Dual three-phase asymmetric stator-type double-permanent-magnet vernier electric motor and improved model predictive current control method therefor

Abstract

The present disclosure proposes a dual three-phase asymmetric stator dual-permanent-magnet vernier machine (DTP-ASDPMVM) and an improved model predictive current control (MPCC) method thereof, where the machine comprises the stator core, stator permanent magnets (SPMs), rotor core, and rotor ferrites. The stator core along the circumferential direction is composed of umbrella-shaped armature teeth and quasi-rectangular armature teeth arranged alternately, the two openings on both sides of the umbrella-shaped armature tooth are affixed with a radially outward magnetized SPM respectively, and a fan-shaped auxiliary slot is opened between the umbrella-shaped armature tooth and the SPM. The shape of the rotor core is similar to that of the gear, the rotor ferrite is embedded in the rotor core by using the consequent-pole (CP) array based on the Halbach, and the length of the rotor ferrite is greater than that of the rotor and stator cores. In addition, the present disclosure adopts the MPCC system based on the virtual voltage vector. A cost function is designed to select the optimal voltage vector with the least number of switching times. The present disclosure can not only improve the torque density but also reduce the switching loss and harmonic current of the model predictive control.

Description

The present disclosure relates to a dual-permanent-magnet machine and a control strategy thereof, and especially to a dual three-phase asymmetric stator dual-permanent-magnet vernier machine (DTP-ASDPMVM) and an improved model predictive current control (MPCC) method thereof. The present disclosure belongs to the technical field of permanent magnet (PM) motors. BACKGROUND With the rapid development of rare-earth PM materials and power electronics technology, permanent magnet synchronous machines (PMSMs) with high power density and high efficiency have a broad application prospect in many fields, such as electric vehicles, rail transportation, agricultural equipment, and so on. Dual-permanent-magnet vernier machine (DPMVM) is a motor with PM distribution on both stator and rotor, and it has high torque density. The unique bidirectional excitation structure has optimized the utilization of the excitation source distribution, so how to further improve the torque density of the machine has become a research topic. The paper “Design and analysis of novel asymmetric-stator-pole flux reversal PM machine (IEEE Transactions on Industrial Electronics, 67(1): 101-114, 2020)” proposed a flux reversal machine (FRM) with asymmetric stator poles. Compared with the traditional FRM, the torque density of this machine has been improved, but it still does not fully utilize the internal space of the machine, and there exists a larger interpole leakage flux. The paper “A novel dual-permanent-magnet-excited machine with non-uniformly distributed permanent-magnets and flux modulation poles on the stator (IEEE Transactions on Vehicular Technology, 69(7): 7104-7115, 2020)” proposed a DPMVM with non-uniformly distributed poles on the stator, and the machine has a consequent decrease in efficiency due to the increase in eddy current losses caused by the more pole pairs of the stator PM and rotor PM of the machine. Meanwhile, the MPCC for PMSM is used to transform every available voltage vector into the corresponding current value, and then the current value is substituted into the current cost function. For the dual three-phase PMSM, the available voltage vectors are 64, and the computation for optimal voltage vector is substantially increased compared with vector control and direct torque control. Thus, it requires high hardware and has poor steady-state performance. Therefore, the candidate voltage vector is reduced through the sector where the stator flux linkage is located and the torque error, and the duty cycle control is also introduced to improve the steady-state performance of the algorithm. Although the method can effectively reduce the computation and improve the control performance, the method still has a higher computation. Therefore, the MPCC strategy needs to be further optimized. To reduce the computation, the paper “Two vectors based model predictive torque control of dual three-phase permanent magnet synchronous motor (Electric Machines and Control, 26(09): 97-107, 2022)” has synthesized 24 large-medium vectors of 64 voltage vectors into 12 virtual voltage vectors, and then reduced the iterative voltage vectors from 12 to 4 based on the sector where the stator flux linkage is located and the torque error to significantly reduce the computation. However, the method still cannot get rid of the disadvantage of high computation load caused by the iteration of the voltage vector into the cost function. In addition, the traditional MPCC directly outputs the voltage vectors into the inverter without considering the loss of power switching devices caused by the switching times. Therefore, MPCC needs to be further investigated to further reduce the computation and improve the steady-state performance of the system under the condition of keeping the dynamic performance unchanged, as well as to optimize the switching times of the voltage source inverter. An objective of the present disclosure is to design a novel PMVM with dual-PM and dual winding structure and a control strategy thereof to make the machine has good dynamic and steady-state performance. SUMMARY Due to the deficiencies and drawbacks of the prior art, the present disclosure proposes a DTP-ASDPMVM and an improved MPCC method thereof. The machine adopts dual PM combined with the dual winding structure to improve the torque density. To further improve torque density and reduce PM consumption, the consequent-pole (CP) concept and asymmetric design are introduced by setting ferrite with Halbach array between adjacent rotor teeth. Thus, the working harmonics in the air-gap flux density is enhanced and the PM utilization is improved. Meanwhile, by simplifying the duty cycle computation and optimizing the average switching frequency, the proposed MPCC strategy in the present invention can reduce the model prediction computation time and achieve excellent dynamic and steady-state performance, as well as a lower computation complexity and a lower average switching frequency. Technical solutions of the present disclosure: A DTP-ASDPMVM includes the stator and rotor, wherein both the stator and the rotor are of a salient-pole structure. The stator includes the stator core (1), dual three-phase windings (DTPWs) (2), and stator permanent magnets (SPMs) (3). The stator core (1) along the circumferential direction is composed of a plurality of umbrella-shaped armature teeth (11) and quasi-rectangular armature teeth (12) arranged alternately. The quasi-rectangular armature teeth (12) are each composed of an iron core, and the umbrella-shaped armature teeth (11) are each composed of a sandwich array, i.e., “PM + iron core + PM”, wherein two SPMs are affixed on openings of both sides of each of the umbrella-shaped armature teeth, respectively. The SPMs (3) are affixed on openings of both sides of the umbrella-shaped armature teeth and are radially outward magnetized by the same polarity. The DTPWs (2) are wound on the umbrella-shaped armature teeth (11) and the quasi-rectangular armature teeth (12), respectively. And the DTPWs are the concentrated winding. The rotor includes the rotor core (5) and the rotor ferrite (6). The shape of the rotor core (5) is similar to a gear, and the rotor ferrite (6) is embedded in the rotor core by using the CP array based on Halbach. The overhang structure is adopted in the rotor ferrite, and the rotor ferrite has a length greater than the rotor core and stator core. The stator core (1) and rotor core (5) are both made of silicon steel sheets laminated together. Further, the umbrella-shaped armature teeth (11) are in the inverted “T” shape on the circumferential section, and fan-shaped slot openings are on both sides of the teeth ends. Further, two SPMs (3) are affixed in the shape of a tile in the openings on both sides of the umbrella-shaped armature teeth respectively, and a fan-shaped auxiliary slot is opened between the umbrella-shaped armature teeth (11) and SPM (3). The auxiliary slot (4) is filled with non-magnetic materials, such as epoxy resin, or without any filler. Further, the SPMs (3) are adopted with high-coercivity rare-earth PM material such as NdFeB magnet and is magnetized radially outward with the same polarity, and the SPMs are adopted with CP-PM array and is affixed on the umbrella-shaped armature teeth at the end openings on both sides. The rotor ferrite (6) is adopted with low-coercivity PM material such as the ferrite magnet and is embedded in the rotor core with the Halbach array formed by radial (61) and parallel (62) magnetization. The concentrated-flux direction of the rotor ferrite is radially pointing towards the stator, and the length of the rotor ferrite is greater than that of the rotor core and stator core. Further, the maximum length along the circumferential direction of the ferrite magnetized in radial (61) and parallel (62) are tr0=(0.25~0.5)*r and tri =(0-0.25)*r, respectively, and t is the pole pith. Further, the DTPWs (2) adopt the concentrated winding structure and are wound on the umbrella-shaped armature teeth (11) and the quasi-rectangular armature teeth (12), respectively. The windings wound on the umbrella-shaped armature teeth are divided into phase-Al, phase-Bl and phase-Cl, respectively; and the phase difference between phase-Al and phase-Bl, phase-Al and phase-Cl, and phase-Bl and phase-Cl is 120°. The windings wound on the quasi-rectangular armature teeth are divided into phase-A2, phase-B2 and phase-C2, respectively; and the phase difference between phase-A2 and phase-B2, phase-A2 and phase-C2, and phase-B2 and phase-C2 is 120°. The phase difference between phase-Al and phase-A2, phase-Bl and phase-B2, and phase-Cl and phase-C2 is 30°. Further, according to the bidirectional excitation structure and bidirectional magnetic field modulation effect, the relationship between the equivalent armature winding pole pairs Pa, stator teeth slots Ns, SPM pole pairs Ps, rotor teeth slots Nr, and rotor ferrite pole pairs Pr, should be satisfied as JjPa = \mPsIp +nNjPa = \kPr ±lNt / s| [j = 1,3,5...; m,k = 1,2,3...; n,l = 0,±l,±2... where p is the pole pairs of the unit stator symmetry period and 5 is the slots of the unit stator symmetry period. The present disclosure provides an improved MPCC method for the DTP-ASDPMVM, including the following steps: step 1, establishing the structure model of DTP-ASDPMVM; step 2, taking the actual speed of the detected DTP-ASDPMVM as the feedback speed n, obtaining the speed error of the machine by making the difference between the given speed n'^and n, obtaining the ^-axis reference current U" by the speed PI controller, setting the J-axis reference current expressing the stator current vector in the d-q frame as Isre^=id^+jiqe^ and making the projection of II" in the a-fi frame as Is e / e^ step 3, using the Clarke transformation matrix to transform the double three-phase currents Iai, hn, ia, iA2, 'a and ic2 into currents ia and ip in the a-fi frame and currents ix and iy in the x-y frame, then using Park transformation matrix to transform ia and ip into ij1 and / / in the d-q frame; step 4, using the base voltage vector to synthesize the virtual voltage vector Wi ( / =1, 2, ..., 12) with the principle of suppressing current in the x-y frame, making the amplitude of Wi as 0 in the x-y frame and 0.596 / V in the a-fi frame, wherein Ude is the DC bus voltage; taking the large voltage vector Vmax and the medium-large voltage vector VmidL with the same direction in the a-P frame as the two base voltage vectors involved in the synthesis of virtual voltage vectors, dividing the virtual voltage vector into sectors in the a~P frame, making the angular bisector of two adjacent virtual voltage vectors as the dividing line of each sector, and locating each virtual voltage vector on the center line of the virtual voltage vector sector; step 5, establishing and discretizing the current equation in the d-q frame of the DTP-ASDPMVM as Djk , ..r ;k\.Ud 1 s \ ;k d — I d ' ‘ tj J ”1” ' *d < A 1 A1 = ZL (-Rjk - aLi* - m ) + + f where i / and iq are the current components in the d-q frame at kT, moment, respectively; z / +1, z / +1, w / +1 and n / ‘ are the components of the current and voltage in the d-q frame at (£+1)7, moment, respectively; R is the stator resistance; L is the stator inductance; co is the electric angular speed; Vf is the PM flux linkage; Ts is the control period; k represents for kT moment; thus, separating the current related terms z; / +1 and z) / +1 and the voltage related terms iud+l and as A1 = + id \ d q ) d T = iK ® )+ I uk^lT ■£+l = lld *s ud j ■k + l _ g s L expressing the current related terms and voltage related terms in vector form in the d-q frame as ISi=iidk+x+jiiq^1 and Isu=iudk+i+jiuqk+i, making the projections of 7 / and Isu in the a~P frame as L+ap and Lujgi, wherein Isu^.p is in the same phasor as the virtual voltage vector and the amplitude of Lu-op is TsIL times of the virtual voltage vector; step 6, according to the reference and feedback currents, calculating the current increment vector Isdb and the phasor angle dap of the current increment vector projection !++<.+ in the a-ft frame as [j =Tref-I ^sDB 1s 1 si \daP where ^Isdb is the phasor angle of han in the d-q frame; 0e is the electrical angle of the rotor; hence, minimizing the angle between Lu ap and Isdb_oP to track the reference current faster; due to the same phasor angle of hu and the virtual voltage vector, determining a sector where ISDB_ap is located based on then making the virtual voltage vector at the centerline of the sector as the desired optimal voltage vector, and thus directly obtaining the optimal voltage vector to avoid the constant iteration of voltage vectors in the cost function; step 7, calculating the actuation duration Tmax, TmidL and To of the base voltage vectors Vmax, VmidL and zero voltage vector corresponding to the optimal voltage vector as X=o. M>=IU , = (1 |G» | <| / s„ | . , P SU I ’ t =o.73i(r-r) max s O' T .., = 0.269(T -Z) k nuaL v s 0' step 8, substituting the current voltage vector and the voltage vector to be acted soon in the system into the switching cost function to reduce the switching times of the power switching devices of the voltage source inverter J = (al,- - al)2 + (61,. - 61)2 + (cl, - cl)2 + (a2,. - a2)2 + (62,. - 62)2 + (c2,. - c2)2 where al, bl, cl, a2, b2 and c2 are the switching states of the power switch devices of the upper bridge arm of the phase-Al, phase-Bl, phase-Cl, phase-A2, phase-B2 and phase-C2 of the voltage source inverter corresponding to the base voltage vector at kl, moment, wherein 0 represents the power switch device (such as msulated-gate bipolar transistor (IGBT)) is off and 1 represents the power switch device (such as IGBT) is on; ah, bh, ch, a2t, b2t and c2, represent the switching states of the six-phase upper bridge arm power switch devices of the voltage source inverter corresponding to the zth base voltage vector at the (k+l)Ts moment; and thus, obtaining a voltage vector to minimize J, and making the voltage vector act preferentially. The present disclosure relates to a DTP-ASDPMVM and an improved MPCC method thereof, which has the following advantages: 1) The DTP-ASDPMVM in the present invention adopts the DPM and dual salient-pole structure, which can maximize the torque density by combining the unique bidirectional flux-modulation effect of the structure with the DTPWs structure. 2) The DTP-ASDPMVM in the present invention adopts the CP concept in combination with the asymmetric design to generate additional working harmonics to increase torque, and introduces the Halbach array in the rotor ferrite to enhance the air-gap effective harmonics, which further improves the torque density and increases the PM utilization. 3) The DTP-ASDPMVM in the present invention adopts the auxiliary slot, which can suppress the leakage flux between the stator poles, thus improving the torque capacity and reducing the torque ripple. 4) The DTP-ASDPMVM in the present invention adopts the DPM structure and introduces the CP concept, asymmetric design, and Halbach array, which greatly expands the high-efficiency operating range and reduces material costs. 5) The rotor ferrite of DTP-ASDPMVM in the present invention adopts the overhang structure, which makes the length of the rotor ferrite greater than that of the rotor and stator cores, then the overhang structure can maximize the internal space utilization of the machine, which further enhances the air-gap effective flux and improves the torque capacity. 6) The improved MPCC strategy for the machine of the present invention is based on the idea of current tracking, where the optimal voltage vector is directly obtained in the a-P frame according to the angle of current increment vector. Thus, effectively avoiding the constant iteration of the voltage vector in the cost function, and greatly reducing the load and complexity of the calculation. 7) The improved MPCC strategy for the machine of the present invention is to calculate the amplitude of the current increment vector and to obtain the duty cycle of the optimal voltage vector based on the current increment vector and the voltage vector. Thus, the calculation method of the duty cycle based on the derivation of current cost function in the traditional model prediction can be simplified, and the calculation burden can be greatly reduced. The combination of zero vector and current increment vector can effectively improve the accuracy and response speed of current tracking. 8) The improved MPCC strategy for the machine of the present invention designs the switching state cost function of the power switch devices on every bridge arm of the voltage source inverter. By adjusting the base voltage vector action order in every control period, the base voltage vector action order with the lowest switching times can be obtained, thus substantially reducing the average switching frequency. Combined with the acquisition of optimal voltage vectors, it can not only reduce the computation burden but also shorten the average switching times of power switch devices (such as IGBT) and the switching losses. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 shows the topologies of the traditional FRM and the proposed machine in the present invention, (a) Traditional FRM. (b) Proposed machine in the present invention. FIG. 2 shows the evolution process of the proposed machine m the present invention. FIG. 3 shows the comparison of Fourier decomposition of the magnetomotive force (MMF) generated by the SPMs of the traditional FRM and the proposed machine in the present invention. FIG. 4 shows the comparison of Fourier decomposition of the air-gap flux density of the traditional FRM and the proposed machine in the present invention. FIG. 5 shows the comparison of harmonic contribution to torque of the traditional FRM and the proposed machine in the present invention with and without the rotor ferrite. FIG. 6 shows the comparison of torque of the traditional FRM and the proposed machine in the present invention with and without the rotor ferrite. FIG. 7 shows the efficiency map of the traditional FRM and the proposed machine in the present invention, (a) Traditional FRM. (b) Proposed machine in the present invention. FIG. 8 shows the cutaway view of the rotor ferrite overhang structure of the proposed machine in the present invention. FIG. 9 shows the block diagram of the improved MPCC for the proposed machine in the present invention. FIG. 10 shows the distribution of large and medium voltage vectors in a-ft and x-y frames. FIG. 11 shows the virtual voltage vector distribution and sector division. FIG. 12 shows the schematic diagram of the current vector control. FIG. 13 shows the flowchart of the optimization algorithm for PWM switching times. FIG. 14 shows the schematic diagram of the optimization effect for PWM switching times. FIG. 15 shows the comparison of optimization effect of PWM switching times under multiple-operation conditions. FIG. 16 shows the Fourier decomposition waveform of phase-Al current of the proposed machine in the present invention. FIG. 17 shows the torque and speed waveforms of proposed machine in the present invention. FIG. 18 shows the torque step waveform of proposed machine under normal operation in the present invention. FIG. 19 shows the speed step waveform of proposed machine under normal operation in the present invention. In the drawings: 1 - stator core, 11 - umbrella-shaped armature tooth, 12 - quasi-rectangular armature tooth, 2 - DTPWs, 3 - SPMs, 4 - auxiliary slot, 5 - rotor core, 6 - rotor ferrite, 61 - radial magnetized ferrite, 62 - parallel magnetized ferrite. DETAILED DESCRIPTION OF THE EMBODIMENTS The following is a clear and complete description of the technical scheme combined with the attached figures of the present invention. To more simply and clearly illustrate the structure characteristics and beneficial effects of DTP-ASDPMVM and an improved MPCC strategy thereof in the present invention, a DTP-ASDPMVM is described as follows. FIG. 1 shows the topologies of the DTP-ASDPMVM of the present invention and traditional FRM, both machines include the stator and rotor, and the 12-slot / 34-pole combinations are adopted. The traditional FRM adopts the conventional N-S pole surface-mounted PM array affixed to the stator surface. Compared with the traditional magnetic flux reversing motor, the umbrella-shaped armature teeth (11) and quasi-rectangle armature teeth (12) of the proposed machine in the present invention are different due to the difference in the PM arrangement, and the two stator teeth are composed of an iron core and a sandwich array, i.e., “PM + iron core + PM”, respectively. The two SPMs (3) affixed on the openings of both sides of the umbrella-shaped armature teeth are both radially outward magnetized by the same polarity, and a fan-shaped auxiliary slot is opened between the two PMs and middle iron core. The width of the auxiliary slot (4) is 2°, and the auxiliary slot (4) is filled with non-magnetic material, such as epoxy resin, or without any filler. The stator along the circumferential direction is arranged in alternating 6 umbrella-shaped armature teeth (11) and 6 quasi-rectangular armature teeth (12). The rotor core (5) is similar to a gear with 17 teeth, the CP ferrite based on Halbach-array is embedded between the adjacent rotor teeth. The rotor ferrite (6) with 17 pole pairs consists of radial (61) and parallel (62) magnetized ferrites, and the concentrated-flux direction of the rotor ferrite (6) is radially pointing towards the stator. The maximum length along the circumferential direction of the ferrites magnetized in radial (61) and parallel (62) is tr0=5.7 mm and trl=3.1 mm, respectively, and the length of the rotor ferrite is 10 mm greater than that of the rotor core and stator core. The DTPWs (2) adopt the concentrated winding structure and is wound on the umbrella-shaped armature teeth (11) and the quasi-rectangular armature teeth (12), respectively. The windings wound on the umbrella-shaped armature teeth (11) are divided into phase-Al, phase-Bl and phase-Cl, respectively, and the phase difference between phase-Al and phase-Bl, phase-Al and phase-Cl, and phase-Bl and phase-Cl is 120°. The windings wound on the quasi-rectangular armature teeth (12) are divided into phase-A2, phase-B2 and phase-C2, respectively, and the phase difference between phase-A2 and phase-B2, phase-A2 and phase-C2, and phase-B2 and phase-C2 is 120°. The angle displacement between the two sets of three-phase wingdings (that means phase-Al phase-Bl phase-Cl and phase-A2 phase-B2 phase-C2) is 30°. FIG. 2 shows the evolution process of the proposed machine structure in the present invention. Compared with the traditional FRM, the SPM of the present invention adopts the CP-PM array with the asymmetric design and the auxiliary slots, and the CP ferrite based on the Halbach array is embedded in the rotor core. According to the bidirectional excitation structure and bidirectional magnetic field modulation effect, the relationship between the equivalent armature winding pole pairs Pa, stator teeth slots Ns, SPM pole pairs P,, rotor teeth slots Nr, and rotor ferrite pole pairs Pr, should be satisfied as jPa = \mPjp ±nNr\,jPa =\kPr+ INjs | / = 1,3,5...; th,A-= 1,2,3...; n, / = 0,±1,±2... ' where p is the pole pairs of the unit stator symmetry period and s is the slots of the unit stator symmetry period. FIG. 3 shows the comparison of the Fourier decomposition of the MMF generated by the SPMs of the traditional FRM and the proposed machine in the present invention. It can be seen from FIG. 3 that, due to the difference in the adjacent stator teeth, the MMF of the proposed machine in the present invention is increased by odd multiples of the 6th harmonics, such as the 6th, 18th, and 30th, and the harmonics are modulated by the rotor teeth to produce some additional harmonics, as shown m FIG. 4. FIG. 5 shows the comparison of harmonic contribution to torque of the traditional FRM and the proposed machine in the present invention with and without the rotor ferrite. It can be seen from FIG. 5 that some additional harmonics, such as the 6th, 11th and 18th, are increased by the stator-pole asymmetric design to generate torque, while the 17th harmonic of the proposed machine in the present invention also contributes to the torque due to the existence of the rotor ferrite. FIG. 6 shows the comparison of the torque of the traditional FRM and the proposed machine in the present invention with and without the rotor ferrite. It can be seen from FIG. 6 that the torque of the traditional FRM is 5.59 Nm with 0.8% torque ripple, the torque of the proposed machine without rotor ferrite in the present invention is 7.53 Nm with 1.6% torque ripple, and the torque of the proposed machine in the present invention is 10.05 Nm with 1.26% torque ripple. The combination of asymmetric stator-pole design and CP concept for the proposed machine without rotor ferrite in the present invention can result in 34.7% torque increment while reducing the consumption of PMs by half. On this basis, the introduction of CP rotor ferrite based on the Halbach array in the proposed machine in the present invention can result in 33.47% torque increment. Meanwhile, the torque ripple is low in all the above three machines due to the existence of DTPWs. The optimal slot width of the auxiliary slot of the machine in the present invention is 2°. FIG. 7 shows the efficiency Map of the traditional FRM and the proposed machine in the present invention. It can be seen from FIG. 7 that the efficiency of the traditional FRM is 81.11% and the efficiency of the proposed machine in the present invention is 87.88% at the rated speed 300 r min and the rated peak current 10 A. Furthermore, the torque performance of the proposed machine in the present invention in the constant torque region is improved and the high-efficiency region in the flux-weakening region is greatly expanded. FIG. 8 shows the cutaway view of the rotor ferrite overhang structure of the proposed machine in the present invention, where the length of the rotor ferrite is greater than that of the rotor and the stator core. Considering the consumption of ferrite material, the optimal overhang length of one end of the rotor ferrite is 5 mm. The control strategy for the proposed machine in the present invention is shown in FIG. 9, and the control strategy adopts the double closed-loop control system composed of the current loop based on model predictive control and PI speed control loop. The speed error is obtained by making the difference between the given speed and the actual speed. The speed error is input into the speed PI controller to obtain the g-axis reference current iqe^, and the J-axis reference current ir' is set 0. The stator current vector in the d-q frame can be expressed as and the projection of Z^in the a~P frame is Is aft' . The Clarke transformation matrix is used to transform the DTPWs currents iai, ibi, ici, iA2, is2 and ic2 into ia and ip in the a-P frame and ix and iy in the x-y frame, where ia and ip are the fundamental currents, and ix and iy are harmonic currents. Then Park transformation matrix is used to transform ia and ip into / / and iqk in the d-q frame. As shown in FIG. 10, the direction of the large voltage vector in the a-P frame is the same as that of the medium-large voltage vector with a smaller amplitude but opposite in the x-y frame. To suppress the current in the x-y frame, a set of the large and medium-large voltage vectors can be output in a control cycle, which can be expressed as 5hl=>M+^^                     (2) Let JAy|=O, then b = 0.73 IT (3) W =0-596½ It can be seen that when the action time ratio of the large voltage vector and the medium-large voltage vector is 0.731:0.269, the effect can be equivalent to a synthetic vector with 0.596½ amplitude in the a-P frame and zero in the x-y frame. For convenience, the synthesized voltage vector is expressed as the virtual voltage vector W} ( / =1, 2, ..., 12). Then, the virtual voltage vectors are divided into sectors in a~p frame, and the angular bisector lines of adjacent virtual voltage vectors are used as sector dividing lines. The synthesized virtual voltage vector diagram is shown in FIG. 11. The current equation in the d-q frame is established and discretized, and then the current prediction equation can be expressed as T ■ k+1 1s ( n-k , T -k\ , 11 d 1 s , -k ld -—[~Rld+a)Llg)+--j--+ ld L L (4) T uk+lT V 'o+1 = — (-Rik- &Lik, -o)iy f]+ ——- + ik q y \ q « r j ! j q where R is the stator resistance. L is the stator inductance, co is the electric angular speed. Tp is the PM flux linkage. Ts is the control period, k stands for kTs time, id, iq, Ud and uq are the components of current and voltage in the d-q frame, respectively. According to (5), the current related terms 7; / +1 and hqk+x and the voltage related terms iu / +x and iuk+x are separated, and then are expressed in vector form as ISj=ii^+x +piqkn and n +jiuq +x in the d-q frame. The projections of ISi and Isu in the a-p frame are Ispop and Lu^ap. It can be seen that the phasor angle of Isu_ap is the same as that of the virtual voltage vector, and the amplitude of Isu ap is TJL times of the virtual voltage vector. A1 =^-(-Rik, +(0Lik\ + ik, id y \ a q } d B । 1 p / t r P \ . z. =— \-Ri„ -o)Li,-(Dwf + z„ iq j^\ q d r J ! q According to (6), the current increment vector Isdb required for tracking Isref and the phasor angle of the vector Ldb^ projected by Isdb in the a-fi frame can be obtained as i =ref-i 1 sDB 1 s 1 si / □ _ / T .n ' * ^ap ^-1 sDB where ^Isdb is the angle of Isdb in the d-q frame. 0e is the electrical angle of the machine. As shown in FIG. 12, to make Isi_ep quickly track Isj^, the angle between Isu_ap and Isdb_«p should be minimized. Because the phasor angle of Isu_ap and the virtual voltage vector is the same, the virtual voltage vector with the smallest angle with Ldbop should be adopted as the optimal voltage vector. In summary, the virtual voltage vector sector where ISDB_ap is located can be judged according to 9ap, and then the virtual voltage vector in the center line of the sector is selected as the optimal voltage vector to achieve current tracking control. The action range of lsu is approximated as a circle with radius \ISU\. Since the amplitude of Isu is TSIL times the virtual voltage vector, thus the amplitude of Isu can be obtained as: 0.596¾ I 1 SU I j x 7 / The base voltage vectors Vmax and VmidL that constitute the optimal voltage vector, are used to calculate the action time Tmax, Tm,dL and To of Vmax, VmidL and zero voltage vector through the following equation. ^0=0, |^|>=|4| 7 = 7 q-AJ) I; I <1 / I <o p । a (g) Tl =0.731(7^-To) ^=0.269(7,-^) It can be concluded that when Isdb exceeds the action range of Isu, i.e.. \Isdb\ is greater than |4M|, then zero voltage vector is not used. When Isdb does not exceed the action range of Isu, i.e., \Isdb\ is less than \ISU\, then zero voltage vector is added, and the duty cycle optimization is adopted to improve the accuracy of current tracking control. The controller must output the large voltage vector, medium-large voltage vector and zero voltage vector, or large voltage vector and medium-large voltage vector to the PWM module in every control period. To reduce the switching times of the power switch device (IGBT) of the voltage source inverter, the acting base voltage vector is substituted into the switching cost function of (9). A base voltage vector that minimizes Jt is obtained and made to act preferentially. The detailed flowchart is shown in Fig. 13. J = (al, - al)2 + (61,. - 61)2 + (cl,. - cl)2 + (a2,. - a2)2 + (62,. -62)2 + (c2, - c2)2 (9) where al, bl, cl, a2, b2 and c2 are the switching states of the IGBTs of the upper bridge arm of the phase-Al, phase-Bl, phase-Cl, phase-A2, phase-B2 and phase-C2 of the voltage source inverter corresponding to the base voltage vector at kTs moment, wherein 0 represents that the IGBT is off and 1 represents that the IGBT is on. ah, bl„ cl,, a2,, b2, and c2, represent the switching states of the six-phase upper bridge arm IGBT of the voltage source inverter at (kl)T, moment. Assuming that the acting voltage vector is F24 at the current moment, and ITio being going to be fed to the PWM module, according to the traditional voltage vector action sequence, the action order of the base voltage vector is F24-F51-F15. If the base voltage vector that is about to act is substituted into (9) to obtain a base voltage vector with the smallest Ji and make the base voltage vector act preferentially, the action order of the base voltage vector will be changed to V24-V15-K51, and the effect is shown in FIG. 14, where the switching times are reduced from 9 to 7. The switching optimization effect of the voltage source inverter is shown in FIG. 15. It can be seen that the switching times were significantly reduced under different operation conditions. When the machine is in steady-state condition, the current harmonics, torque and speed waveforms are shown in Figs. 16 and 17. It can be seen that the speed is smooth and the harmonic content of phase-Al current is 0.17%. It is proved that the proposed control algorithm has good steady-state performance and can successfully suppresses current harmonics. The waveforms with load step and speed step are shown in Figs. 18 and 19. It can be seen that when the load step from 20 Nm to 30 Nm, the torque can quickly and accurately track the load. When the speed step from 200 r / min to 300 r / min, the feedback speed has no obvious overshoot. Therefore, the proposed machine with the proposed control method has excellent steady-state and dynamic performance. Although the present invention has been made public as the above implement example, the example is not used to limit the present invention. Any equivalent change or polish within the spirit and field of the present invention belongs to the protective range of the present disclosure. In the description of the introduction, the description of the reference terms “an embodiment”, “some embodiments” and “schematic embodiments”, “example”, “specific example”, or “some examples”, etc. means that the specific characteristics, structures, materials, or features described in combination with embodiment or example are included in at least one embodiment or example of the present invention. In the introduction, schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific characteristics, structures, materials, or features described may be combined appropriately in any one or more embodiments or examples. Although the embodiment of the present invention has been disclosed and described, it will be appreciated by those skilled in the art that the present invention can be embodied in multiple variations, modifications, replacements and variants of the embodiment without departing from the principle and purpose thereof, and the scope of the present disclosure is restricted by the claims and equivalents thereof.

Claims

1. A dual three-phase asymmetric stator dual-permanent-magnet vernier machine (DTP-ASDPMVM), comprising a stator and a rotor, wherein both the stator and the rotor are of a salient-pole structure; the stator comprises a stator core (1), dual three-phase windings (DTPWs) (2), and stator permanent magnets (SPMs) (3); the stator core (1) along a circumferential direction is composed of a plurality of umbrella-shaped armature teeth (11) and quasi-rectangular armature teeth (12) arranged alternately; the quasi-rectangular armature teeth (12) are each composed of an iron core, and the umbrella-shaped armature teeth (11) are each composed of a sandwich array, i.e., “permanent magnet (PM) + iron core + PM”, wherein two of the SPMs are affixed on openings of both sides of each of the umbrella-shaped armature teeth, respectively; the SPMs (3) are affixed on openings of both sides of the umbrella-shaped armature teeth and are radially outward magnetized by a same polarity; the DTPWs (2) are wound on the umbrella-shaped armature teeth (11) and the quasi-rectangular armature teeth (12), respectively, and the DTPWs are concentrated windings; the rotor comprises a rotor core (5) and a rotor ferrite (6); a shape of the rotor core (5) is similar to a gear, and the rotor ferrite (6) is embedded in the rotor core through a consequent pole (CP) array based on Halbach; an overhang structure is adopted in the rotor ferrite, and the rotor ferrite has a length greater than the rotor core and the stator core; and the stator core (1) and the rotor core (5) are both made of silicon steel sheets laminated together.

2. The DTP-ASDPMVM as claimed in Claim 1, wherein the umbrella-shaped armature teeth (11) are each in an inverted “T” shape on a circumferential section, and fan-shaped slot openings are on both sides of the tooth ends.

3. The DTP-ASDPMVM as claimed in Claim 1, wherein two SPMs (3) are affixed in the shape of a tile in the end openings on both sides of the umbrella-shaped armature tooth, respectively, a fan-shaped auxiliary slot is opened between the umbrella-shaped armature teeth (11) and the SPMs (3); and an auxiliary slot (4) is filled with non-magnetic materials, such as an epoxy resin, or without any filler.

4. The DTP-ASDPMVM as claimed in Claim 1, wherein the SPMs (3) are adopted with high-coercivity rare-earth PM material such as NdFeB magnet and are magnetized radially outward with the same polarity, and the SPMs are adopted with CP-PM array and are affixed on the umbrella-shaped armature teeth at the end openings on both sides; the rotor ferrite (6) is adopted with low-coercivity PM material such as ferrite magnet and is embedded in the rotor core with a Halbach array formed by a radial (61) and parallel (62) magnetization, a concentrated-flux direction of the rotor ferrite is radially pointing towards the stator, and the rotor ferrite has a length greater than the rotor core and the stator core.

5. The DTP-ASDPMVM as claimed in Claim 1, wherein a maximum length along the circumferential direction of two ferrites with radial (61) and parallel (62) magnetization are tr0=(0.25~0.5)*r and trl=(0~0.25)*T, respectively, and t is the pole pith.

6. The DTP-ASDPMVM as claimed in Claim 1, wherein the DTPWs (2) adopt the concentrated winding structure and are wound on the umbrella-shaped armature teeth (11) and the quasi-rectangular armature teeth (12), respectively; the windings wound on the umbrella-shaped armature teeth are divided into phase-Al, phase-Bl and phase-Cl, respectively, and the phase difference between phase-Al and phase-Bl, phase-Al and phase-Cl, and phase-Bl and phase-Cl is 120°; the windings wound on the quasi-rectangular armature teeth are divided into phase-A2, phase-B2 and phase-C2, respectively, and the phase difference between phase-A2 and phase-B2, phase-A2 and phase-C2, and phase-B2 and phase-C2 is 120°; the angle displacement between the two sets of three-phase wingdings (that means phase-Al phase-Bl phase-Cl and phase-A2 phase-B2 phase-C2) is 30°.

7. The DTP-ASDPMVM as claimed in Claim 1, wherein according to a bidirectional excitation structure and bidirectional magnetic field modulation effect, the relationship between an equivalent armature winding pole-pair number Pa, stator tooth slot number Ns, SPM pole-pair number Ps, rotor tooth slot number M, and rotor ferrite pole-pair number Pr, should be satisfied as[ JP = \mP IP- | ’ jP = |^r ± / ' |V = 1,3,5...; m,k = 1,2,3...; n,l = 0,±l,±2...wherein p is a pole-pair number of a unit stator symmetry period and 5 is a slot number of the unit stator symmetry period.

8. An improved model predictive current control (MPCC) method for the DTP-ASDPMVM as claimed in any one of Claim 1 to Claim 7, comprising the following steps:step 1, establishing a structure model of DTP-ASDPMVM;step 2, taking an actual speed of the DTP-ASDPMVM as a feedback speed w; obtaining a speed error by making the difference between a given speed and / / ; obtaining a < / -axis reference current iq* by speed PI controller; setting a <7-axis reference current z / ^O; expressing a stator current vector in a d-q frame as and making a projection ofin the a~P frame asis of ,step 3, using Clarke transformation matrix to transform double three-phase currents iA\, ibi, ici, iA2, iB2 and ic2 into currents ia and ip in a a-P frame, and currents ix and iy in a x-y frame; then using Park transformation matrix to transform ia and ip into id and iq in the d-q frame;step 4, using two base voltage vectors to synthesize a virtual voltage vector W} (z=l, 2, ..., 12) with a principle of suppressing current in the x-y frame, making an amplitude of PTS as 0 in the x-y frame and 0.596f / <* in the a~P frame, wherein Ude is a DC bus voltage; taking a large voltage vector Vl and a medium-large voltage vector VmL with a same direction in the a-P frame as the two base voltage vectors involved in the synthesis of virtual voltage vectors, dividing the virtual voltage vectors into sectors in the a-P frame, making an angular bisector of two adjacent virtual voltage vectors as a dividing line of each sector, and locating each virtual voltage vector on a center line of the virtual voltage vector sector;step 5, establishing and discretizing a current equation in the d-q frame of the DTP-ASDPMVM asTik+1 = + o)Lit j + +T uk+lTA1 = (-Rjk © I / / J + + / *<7 L \ q a r J ' J qwherein i / and z / are the components of the current at kTs time in the d-q frame, respectively; z / +1, z / +1, w / +1 and zz / +1 are the components of the current and voltage at (k+l)Ts time in the d-q frame, respectively; R is a stator resistance; L is a stator inductance; co is an electric angular speed; 'A is a PM flux linkage; Ts is a control period; k represents a kTs moment; thus, separating current related terms z; / +1 and z, / +1 and voltage related terms zu / +1 and iUq+i as'ul =-r(-Rij + ®Lik) + ik¢+1 ~®Lik -(0^ + ik\+1L, , uk+1T■ k + Y _ q Sexpressing the current related terms and voltage related terms in vector form in the d-q frame as and Isu=iudk+l+jiuq+l', making the projections of Lb and Jsu in the a~P frame as Isi^ and Isu_ap, wherein Isu_ap is in a same phasor angle as the virtual voltage vector and the amplitude of ISu_ap is TSIL times of the virtual voltage vector;step 6, according to the reference current and feedback current, calculating a current increment vector Isdb and a phasor angle dap of the current increment vector projection Isdb_oP in the a-P frame as(j j^sDB — 1s 1 siwherein Thun is the phasor angle of Isdb in the d-q frame; de is an electrical angle of the rotor; hence, minimizing the angle between ISu_ap and Isdb_op to track the reference current faster; due to the same phasor angle of ku „p and the virtual voltage vector, determining a sector wherein ISDB_ap is located based on dap, then making the virtual voltage vector at a centerline of the sector as an desired optimal voltage vector, and thus directly obtaining an optimal voltage vector to avoid the constant iteration of voltage vectors in a cost function;step 7, calculating the action time Tmax, TmidL and To of the base voltage vectors Vmax, VmtdL, and zero voltage vector corresponding to the optimal voltage vector asir=o. | / ™l>=k I0 I sDB I | su |l^l<K.I;and^=o.73i(C-o ^=0.269(7)-7))step 8, substituting a current voltage vector and the voltage vector to be acted soon in the system into a switching cost function to reduce the switching times of a power switching device in a voltage source inverterJ = (alt - al)2 + (61,. - 61)2 + (cl,. - cl)2 + (a2, - a2)2 + (62,. -62)2 + (c2, - c2)2wherein al, bl, cl, a2, b2 and c2 are the switching states of the power switch devices of the upper bridge arm of the phase-Al, phase-Bl, phase-Cl, phase-A2, phase-B2 and phase-C2 of the voltage source inverter corresponding to the base voltage vector at kTs moment, wherein 0 represents that the power switch device is off and 1 represents that the power switch device is on; ah, bh, ch, a2i, b2t and c2; represent the switching states of the six-phase upper bridge arm power switch devices of the voltage source inverter corresponding to the zth base voltage vector at the (k+h)Ts moment; thus, obtaining a voltage vector to minimize Ji, and making the voltage vector act preferentially.

9. The improved MPCC method for the DTP-ASDPMVM as claimed in Claim 8, wherein the step 4 comprises:step 4.1, expressing the per-phase voltage corresponding to the switching state of the power switch devices of upper bridge arm of the voltage source inverter of the DTP-ASDPMVM as«„1 5 -1 -1 -1 -1 -1 al -1 5 -1 -1 -1 -1 61 -1 -1 5 -1 -1 -1 cl 6 -1 -1 —1 5 -1 -1 a2 Ub2 -1 -1 -1 -1 5 -1 62 uc2 -1 -1 -1 -1 -1 5 c2step 4.2, using Clarke transformation matrix to transform the six-phase voltages to a two-phase stationary frame; thus, obtaining the base voltage vectors in the a-fi and x-y frames aswherein the amplitude of the large voltage vectors V44, V64, V66, V26, V22, V32, V33, his, Vn, V51, V55 and V45 in the a-fi frame is 0.644( / 7.; the amplitude of the medium-large voltage vectors V65, V46, V24, hs2, V36, V23, V12, V31, V53, V15, V41 and V54 is 0.471 Ude', the amplitude of the medium-smallvoltage vectors V04, V74, Voo, V67, V06, V76, V20, V27, V02, V72, V30, V37, Vos, V73, V 10, Vn, V01, V71, V50, V57, Vos, V75, V40 and V47 is Q.333Udc', the amplitude of the small voltage vectors V56, V25, V42, V34, V63, V16, V21, V52, V35, V43, V14 and V6I is 0.113Udc;step 4.3, due to the large voltage vector Vmax and medium-large voltage vector Vm / dL with the same direction in the a-p frame but the opposite direction in the x-y frame, selecting the two base voltage vectors in the control period and then allocating an appropriate action time to equate the two base voltage vectors to a virtual voltage vector; making the amplitude of the virtual voltage vector in the a-P and x-y frames satisfied asLet |=0, then / = 0.7317;'1^1 = 0.596^wherein t is an action time of Vmax in a control period; \Vap\ is the amplitude of a synthesized voltage vector in the a-P frame; | Vxy\ is the amplitude of the synthesized voltage vector in the x-y frame; thus, making the amplitude of the synthesized voltage vector as 0.59677* in the a-p frame and 0 in the x-y frame when the action time ratio of Vmax and VmidL is 0.731:0.269 in a control period; expressing the synthesized voltage vector as the virtual voltage vector W ( / =1, 2, ..., 12) for convenience; andstep 4.4, dividing the virtual voltage vectors into sectors in a-P frame; making the angular bisector of two adjacent virtual voltage vectors as the dividing line of each sector, and making each virtual voltage vector at the center line of each sector.

10. The improved MPCC method for the DTP-ASDPMVM as claimed in Claim 8, wherein in the step 7, an action range of the virtual voltage vector on the current is approximated to a circle with radius | / OT|; when Isdb exceeds the action range of the virtual voltage vector, zero voltage vector is not used to track when Isdb does not exceed the action range of the virtual voltage vector, zero voltage vector is introduced and duty cycle optimization is adopted, and the proportion of zero voltage vector action time is 1 -\ISu\ / \Isdb|.INTERNATIONAL SEARCH REPORT International application No. PCT / CN2024 / 107604A. CLASSIFICATION OF SUBJECT MATTER H02P21 / 14(2016.01)i; H02Kl / 14(2006.01)i; H02Kl / 17(2006.01)i; H02Kl / 24(2006.01)i; H02Kl / 27(2022.01)i According to International Patent Classification (IPC) or to both national classification and IPC B. FIELDS SEARCHED Minimum documentation searched (classification system followed by classification symbols) IPC:H02K,H02P Documentation searched other than minimum documentation to the extent that such documents are included in the fields searched Electronic data base consulted during the international search (name of data base and, where practicable, search terms used) VEN, CNABS, CNTXT, WOTXT, EPTXT, USTXT, CNKI, IEEE: ¢,^1= / 1, ¢,1= / 1, halbach, motor, double winding, double salient pole, teeth, permanent magnet, iron core, ferrite C. DOCUMENTS CONSIDERED TO BE RELEVANT Category* Citation of document, with indication, where appropriate, of the relevant passages Relevant to claim No. A CN 114123708 A (JIANGSU UNIVERSITY) 01 March 2022 (2022-03-01) description, paragraphs [0030]-[0033], and figures 1-5 1-10 A A A CN 103683564 A (NANJING UNIVERSITY OF INFORMATION SCIENCE &TECHNOLOGY) 26 March 2014 (2014-03-26) entire document CN 107134866 A (TIANJIN UNIVERSITY) 05 September 2017 (2017-09-05) entire document CN 115173591 A (NANTONG UNIVERSITY) 11 October 2022 (2022-10-11) entire document 1-10 1-10 1-10 A CN 115549335 A (JIANGSU UNIVERSITY) 30 December 2022 (2022-12-30) entire document 1-10 A A CN 115833671 A (JIANGSU UNIVERSITY) 21 March 2023 (2023-03-21) entire document CN 117277629 A (JIANGSU UNIVERSITY) 22 December 2023 (2023-12-22) entire document 1-10 1-10 | J | Further documents are listed in the continuation of Box C. | J | See patent family annex. * Special categories of cited documents: “p” later document published after the international filing date or priority “A" document defining the general state of the art which is not considered date and not in conflict with the application but cited to understand the to be of particular relevance principle or theory underlying the invention “D" document cited by the applicant in the international application “X” document of particular relevance; the claimed invention cannot be “E" earlier application orpatent but published on or after the international considered novel or cannot be considered to involve an inventive step filing date when the document is taken alone “L" document which may throw doubts on priority claim(s) or which is “Y” document of particular relevance; the claimed invention cannot be cited to establish the publication date of another citation or other considered to involve an inventive step when the document is special reason (as specified) combined with one or more other such documents, such combination “O” document referring to an oral disclosure, use, exhibition or other being obvious to a person skilled in tile art means “&” document member of the same patent family “P” document published prior to the international filing date but later than the priority date claimed Date of the actual completion of the international search 22 October 2024 Date of mailing of the international search report 28 October 2024 Name and mailing address of the ISA / CN China National Intellectual Property Administration (ISA / CN) China No. 6, Xitucheng Road, Jimenqiao, Haidian District, Beijing 100088 Authorized officer Telephone No.INTERNATIONAL SEARCH REPORT International application No. PCT / CN2024 / 107604c.DOCUMENTS CONSIDERED TO BE RELEVANTCategory* Citation of document, with indication, where appropriate, of the relevant passages Relevant to claim No. A CN 109378918 A (XI'AN JIAOTONG UNIVERSITY) 22 February 2019 (2019-02-22) entire document 1-10 A CN 113489274 A (NANJING UNIVERSITY OF AERONAUTICS AND ASTRONAUTICS) 08 October 2021 (2021-10-08) entire document 1-10 A JP 2001186695 A (FUJITSU GENERAL LTD.) 06 July 2001 (2001-07-06) entire document 1-10 A JP 2013198171 A (ASMO CO., LTD.) 30 September 2013 (2013-09-30) entire document 1-10INTERNATIONAL SEARCH REPORT Information on patent family membersInternational application No.PCT / CN2024 / 107604Patent document cited in search report Publication date (day / month / year) Patent family member(s) Publication date (day / month / year) CN 114123708 A 01 March 2022 None CN 103683564 A 26 March 2014 None CN 107134866 A 05 September 2017 None CN 115173591 A 11 October 2022 None CN 115549335 A 30 December 2022 None CN 115833671 A 21 March 2023 WO 2024065986 Al 04 April 2024 GB 2623189 A 10 April 2024 CN 117277629 A 22 December 2023 None CN 109378918 A 22 February 2019 None CN 113489274 A 08 October 2021 None JP 2001186695 A 06 July 2001 None JP 2013198171 A 30 September 2013 None