Method for correcting detection deviation of radial displacement sensor of bearingless permanent magnet slice motor

Abstract

The invention discloses a method for correcting the detection deviation of a radial displacement sensor of a bearingless permanent magnet slice motor, which can accurately calculate the eccentric direction and distance through suspension current and quickly correct the eccentric direction and distance. Mathematical models of unbalanced magnetic pulling force caused by controllable suspension force and eccentricity of the bearingless permanent magnet slice motor under a rotor coordinate system are deduced respectively; the suspension current is divided into an anti-eccentricity type and an anti-disturbance type, and decoupling between an eccentric distance and a corresponding suspension current component is achieved; in order to eliminate reverse eccentric current, the eccentric direction and size are judged by extracting m-axis and n-axis suspension current components to determine a compensation point, and a cyclic algorithm of stepping tracking minimum current is introduced near the compensation point. The rotor static eccentricity phenomenon caused by the radial displacement sensor can be corrected, the system efficiency is effectively improved, the temperature rise speed is reduced, and compared with an existing method, the method has the advantages of being short in consumed time, low in cost and high in accuracy.

Description

Technical Field

[0001] The invention relates to the field of optimization of bearingless motor control systems, and in particular to a method for correcting detection deviations of radial displacement sensors of bearingless permanent magnet thin-film motors. Background Art

[0002] The rotor of the bearingless permanent magnet thin-film motor can achieve full suspension with five degrees of freedom by electromagnetic force. Compared with traditional motors, its bearingless design makes it easy to seal the rotor to isolate it from external pollution. This type of motor has been widely used in places with high purity requirements such as ventricular assist devices and semiconductor cleaning pumps.

[0003] The axial length of the rotor of the bearingless permanent magnet thin-film motor is smaller than its outer diameter. Therefore, active control is not required in one axial and two torsional degrees of freedom, and stable suspension can be achieved by relying on magnetic resistance. In the two radial degrees of freedom, displacement closed-loop control needs to be applied to ensure that the rotor is suspended at the geometric center of the air gap. In the displacement closed-loop control system, the devices for detecting the rotor position generally use eddy current sensors and linear Hall sensors. However, the output of the displacement sensor is susceptible to changes due to temperature, electromagnetic and other interferences; at the same time, the application of multiple sensors makes it difficult to fully calibrate the output, and it is easy to have problems such as asymmetric assembly, thereby introducing displacement detection deviation, so that the rotor is not in the geometric center of the stator when it is stably suspended, but in a static eccentric state. This eccentricity will lead to uneven distribution of magnetic flux density, generating unbalanced magnetic pull pointing to the minimum direction of the air gap. In order to balance, the displacement closed-loop control system will introduce redundant suspension current components, resulting in reduced motor power density, increased noise, and increased heat generation. In actual experiments, it was measured that an eccentricity of only 0.1mm in a single direction will cause the system input power to increase by 96% compared with the non-eccentric state, so it is necessary to correct it.

[0004] However, there are still few related studies, and there are different defects. For example, the document "Eccentricity compensation for bearingless motor based on the tracking of the minimum antieccentric levitation current" compensates for the detection deviation by cyclically searching for the minimum levitation current. However, this method can only determine the quadrant in which the rotor is located, but cannot identify the specific eccentricity direction and size of the rotor. In addition, the cyclic algorithm used is complex, the correction time is long, and the resource usage is high. Summary of the invention

[0005] Purpose of the invention: In view of the problems existing in the above-mentioned background technology, the present invention proposes a method for correcting the detection deviation of a radial displacement sensor of a bearingless permanent magnet thin-film motor, which can accurately calculate the eccentricity direction and distance through the suspension current and quickly correct it. First, based on the Maxwell stress tensor method and the virtual displacement method in the rotor coordinate system, the mathematical model of the motor is derived, so as to establish the functional relationship between the suspension current and the controllable suspension force as well as the eccentricity distance and the unbalanced magnetic pull; then the suspension current is divided into two categories: anti-eccentricity and anti-disturbance, so as to realize the decoupling between the eccentricity distance and the corresponding suspension current component; secondly, with the goal of eliminating the anti-eccentricity current, the eccentricity direction and size are quickly and directly determined by extracting the m-axis and n-axis suspension current components to determine the compensation point, and the minimum suspension current is tracked in a step-by-step cycle near the compensation point to determine the compensation value; the compensation value is fed back to the displacement closed-loop control system, so as to realize the correction of the deviation. The present invention can accurately determine the direction and size of the static eccentricity of the rotor caused by the radial displacement sensor and quickly correct it, effectively improve the system efficiency and reduce the temperature rise rate, and has the characteristics of short time consumption, low cost and high accuracy compared with the existing methods.

[0006] Technical conditions: To achieve the above purpose, the technical solution of the present invention is: a method for correcting the detection deviation of a radial displacement sensor of a bearingless permanent magnet thin-film motor, comprising:

[0007] The coordinate system when the rotor is located at the geometric center of the air gap, that is, when it is not eccentric, is defined as the stator coordinate system xOy. According to the Maxwell stress tensor method, the x-axis and y-axis component expressions of the controllable suspension force F on the rotor can be obtained by integrating along the air gap circumference Γ:

[0008]

[0009]

[0010] Where r is the rotor radius; l is the axial length of the rotor; μ0 is the vacuum magnetic permeability; B is the air gap flux density; θ is the air gap circumferential angle.

[0011] When the motor is operating normally, there are two magnetic fields, torque and suspension, in the air gap at the same time. Then B can be further expressed as:

[0012] B(θ,t)=B PM,T cos(p T θ-ω T t+λ)+B S cos(p S θ-ω S t+μ)

[0013] In the formula, B PM,T is the composite air gap magnetic flux amplitude of the permanent magnet and the torque magnetic field; B S is the air gap magnetic flux density amplitude of the suspension magnetic field; p Tis the number of torque winding pole pairs, which should be equal to the number of permanent magnet pole pairs; p S is the number of pole pairs of the suspension winding, which should satisfy p S =p T ±1;ω S With ω T are the electrical angular velocities of the torque magnetic field and the suspension magnetic field respectively; λ and μ are the initial phase angles.

[0014] Combining the above two equations, we can get F x With F y The expression is:

[0015] F x =F M cos(μ-λ)

[0016] F y =F M sin(μ-λ)

[0017] In the formula, F M is the Maxwell force amplitude.

[0018] It can be seen from the above formula that the radial suspension force in the stator coordinate system is related to the rotor position, which is not convenient for application. To further decouple, define the rotor coordinate system mO'n, and then transform the above formula to obtain:

[0019] F x = k(i Sm ψ f +i Sn ψ Tn )

[0020] F y =-k(i Sn ψ f -i Sm ψ Tn )

[0021] In the formula, L S is the mutual inductance of the torque winding; N T With N S are the turns of the torque winding and the suspension winding respectively; i Sm with i Sn are the m-axis and n-axis components of the suspension current; Ψ f is the permanent magnet flux; Tn is the n-axis component of the torque winding flux.

[0022] When the radial displacement sensor produces a detection deviation, the permanent magnet rotor of the motor will no longer be suspended at the geometric center of the stator, but will be located at the detection center. This eccentricity will cause the self-inductance of the torque winding to increase, and its expression can be written as:

[0023]

[0024] Where, L Tm,n is the self-inductance component of the torque winding m-axis and n-axis; x e With y e is the rotor eccentricity in the stator coordinate system; k e , C is a constant.

[0025] According to the virtual displacement principle, the unbalanced magnetic pull on the rotor is the partial derivative of the torque winding inductance energy storage and the radial displacement, which is 0 only when the rotor is located at the geometric center of the stator. Tm =0 control, so the expression of unbalanced magnetic pull is:

[0026]

[0027] Where W m I is the energy storage of the torque winding inductance; f is the equivalent excitation current of the permanent magnet; i Tn is the n-axis component of the torque winding current.

[0028] According to the above formula, the unbalanced magnetic pull on the rotor is only related to the n-axis component of the torque winding current and the eccentricity, but has nothing to do with the suspension current. Taking the x-axis as an example, when the torque winding is not flowing with current, the unbalanced magnetic pull F on the rotor can be obtained through finite element simulation. ex Mean and eccentricity x e is a linear relationship, so the corresponding suspension current m-axis component i Sm However, when the motor is actually running, random external disturbances also act on the rotor, causing fluctuations in the m-axis and n-axis components of the suspension current. In order to accurately identify the displacement detection deviation, the suspension current needs to be classified. The suspension current that resists the unbalanced magnetic pull is defined as the anti-eccentricity current i e When the rotor is in a static eccentric state, the anti-eccentric current is a constant DC component, which is easy to extract; the current that resists external disturbances is the anti-disturbance current i r Since the external disturbance force is random, the anti-disturbance current that counteracts it is a fluctuating quantity and needs to be eliminated by low-pass filtering.

[0029] At the same time, since the rotor volume of the bearingless motor is generally small and the load is light, its torque current n-axis component i Tn Much smaller than the permanent magnet equivalent excitation current I f To simplify the analysis, we ignore i Tn The anti-eccentricity current is obtained by low-pass filtering the detected m-axis and n-axis components of the suspension current and extracting the DC component. The anti-eccentricity current is then combined with the above theoretical analysis and simulation results to calculate the specific direction and size of the rotor eccentricity.

[0030] The smaller the eccentric distance, the smaller the corresponding anti-eccentric current. When the rotor is located at the geometric center of the stator, that is, not eccentric, the anti-eccentric current is zero, and the system reaches the optimal operating state. The hysteresis comparator is introduced into the anti-eccentric current elimination algorithm. h is the error limit value, which is only valid when the absolute value of the reverse eccentricity current is greater than i h Continue to execute the subsequent algorithm; j and j h The number of fluctuations and the limit value of the number of fluctuations are respectively. h The number of times is greater than j h Continue to execute the subsequent algorithm. Set i h With j h The purpose is to prevent the system from repeatedly fluctuating in order to eliminate the anti-eccentricity current. The complete correction algorithm is executed only when the absolute value of the anti-eccentricity current and the number of fluctuations exceed the limit value, and the new displacement detection deviation is calculated and compensated. Otherwise, the output result of the previous cycle remains unchanged.

[0031] The displacement detection deviation correction algorithm proposed in this invention is based on the above theoretical derivation and simulation. It is relatively accurate in identifying the eccentricity direction, but it depends on the model parameters when calculating the eccentricity. It is possible that the rotor center is still not located at the stator geometric center after compensation. To solve the above problem, a loop algorithm for stepping and tracking the minimum anti-eccentricity current is added near the compensation point. d With y d is the expansion range; h is the number of cycles; x h With y h Represents the displacement step length increased in each cycle. The implementation method of this loop algorithm is not to directly use the displacement compensation value x calculated by the previous stage e With y e Feedback to the displacement given, but x e ±x d With y e ±y d Feedback is returned (when the eccentricity is in the positive direction, it is increased, and when it is in the opposite direction, it is decreased), and the number of cycles h is set to zero. After the displacement feedback is applied, the previous stage judgment part is skipped in the next cycle, and the cycle algorithm is directly entered. The number of cycles h is increased by 1, and the anti-eccentricity current of this cycle is sampled. If the sampling result is less than the current value of the previous cycle, the cycle continues. At this time, the displacement feedback value is and When the anti-eccentricity current of a certain cycle is greater than the current of the previous cycle, the cycle is exited and the compensation value is determined.

[0032] Compared with the algorithm in the document “Eccentricity compensation for bearingless motor based on the tracking of the minimum antieccentric levitation current”, the detection deviation correction method of the radial displacement sensor for the bearingless permanent magnet thin-film motor proposed in the present invention is to quickly determine the specific direction and size of the eccentricity through the levitation current offset, and only add a step tracking cycle algorithm near the compensation point. Therefore, it is possible to accurately suspend the rotor at the geometric center of the stator through compensation while ensuring rapidity, so that the motor can run in the best state. At this time, the unbalanced magnetic pull is 0, the levitation current is minimum, and the system efficiency is highest.

[0033] The present invention also provides a radial displacement sensor detection deviation correction hardware system for a bearingless permanent magnet thin-film motor, comprising: a power module, a control module and a bearingless permanent magnet thin-film motor; the power module is composed of a sampling circuit and two sets of three-phase full-bridge inverters; the sampling circuit has 6 sets of current sampling channels, which are responsible for sampling three-phase suspension winding current and three-phase torque winding current; the two sets of three-phase full-bridge inverters respectively realize the control of the suspension winding and the torque winding; the control module includes a DSP main control chip and its peripheral circuits, a conditioning circuit and an external AD circuit; the conditioning circuit converts the level of the current sensor output to within 0-3V so as to safely input it into the DSP internal ADC module; the external AD circuit adopts a 12-bit ADC chip to realize high-precision sampling of the Hall sensor output, and transmits the result to the DSP; the DSP, as the main control chip, collects all external data and generates 6 sets of PWM waves through algorithm calculation to control the two sets of inverters.

[0034] The present invention also provides a bearingless permanent magnet thin-film motor, which adopts the above-mentioned radial displacement sensor detection deviation correction method of the bearingless permanent magnet thin-film motor to correct the radial displacement sensor detection deviation.

[0035] Beneficial effects: The present invention proposes a radial displacement sensor detection deviation correction method suitable for bearingless permanent magnet thin-film motors. According to the proposed method, the rotor static eccentricity caused by the radial displacement sensor can be effectively corrected, the system efficiency can be effectively improved, and the temperature rise rate can be reduced. Compared with the existing methods, it has the characteristics of short time consumption, low cost and high accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of a method for correcting a radial displacement sensor detection deviation applicable to a bearingless permanent magnet thin-film motor proposed by the present invention;

[0037] Figure 2 It is a schematic diagram of coordinate system definition;

[0038] Figure 3 is a relationship diagram between the unbalanced magnetic pull force, the eccentric distance and the corresponding anti-eccentric current obtained by finite element simulation;

[0039] Figure 4 This is a schematic diagram of the rotor forces when the motor is operating normally;

[0040] Figure 5 It is a block diagram of the bearingless permanent magnet thin-film motor suspension control system that introduces displacement detection deviation identification and correction algorithms;

[0041] Figure 6 It is the displacement simulation waveform;

[0042] Figure 7 It is the simulated waveform of reverse eccentricity current;

[0043] Figure 8 It is a diagram of the experimental platform;

[0044] Fig. 9 This is the static suspension displacement waveform before correction;

[0045] Fig.10 This is the static suspension current waveform before correction;

[0046] Fig.11 It is the waveform diagram of static suspension displacement after correction;

[0047] Fig.12 It is the waveform of the static suspension current after correction;

[0048] Fig.13 This is the 3000rpm suspension displacement waveform;

[0049] Fig.14 It is the 3000rpm suspension current waveform;

[0050] Fig.15 This is a time-consuming comparison chart of the algorithms. DETAILED DESCRIPTION

[0051] The present invention is further described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] The present invention proposes a radial displacement sensor detection deviation correction method suitable for a bearingless permanent magnet thin-film motor. Figure 1 A flowchart of the method is shown in FIG.

[0053] Embodiment 1

[0054] This embodiment adopts the following bearingless permanent magnet thin-film motor structure: six stator teeth, one pair of poles of the rotor permanent magnet, two pairs of poles of the suspension winding, and the winding structure adopts a centralized type.

[0055] The coordinate system when the rotor is located at the geometric center of the air gap, that is, when it is not eccentric, is defined as the stator coordinate system xOy, such as Figure 2 According to the Maxwell stress tensor method, the integral along the air gap circumference Γ can be used to obtain the x-axis and y-axis component expressions of the controllable suspension force F on the rotor:

[0056]

[0057]

[0058] Where r is the rotor radius; l is the axial length of the rotor; μ0 is the vacuum magnetic permeability; B is the air gap flux density; θ is the air gap circumferential angle.

[0059] When the motor is operating normally, there are two magnetic fields, torque and suspension, in the air gap at the same time. Then B can be further expressed as:

[0060] B(θ,t)=B PM,T cos(p T θ-ω T t+λ)+B S cos(p S θ-ω S t+μ)

[0061] In the formula, B PM,T is the composite air gap magnetic flux amplitude of the permanent magnet and the torque magnetic field; B S is the air gap magnetic flux density amplitude of the suspension magnetic field; p T is the number of torque winding pole pairs, which should be equal to the number of permanent magnet pole pairs; p S is the number of pole pairs of the suspension winding, which should satisfy p S =p T ±1;ω S With ω T are the electrical angular velocities of the torque magnetic field and the suspension magnetic field respectively; λ and μ are the initial phase angles.

[0062] Substituting equation (2) into equation (1) yields F x With F y The expression is:

[0063] F x =F M cos(μ-λ)

[0064] F y =F M sin(μ-λ)

[0065] In the formula, F Mis the Maxwell force amplitude.

[0066] It can be seen from the above formula that the radial suspension force in the stator coordinate system is related to the rotor position, which is not convenient for application. To further decouple, define the rotor coordinate system mO'n as Figure 2 , then transform the above formula to get:

[0067] F x = k(i Sm ψ f +i Sn ψ Tn )

[0068] F y =-k(i Sn ψ f -i Sm ψ Tn )

[0069] In the formula, L S is the mutual inductance of the torque winding; N T With N S are the turns of the torque winding and the suspension winding respectively; i Sm with i Sn are the m-axis and n-axis components of the suspension current; Ψ f is the permanent magnet flux; Tn is the n-axis component of the torque winding flux.

[0070] When the radial displacement sensor produces a detection deviation, the permanent magnet rotor of the motor will no longer be suspended at the stator geometric center point O, but will be located at the detection center point O'. This eccentricity will cause the self-inductance of the torque winding to increase, and its expression can be written as:

[0071]

[0072] Where, L Tm,n is the self-inductance component of the torque winding m-axis and n-axis; x e With y e is the rotor eccentricity in the stator coordinate system; k e , C is a constant.

[0073] According to the virtual displacement principle, the unbalanced magnetic pull on the rotor is the partial derivative of the torque winding inductance energy storage and the radial displacement, which is 0 only when the rotor is located at the geometric center of the stator. Tm =0 control, so the expression of unbalanced magnetic pull is:

[0074]

[0075] Where W m I is the energy storage of the torque winding inductance;f is the equivalent excitation current of the permanent magnet; i Tn is the n-axis component of the torque winding current.

[0076] According to the above formula, the unbalanced magnetic pull on the rotor is only related to the n-axis component of the torque winding current and the eccentricity, but has nothing to do with the suspension current. A finite element simulation model is built using a 6-slot 2-pole bearingless permanent magnet thin-film motor as an example, and its parameters are shown in Table 1.

[0077] Table 1 Motor parameters

[0078]

[0079] Taking the x-axis as an example, when the torque winding is not flowing with current, the simulation results of the relationship between the unbalanced magnetic pull on the rotor, the eccentric distance and the corresponding anti-eccentric current are as follows: Figure 3 As shown. It can be seen that F ex Mean and eccentricity x e is a linear relationship, so the corresponding suspension current m-axis component i Sm However, when the motor is actually running, random external disturbances also act on the rotor, causing fluctuations in the m-axis and n-axis components of the suspension current. The rotor force diagram is shown in the figure Figure 4 As shown. In order to accurately identify the displacement detection deviation, the suspension current needs to be classified. The suspension current that resists the unbalanced magnetic pull is defined as the anti-eccentric current i e When the rotor is in a static eccentric state, the anti-eccentric current is a constant DC component, which is easy to extract; the current that resists external disturbances is the anti-disturbance current i r Since the external disturbance force is random, the anti-disturbance current that counteracts it is a fluctuating quantity and needs to be eliminated by low-pass filtering.

[0080] At the same time, since the rotor volume of the bearingless motor is generally small and the load is light, its torque current n-axis component i Tn Much smaller than the permanent magnet equivalent excitation current I f To simplify the analysis, we ignore i Tn The anti-eccentricity current is obtained by low-pass filtering the detected m-axis and n-axis components of the suspension current and extracting the DC component. The anti-eccentricity current is then combined with the above theoretical analysis and simulation results to calculate the specific direction and size of the rotor eccentricity.

[0081] according to Figure 3 The smaller the eccentric distance, the smaller the corresponding anti-eccentric current. When the rotor is located at the geometric center of the stator, that is, not eccentric, the anti-eccentric current is zero. At this time, the system reaches the best operating state. The block diagram of the bearingless permanent magnet thin-film motor suspension control system that introduces the displacement detection deviation identification and correction algorithm is as follows: Figure 5As shown. Among them, LPF stands for low-pass filter; the hysteresis comparator is introduced into the algorithm for eliminating the reverse eccentricity current, i h is the error limit value, which is only valid when the absolute value of the reverse eccentricity current is greater than i h Continue to execute the subsequent algorithm; j and j h The number of fluctuations and the limit value of the number of fluctuations are respectively. h The number of times is greater than j h Continue to execute the subsequent algorithm. Set i h With j h The purpose is to prevent the system from repeatedly fluctuating in order to eliminate the anti-eccentricity current. The complete correction algorithm is executed only when the absolute value of the anti-eccentricity current and the number of fluctuations exceed the limit value, and the new displacement detection deviation is calculated and compensated. Otherwise, the output result of the previous cycle remains unchanged.

[0082] The displacement detection deviation correction algorithm proposed in the present invention is based on the above theoretical derivation and simulation. It is relatively accurate in identifying the eccentricity direction, but it depends on the model parameters when calculating the eccentricity. It is possible that the rotor center is still not located at the stator geometric center after compensation. To solve the above problem, a cyclic algorithm for stepping and tracking the minimum anti-eccentricity current is added near the compensation point, that is, Figure 5 The lower right part of the box diagram. d With y d is the expansion range; h is the number of cycles; x h With y h Represents the displacement step length increased in each cycle. The implementation method of this loop algorithm is not to directly use the displacement compensation value x calculated by the previous stage e With y e Feedback to the displacement given, but x e ±x d With y e ±y d Feedback is returned (when the eccentricity is in the positive direction, it is increased, and when it is in the opposite direction, it is decreased), and the number of cycles h is set to zero. After the displacement feedback is applied, the previous stage judgment part is skipped in the next cycle, and the cycle algorithm is directly entered. The number of cycles h is increased by 1, and the anti-eccentricity current of this cycle is sampled. If the sampling result is less than the current value of the previous cycle, the cycle continues. At this time, the displacement feedback value is and When the anti-eccentricity current of a certain cycle is greater than the current of the previous cycle, the cycle is exited and the compensation value is determined.

[0083] Compared with the algorithm in the document “Eccentricity compensation for bearingless motor based on the tracking of the minimum antieccentric levitation current”, the detection deviation identification and correction method of the radial displacement sensor for bearingless permanent magnet thin-film motor proposed in the present invention is to directly and quickly determine the specific direction and size of the eccentricity by utilizing the levitation current offset, and then only add a step tracking cycle algorithm near the compensation point. Therefore, it is possible to accurately suspend the rotor at the geometric center of the stator through compensation under the premise of ensuring rapidity, so that the motor can run in the best state. At this time, the unbalanced magnetic pull is 0, the levitation current is minimum, and the system efficiency is highest.

[0084] Based on Simulink, a simulation model of the bearingless permanent magnet thin-film motor control system was built to verify the rapidity and effectiveness of the proposed displacement detection deviation identification and correction algorithm. The rotor eccentricity error caused by the displacement sensor causes the rotor to be located not at the geometric center of the stator during stable operation, but at the detection center. To simulate this deviation, an offset is added to the radial displacement feedback loop, with an x-axis offset of -0.2mm and a y-axis offset of 0.1mm, respectively, so that the rotor runs at the position of (0.2mm, -0.1mm). At the same time, a random disturbance force is added to the displacement calculation module to simulate external disturbances. Figure 6 The displacement waveform is shown. The error limit value of the correction algorithm in the simulation is i h Take 0.02A, the fluctuation number limit value j h Take 10, and take x for the external expansion range d With y d Take 0.1mm, displacement increase step x h With y h Take 0.01mm and the cycle period as 0.1s. It can be seen that the rotor runs stably at the zero point of the stator geometric center before the displacement detection deviation occurs. When the deviation occurs at 0.7s, the rotor turns to run stably at the position of (0.2mm, -0.1mm), and the anti-eccentricity current increases. The correction algorithm is started at 1.5s, and the anti-eccentricity current is detected to be greater than the limit value i in 10 consecutive control cycles. h Then the eccentricity direction and size are determined, and an external expansion range of 0.1mm is added, that is, the x-axis and y-axis compensation values ​​at this moment are -0.1mm and 0mm respectively. Then the loop algorithm is started, and the x-axis displacement compensation value is reduced by 0.01mm and the y-axis displacement compensation value is increased by 0.01mm in each cycle, and the anti-eccentricity current is detected until it reaches the allowable error range of the hysteresis comparator.

[0085] The waveforms of the m-axis and n-axis components of the anti-eccentricity current after the correction algorithm is started are as follows: Figure 7As shown in the figure, it can be seen that both the x-axis and the y-axis reach the minimum anti-eccentric current point after about 10 cycles, and the suspension current gradually decreases in each cycle. d Therefore, the compensation value of the first cycle is 0mm, that is, no compensation. In the next cycle, the compensation value is given as 0.01mm and i is monitored. en The fluctuation of the anti-eccentricity current component is due to the fact that the low-pass filter cannot completely filter out the pulsation, so that part of the anti-disturbance current is mixed into the anti-eccentricity current. However, as long as the limit value i is reasonably set, h If the fluctuation is within the error range allowed by the hysteresis comparator, the correction algorithm will not be triggered, thus maintaining the stability of the system.

[0086] based on Figure 8 The experimental platform shown in the figure verifies the proposed displacement detection deviation identification and correction algorithm. The specific parameters of the experimental prototype have been given in the previous article. When the correction algorithm does not work and the speed is set to 0rpm, that is, static suspension, the rotor displacement waveform and the related suspension current waveform are as follows: Fig. 9 and Fig.10 As shown. It can be seen that the x-axis and y-axis displacements of the rotor can be stabilized near the detection zero point. However, from the suspension current waveform, it can be obtained that the m-axis component i of the anti-eccentric current at this time em With n-axis component i en The magnitudes are -0.36A and -0.13A respectively; taking the suspension winding phase A as an example, its current i SA The average value is about -0.37A, that is, the rotor is actually in a static eccentric state, which introduces redundant current components into the suspension control system. As mentioned above, the cause of this phenomenon may be assembly error, incomplete calibration of each Hall output, or temperature and electromagnetic interference. After adding the correction algorithm proposed by the present invention to the program, the rotor displacement waveform and the related suspension current waveform during stable suspension are as follows: Fig.11 and Fig.12 It can be seen that after the correction algorithm is applied, the rotor is not suspended at the detection zero point, but the average value of the suspended A-phase current is reduced by 0.37A compared with that before correction, the anti-eccentricity current is about 0A, and the rotor is located at the true stator geometric center.

[0087] Through the above static suspension experiment, it has been basically determined that the displacement detection deviation of this set of Hall sensors caused by assembly error and incomplete output calibration is 0.168mm eccentricity to the x-axis and -0.061mm eccentricity to the y-axis. After correction, the rotor can theoretically run stably at the geometric center of the stator, but the output of the Hall sensor is still susceptible to external interference and changes. For simulation, the x-axis displacement detection feedback loop and the y-axis displacement detection feedback loop are given offsets of +0.2mm and -0.2mm respectively. When the given speed is 3000rpm, the rotor displacement waveform and the related suspension current waveform after the correction algorithm is started at a certain moment are as follows Fig.13 and Fig.14 It can be seen that the algorithm proposed in the present invention recognizes that the reverse eccentricity current i is greater than the limit value. em with i en After that, the eccentricity direction and magnitude can be effectively determined, and steps are made near the compensation point every 0.1s cycle, so that the anti-eccentricity current gradually decreases to about 0A, and the peak-to-peak value of the suspension A phase current decreases by 1A. The total time consumption of the identification and correction algorithm in the experiment is about 2s, which is 90% shorter than the cycle algorithm time consumption of the literature "Eccentricity compensation for bearingless motor based on the tracking of the minimum antieccentric levitation current". Fig.15 As shown, the algorithm is more convenient and easy to implement, which verifies its speed and effectiveness.

[0088] The present invention also provides a radial displacement sensor detection deviation correction hardware system for a bearingless permanent magnet thin-film motor, comprising: a power module, a control module and a bearingless permanent magnet thin-film motor; the power module is composed of a sampling circuit and two sets of three-phase full-bridge inverters; the sampling circuit has 6 sets of current sampling channels, which are responsible for sampling three-phase suspension winding current and three-phase torque winding current; the two sets of three-phase full-bridge inverters respectively realize the control of the suspension winding and the torque winding; the control module includes a DSP main control chip and its peripheral circuits, a conditioning circuit and an external AD circuit; the conditioning circuit converts the level of the current sensor output to within 0-3V so as to safely input it into the DSP internal ADC module; the external AD circuit adopts a 12-bit ADC chip to realize high-precision sampling of the Hall sensor output, and transmits the result to the DSP; the DSP, as the main control chip, collects all external data and generates 6 sets of PWM waves through algorithm calculation to control the two sets of inverters.

[0089] The present invention also provides a bearingless permanent magnet thin-film motor, which adopts the above-mentioned radial displacement sensor detection deviation correction method of the bearingless permanent magnet thin-film motor to correct the radial displacement sensor detection deviation.

[0090] The above are only preferred specific implementation methods of the present application, but the protection scope of the present application is not limited thereto. It should be pointed out that any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.

Claims

1. A method for correcting the detection deviation of a radial displacement sensor of a bearingless permanent magnet thin-film motor, characterized in that: The steps include: Based on the Maxwell stress tensor method and the virtual displacement method, the mathematical models of the controllable suspension force of the bearingless permanent magnet thin-film motor and the unbalanced magnetic pull caused by eccentricity are derived respectively in the rotor coordinate system, so as to establish the functional relationship between the suspension current and the controllable suspension force, and the eccentric distance and the unbalanced magnetic pull. The suspension current is divided into two categories: anti-eccentricity and anti-disturbance, to achieve decoupling between the eccentric distance and the corresponding suspension current component; With the goal of eliminating the reverse eccentricity current, the eccentricity direction and magnitude are determined by extracting the m-axis and n-axis suspension current components to determine the compensation point; The minimum suspension current is tracked in a step-by-step cycle near the compensation point to determine the compensation value; The compensation value is fed back to the displacement closed-loop control system to correct the deviation.

2. The method for correcting the detection deviation of the radial displacement sensor of the bearingless permanent magnet thin-film motor according to claim 1 is characterized in that: The process of deriving the expression of the controllable radial suspension force of the bearingless permanent magnet thin-film motor in the rotor coordinate system based on the Maxwell stress tensor method includes: based on the Maxwell stress tensor method, integrating along the circumference of the air gap to obtain the expressions of the x-axis and y-axis components of the controllable suspension force on the rotor; determining the magnetic field existing in the air gap when the motor is operating normally; combining the above two equations to derive the expression of the controllable radial suspension force in the stator coordinate system xOy; defining the rotor coordinate system mO'n according to the position of the center of the circle when the rotor is statically eccentric; and deriving the expression of the controllable radial suspension force in the rotor coordinate system through coordinate transformation.

3. The method for correcting the detection deviation of the radial displacement sensor of the bearingless permanent magnet thin-film motor according to claim 1 is characterized in that: The process of deriving the expression of the unbalanced magnetic pull on the rotor of a bearingless permanent magnet thin-film motor when it is running statically eccentrically in the rotor coordinate system based on the virtual displacement method includes: determining the relationship between the eccentric distance and the self-inductance of the motor torque winding; deriving the expression of the inductance energy storage of the torque winding; and deriving the expression of the unbalanced magnetic pull by taking the partial derivative between the inductance energy storage of the torque winding and the radial displacement based on the virtual displacement method.

4. The method for correcting the detection deviation of the radial displacement sensor of a bearingless permanent magnet thin-film motor according to claim 1, characterized in that: The process of dividing the suspension current into two categories, namely, anti-eccentricity and anti-disturbance, and realizing decoupling between the eccentric distance and the corresponding suspension current component includes: defining the suspension current that resists the unbalanced magnetic pull as the anti-eccentric current, and when the rotor is in a static eccentric state, the anti-eccentric current is a constant DC component; the current that resists external disturbances is the anti-disturbance current, and this current component is a fluctuation quantity, which is eliminated by using a low-pass filter; and realizing decoupling between the eccentric distance and the corresponding suspension current component according to the fact that the magnitude of the anti-eccentric current that resists the unbalanced magnetic pull is proportional to the eccentric distance.

5. The method for correcting the detection deviation of the radial displacement sensor of a bearingless permanent magnet thin-film motor according to claim 1, characterized in that: The process of eliminating the reverse eccentricity current as the goal, judging the eccentricity direction and magnitude by extracting the m-axis and n-axis suspension current components and determining the compensation point includes: based on coordinate transformation, transforming the detected three-phase suspension current of the motor into the suspension current in a two-phase rotating coordinate system; performing low-pass filtering on the current to extract the m-axis and n-axis reverse eccentricity current components; based on finite element simulation, obtaining the corresponding relationship between the controllable radial force and the suspension current; based on finite element simulation, obtaining the corresponding relationship between the unbalanced magnetic pull and the eccentricity distance; combining the above two corresponding relationships obtained by finite element simulation with the detected m-axis and n-axis components of the reverse eccentricity current, solving the magnitude of the rotor eccentricity in the x-axis and y-axis directions, and determining the compensation point.

6. The method for correcting the detection deviation of the radial displacement sensor of a bearingless permanent magnet thin-film motor according to claim 1, characterized in that: A cyclic algorithm of step-tracking minimum current is introduced near the calculated compensation point to determine the optimal compensation value, including: the implementation method of the cyclic algorithm is not to directly feed back the displacement compensation value calculated by the previous stage to the displacement setting, but to add or subtract an external expansion range from the calculated value and then feed it back. When the eccentricity is in the positive direction, it is added, and when it is opposite, it is subtracted. At this time, the number of cycles is set to zero; after the displacement feedback takes effect, the previous stage calculation part is skipped in the next cycle period, and the cyclic algorithm is directly entered. The number of cycles is increased by 1, and the anti-eccentricity current of this cycle is sampled. If the sampling result is less than the current value of the previous cycle period, the cycle continues. At this time, the displacement feedback value is the action value of the previous cycle period minus or plus the number of cycles multiplied by the step size, until the anti-eccentricity current of a certain cycle period is greater than the current of the previous cycle, then the cycle is jumped out to determine the compensation value.

7. A hardware system for correcting radial displacement sensor detection deviation of a bearingless permanent magnet thin-film motor based on the method described in any one of claims 1 to 6, characterized in that: include: Power module, control module and bearingless permanent magnet thin-film motor; the power module consists of a sampling circuit and two sets of three-phase full-bridge inverters; the sampling circuit has 6 sets of current sampling channels, which are responsible for sampling the three-phase suspension winding current and the three-phase torque winding current; the two sets of three-phase full-bridge inverters respectively control the suspension winding and the torque winding; the control module includes a DSP main control chip and its peripheral circuits, a conditioning circuit and an external AD circuit; the conditioning circuit converts the level of the current sensor output to within 0-3V so as to safely input the ADC module inside the DSP; the external AD circuit uses a 12-bit ADC chip to achieve high-precision sampling of the Hall sensor output and transmit the result to the DSP; DSP is the main control chip, which collects all external data and generates 6 groups of PWM waves through algorithm calculation to control two sets of inverters.

8. A bearingless permanent magnet thin-film motor, characterized in that: It adopts the method described in any one of claims 1-6 to correct the detection deviation of the radial displacement sensor.

Images