PERMANENT-MAGNET FAULT-TOLERANT IN-WHEEL MOTOR BASED ON ACTIVE SENSORLESS STRATEGY AND DRIVE AND DESIGN METHODS THEREOF

Abstract

The present disclosure provides a permanent-magnet fault-tolerant in-wheel motor based on an active sensorless strategy and drive and design methods thereof. The present disclosure proposes the permanent-magnet fault-tolerant in-wheel motor drive system based on an active sensorless strategy by considering sensorless operation performance in advance in a motor design stage. The present disclosure adopts fractional-slot concentrated windings, and ingeniously arranges alternating poles, a multi-layer magnetic barrier, and auxiliary permanent magnets, thus improving a sensorless operation accuracy of the motor while ensuring fault tolerance of the motor. The present disclosure proposes a frequency-band-adaptive secondary harmonic suppression strategy at a control layer to suppress an influence of a secondary salient harmonic on position observation and improve dynamic response performance of a system.

Claims

1. A permanent-magnet fault-tolerant in-wheel motor based on an active sensorless strategy, comprising a rotating shaft, a stator, and a rotor that are arranged in order from inside to outside, wherein the stator is composed of a stator yoke, armature teeth, stator slots, and fault-tolerant teeth; the armature teeth and the fault-tolerant teeth are evenly spaced along an outer circumference of the stator; the armature teeth are wound with armature windings, respectively; two adjacent ones of the armature teeth are isolated by one of the fault-tolerant teeth; main permanent magnets and core poles are evenly spaced along an inner circumference of the rotor; the main permanent magnets are surface-embedded arc permanent magnets, and each of surface-embedded main permanent magnet poles forms a pair of magnetic poles with adjacent one of the core poles; a q-axis magnetic barrier close to an air gap side is formed between one of the surface-embedded main permanent magnet poles and one of the core poles, and the q-axis magnetic barrier forms an uneven air gap with the outer circumference of the stator; a multi-layer arc magnetic barrier close to an outer circumference side is formed between one of the surface-embedded main permanent magnet poles and one of the core poles, and a magnetic bridge is formed between layers of the multi-layer arc magnetic barrier; a rectangular auxiliary permanent magnet is provided between the q-axis magnetic barrier and the arc magnetic barrier close to the air gap; and the auxiliary permanent magnet is connected in series with one of the main permanent magnets.

2. The permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, wherein the stator and the rotor each are formed by laminating a magnetically conductive material such as a silicon steel sheet, with a lamination factor of 0.96; and the armature windings are made of an enameled copper conductor material, and the armature windings are single-layer concentrated windings.

3. The permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, wherein a number of stator teeth is a multiple of 2m, and a difference between the number of the stator teeth and a number of rotor poles is 2, m denoting a phase number of the motor; a sum of a number of the permanent magnet poles and a number of the core poles is P.sub.s; the number of the rotor poles is P.sub.s, a number of the main permanent magnet poles is P.sub.s, a number of auxiliary permanent magnet poles is P.sub.a, and the number of the core poles is P.sub.f, wherein P.sub.m+P.sub.f=P.sub.s=P.sub.a.

4. The permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, wherein the multi-layer arc magnetic barriers are each located between one of the main permanent magnet poles and one of the core poles and are evenly distributed along a circumference of the rotor; the auxiliary permanent magnets are each located between one of the arc magnetic barriers and one of the q-axis magnetic barriers and are evenly distributed along an outer circumference of the rotor; the q-axis magnetic barriers each are centered on a point of O.sub.1 and have a radius of R.sub.1; and the multi-layer arc magnetic barriers each are centered on a point of O.sub.2, and have radii of R.sub.2 and R.sub.3 and a thickness of H.sub.0.

5. The permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, wherein the main permanent magnets and the auxiliary permanent magnets are both made of neodymium-iron-boron permanent magnet steel; magnetizing directions of the main permanent magnets each point to a center of a circle, and magnetizing directions of the auxiliary permanent magnets are each along a circumference of the circle; and the magnetizing directions of two adjacent ones of the auxiliary permanent magnets are opposite.

6. A design method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, comprising the following steps: step 1-1: initially determining, by a slot-pole combination design method of a fault-tolerant motor, a number of teeth of the stator and a number of pole pairs of the rotor; and determining slot vectoring based on a principle of a maximum fundamental resultant vector; step 1-2: arranging the alternating surface-embedded main permanent magnet poles, wherein the main permanent magnet poles have magnetizing directions pointing to the air gap and form a closed magnetic circuit with the core poles, to improve utilization of the permanent magnets; step 1-3: providing the q-axis magnetic barriers and the multi-layer arc magnetic barriers between the main permanent magnet poles and the core poles to increase a quadrature-axis reluctance, to realize a negative saliency of the motor; step 1-4: providing the rectangular auxiliary permanent magnets with circumferential magnetizing directions between the q-axis magnetic barriers and the multi-layer arc magnetic barriers to increase the quadrature-axis reluctance while providing an auxiliary magnetic field, to reduce a quadrature-axis inductance and further increase a negative saliency effect of the motor; and step 1-5: optimizing parameters of the main permanent magnets, the auxiliary permanent magnets, the q-axis magnetic barriers, and the multi-layer arc magnetic barriers, to obtain a desired negative saliency; and considering sensorless operation performance in advance in a motor design stage, to complete the design of the permanent-magnet fault-tolerant in-wheel motor with active sensorless operation.

7. A drive method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, comprising the following steps: step 1: designing the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 1, wherein sensorless operation performance is comprehensively considered in a design stage, to obtain desired sensorless operation performance and fault tolerance; step 2: in order to give full play to the superior sensorless operation performance of the permanent-magnet fault-tolerant in-wheel motor in the step 1, proposing a frequency-band-adaptive secondary harmonic suppression and position error signal extraction algorithm, to suppress an influence of a secondary salient harmonic on position observation, and improve dynamic response performance of sensorless control; and step 3: constructing a sensorless drive control system for a five-phase permanent-magnet fault-tolerant in-wheel motor according to the new permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy in the step 1 and a sensorless control method in the step 2.

8. The drive method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 7, wherein the step of proposing the frequency-band-adaptive secondary harmonic suppression and position error signal extraction algorithm to suppress the influence of the secondary salient harmonic on the position observation specifically comprises: adjusting, by an adaptive linear neuron filter based on a recursive least squares algorithm, a filter coefficient through the adaptive algorithm to suppress a specific sub-harmonic in a position error signal; processing a harmonic reference signal r(n) and an adjustable weight component x(k) to obtain a desired output signal y(n) of the filter; and subtracting the desired output signal y(n) of the filter from an input signal U(n) to obtain a desired fundamental signal Y(n), wherein calculations by the recursive least squares algorithm are as follows: $\{\begin{array}{c}{y}_1\left(n\right)=x_{11}\left(n-1\right)r_{11}\left(n\right)+x_{21}\left(n-1\right)r_{21}\left(n\right)\\ Y\left(n\right)=U\left(n\right)-y_1\left(n\right)\end{array}$ wherein, y.sub.1(n) denotes a harmonic estimator; x.sub.11(n?1) and x.sub.21(n?1) each denote an estimated amplitude of a harmonic component; r.sub.11(n)=sin (2{circumflex over (?)}.sub.e) and r.sub.21(n)=cos (2{circumflex over (?)}.sub.e) each denote a harmonic reference signal; {circumflex over (?)}.sub.e denotes an estimated rotor position; Y(n) denotes a filtered output; U(n) denotes a filtered input; and adjustable filter coefficients x.sub.11(n) and x.sub.21(n) are updated online based on the harmonic reference signal, and are expressed as follows: $\{\begin{array}{c}{x}_{11}\left(n\right)=x_{11}\left(n-1\right)+k_{11}\left(n\right)Y\left(n\right)\\ x_{21}\left(n\right)=x_{21}\left(n-1\right)+k_{21}\left(n\right)Y\left(n\right)\end{array}$ wherein, gain coefficients k.sub.11(n) and k.sub.21(n) are expressed as follows: $\{\begin{array}{c}{k}_{11}\left(n\right)=\frac{H_{11}\left(n-1\right)r_{11}\left(n\right)}{?+H_{11}\left(n-1\right)r_{11}^2\left(n\right)}\\ k_{21}\left(n\right)=\frac{H_{21}\left(n-1\right)r_{21}\left(n\right)}{?+H_{21}\left(n-1\right)r_{21}^2\left(n\right)}\end{array}$ wherein, ? denotes a forgetting factor, 0<?<1; an inverse of an autocorrelation matrix H.sub.1(n) is converted into two scales, namely H.sub.11(n) and H.sub.21(n), to realize a simple and fast implementation of the recursive least squares algorithm; and the H.sub.11(n) and H.sub.21(n) are expressed as follows: $\{\begin{array}{c}{H}_{11}\left(n\right)=\frac{H_{11}\left(n-1\right)-k_{11}\left(n\right)H_{11}\left(n-1\right)r_{11}\left(n\right)}{?}\\ H_{21}\left(n\right)=\frac{H_{21}\left(n-1\right)-k_{21}\left(n\right)H_{21}\left(n-1\right)r_{21}\left(n\right)}{?}\end{array}.$

9. The drive method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 8, further comprising: performing linear operation, based on a characteristic that an all-pass network filter only changes a signal phase, on signals before and after filtering, to construct an adaptive band-pass filter and an adaptive band-notch filter, wherein a transfer function of a typical second-order all-pass network filter is: $A\left(z\right)=\frac{d-p\left(1+d\right)z^{-1}+z^{-2}}{1-p\left(1+d\right)z^{-1}+d{z}^{-2}}$ wherein, d denotes a correlation coefficient of a filter bandwidth and $d=\frac{1-\tan \left({?}_m{T}_s/2\right)}{1+\tan \left({?}_m{T}_s/2\right)};$ ?.sub.m denotes the filter bandwidth with a 3 dB attenuation; T.sub.s denotes a digital sampling period; p denotes a correlation coefficient of a filtering frequency and p=cos (?.sub.nT.sub.s); ?.sub.n denotes a resonance frequency point; and a resonance frequency in the adaptive band filter is set as:
?.sub.n=?.sub.c+{circumflex over (?)}.sub.r wherein, ?.sub.c denotes a frequency of an injected high-frequency signal; {circumflex over (?)}.sub.r denotes an estimated motor speed; the resonance frequency is automatically adjusted with a motor speed to reduce a phase delay caused by the filter; and the filter bandwidth is set as:
?.sub.m=?.sub.b+?|?.sup.*?{circumflex over (?)}.sub.r| wherein, ?.sub.b denotes an adjustable bandwidth; ? denotes a dynamic adjustment factor; ?.sub.* denotes a given speed; when the motor runs stably, the dynamic adjustment factor does not work, and the filter bandwidth depends on ?.sub.b; when the motor runs at a variable speed, the dynamic adjustment factor works, and the filter bandwidth is adjusted adaptively according to an error between an actual speed and the given speed to improve the dynamic response performance of the sensorless control; and a modulated current is expressed as:
?.sub.q1.sup.*=?.sub.q1h*sin (?.sub.ht)=f.sub.k({tilde over (?)}.sub.e)+HF(2?.sub.ht) wherein, ?.sub.q1h denotes a q.sub.1-axis high-frequency response current; HF(2?.sub.ht)=?.sub.q1h cos (2?.sub.ht); ?.sub.h denotes an angular frequency of the injected high-frequency signal; ?.sub.q1h denotes an amplitude of a 2.sup.nd-order high-frequency injected harmonic; f.sub.k({tilde over (?)}.sub.e) denotes a position error function; {tilde over (?)}.sub.e denotes an estimated angular position error; and the modulated current comprises the 2.sup.nd-order high-frequency injected harmonic, so ?.sub.c in the adaptive band-notch filter is set as 2?.sub.h, to obtain the position error signal f.sub.k({tilde over (?)}.sub.e).

10. The drive method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to claim 7, wherein the constructing the sensorless drive control system of the five-phase permanent-magnet fault-tolerant in-wheel motor specifically comprises: configuring the five-phase permanent-magnet fault-tolerant in-wheel motor as a system drive motor module to output an electromagnetic torque T.sub.e and five-phase currents i.sub.abcde; configuring a Park transform module to obtain feedback quadrature- and direct-axis current signals i.sub.d1g1d3g3 based on the five-phase currents i.sub.abcde; configuring a sensorless control module based on the frequency-band-adaptive secondary harmonic suppression and position error signal extraction algorithm to estimate a rotor position {circumflex over (?)}.sub.e and a speed {circumflex over (n)}.sub.e based on the feedback current signal i.sub.q1; configuring a maximum torque per ampere (MTPA) control module based on an equation method to optimally distribute a difference between a given speed n and an estimated speed {circumflex over (n)}.sub.e through a given torque output by a proportional integral (PI) controller to obtain optimal given quadrature- and direct-axis currents i.sub.d1q1.sup.*; configuring the PI controller to adjust a deviation between given currents i.sub.d1q1d3q3.sup.* and feedback currents i.sub.d1q1d3q3 to obtain given quadrature- and direct-axis voltage signals U.sub.d1q1d3q3; configuring an inverse-Park transform module to inversely transform the given quadrature- and direct-axis voltage signals U.sub.d1q1d3q3 to obtain voltage signals U.sub.?1?1?3?3 in a static coordinate system; configuring a space vector pulse width modulation (SVPWM) module to modulate the voltage signals U.sub.?1?1?3?3 in a given two-phase static coordinate system into a ten-channel pulse width modulation (PWM) signal required for driving the motor; and configuring an inverter module to output a five-phase voltage signal through the ten-channel PWM signal to provide power to the five-phase permanent-magnet fault-tolerant in-wheel motor.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0052] FIG. 1 is a block diagram of a permanent-magnet fault-tolerant in-wheel motor drive system based on an active sensorless strategy according to the present disclosure.

[0053] FIG. 2 is a structure diagram of a permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to the present disclosure.

[0054] FIG. 3 is a structure diagram of a stator and a distribution diagram of an armature winding, where the stator and the armature winding are shown in FIG. 1.

[0055] FIG. 4 is an enlarged partial structure diagram, with geometric dimensions, of a rotor shown in FIG. 1.

[0056] FIG. 5 is a schematic diagram of magnetization of neodymium-iron-boron permanent magnets on the rotor shown in FIG. 1.

[0057] FIG. 6 is a schematic diagram of a direct axis and a quadrature axis of the rotor shown in FIG. 1.

[0058] FIG. 7 is a block diagram of a design method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to the present disclosure.

[0059] FIG. 8 shows a back electromotive force waveform of the motor according to the present disclosure.

[0060] FIG. 9 shows an inductance waveform of the motor according to the present disclosure.

[0061] FIG. 10 shows d- and q-axis inductance waveforms of the motor according to the present disclosure under load.

[0062] FIG. 11 is a schematic diagram of a rotor position error angle of the motor changingwith a q-axis current according to the present disclosure.

[0063] FIG. 12 is a simplified block diagram of a sensorless control module according to the present disclosure.

[0064] FIG. 13 is a schematic block diagram of an adaptive linear neuron filter according to the present disclosure.

[0065] FIG. 14 is a schematic block diagram of adaptive band filtering according to the present disclosure.

[0066] Reference Numerals: 1. five-phase permanent-magnet fault-tolerant in-wheel motor; 2. Park transform module; 3. sensorless control module; 4. MTPA control module; 5. PI module; 6. inverse-Park transform module; 7. SVPWM module; 8. inverter module; 9. stator; 10. rotor; 11. rotating shaft; 12. armature winding; 13. fault-tolerant tooth; 14. armature tooth; 15. main permanent magnet; 16. core pole; 17. q-axis magnetic barrier; 18. auxiliary permanent magnet; 19. multi-layer arc magnetic barrier; 20. magnetic bridge; 21. stator slot; and 22. stator yoke.

DETAILED DESCRIPTION OF THE EMBODIMENTS

[0067] To make the objectives, technical solutions, and advantages of the present disclosure clearer, the present invention is further described below in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are merely intended to explain the present disclosure, rather than to limit the present disclosure.

[0068] FIG. 1 is a block diagram of a new permanent-magnet fault-tolerant in-wheel motor drive system, with zero- and low-speed operation, based on an active sensorless strategy according to the present disclosure. A five-phase permanent-magnet fault-tolerant in-wheel motor (1) is configured to serve as a system drive motor. A Park transform module (2) and an inverse-Park transform module (6) are configured to decouple a mathematical model in a natural coordinate system. A sensorless control module (3) is configured to estimate a rotor position {circumflex over (?)}.sub.e and a speed {circumflex over (n)}.sub.e. A MTPA control module (4) based on an equation method is configured to distribute an optimal quadrature- and direct-axis currents i.sub.d1q1.sup.* under a given torque according to quadrature- and direct-axis reference current calculation equations

[00007]$i_{d1}^{*}=\frac{-{?}_f+\sqrt{{?}_f^2+8{\left(L_{d1}-L_{q1}\right)}^2{i}_s^2}}{4\left(L_{d1}-L_{q1}\right)}\mathrm{and}{i}_{q1}^{*}=\sqrt{i_s^2-i_{d1}^{*2}},$

where ?.sub.f, i.sub.s, L.sub.d1, and L.sub.q1 denote a permanent-magnet flux linkage, a stator current amplitude, a direct-axis inductance, and a quadrature-axis inductance, respectively. A PI controller (5) is configured to adjust a deviation between given currents and a feedback current to obtain a given voltage signal. A SVPWM module (7) is configured to modulate the voltage signals in a given two-phase static coordinate system into a desired PWM signal. An inverter module (8) is configured to output a five-phase voltage signal to provide power to the five-phase permanent-magnet fault-tolerant in-wheel motor (1).

[0069] FIG. 2 is a structure diagram of a permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to the present disclosure. The motor adopts a slot-pole combination solution of a traditional fault-tolerant motor, and a five-phase motor adopts a combination of 20 slots/18 poles. The permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy includes a rotating shaft (11), a stator (9), and a rotor (10) that are arranged in order from inside to outside. The stator (9) is composed of a stator yoke (22), armature teeth (14), stator slots (21), and fault-tolerant teeth (13); the armature teeth (14) and the fault-tolerant teeth (13) are evenly spaced along an outer circumference of the stator (9); the armature teeth (14) are wound with armature windings (12), respectively; each two adjacent armature teeth (14) are isolated by the fault-tolerant tooth (13); main permanent magnets (15) and core poles (16) are evenly spaced along an inner circumference of the rotor (10); the main permanent magnets (15) are surface-embedded arc permanent magnets, and each of surface-embedded main permanent magnet poles forms a pair of magnetic poles with the adjacent core pole (16); a q-axis magnetic barrier (17) close to an air gap side is formed between the surface-embedded main permanent magnet pole and the core pole (16), and the q-axis magnetic barrier (17) forms an uneven air gap with the outer circumference of the stator (9); a multi-layer arc magnetic barrier (19) close to an outer circumference side is formed between the surface-embedded main permanent magnet pole and the core pole (16), and each layer of the multi-layer arc magnetic barrier forms a magnetic bridge (20); a rectangular auxiliary permanent magnet (18) is provided between the q-axis magnetic barrier (17) and the arc magnetic barrier close to the air gap; and the auxiliary permanent magnet (18) is connected in series with the main permanent magnet (15). The stator (9) and the rotor (10) each are formed by laminating a magnetically conductive material such as a silicon steel sheet, with a lamination factor of 0.96; and the armature windings (12) are made of an enameled copper conductor material.

[0070] FIG. 3 is a structure diagram of the stator (9) and a distribution diagram of the armature winding (12), where the stator and the armature winding are shown in FIG. 1. The stator (9) is provided with 20 slots, which are semi-closed flat-bottom slots. The armature windings (12) are wound on the armature teeth (14) and are single-layer concentrated windings. The armature windings (12) are arranged in an order indicated in FIG. 2. + indicates an incoming direction of the winding, and ? indicates an outgoing direction of the winding. The armature windings (12) are isolated by the fault-tolerant teeth (13).

[0071] FIG. 4 is an enlarged partial structure diagram, with geometric dimensions, of the rotor (10) shown in FIG. 1. The rotor (10) has an inner radius of R.sub.i and an outer radius of R.sub.0. The surface-embedded arc main permanent magnets (15) are alternately distributed on the inner circumference of the rotor (10). The main permanent magnet (15) and the core pole (16) form a pair of magnetic poles. The q-axis magnetic barrier (17) close to an inner circumference side is formed between the main permanent magnet (15) and the core pole (16), and the q-axis magnetic barrier (17) is centered on a point O.sub.1 of a circle and has a radius of R.sub.1. The multi-layer arc magnetic barrier (19) close to the outer circumference side is formed between the main permanent magnet (15) and the core pole (16), and the multi-layer arc magnetic barrier is centered on a point O.sub.2 of a circle and has radii of R.sub.2 and R.sub.3. The arc magnetic barrier has a thickness of H.sub.0, effectively increasing a q-axis reluctance. The multi-layer arc magnetic barrier (19) forms a magnetic bridge (20), which reduces an influence on a d-axis flux. The rectangular auxiliary permanent magnet (18) close to the air gap is provided between the q-axis magnetic barrier (17) and the arc magnetic barrier. Due to a low permeability of the permanent magnet, the auxiliary permanent magnet acts as the q-axis magnetic barrier (17) while providing an auxiliary magnetic field, to improve a negative saliency effect while increasing the torque.

[0072] FIG. 5 is a schematic diagram of magnetization of the neodymium-iron-boron permanent magnets on the rotor (10) shown in FIG. 3. The magnetizing direction of the alternating surface-embedded main permanent magnets (15) on the inner circumference of the rotor (10) points to the center of the circle, and the main permanent magnet forms a pair of magnetic poles with the core pole (16) nearby. The auxiliary permanent magnets (18) located on both sides of the main permanent magnet (15) are magnetized in a reverse direction along the circumference, forming a magnetic circuit in series with the main permanent magnet (15).

[0073] FIG. 6 is a schematic diagram of the direct axis and the quadrature axis of the rotor (10) shown in FIG. 1. A pole centerline of the main permanent magnet (15) is along the direct axis, and a centerline between the main permanent magnet (15) and the core pole (16) is along the quadrature axis. A difference between the direct axis and the quadrature axis is a 90-degree electrical angle. An inductance corresponding to the quadrature axis of the motor is called a quadrature-axis inductance, and an inductance corresponding to the direct axis of the motor is called a direct-axis inductance.

[0074] FIG. 7 is a block diagram of a design method of the permanent-magnet fault-tolerant in-wheel motor based on the active sensorless strategy according to the present disclosure. The design method specifically includes the following steps.

[0075] Step 1.1). A number of teeth of the stator (9) and a number of pole pairs of the rotor (10) are initially determined by a slot-pole combination design method of a traditional fault-tolerant motor, and slot vectoring is determined based on a principle of a maximum fundamental resultant vector.

[0076] Step 1.2). The alternating surface-embedded main permanent magnet poles are arranged, where the main permanent magnet poles have magnetizing directions pointing to the air gap and form a closed magnetic circuit with the core poles, to improve utilization of the permanent magnets.

[0077] Step 1.3). The q-axis magnetic barrier (17) and the multi-layer arc magnetic barrier (19) are provided between the main permanent magnet pole and the core pole (16) to increase a quadrature-axis reluctance, to realize a negative saliency of the motor.

[0078] Step 1.4). The rectangular auxiliary permanent magnet (18) with a circumferential magnetizing direction is provided between the q-axis magnetic barrier (17) and the multi-layer arc magnetic barrier (19) to increase the quadrature-axis reluctance while providing an auxiliary magnetic field, to reduce a quadrature-axis inductance and further increase a negative saliency effect of the motor.

[0079] Step 1.5). The parameters of the main permanent magnet (15), the auxiliary permanent magnet (18), the q-axis magnetic barrier (17), and the multi-layer arc magnetic barrier (19) are optimized, to obtain a desired negative saliency; and sensorless operation performance is considered in advance in a motor design stage, to complete the design of the permanent-magnet fault-tolerant in-wheel motor with active sensorless operation.

[0080] FIG. 8 shows a back electromotive force waveform of the motor according to the present disclosure. A maximum back electromotive force is about 78 V, and a harmonic distortion rate is 2.3%. The back electromotive force of the motor has a desired sinusoidal amplitude, which is easy for the corresponding drive control and can reduce a cogging torque of the motor.

[0081] FIG. 9 shows an inductance waveform of the motor according to the present disclosure. The motor has a high self-inductance but a low mutual inductance. The mutual inductance accounts for 2.8% of the self-inductance and has certain short-circuit current suppression capability. The magnetic coupling between phases is small, and magnetic isolation between phases can be realized to improve the fault tolerance of the motor.

[0082] FIG. 10 shows d- and q-axis inductance waveforms of the motor according to the present disclosure under load. The d-axis inductance of the motor is greater than the q-axis inductance. A ratio of the d-axis inductance to the q-axis inductance is about 1.3, which can achieve desired negative saliency. It is easy for the zero- and low-speed sensorless operation of the motor, and can reduce the risk of irreversible demagnetization of the permanent magnet during high-speed flux weakening, to improve the reliability of high-speed cruise.

[0083] FIG. 11 is a schematic diagram of a rotor position error angle of the motor changing with a q-axis current according to the present disclosure. When the q-axis current of the motor changes, the rotor position error angle of the motor changes slightly and has high stability, which can effectively improve the sensorless control accuracy of the motor.

[0084] FIG. 12 is a simplified block diagram of a sensorless control module according to the present disclosure. The frequency-band-adaptive secondary harmonic suppression and position error signal extraction algorithm includes an adaptive band filter and an adaptive linear neuron filter.

[0085] FIG. 13 is a schematic block diagram of the adaptive linear neuron filter according to the present disclosure. A filter coefficient is adjusted through an adaptive algorithm to suppress a specific sub-harmonic in a position error signal. A harmonic reference signal r(n) and an adjustable weight component x(k) are processed to obtain a desired output signal y(n) of the filter. The desired output signal y(n) of the filter is subtracted from an input signal U(n) to obtain a desired fundamental signal Y(n). Calculations by the RLS algorithm are as follows:

[00008]$\{\begin{array}{c}{y}_1\left(n\right)=x_{11}\left(n-1\right)r_{11}\left(n\right)+x_{21}\left(n-1\right)r_{21}\left(n\right)\\ Y\left(n\right)=U\left(n\right)-y_1\left(n\right)\end{array}$ [0086] where, y.sub.1(n) denotes a harmonic estimator; x.sub.11(n?1) and x.sub.21(n?1) each denote an estimated amplitude of a harmonic component; r.sub.11(n)=sin (2{circumflex over (?)}.sub.e) and r.sub.21(n)=cos (2{circumflex over (?)}.sub.e) each denote a harmonic reference signal; {circumflex over (?)}.sub.e denotes an estimated rotor position; Y(n) denotes a filtered output; U(n) denotes a filtered input; and adjustable filter coefficients x.sub.11(n) and x.sub.21(n) are updated online based on the harmonic reference signal, and are expressed as follows:

[00009]$\{\begin{array}{c}{x}_{11}\left(n\right)=x_{11}\left(n-1\right)+k_{11}\left(n\right)Y\left(n\right)\\ x_{21}\left(n\right)=x_{21}\left(n-1\right)+k_{21}\left(n\right)Y\left(n\right)\end{array}$ [0087] where, gain coefficients k.sub.11(n) and k.sub.21(n) are expressed as follows:

[00010]$\{\begin{array}{c}{k}_{11}\left(n\right)=\frac{H_{11}\left(n-1\right)r_{11}\left(n\right)}{?+H_{11}\left(n-1\right)r_{11}^2\left(n\right)}\\ k_{21}\left(n\right)=\frac{H_{21}\left(n-1\right)r_{21}\left(n\right)}{?+H_{21}\left(n-1\right)r_{21}^2\left(n\right)}\end{array}$ [0088] where, ? denotes a forgetting factor, .sub.0<?<1; An inverse of an autocorrelation matrix H.sub.1(n) is converted into two scales, namely H.sub.11(n) and H.sub.21(n), to realize a simple and fast implementation of the RLS algorithm; and H.sub.11(n) and H.sub.21(n) are expressed as follows:

[00011]$\{\begin{array}{c}{H}_{11}\left(n\right)=\frac{H_{11}\left(n-1\right)-k_{11}\left(n\right)H_{11}\left(n-1\right)r_{11}\left(n\right)}{?}\\ H_{21}\left(n\right)=\frac{H_{21}\left(n-1\right)-k_{21}\left(n\right)H_{21}\left(n-1\right)r_{21}\left(n\right)}{?}\end{array}.$

[0089] FIG. 14 is a schematic block diagram of adaptive band filtering according to the present disclosure. A signal before and after filtering is linearly calculated based on a characteristic that an all-pass network filter only changes a signal phase, and an adaptive band-pass filter and an adaptive band-notch filter are constructed, where a transfer function of a typical second-order all-pass network filter is:

[00012]$A\left(z\right)=\frac{d-p\left(1+d\right)z^{-1}+z^{-2}}{1-p\left(1+d\right)z^{-1}+d{z}^{-2}}$ [0090] where, d denotes a correlation coefficient of a filter bandwidth and

[00013]$d=\frac{1-\tan \left({?}_m{T}_s/2\right)}{1+\tan \left({?}_m{T}_s/2\right)};$

?.sub.m denotes the filter bandwidth with a 3 dB attenuation; T.sub.s denotes a digital sampling period; p denotes a correlation coefficient of a filtering frequency and p=cos (?.sub.nT.sub.s); ?.sub.n denotes a resonance frequency point; and a resonance frequency in the adaptive band filter is set as:

?.sub.n=?.sub.c+{circumflex over (?)}.sub.r [0091] where, ?.sub.c denotes a frequency of an injected high-frequency signal; {circumflex over (?)}.sub.r denotes an estimated motor speed; the resonance frequency is automatically adjusted with the motor speed to reduce a phase delay caused by the filter; and the filter bandwidth is set as:

?.sub.m=?.sub.b+?|?.sup.*?{circumflex over (?)}.sub.r| [0092] where, ?.sub.b denotes an adjustable bandwidth; ? denotes a dynamic adjustment factor; ?.sup.* denotes a given speed; when the motor runs stably, the dynamic adjustment factor does not work, and the filter bandwidth depends on ?.sub.b; when the motor runs at a variable speed, the dynamic adjustment factor works, and the filter bandwidth is adjusted adaptively according to an error between an actual speed and the given speed to improve the dynamic response performance of the sensorless control; and a modulated current is expressed as:

?.sub.q1.sup.*=?.sub.q1h*sin (?.sub.ht)=f.sub.k({tilde over (?)}.sub.e)+HF(2?.sub.ht) [0093] where, ?.sub.q1h denotes a q.sub.1-axis high-frequency response current; HF(2?.sub.ht)=?.sub.q1h cos (2?.sub.ht); ?.sub.h denotes an angular frequency of the injected high-frequency signal; ?.sub.q1h denotes an amplitude of a 2.sup.nd-order high-frequency injected harmonic; f.sub.k({tilde over (?)}.sub.e) denotes a position error function; {tilde over (?)}.sub.e denotes an estimated angular position error; and the modulated current includes the 2.sup.nd-order high-frequency injected harmonic, so ?.sub.c in the adaptive band-notch filter is set as 2?.sub.h, to obtain the position error signal f.sub.k({tilde over (?)}.sub.e).