阅读说明：本技术 三电平逆变器驱动永磁同步电机系统有限集预测控制方法 (Three-level inverter driving permanent magnet synchronous motor system finite set prediction control method ) 是由 谷鑫 李阳 王志新 高乐 张国政 金雪峰 于 2020-12-09 设计创作，主要内容包括：一种三电平逆变器驱动永磁同步电机系统有限集预测控制方法：通过对定子磁链信息(定子磁链位置角和所处扇区)的观测,进而根据转矩和磁链的变化量,通过滞环控制器,对所有基本电压矢量,进行筛选,筛选过后得到备选矢量集1,再通过中点电位滞环去掉产生中点电位偏移的小矢量,得到备选矢量集2,最后带入价值函数中进行选最优矢量,来解决现有中点电位有偏移,动态性能较差、算法复杂等问题。本发明解决了现有动态性能较差、计算负担较大、中点电位不平衡等问题,与已有方法相比算法简单,计算速度快,动态性能良好。(A finite set prediction control method for a three-level inverter driving permanent magnet synchronous motor system comprises the following steps: through observation of stator flux linkage information (stator flux linkage position angle and located sector), all basic voltage vectors are screened through a hysteresis controller according to torque and flux linkage variation, a candidate vector set 1 is obtained after screening, small vectors generating midpoint potential offset are removed through midpoint potential hysteresis to obtain a candidate vector set 2, and finally the candidate vector set is brought into a cost function to select optimal vectors, so that the problems that existing midpoint potential has offset, dynamic performance is poor, algorithm is complex and the like are solved. The invention solves the problems of poor dynamic performance, large calculation burden, unbalanced midpoint potential and the like in the prior art, and has simple algorithm, high calculation speed and good dynamic performance compared with the prior art.)

1. A limited set prediction control method for a three-level inverter driving permanent magnet synchronous motor system is characterized by comprising the following steps:

1) sampling the position angle theta of the motor rotor by using a position sensor, and calculating the sampled position angle to obtain the motor rotorElectrical angular velocity omega_eAnd a motor speed n; setting the motor to a given speed n^refMaking difference with the calculated motor rotating speed n, and taking the obtained rotating speed difference value as the input quantity of a rotating speed outer ring PI controller, wherein the output quantity of the PI controller is an electromagnetic torque reference value T_e ^ref(ii) a Setting permanent magnet flux linkage as stator flux linkage reference value psi_s ^refThe reference value is a constant value;

2) acquisition of inverter DC bus voltage U by sensor_dcAnd a three-phase stator current i of the permanent magnet synchronous motor_a、i_b、i_cConverting the three-phase stator current from a three-phase stationary coordinate system, namely an a-b-c coordinate system, to a two-phase synchronous rotating coordinate system, namely a d-q coordinate system, and obtaining a stator current d-axis component i in the synchronous rotating coordinate system_dAnd q-axis component i_q；

3) According to the d-axis and q-axis components i of the motor stator current at the moment k_d ^k、i_q ^kD-axis and q-axis components u of sum voltage vector_d ^k、u_q ^kAnd the electrical angular velocity ω of the rotor of the machine at time k_e ^kObtaining the predicted value i of the d-axis and q-axis currents at the moment of k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1The electric angular speed of the rotor of the motor is equal at the moment k and the moment k +1, and the u_d ^k、u_q ^kIs a voltage vector u applied to the stator side for one control period from the time k-1^kAfter Park conversion is carried out, the partial conversion is carried out; according to the current predicted value i_d ^k+1And i_q ^k+1And calculating the component psi of the stator flux linkage between the d axis and the q axis at the moment k +1_d ^k+1、ψ_q ^k+1And electromagnetic torque T_e ^k+1；

Wherein, d-axis and q-axis current predicted values i at the moment of k +1_d ^k+1、i_q ^k+1The method is obtained by using a motor prediction model as follows:

in the formula, R_sIs the stator resistance value, T_sFor the system control period, L_d、L_qFor stator d-and q-axis inductances,. psi_fPredicting value i of d-axis and q-axis current according to k +1 time for permanent magnet flux linkage_d ^k+1、i_q ^k+1The component ψ of the stator flux linkage in the d-axis and q-axis at the end of the control period is calculated from the following equation_d ^k+1，ψ_q ^k+1：

Knowing the number p of pole pairs of the motor, the stator flux linkage and the electromagnetic torque psi at the moment k +1 are calculated by the following formula_s ^k+1，T_e ^k+1：

4) Establishing an x-y coordinate system taking the stator flux linkage as a reference, transforming a stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system, and expressing the stator flux linkage psi of the motor by using a stator voltage vector in the x-y coordinate system_sAnd electromagnetic torque T_eBy changing the amount of change of stator flux linkage psi of the motor_sAnd electromagnetic torque T_eDerivative time, which is obtained by the formula after derivation: flux linkage variation is determined by the x-axis component u of the stator voltage_sxDetermining the variation of the voltage; the variation of torque being determined by the y-component u of the stator voltage_syDetermining the variation of the voltage;

5) the electromagnetic torque T at the moment k +1_e ^k+1With reference value T of electromagnetic torque_e ^refAnd the stator flux linkage psi at the time k +1_s ^k+1With reference value psi of stator flux linkage_s ^refThe input signals are respectively used as the input of a torque hysteresis comparator and a flux linkage hysteresis comparator; the electromagnetic torque is changed by an amount Delta T according to the output of the hysteresis comparator_eAnd defineVariation amount Δ ψ of sub flux linkage_sThe combination of (a) is divided into four cases:

a)ΔT_e＞0、Δψ_s＞0；

b)ΔT_e＞0、Δψ_s＜0；

c)ΔT_e＜0、Δψ_s＜0；

d)ΔT_e＜0、Δψ_s＞0

and selecting a first candidate voltage vector set corresponding to four conditions by combining the flux linkage sector where the stator flux linkage vector is located, wherein the flux linkage sector where the stator flux linkage vector is located passes through a stator flux linkage vector angle theta_sJudging;

6) according to the capacitance C on the DC side₁And a lower capacitor C₂The voltage difference between the capacitors and the direction of the inverter midpoint current under the action of the redundancy vector, the first set of candidate voltage vectors is optimized, and the capacitor C on the DC side is discarded₁And a lower capacitor C₂Obtaining a second alternative voltage vector set by the redundant vectors with increased voltage difference;

7) and collecting the second candidate voltage vector set and predicted values i of d-axis and q-axis currents at the moment k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1Substituting the motor prediction model to predict stator flux linkage psi corresponding to different voltage vectors in the second alternative voltage vector set at the moment of k +2_sN ^k+2And electromagnetic torque T_eN ^k+2The subscript N is a basic voltage vector sequence number in the second candidate voltage vector set, the sequence number is substituted into a cost function for calculation, and a basic voltage vector corresponding to the minimum value of the cost function is selected to be the optimal voltage vector; the cost function g is as follows:

in the formula, λ_ψIs the stator flux linkage weight term, λ_TIs an electromagnetic torque weight term;

8) and determining the switching state corresponding to the three-level inverter according to the optimal voltage vector, thereby determining the driving signal of each power device of the three-phase bridge arm of the inverter at the current moment, further enabling the inverter to output the corresponding three-phase voltage, and acting on the motor.

2. The finite set predictive control method for a three-level inverter-driven permanent magnet synchronous motor system according to claim 1, wherein the transformation of the stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system in step 4), specifically, the x-y coordinate system based on the stator flux linkage is established, and the stator voltage vector equation in the a-b-c coordinate system is transformed into the x-y coordinate system, with the transformation relationship:

wherein theta is_s＝θ_e+δ，θ_eIs the electrical angle, δ is the load angle, so the stator voltage vector equation is rewritten as:

wherein u is_sx、u_syRespectively the component of the stator voltage in the x-y coordinate system, i_sx，i_syThe component of the stator current in the x-y coordinate system, #_sx，ψ_syIs the component of the stator flux linkage in the x-y coordinate system, R_sIs the stator resistance.

3. The limited set predictive control method for a three-level inverter-driven permanent magnet synchronous motor system according to claim 1, wherein the step 4) is performed by linking the stator flux ψ of the motor_sAnd electromagnetic torque T_eDerivative time, which is obtained by the formula after derivation: flux linkage variation is determined by the x-axis component u of the stator voltage_sxDetermining the variation of the voltage; the variation of torque being determined by the y-component u of the stator voltage_syIs determined.

Neglecting the influence of stator resistance, the stator flux linkage psi is expressed by a feedforward Euler formula through a mathematical model of the motor_s ^k+1Obtaining:

wherein psi_s ^k+1，ψ_s ^kStator flux linkages at the time of k +1 and k respectively, and motor stator flux linkage psi_sAnd electromagnetic torque T_eThe time is derived as follows:

in the above formula, L_sIs stator inductance, delta^k+1And δ is the load angle at times k +1 and k, u, respectively_sx ^k+1、u_sx ^kStator voltage x-axis component, i, at times k +1 and k, respectively_sx ^k+1、i_sx ^kStator current x-axis component at time k +1 and k, respectively, per unit control period T_sInner, rotor flux linkage angle constant theta_r ^k+1＝θ_r ^kSelecting a basic voltage vector to act on the motor, and obtaining:

θ_s ^k+1-θ_r ^k+1-θ_s ^k+θ_r ^k＝Δθ_s (10)

Δθ_srepresenting the angle variation of the stator flux linkage from k to k + 1;

the increment of the y-axis component of the stator flux linkage in a unit control period is represented by Δ ψ_y＝ψ_ssinΔθ_s≈ψ_sΔθ_s，ψ_sΔθ_s＝u_sySo torque and magnetismThe amount of change in the chain, and the flux linkage and torque at time k +1 are expressed as:

therefore, it can be seen that the torque variation is represented by the stator voltage y-axis component u_syDetermining the variation of the voltage; flux linkage variation is determined by the x-axis component u of the stator voltage_sxIs determined.

4. The finite set prediction control method of the three-level inverter-driven permanent magnet synchronous motor system according to claim 1, wherein the flux linkage sectors in step 5) are divided into 12 flux linkage sectors in the counterclockwise direction by pi/6 intervals based on the a axis; after the flux linkage sector where the flux linkage is located is judged, the space vector plane is divided into 4 voltage vector sectors by taking a straight line where the flux linkage is located and a direction perpendicular to the flux linkage as a boundary, the voltage vector sectors are sequentially ordered anticlockwise from the flux linkage position, and each voltage vector sector contains 8 voltage vectors to be selected, wherein:

the voltage vector in the 1 st voltage vector sector satisfies: delta T_e＞0，Δψ_s＞0

The voltage vector in the 2 nd voltage vector sector satisfies: delta T_e＞0，Δψ_s＜0

The voltage vector in the 3 rd voltage vector sector satisfies: delta T_e＜0，Δψ_s＜0

The voltage vector in the 4 th voltage vector sector satisfies: delta T_e＜0，Δψ_s＞0。

5. The finite set prediction control method of the three-level inverter-driven permanent magnet synchronous motor system according to claim 4, wherein the flux linkage sector where the stator flux linkage is located is known, the voltage vector sector is divided, and the corresponding voltage vector sector is selected according to four torque flux linkage variations output by the torque hysteresis comparator and the flux linkage hysteresis comparator to construct the first candidate voltage vector set.

6. The limited set predictive control method for a three-level inverter-driven PMSM system according to claim 1, wherein said step 6) of optimizing said first set of candidate voltage vectors abandons the capacitor C on the DC side₁And a lower capacitor C₂The redundant small vectors with increased voltage difference between them are used to obtain the second alternative voltage vector set, and the two redundant small vectors have the same amplitude and direction in the space vector formed by inverter, and have the same effect on load, but have opposite effect on the medium-point voltage difference, and the upper capacitor voltage U is set_C1And lower capacitor voltage U_C2Voltage difference v between_oFor the midpoint voltage difference, when two redundant small vectors act, one phase current flows into or out of the midpoint, generating a midpoint current i_oThe positive direction of the phase current is defined as the direction toward the load side, the positive direction of the midpoint current is defined as the direction toward the midpoint of the outflow, and when v is the phase current_o>0, i.e. the voltage of the upper capacitor is large, the midpoint current is selected to be opposite to the phase current, the phase current flows to the load side, and the upper capacitor C₁Discharge, lower capacitance C₂One redundant small vector is kept, and the other redundant small vector is eliminated; v. of_o<0, i.e. the lower capacitor voltage is large, the midpoint current is selected to be the same as the phase current, the phase current flows to the load side, and the upper capacitor C₁Charging, lower capacitance C₂One discharged redundant small vector is kept, and the other redundant small vector is removed; and (4) removing 1 or 2 small vectors from the first candidate voltage vector set to form a second candidate voltage vector set.

Technical Field

The invention relates to a limited set prediction torque control method for a permanent magnet synchronous motor. In particular to a finite set prediction control method for a three-level inverter driving permanent magnet synchronous motor system.

Background

In recent years, the fields of traction transmission and foundation construction in China are rapidly developed, and the motor used for converting electric energy and mechanical energy is an essential part in the engineering field. Meanwhile, the topology structure of the Neutral Point Clamped (NPC) three-level inverter is developed well, and the Neutral Point Clamped (NPC) three-level inverter has the advantages of higher voltage withstanding level, better output voltage harmonic wave and the like. The Neutral Point Clamped (NPC) three-level inverter is applied to the field of motor control, and has far-reaching significance particularly in the field of high-voltage alternating-current motor control and speed regulation.

In the motor control, Model Predictive Torque Control (MPTC) may be employed. The MPTC can well combine basic voltage vectors generated by the inverter with control targets, is easy to process system constraints, and can realize control over a plurality of variables. When the MPTC is applied to a motor speed regulation system driven by an NPC three-level inverter, the three-level inverter has a problem of unbalanced dc side capacitor voltage due to the particularity of the topology structure of the NPC three-level inverter (shown in fig. 1). The operation requirement of the NPC three-level inverter can be met by adding the midpoint potential weight item in the evaluation function, but the problem of complex weight coefficient setting is also increased, and as more three-level basic voltage vectors are provided, as shown in FIG. 2, traversing all the basic voltage vectors increases the calculation burden, affects the dynamic performance of the control system, cannot track the dynamic change of the system in time, and causes adverse effect on the system, even damages the system.

Disclosure of Invention

The invention aims to solve the technical problem of providing a limited set prediction control method for a three-level inverter driving permanent magnet synchronous motor system, which has good dynamic response characteristics and utilizes the characteristics of the inverter.

The technical scheme adopted by the invention is as follows: a limited set prediction control method for a three-level inverter driving permanent magnet synchronous motor system comprises the following steps:

1) sampling the position angle theta of the motor rotor by using a position sensor, and calculating the sampled position angle to obtain the electrical angular velocity omega of the motor rotor_eAnd a motor speed n; setting the motor to a given speed n^refMaking difference with the calculated motor rotating speed n, and taking the obtained rotating speed difference value as the input quantity of a rotating speed outer ring PI controller, wherein the output quantity of the PI controller is an electromagnetic torque reference value T_e ^ref(ii) a Setting permanent magnet flux linkage as stator flux linkage reference value psi_s ^refThe reference value is a constant value;

2) acquisition of inverter DC bus voltage U by sensor_dcAnd a three-phase stator current i of the permanent magnet synchronous motor_a、i_b、i_cConverting the three-phase stator current from a three-phase stationary coordinate system, namely an a-b-c coordinate system, to a two-phase synchronous rotating coordinate system, namely a d-q coordinate system, and obtaining a stator current d-axis component i in the synchronous rotating coordinate system_dAnd q-axis component i_q；

3) According to the d-axis and q-axis components i of the motor stator current at the moment k_d ^k、i_q ^kD-axis and q-axis components u of sum voltage vector_d ^k、u_q ^kAnd the electrical angular velocity ω of the rotor of the machine at time k_e ^kObtaining the predicted value i of the d-axis and q-axis currents at the moment of k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1The electric angular speed of the rotor of the motor is equal at the moment k and the moment k +1, and the u_d ^k、u_q ^kIs a voltage vector u applied to the stator side for one control period from the time k-1^kAfter Park conversion is carried out, the partial conversion is carried out; according to the current predicted value i_d ^k+1And i_q ^k+1And calculating the component psi of the stator flux linkage between the d axis and the q axis at the moment k +1_d ^k+1、ψ_q ^k+1And electromagnetic torque T_e ^k+1；

Wherein, d-axis and q-axis current predicted values i at the moment of k +1_d ^k+1、i_q ^k+1The method is obtained by using a motor prediction model as follows:

in the formula, R_sIs the stator resistance value, T_sFor the system control period, L_d、L_qFor stator d-and q-axis inductances,. psi_fPredicting value i of d-axis and q-axis current according to k +1 time for permanent magnet flux linkage_d ^k+1、i_q ^k+1The component ψ of the stator flux linkage in the d-axis and q-axis at the end of the control period is calculated from the following equation_d ^k+1，ψ_q ^k+1：

Knowing the number p of pole pairs of the motor, the stator flux linkage and the electromagnetic torque psi at the moment k +1 are calculated by the following formula_s ^k+1，T_e ^k+1：

4) Establishing an x-y coordinate system taking the stator flux linkage as a reference, transforming a stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system, and expressing the stator flux linkage psi of the motor by using a stator voltage vector in the x-y coordinate system_sAnd electromagnetic torque T_eBy changing the amount of change of stator flux linkage psi of the motor_sAnd electromagnetic torque T_eDerivative time, which is obtained by the formula after derivation: flux linkage variation is determined by the x-axis component u of the stator voltage_sxDetermining the variation of the voltage; the variation of torque being determined by the y-component u of the stator voltage_syDetermining the variation of the voltage;

5) the electromagnetic torque T at the moment k +1_e ^k+1With reference value T of electromagnetic torque_e ^refAnd the stator flux linkage psi at the time k +1_s ^k+1With reference value psi of stator flux linkage_s ^refThe input signals are respectively used as the input of a torque hysteresis comparator and a flux linkage hysteresis comparator; the electromagnetic torque is changed by an amount Delta T according to the output of the hysteresis comparator_eVariation quantity delta psi of stator flux linkage_sThe combination of (a) is divided into four cases:

a)ΔT_e＞0、Δψ_s＞0；

b)ΔT_e＞0、Δψ_s＜0；

c)ΔT_e＜0、Δψ_s＜0；

d)ΔT_e＜0、Δψ_s＞0

and selecting a first candidate voltage vector set corresponding to four conditions by combining the flux linkage sector where the stator flux linkage vector is located, wherein the flux linkage sector where the stator flux linkage vector is located passes through a stator flux linkage vector angle theta_sJudging;

6) according to the capacitance C on the DC side₁And a lower capacitor C₂The voltage difference between the capacitors and the direction of the inverter midpoint current under the action of the redundancy vector, the first set of candidate voltage vectors is optimized, and the capacitor C on the DC side is discarded₁And a lower capacitor C₂Obtaining a second alternative voltage vector set by the redundant vectors with increased voltage difference;

7) and collecting the second candidate voltage vector set and predicted values i of d-axis and q-axis currents at the moment k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1Substituting the motor prediction model to predict stator flux linkage psi corresponding to different voltage vectors in the second alternative voltage vector set at the moment of k +2_sN ^k+2And electromagnetic torque T_eN ^k+2The subscript N is a basic voltage vector sequence number in the second candidate voltage vector set, the sequence number is substituted into a cost function for calculation, and a basic voltage vector corresponding to the minimum value of the cost function is selected to be the optimal voltage vector; the cost function g is as follows:

g＝λ_ψ|ψ_s|^ref-|ψ_s|^k+2|+λ_T|T_e ^ref-T_e ^k+2 (4)

in the formula, λ_ψIs the stator flux linkage weight term, λ_TIs an electromagnetic torque weight term;

8) and determining the switching state corresponding to the three-level inverter according to the optimal voltage vector, thereby determining the driving signal of each power device of the three-phase bridge arm of the inverter at the current moment, further enabling the inverter to output the corresponding three-phase voltage, and acting on the motor.

The invention discloses a finite set prediction control method of a three-level inverter driving permanent magnet synchronous motor system, which observes a flux linkage through a flux linkage observer, screens and optimizes candidate voltage vectors through a torque, flux linkage and a midpoint potential hysteresis controller, selects a final candidate vector set, selects an optimal voltage vector through a value function, reduces the number of the candidate voltage vectors, quickly controls the torque and the flux linkage, maintains the stability of the torque and the flux linkage, better and faster controls the motor, maintains the dynamic performance, balances the midpoint potential of an inverter, maintains the balance of the inverter and has good dynamic response characteristics. The invention solves the problems of poor dynamic performance, large calculation burden, unbalanced midpoint potential and the like in the prior art, and has simple algorithm, high calculation speed and good dynamic performance compared with the prior art.

Drawings

Fig. 1 is a main circuit topology structure diagram of a Neutral Point Clamped (NPC) three-level inverter;

fig. 2 is a space vector diagram of a Neutral Point Clamped (NPC) three-level inverter;

FIG. 3 is a control block diagram of a finite set prediction of a Neutral Point Clamped (NPC) three level inverter;

FIG. 4 is a schematic diagram of torque and flux linkage increments in an x-y coordinate system;

FIG. 5a is a graph satisfying Δ T_e＞0、Δψ_sThe sector where the voltage vector is greater than 0;

FIG. 5b is a graph satisfying Δ T_e＞0，Δψ_sSector where voltage vector < 0 is located;

FIG. 5c is a graph satisfying Δ T_e＜0，Δψ_sSector where voltage vector < 0 is located;

FIG. 5d is a graph satisfying Δ T_e＜0，Δψ_sThe sector where the voltage vector is greater than 0;

fig. 6 is a diagram of the midpoint current in the small vector POO state.

Detailed Description

The present invention will be described in detail with reference to the following embodiments and accompanying drawings.

The invention discloses a finite set prediction control method of a three-level inverter driving permanent magnet synchronous motor system, which comprises the steps of observing stator flux linkage information (a stator flux linkage position angle and a located sector), screening all basic voltage vectors through a hysteresis controller according to torque and flux linkage variation, obtaining a first candidate voltage vector set after screening, removing a small vector generating midpoint potential offset through a midpoint potential hysteresis loop, obtaining a second candidate vector set, and finally substituting into a value function to select an optimal vector, so that the problems of offset of the existing midpoint potential, poor dynamic performance, complex algorithm and the like are solved.

As shown in fig. 3, the finite set predictive control method for a three-level inverter-driven permanent magnet synchronous motor system of the present invention includes the following steps:

1) sampling the position angle theta of the motor rotor by using a position sensor, and calculating the sampled position angle to obtain the electrical angular velocity omega of the motor rotor_eAnd a motor speed n; setting the motor to a given speed n^refMaking difference with the calculated motor rotating speed n, and taking the obtained rotating speed difference value as the input quantity of a rotating speed outer ring PI controller, wherein the output quantity of the PI controller is an electromagnetic torque reference value T_e ^ref(ii) a Setting permanent magnet flux linkage as stator flux linkage reference value psi_s ^refThe reference value is a constant value;

2) acquisition of inverter DC bus voltage U by sensor_dcAnd a three-phase stator current i of the permanent magnet synchronous motor_a、i_b、i_cConverting the three-phase stator current from a three-phase stationary coordinate system, namely an a-b-c coordinate system, to a two-phase synchronous rotating coordinate system, namely a d-q coordinate system to obtainStator current d-axis component i in synchronous rotating coordinate system_dAnd q-axis component i_qThe transformation matrix is represented as:

in the formula [ theta ]_eIs an electrical angle, θ_eP is the number of pole pairs of the motor;

3) according to the d-axis and q-axis components i of the motor stator current at the moment k_d ^k、i_q ^kD-axis and q-axis components u of sum voltage vector_d ^k、u_q ^k(k-1 time point A voltage vector u applied to the stator side for one control cycle^kObtained after Park conversion), and the electrical angular speed ω of the rotor of the machine at time k_e ^kObtaining the predicted value i of the d-axis and q-axis currents at the moment of k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1The electric angular speed of the motor rotor at the time k is equal to that at the time k + 1; according to the current predicted value i_d ^k+1And i_q ^k+1And calculating the component psi of the stator flux linkage between the d axis and the q axis at the moment k +1_d ^k+1、ψ_q ^k+1And electromagnetic torque T_e ^k+1；

Wherein, d-axis and q-axis current predicted values i at the moment of k +1_d ^k+1、i_q ^k+1The method is obtained by using a motor prediction model as follows:

in the formula u_d ^kAnd u_q ^kIs stator voltage vector u^kComponent on d-and q-axes, R_sIs the stator resistance value, T_sFor the system control period, L_d、L_qFor stator d-and q-axis inductances,. psi_fPredicting value i of d-axis and q-axis current according to k +1 time for permanent magnet flux linkage_d ^k+1、i_q ^k+1The component ψ of the stator flux linkage in the d-axis and q-axis at the end of the control period is calculated from the following equation_d ^k+1，ψ_q ^k+1；

Knowing the number p of pole pairs of the motor, the stator flux linkage and the electromagnetic torque psi at the moment k +1 are calculated by the following formula_s ^k+1，T_e ^k+1：

4) Establishing an x-y coordinate system taking the stator flux linkage as a reference, transforming a stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system, and expressing the stator flux linkage psi of the motor by using a stator voltage vector in the x-y coordinate system_sAnd electromagnetic torque T_eBy changing the amount of change of stator flux linkage psi of the motor_sAnd electromagnetic torque T_eDerivative time, which is obtained by the formula after derivation: flux linkage variation is determined by the x-axis component u of the stator voltage_sxDetermining the variation of the voltage; the variation of torque being determined by the y-component u of the stator voltage_syDetermining the variation of the voltage; the incremental schematic is shown in fig. 4, wherein,

(1) the transformation of the stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system is specifically to establish the x-y coordinate system based on the stator flux linkage, and transform the stator voltage vector equation in the a-b-c coordinate system into the x-y coordinate system, wherein the transformation relationship is as follows:

wherein theta is_s＝θ_e+δ，θ_eIs the electrical angle, delta is the load angle, which is defined as the d-axis component psi of the stator flux linkage_dWith q-axis component psi_qSo stator voltage vector equation is rewrittenComprises the following steps:

wherein u is_sx、u_syRespectively the component of the stator voltage in the x-y coordinate system, i_sx，i_syThe component of the stator current in the x-y coordinate system, #_sx，ψ_syIs the component of the stator flux linkage in the x-y coordinate system, R_sIs the stator resistance.

(2) The stator flux linkage psi of the motor_sAnd electromagnetic torque T_eDerivative time, which is obtained by the formula after derivation: flux linkage variation is determined by the x-axis component u of the stator voltage_sxDetermining the variation of the voltage; the variation of torque being determined by the y-component u of the stator voltage_syIs determined.

Neglecting the influence of stator resistance, the stator flux linkage psi is expressed by a feedforward Euler formula through a mathematical model of the motor_s ^k+1Obtaining:

wherein psi_s ^k+1，ψ_s ^kStator flux linkages at the time of k +1 and k respectively, and motor stator flux linkage psi_sAnd electromagnetic torque T_eThe time is derived as follows:

in the above formula, L_sIs stator inductance, delta^k+1And delta^kLoad angle at times k +1 and k, u_sx ^k+1、u_sx ^kStator voltage x-axis at times k +1 and k, respectivelyComponent i_sx ^k+1、i_sx ^kStator current x-axis component at time k +1 and k, respectively, per unit control period T_sInner, rotor flux linkage angle constant theta_r ^k+1＝θ_r ^kSelecting a basic voltage vector to act on the motor, and obtaining:

θ_s ^k+1-θ_r ^k+1-θ_s ^k+θ_r ^k＝Δθ_s (10)

Δθ_srepresenting the angle variation of the stator flux linkage from k to k + 1;

the increment of the y-axis component of the stator flux linkage in a unit control period is represented by Δ ψ_y＝ψ_ssinΔθ_s≈ψ_sΔθ_s，ψ_sΔθ_s＝u_sySo the amount of change in torque and flux linkage, and flux linkage and torque at time k +1 are expressed as:

therefore, it can be seen that the torque variation is represented by the stator voltage y-axis component u_syDetermining the variation of the voltage; flux linkage variation is determined by the x-axis component u of the stator voltage_sxIs determined.

5) The electromagnetic torque T at the moment k +1_e ^k+1With reference value T of electromagnetic torque_e ^refAnd the stator flux linkage psi at the time k +1_s ^k+1With reference value psi of stator flux linkage_s ^refThe input signals are respectively used as the input of a torque hysteresis comparator and a flux linkage hysteresis comparator; the electromagnetic torque is changed by an amount Delta T according to the output of the hysteresis comparator_eVariation quantity delta psi of stator flux linkage_sThe combination of (a) is divided into four cases:

a)ΔT_e＞0、Δψ_s＞0；

b)ΔT_e＞0、Δψ_s＜0；

c)ΔT_e＜0、Δψ_s＜0；

d)ΔT_e＜0、Δψ_s＞0

and selecting a first candidate voltage vector set corresponding to four conditions by combining the flux linkage sector where the stator flux linkage vector is located, wherein the flux linkage sector where the stator flux linkage vector is located passes through a stator flux linkage vector angle theta_sJudging;

the flux linkage sectors are based on a space vector plane, as shown in fig. 2, the space vector plane is divided into 12 flux linkage sectors in the counterclockwise direction by taking an a axis as a reference and pi/6 as an interval; after the flux linkage sector where the flux linkage is located is judged, the space vector plane is divided into 4 voltage vector sectors by taking the straight line where the flux linkage is located and the direction perpendicular to the flux linkage as a boundary, and the voltage vector sectors are sequentially ordered anticlockwise from the flux linkage position, each voltage vector sector contains 8 voltage vectors to be selected, as shown in fig. 5, wherein:

the voltage vector in the 1 st voltage vector sector satisfies: delta T_e＞0，Δψ_s＞0

The voltage vector in the 2 nd voltage vector sector satisfies: delta T_e＞0，Δψ_s＜0

The voltage vector in the 3 rd voltage vector sector satisfies: delta T_e＜0，Δψ_s＜0

The voltage vector in the 4 th voltage vector sector satisfies: delta T_e＜0，Δψ_s＞0。

And dividing a voltage vector sector according to the known flux linkage sector where the stator flux linkage is located, selecting a corresponding voltage vector sector according to four torque flux linkage variable quantities output by a torque hysteresis comparator and a flux linkage hysteresis comparator, and constructing a first candidate voltage vector set.

6) According to the capacitance C on the DC side₁And a lower capacitor C₂The voltage difference between them (i.e. midpoint potential shift) and the inverter midpoint current direction under the action of redundancy vector, and the first candidate voltage vector set is optimized, and the capacitor C on the DC side is abandoned₁And a lower capacitor C₂Increase in voltage difference therebetweenObtaining a second alternative voltage vector set by using the large redundant vector;

optimizing the first alternative voltage vector set and abandoning the capacitor C on the direct current side₁And a lower capacitor C₂And obtaining a second candidate voltage vector set by the increased redundant small vectors of the voltage difference. In the topology structure of the three-level inverter, four power switching tubes on each phase of bridge arm are respectively S_n1、S_n2、S_n3And S_n4N is equal to { A, B, C }. According to the switching state combination of each phase of switching tube, the three-level inverter can output the following three levels, taking phase A as an example: when S is_A1And S_A2Conduction, S_A3And S_A4When the inverter is turned off, the A-phase load is directly connected with the positive end P of the DC bus of the inverter, and the A-phase output voltage is V_dc/2, the output level state is marked as "P" or "1"; when S is_A2And S_A3Conduction, S_A1And S_A4When the circuit is turned off, the clamping diode clamps the A-phase output level on a point O of the middle point of the capacitor, the A-phase output voltage is 0 at the moment, and the output level state is marked as 'O' or '0'; s_A3And S_A4Conduction, S_A1And S_A2When the inverter is turned off, the A-phase load is directly connected with the negative end N of the DC bus of the inverter, and the output voltage is-V_dcAnd 2, the output level state is marked as N or' -1. These basic voltage vectors can be classified into four categories according to the magnitude difference: large, medium, small, and zero vectors. Forming a space vector as shown in figure 2. The two redundant small vectors have the same amplitude and direction, the effect on the load is the same, but the effect on the intermediate voltage difference is opposite, and the capacitor voltage U is set_C1And lower capacitor voltage U_C2Voltage difference v between_oFor the midpoint voltage difference, when two redundant small vectors act, one phase current flows into or out of the midpoint, generating a midpoint current i_oThe positive direction of the phase current is defined as the direction toward the load side, the positive direction of the midpoint current is defined as the direction toward the midpoint of the outflow, and when v is the phase current_o>0, i.e. the voltage of the upper capacitor is large, the midpoint current is selected to be opposite to the phase current, the phase current flows to the load side, and the upper capacitor C₁Discharge, lower capacitance C₂Charged with electricityOne redundant small vector is kept, and the other redundant small vector is removed; v. of_o<0, i.e. the lower capacitor voltage is large, the midpoint current is selected to be the same as the phase current, the phase current flows to the load side, and the upper capacitor C₁Charging, lower capacitance C₂One discharged redundant small vector is kept, and the other redundant small vector is removed; such as the current loop corresponding to the small vector POO shown in fig. 6. And (4) removing 1 or 2 small vectors from the first candidate voltage vector set to form a second candidate voltage vector set.

7) And collecting the second candidate voltage vector set and predicted values i of d-axis and q-axis currents at the moment k +1_d ^k+1、i_q ^k+1And the electrical angular velocity ω of the rotor of the machine at the moment k +1_e ^k+1Substituting the motor prediction model to predict stator flux linkage psi corresponding to different voltage vectors in the second alternative voltage vector set at the moment of k +2_sN ^k+2And electromagnetic torque T_eN ^k+2The subscript N is a basic voltage vector sequence number in the second candidate voltage vector set, the sequence number is substituted into a cost function for calculation, and a basic voltage vector corresponding to the minimum value of the cost function is selected to be the optimal voltage vector; the cost function g is as follows:

g＝λ_ψ|ψ_s|^ref-|ψ_s|^k+2|+λ_T|T_e ^ref-T_e ^k+2 (13)

in the formula, λ_ψIs the stator flux linkage weight term, λ_TIs an electromagnetic torque weight term;

8) and determining the switching state corresponding to the three-level inverter according to the optimal voltage vector, thereby determining the driving signal of each power device of the three-phase bridge arm of the inverter at the current moment, further enabling the inverter to output the corresponding three-phase voltage, and acting on the motor.