial displacement sensor signal processing circuit (2_1), an electromagnetic bearing current PWM power amplifier (2_2), a magnetic suspension rotor assembly angular displacement signal processing circuit (2_3), a magnetic suspension position closed-loop controller (2_4) and a motor PWM driving and controller (2_5).

4. The device for compensating for the radial runout component of a magnetically suspended rotor position of claim 1, wherein: the magnetically levitated rotor assembly (1_1) includes:

the method comprises the following steps of (1) a turbine impeller (1 \u1), a radial magnetic suspension bearing rotor magnetic circuit (1 _1 \u2), a motor rotor (1 _1 _3), a rotor assembly radial displacement sensor electromagnetic target ring (1 _1 _4) and an axial magnetic suspension bearing rotor (1 _1 _5); the magnetic suspension rotor assembly (1_1) is a rigid body rotator neglecting a flexible mode.

5. A device for compensating the position runout component of a magnetically levitated rotor as claimed in claim 3, wherein: the radial electromagnetic bearing (1_2), the rotor assembly radial displacement sensor (1_3), the magnetic suspension position closed-loop controller (2_4) and the attached sensor signal processing circuit and power amplification circuit form a magnetic suspension radial position control system of the magnetic suspension rotor assembly (1_1).

6. The device for compensating for the radial runout component of a magnetically suspended rotor position of claim 1, wherein: the dynamic mathematical model (3) of the high-speed magnetic suspension motor comprises:

7. The device for compensating for the radial run-out component of a magnetic levitation rotor position of claim 1, wherein: the hardware-in-the-loop simulation system further comprises:

an inverse dynamics control algorithm which is a radial position closed loop control algorithm of the magnetic levitation rotor assembly (1_1).

8. A method of a device for compensating the radial run-out component of a magnetic levitation rotor position as claimed in claim 1, characterized in that: the method comprises the following steps:

s1: determining characteristic parameters of position jumping components in output signals of a radial position sensor of a rotor of the magnetic suspension motor by operating the hardware in a ring simulation system;