Non-linear harmonic adaptive neural network multi-dead-corner PDS controlled bearingless permanent magnet synchronous motor

Abstract

The invention belongs to the technical field of motors, and particularly relates to a non-linear harmonic adaptive neural network multi-dead-angle PDS controlled bearingless permanent magnet synchronous motor, which comprises a stator, the stator comprises a stator core, a suspension winding and a torque winding are mounted on the surface of the stator core, a rotor is arranged in the stator, the rotor comprises a rotor core, a magnetic bearing is mounted on the stator, and the magnetic bearing is connected with the suspension winding and the torque winding. And a permanent magnet is mounted on the magnetic bearing. A mechanical bearing in a traditional motor is omitted, mechanical abrasion does not exist, the service life of the motor is longer, the maintenance cost is lower, and noise is low during operation; a magnetic suspension technology is adopted in the motor, so that current of the electromagnet can be better controlled, electric energy loss is reduced, and energy consumption efficiency is improved; a non-linear harmonic self-adaptive neural network multi-dead-angle PDS control algorithm is applied to a sensor in the motor, the motion displacement condition of the rotor during working can be monitored in an all-around mode, the condition of the rotor can be obtained in time, and the direction of the rotor can be regulated and controlled in time.

Description

Technical Field

[0001] The present invention relates to the technical field of motors, and in particular to a bearingless permanent magnet synchronous motor controlled by a nonlinear harmonic adaptive neural network with multiple dead angles (PDS). Background Art

[0002] The increasing popularity of new energy vehicles has increased the demand for high-efficiency motor technology. In recent years, switched reluctance motors (SRMs) have been widely used due to their basic structure, large starting torque, flexible control, and good fault tolerance. Although permanent magnet synchronous motors (PMSMs) currently dominate the market with excellent performance characteristics, they face significant challenges, including rising material costs and the risk of high-temperature demagnetization associated with permanent magnet materials. In contrast, PDS-BMSs can achieve a wide range of speed regulation and maintain high torque output at low speeds, which makes them of great application value in electric vehicles, aerospace, industrial drives and other fields.

[0003] Although the PDS-bearingless permanent magnet synchronous motor has the advantages of both integration and permanent magnet synchronous motors, the additional three-phase winding in the motor also brings greater challenges to its control strategy. Significant current harmonics are the most prominent problem. On the one hand, the PDS-bearingless permanent magnet synchronous motor is more susceptible to harmonic effects caused by factors such as inverter nonlinearity and back electromotive force; on the other hand, due to the existence of independent low-impedance harmonic subspaces, the harmonic currents in the xy subspace will be further amplified. The huge harmonic current will cause significant copper loss, reduce motor efficiency, and bring trouble to the design of the drive solution. Therefore, advanced harmonic compensation strategies have become a new research hotspot in the field of multiphase drives. Existing technologies are in urgent need of solving the following problems:

[0004] Ordinary permanent magnet motors are prone to high temperatures when operating under high load conditions, affecting the performance and life of the motor. The permanent magnet material may also experience magnetic field attenuation, resulting in reduced motor performance and weakened rotor suspension control capabilities. The motor cannot respond quickly and promptly to changes in rotor orientation, resulting in rotor instability and large displacement inside the motor. Therefore, we propose a bearingless permanent magnet synchronous motor with multi-dead-angle PDS control using a nonlinear harmonic adaptive neural network to solve the above problems. Summary of the Invention

[0005] (1) Technical problems solved

[0006] In view of the deficiencies in the prior art, the present invention provides a bearingless permanent magnet synchronous motor controlled by a nonlinear harmonic adaptive neural network with multiple dead angles (PDS), which solves the problems raised in the above background technology.

[0007] (2) Technical solution

[0008] In order to achieve the above-mentioned purpose, the present invention specifically adopts the following technical solutions:

[0009] A bearingless permanent magnet synchronous motor controlled by a nonlinear harmonic adaptive neural network and multiple dead angle PDS includes a stator, the stator includes a stator core, a suspension winding and a torque winding are installed on the surface of the stator core, a rotor is arranged inside the stator, the rotor includes a rotor core, a magnetic bearing is installed on the stator, a permanent magnet is installed on the magnetic bearing, and a rotating wind plate is installed at one end of the rotor.

[0010] Furthermore, the stator winding consists of two sets of symmetrical Y-connected windings, with a spatial phase difference of 30°. Through vector space decoupling (VSD) coordinate transformation, each component can be decomposed into three orthogonal subspaces: α-β subspace, xy subspace, and o1-o2 subspace, to establish the mathematical model of the motor:

[0011]

[0012] In As, Axy are the bidirectional inductances of dq axis respectively, Ms is the magnetic field that hinders the current in the stator, Q m is the permanent magnet quantity, ω is the rotor speed, t and n are the voltage and current, and the subscripts “dq” and “xy” denote variables in the dq subspace and xy subspace, respectively.

[0013] Furthermore, the mathematical model of the nonlinear error of the permanent magnet synchronous motor inverter can be expressed as:

[0014]

[0015] The superscript “qbz” indicates the reference value. Δtd, Δtq, Δtx, and Δty are the dq-axis and xy-axis voltage errors caused by the inverter nonlinearity, respectively.

[0016] The traditional DPCC method is based on a discretized motor model. According to (4), the discrete domain voltage equation considering the VSI nonlinearity in the xy subspace can be expressed as:

[0017]

[0018] Where Ixy=[tx ty]Tt xy=[tx ty]T.Δtxy=[ΔtxΔty]T

[0019] According to (5), considering the current and total disturbance Ixy as estimation objects, the discretized ESO can be designed as:

[0020]

[0021] in and are the estimated values ​​of ixy and dxy respectively. β1 and β2 are the gains of ESO.

[0022] It can be further calculated as

[0023]

[0024] From Equation (6), we can see that both voltage txy and current ixy can be considered finite because they are limited by the DC bus and inverter. At the same time, dxy(k) and dxy(k+1) have very small variations in the sampling period, so Dxy can be considered a constant with a bounded value.

[0025] Furthermore, as the leader of the pack, the alpha wolf is less adaptable to the environment than other coyotes. The alpha wolf and cultural trends in the wolf pack can be described as

[0026]

[0027] in, and is the median of the e-dimensional variable of all coyotes in the entire p group when Nc takes different values ​​at the tth time.

[0028] Coyotes go through a natural life cycle of birth and death. To simulate genetic influences, the emergence of newborn coyotes is marked It is characterized by a convergence of parents' social status and environmental factors.

[0029]

[0030] where m1 and m2 are random individual coyotes in the wolf pack x, e1 and e2 are arbitrary two-dimensional representations of the problem to be solved, and Ge and gande range between 0 and 1.

[0031] (3) Beneficial effects

[0032] Compared with the prior art, the present invention provides a bearingless permanent magnet synchronous motor with a nonlinear harmonic adaptive neural network multi-dead angle PDS control, which has the following beneficial effects:

[0033] The present invention eliminates the mechanical bearings in traditional motors, and there is no mechanical wear, which makes the motor have a longer service life, lower maintenance costs, and low noise during operation; the use of magnetic levitation technology inside the motor can better control the current of the electromagnet, reduce power loss, and improve energy efficiency; the use of a nonlinear harmonic adaptive neural network multi-dead angle PDS control algorithm on the internal sensor of the motor can fully monitor the motion displacement of the rotor during operation, and obtain the rotor status in time so as to timely adjust the rotor position. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0035] Figure 2 This is a two-dimensional planar magnetic circuit magnetic field analysis diagram of the bearingless permanent magnet synchronous motor of the present invention.

[0036] In the figure: 1. Stator; 2. Torque winding; 3. Suspension winding; 4. Permanent magnet; 5. Rotating wind vane; 6. Magnetic bearing; 7. Rotor core; 8. Stator core; 9. Rotor. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0038] Example

[0039] like Figure 1-2 As shown, a bearingless permanent magnet synchronous motor with multi-dead-angle PDS control of a nonlinear harmonic adaptive neural network proposed in one embodiment of the present invention includes a stator 1, the stator 1 includes a stator core 8, a suspension winding 3 and a torque winding 2 are installed on the surface of the stator core 8, a rotor 9 is arranged inside the stator 1, the rotor 9 includes a rotor core 7, a magnetic bearing 6 is installed on the stator 1, a permanent magnet 4 is installed on the magnetic bearing 6, and a rotating wind plate 5 is installed at one end of the rotor 9.

[0040] A permanent magnet 4 is added to the motor rotor 9. The permanent magnet 4 generates a magnetic field, so that the rotor 9 is subjected to the Lorentz force and the Maxwell force in all directions. The stator is made of silicon steel sheet material to improve efficiency and stability, reduce energy consumption, and improve corrosion resistance. A distributed ring winding is used on the stator 1 to reduce harmonic loss, optimize the control performance of the magnetic field, enhance the mechanical strength, and improve the heat dissipation performance of the motor. The winding is divided into a torque winding 2 and a suspension force winding 3, and a stable magnetic field and suspension force winding are provided by the permanent magnet 4. A force perpendicular to the rotation axis is generated, and the windings and the permanent magnets 4 generate magnetic forces acting on the rotor 9. The windings are annular, and the annular magnetic force generated by the windings can offset and balance the magnetic force, Lorentz force and Maxwell force generated by the windings corresponding to the annular ring, overcome the effect of gravity, and enable the rotor 9 to be more stably suspended in the middle of the stator 1, so that the rotor 9 can operate without contact, and generate a rotational torque through the torque winding 2 to drive the rotor 9 to rotate. The two windings complement each other with the magnetic field force to improve the efficiency, stability and high-precision control of the motor.

[0041] The stator windings of a dual three-phase bearingless permanent magnet synchronous motor consist of two sets of symmetrical Y-connected windings, with a 30° spatial phase difference. Using vector space decoupling (VSD) coordinate transformation, the components can be decomposed into three orthogonal subspaces: the α-β subspace, the xy subspace, and the o1-o2 subspace, thereby establishing a mathematical model for the motor:

[0042]

[0043]

[0044] In As, Axy are the bidirectional inductances of dq axis respectively, Ms is the magnetic field that hinders the current in the stator, Q m is the permanent magnet quantity, ω is the rotor speed, t and n are the voltage and current. The subscripts “dq” and “xy” denote variables in the dq subspace and xy subspace, respectively.

[0045] Traditional predictive control performance relies on an ideal mathematical model of the motor. However, in practical applications, interference factors are unavoidable. For dual three-phase motors, these interference factors can cause significant harmonic currents in the xy plane, further deteriorating control performance. Inverter nonlinearity is a major contributor to harmonics. Therefore, this section proposes a method for online compensation of inverter nonlinearity.

[0046] The mathematical model considering the nonlinear error of the PDS-bearingless permanent magnet synchronous motor inverter can be expressed as

[0047]

[0048] The superscript “qbz” indicates the reference value. Δtd, Δtq, Δtx, and Δty are the dq-axis and xy-axis voltage errors caused by the inverter nonlinearity, respectively.

[0049] The traditional DPCC method is based on a discretized motor model. According to (4), the discrete domain voltage equation considering the VSI nonlinearity in the xy subspace can be expressed as:

[0050]

[0051] Where Ixy=[txty]T.txy=[txty]T.Δtxy=[ΔtxΔty]T

[0052] According to (5), considering the current and total disturbance Ixy as estimation objects, the discretized ESO can be designed as:

[0053]

[0054] in and are the estimated values ​​of ixy and dxy respectively. β1 and β2 are the gains of ESO.

[0055] It can be further calculated as

[0056]

[0057] From Equation (6), we can see that both voltage txy and current ixy can be considered finite because they are limited by the DC bus and inverter. At the same time, dxy(k) and dxy(k+1) have very small variations in the sampling period, so Dxy can be considered a constant with a bounded value.

[0058] As the leader of the pack, the alpha wolf is less adaptable to the environment than other coyotes. The alpha wolf and cultural trends in the wolf pack can be described as

[0059]

[0060] in, and is the median of the e-dimensional variable of all coyotes in the entire p group when Nc takes different values ​​at the tth time.

[0061] Coyotes go through a natural life cycle of birth and death. To simulate genetic influences, the emergence of newborn coyotes is marked It is characterized by a convergence of parents' social status and environmental factors.

[0062]

[0063] where m1 and m2 are random individual coyotes in the wolf pack x, e1 and e2 are arbitrary two-dimensional representations of the problem to be solved, and Ge and gande range between 0 and 1.

[0064] Through the above eight formulas, the Coyote Genetic Algorithm can be used to perform multi-angle control of the rotor operation in the nonlinear PDS-bearingless permanent magnet synchronous motor.

[0065] By adding this nonlinear harmonic adaptive neural network multi-dead-angle PDS control algorithm to the motor, the motion displacement of the rotor during operation can be fully monitored, and the rotor status can be obtained in time, so as to timely adjust the rotor position, quickly respond and adjust the rotor, and improve the working efficiency of the motor. The entire motor system has been stably designed and optimized, the speed has been increased, and the precision error has been controlled.

[0066] like Figure 1The figure shows a three-dimensional model of a nonlinear bearingless permanent magnet synchronous motor. Through vector space decoupling (VSD) coordinate transformation, each component can be decomposed into three orthogonal subspaces: the α-β subspace, the xy subspace, and the o1-o2 subspace, thereby establishing the mathematical model of the motor:

[0067] like Figure 2 The figure shows a two-dimensional planar magnetic circuit analysis diagram of a bearingless permanent magnet synchronous motor. When the current runs along the positive X direction, during the stable rotation of the rotor, there will be positions at multiple angles where the dead point magnetic field cannot supply the position. Using the nonlinear adaptive coyote compensation algorithm, when the motor magnetic field suspension force is insufficient, the coyote algorithm will provide the current passing through the dead point rotation to apply the displacement to the dead point.

[0068] Predictive control performance relies on an ideal mathematical model of the motor. In practical applications, interference factors are unavoidable. For dual three-phase motors, these interferences can cause large amounts of harmonic currents in the xy plane, further deteriorating control performance. Inverter nonlinearity is one of the primary factors contributing to harmonics. Therefore, this section proposes a method for online compensation of inverter nonlinearity. This algorithmic compensation stabilizes the rotor's displacement waveform. The longer the harmonics run, the greater the radial suspension force generated. Therefore, compensating for a stable harmonic magnetic field requires a larger current parameter, but the current parameter value exhibits a certain degree of adaptability.

[0069] Considering the nonlinear errors of the PDS-bearingless permanent magnet synchronous motor inverter, a self-test control module method is constructed using a mathematical model. The inverter output error value depends on the smoothness of the rotor output torque. Superscript reference values. Δud, Δuq, Δux, and Δuy are the dq-axis and xy-axis voltage errors caused by the inverter nonlinearity, respectively.

[0070] Traditional DPCC methods are based on a discretized motor model. By considering the discrete-domain voltage equation for the VSI nonlinearity in the xy subspace, a stable domain for anti-dead-spot control is constructed, with current and total disturbance as the estimation targets. The discretized ESO can be designed as the algorithm-based stable domain for overcoming magnetic field blind spots, achieving a gain effect.

[0071] Both voltage and current can be considered finite because they are limited by the DC bus and inverter. Under the condition that all index parameters are stable, the sampling period changes smoothly and takes the minimum value, so the current parameter of the magnetic field is bounded.

[0072] Based on coyote adaptation, the median of the j-dimensional variable across all coyotes in the dataset is calculated when different values ​​are taken. Coyotes undergo a natural life cycle of birth and death. To simulate genetic influences, newborn coyotes are labeled with names, which are characterized by a combination of parental social status and environmental factors. Ultimately, through a process of survival of the fittest among different coyotes, only the strongest indicator parameters can adapt.

[0073] The random individual coyote in the wolf pack is an arbitrary two-dimensional problem to be solved, and finally the nonlinear error of the magnetic field dead angle can be cracked so that the error of the constructed stable mathematical model is minimized and the output stability of the rotor magnetic field dead point is optimized.

[0074] Through the above eight formulas, the Coyote Genetic Algorithm can be used to perform multi-angle control of the rotor operation in the nonlinear PDS-bearingless permanent magnet synchronous motor.

[0075] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art will be able to modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. Bearingless permanent magnet synchronous motor with multi-dead angle PDS control using nonlinear harmonic adaptive neural network, characterized by: The invention comprises a stator, which comprises a stator core, a surface of which is provided with a suspension winding and a torque winding, an interior of the stator is provided with a rotor, which comprises a rotor core, a magnetic bearing mounted on the stator, a permanent magnet mounted on the magnetic bearing, and a rotating wind plate mounted at one end of the rotor.

2. The bearingless permanent magnet synchronous motor with nonlinear harmonic adaptive neural network multi-dead angle PDS control according to claim 1 is characterized in that: The stator winding consists of two sets of symmetrical Y-connected windings, with a spatial phase difference of 30°. Using vector space decoupling (VSD) coordinate transformation, each component can be decomposed into three orthogonal subspaces: the α-β subspace, the xy subspace, and the o1-o2 subspace. This allows the mathematical model of the motor to be established: In As, Axy are the bidirectional inductances of dq axis respectively, Ms is the magnetic field that hinders the current in the stator, Q m is the permanent magnet quantity, ω is the rotor speed, t and n are the voltage and current, and the subscripts "dq" and "xy" denote variables in the dq subspace and xy subspace, respectively.

3. The bearingless permanent magnet synchronous motor with nonlinear harmonic adaptive neural network multi-dead angle PDS control according to claim 1, characterized in that: The mathematical model of the nonlinear error of the permanent magnet synchronous motor inverter can be expressed as: The superscript "qbz" represents the reference value, Δtd, Δtq, Δtx, and Δty are the dq-axis and xy-axis voltage errors caused by the inverter nonlinearity, respectively; The traditional DPCC method is based on a discretized motor model; according to (4), the discrete domain voltage equation considering the VSI nonlinearity in the xy subspace can be expressed as: Where Ixy=[tx ty]Tt xy=[tx ty]T.Δtxy=[ΔtxΔty]T According to (5), considering the current and total disturbance Ixy as estimation objects, the discretized ESO can be designed as: in and are the estimated values ​​of ixy and dxy respectively; β1 and β2 are the gains of ESO; It can be further calculated as From equation (6), it can be seen that both voltage txy and current ixy can be considered finite because they are limited by the DC bus and inverter. At the same time, dxy(k) and dxy(k+1) have very small changes in the sampling period, so Dxy can be regarded as a constant bounded value.

4. The bearingless permanent magnet synchronous motor with nonlinear harmonic adaptive neural network multi-dead angle PDS control according to claim 1, characterized in that: As the leader of the pack, the alpha wolf is less adaptable to the environment than other coyotes; the leader wolf and cultural trends in the wolf pack can be described as in, and is the median of the e-dimensional variable of all coyotes in the entire p group when Nc takes different values ​​at the tth time; Coyotes go through a natural life cycle of birth and death; to simulate genetic influences, the emergence of newborn coyotes is marked It is characterized by a convergence of parents' social status and environmental factors; where m1 and m2 are random individual coyotes in the wolf pack x, e1 and e2 are arbitrary two-dimensional representations of the problem to be solved, and Ge and gande range between 0 and 1.

Images