CN112583315B - Three-vector model prediction torque control method for three-level permanent magnet synchronous motor - Google Patents

Abstract

The invention discloses a three-level permanent magnet synchronous motor three-vector model prediction torque control method. The vector partition method is adopted to select the voltage vector, and the large vector in the three-level inverter candidate vector is substituted into the value function. The first optimal voltage vector is determined by searching for the optimal voltage vector, the sector is determined by the position of the first optimal voltage vector, and the second optimal voltage vector is substituted from the three non-zero vectors in the sector where the first optimal voltage vector is located. The value function is optimized, and the action time of the first optimal voltage vector, the second optimal voltage vector and the zero vector is combined to control the action of the three-level inverter, so that the first optimal voltage vector is optimized from the original 27 times. It is reduced to 6 times. This optimization process not only considers the selection of the two optimal voltage vectors, but also combines its action time to ensure that the selected two optimal voltage vectors can ensure the balance of the mid-point potential and lower switching. frequency.

Description

A three-level permanent magnet synchronous motor three-vector model predictive torque control method

technical field

The invention relates to torque control of permanent magnet synchronous motors, in particular to a three-level permanent magnet synchronous motor three-vector model predictive torque control method.

Background technique

Due to the advantages of simple structure, small size, high power density, high efficiency, and excellent operating performance, permanent magnet synchronous motors have been widely used in recent years, such as in the field of transportation, the field of love, the driving field of vehicles and ships, and as a pump Class, compressor, high-precision servo system and other aspects of the rotating mechanism. Therefore, there are more and more studies on the control of permanent magnet synchronous motors.

Existing permanent magnet synchronous motor control uses model predictive torque control. Model predictive torque control is a classic control method in finite set prediction algorithms. It is very suitable for permanent magnet synchronous motor control systems. It has simple control strategies and dynamic The advantages of good performance are more suitable for multi-scalar, multi-constraint, and nonlinear systems. The traditional model predictive torque control obtains the predictive model of the system, traverses a known number of state vectors, and then sets an appropriate value function according to the expected control target. By comparing the value function, the switching sequence that the power device needs to execute in the future can be obtained. .

Compared with the two-level inverter, the three-level inverter has a smaller switching frequency and a wider range of vector selection. However, in practical applications, the balance of the mid-point potential needs to be considered. In the existing model predictive control strategy, the midpoint potential balance term is introduced into the value function, and the midpoint potential balance problem is considered in the optimization process, but it also increases the difficulty of implementing the control strategy.

Disadvantages of the existing technology:

(1) In the existing model predictive control strategy, only one voltage vector with a fixed size and a fixed direction can be obtained in each control cycle, but the actual reference voltage vector to be tracked is uncertain, so the output voltage vector cannot be Effectively makes the feedback torque track the reference torque, which directly leads to the large magnetic flux pulsation of the permanent magnet synchronous motor, the unstable output torque pulsation, and the poor steady-state performance of the motor.

(2) The number of switch states of devices with three-level or more-level topology is too large, the calculation cost is too large, and there are complex redundant vectors, so it is difficult to determine the switch sequence combination;

(3) The value function contains the mid-point potential control term, and the setting of the weight factor becomes more complicated, resulting in low control accuracy and complicated calculation;

(4) Multi-vector predictive control causes higher switching frequency and increases switching loss.

SUMMARY OF THE INVENTION

Purpose of the invention: In view of the above shortcomings, the present invention provides a three-vector model predictive torque control method for a three-level permanent magnet synchronous motor that can quickly screen vectors and reduce switching losses.

Technical solution: In order to solve the above problems, the present invention adopts a three-level permanent magnet synchronous motor three-vector model predictive torque control method, which includes the following steps:

(1) According to the state space model of permanent magnet synchronous motor, the state equations of current and magnetic flux are constructed, and the prediction model of torque and magnetic flux is obtained according to the state equations of current and magnetic flux. According to the DC side capacitance of the three-level inverter obtaining a first capacitor voltage prediction model from the voltage model; constructing a first value function according to the first capacitor voltage prediction model;

(2) Sampling to obtain the current at time k;

(3) Substitute the data obtained by sampling into the prediction model of torque and magnetic flux, and obtain the torque and flux linkage at the next moment k+1;

(4) Substitute the large vector in the three-level inverter space vector diagram into the first value function as the candidate vector of the first optimal voltage vector, and select the first optimal voltage vector;

(5) Three non-zero vectors in the sector corresponding to the first optimal voltage vector are used as candidate vectors for the second optimal voltage vector, and the zero vector is used as the third voltage vector;

(6) Select three candidate vectors in turn as the second optimal voltage vector, calculate the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, and obtain the duty ratio. The empty ratio modulates a new voltage vector;

(7) establishing a second capacitance voltage prediction model according to the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, and building a second cost function according to the second capacitance voltage prediction model;

(8) Substitute the three new voltage vectors modulated in step 6 into the second value function to select the optimal second optimal voltage vector;

(9) Input the switching sequence corresponding to the selected first optimal voltage vector, the second optimal voltage vector and the third voltage vector and the corresponding action time into the pulse modulator to control the action of the three-level inverter.

Further, the state equations of the current and the magnetic flux in the step 1 are:

Among them, i _α and i _β represent the components of the stator current vector on the α and β axes respectively; L _s represents the stator inductance; u _α and u _β represent the components of the stator voltage vector on the α and β axes respectively; R _s represents the stator resistance ; ω _r represents the rotor electrical angular velocity of the motor; ψ _α , ψ _β represent the components of the electron flux linkage vector under the α and β axes, respectively.

Further, in the step 1, the prediction model of torque and magnetic flux is:

ψ _s (k+1)=ψ _s (k)+T _s [u _s (k)-R _s i _s (k)]

为叉乘符号。Among them, T _s is the sampling period; _{i s} (k+1), ψ _s (k+1), and Te (k+1) are the stator current vector, stator flux linkage and torque vector at time k+1, respectively; i _s (k), ψ _s (k), u _s (k), and ω _r (k) are the stator current vector, stator flux linkage, stator voltage vector and the electrical angular velocity of the rotor at time k, respectively; j is an imaginary unit, and p is Number of pole pairs,

is the cross product symbol.

Further, in the step 1, the capacitance voltage prediction model is:

Among them, v _c1 (k) and v _c2 (k) are the capacitor voltages sampled at time k; v _c1 (k+1) and v _c2 (k+1) are the predicted capacitor voltages at time k+1; i _c1 (k ) and i _c2 (k) are defined values according to the switching state and output current of the inverter at time k, C ₁ and C ₂ are the two capacitors on the DC side.

Further, in the step 4, the first value function is:

为转矩给定值，

为定子磁通给定值，λ₁、λ₂、λ₃为权重系数，F_sw为一个控制周期内开关管发生开通和关断的动作变换次数。in,

is the torque given value,

is the given value of the stator magnetic flux, λ ₁ , λ ₂ , λ ₃ are the weight coefficients, and F _sw is the switching times of switching on and off in a control cycle.

Further, in the step 6, the action times of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector are respectively t ₁ , t ₂ and t ₀ :

t ₀ =T _s -t ₁ -t ₂

S为一个量，且

为转矩和磁通在第一最优电压矢量作用下的变化率；

为转矩和磁通在第二最优电压矢量作用下的变化率；

为转矩和磁通在零矢量作用下的变化率。in,

S is a quantity, and

is the rate of change of torque and magnetic flux under the action of the first optimal voltage vector;

is the rate of change of torque and magnetic flux under the action of the second optimal voltage vector;

is the rate of change of torque and flux under the action of zero vector.

Further, t ₁ , t ₂ , and t ₀ are in the range of 0 to T _s , and to ensure that t ₁ , t ₂ , and t ₀ are in the range of 0 to T _s , the following processing is required:

(6.1) When any two of the action times t ₁ , t ₂ and t ₀ are less than 0, and the other action time is greater than T _s , then the action time less than 0 is equal to 0, and the action time greater than T _s is equal to T _s ;

(6.2) When the action time t ₁ +t ₂ ≤T _s , the action time of each vector remains unchanged;

(6.3) When the action time t ₁ +t ₂ ≥T _s , the action time of the three vectors needs to be substituted into the following formula for correction:

Wherein: t ₁ ', t' ₂ and t' ₀ respectively represent the action time of the corrected first optimal voltage vector, the second optimal voltage vector and the third voltage vector.

Further, in the step 6, the modulation formula of the new voltage vector is:

Among them, u _α2 and u _β2 are the α and β axis components of the new voltage vector, u _{opt1_α} and u _{opt1_β} are the α and β axis components of the first optimal voltage vector, and u _{opt2_n_α} and u _{opt2_n_β} are the second optimal voltage vector set In the α and β axis components of the nth voltage vector, the duty ratios of the first optimal voltage vector u _opt1 , the second optimal voltage vector u _{opt2_n} and the zero vector u ₀ are η _opt1 , η _{opt2_n} , and η ₀ , respectively.

Further, the second capacitor voltage prediction model in the step 7 is:

Among them, v' _c1 (k+1), v' _c2 (k+1) are the two voltages on the DC side at time k+1 under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector Capacitor voltage, v' _c1 (k), v' _c2 (k) are the two capacitor voltages at time k under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, i _c11 (k) , _ic21 (k), _ic31 (k), _ic12 (k), _ic22 (k), _ic32 (k) are the first optimal voltage vector, the second optimal voltage vector and the third voltage vector respectively Output current value when active.

Further, in the step 7, the second value function g _n is:

Among them, g _n is the second value function.

Beneficial effect: Compared with the prior art, the present invention has the significant advantage that the sector is determined by the position of the first optimal voltage vector to select the second optimal voltage vector, which reduces the number of optimizations, and combines the two optimal voltage vectors and the action time of the zero vector to ensure the neutral point potential balance and lower switching frequency; using the three-vector model predictive control, the synthesized vector can not only change in direction, but also change its size to achieve the best control effect. .

Description of drawings

1 is a topology diagram of an NPC three-level inverter of the present invention;

Fig. 2 is the three-level basic space vector distribution diagram of the vector partition provided by the present invention;

Fig. 3 is the block diagram of the three-vector model predictive control of the three-level inverter-driven permanent magnet synchronous motor of the present invention;

Figure 4 is a flow chart of the vector selection of the present invention.

Detailed ways

In this embodiment, a three-level permanent magnet synchronous motor three-vector model predicted torque control method, wherein

According to the state space model of permanent magnet synchronous electromotive force, the state equations of current and magnetic flux are constructed as:

Among them, i _α and i _β represent the components of the stator current vector on the α and β axes respectively; L _s represents the stator inductance; u _α and u _β represent the components of the stator voltage vector on the α and β axes respectively; R _s represents the stator resistance ; ω _r represents the rotor electrical angular velocity of the motor; ψ _α , ψ _β represent the components of the electron flux linkage vector under the α and β axes, respectively.

According to the state equation of the permanent magnet synchronous motor current and magnetic flux, the differential term is discretized, and the prediction model of the permanent magnet synchronous motor torque and magnetic flux is calculated as follows:

ψ _s (k+1)=ψ _s (k)+T _s [u _s (k)-R _s i _s (k)]

为叉乘符号。Among them, T _s is the sampling period; _{i s} (k+1), ψ _s (k+1), and Te (k+1) are the stator current vector, stator flux linkage and torque vector at time k+1, respectively; i _s (k), ψ _s (k), u _s (k), and ω _r (k) are the stator current vector, stator flux linkage, stator voltage vector and the electrical angular velocity of the rotor at time k, respectively; j is an imaginary unit, and p is Number of pole pairs,

is the cross product symbol.

As shown in Figure 1, the three-level inverter has three bridge arms, which are three-phase A, B, and C, and each phase has four switching tubes. During normal operation, each phase has three switching states. For phase A, the "P" state corresponds to the conduction of Sa1 and Sa2, and the conduction of Sa3 and Sa4; the state of "O" corresponds to the conduction of Sa2 and Sa3, and the conduction of Sa1 and Sa4; The "N" state corresponds to that Sa3 and Sa4 are turned on, and Sa1 and Sa2 are turned off. Therefore, the three-phase bridge arm has 27 switching states and 19 different voltage vectors. Different voltage vectors have different effects on the midpoint potential, so it is necessary to set the midpoint potential balance control in the control system. According to the DC side capacitor voltage model of the three-level inverter, the first capacitor voltage prediction model is obtained as:

Among them, v _c1 (k) and v _c2 (k) are the capacitor voltages sampled at time k; v _c1 (k+1) and v _c2 (k+1) are the predicted capacitor voltages at time k+1; i _c1 (k ) and i _c2 (k) are defined values according to the switching state and output current of the inverter at time k, C ₁ and C ₂ are the two capacitors on the DC side, i _a (k), i _b (k), _ic ( k) is the three-phase stator current, i _dc (k) is the DC bus current, H _1a , H _1b , H _1c , H _2a , H _2b , H _2c are the three-phase state values, for phase A, "P" In the state, H _1a is 1 and the rest are 0; in the "O" state, H _1b is 1 and the rest are 0; in the "N" state, H _1c is 1 and the rest are 0.

The target first value function g is constructed by the absolute value of the difference between the electromagnetic torque and magnetic flux predicted value of the permanent magnet synchronous motor and the reference value, the absolute value of the difference between the DC side capacitor voltage of the inverter and the switching frequency limit:

为转矩给定值，

为定子磁通给定值，λ₁、λ₂、λ₃为权重系数，F_sw为一个控制周期内开关管发生开通和关断的动作变换次数，具体计算如下：in,

is the torque given value,

is the given value of the stator magnetic flux, λ ₁ , λ ₂ , and λ ₃ are the weight coefficients, and F _sw is the switching times of switching on and off in a control cycle. The specific calculation is as follows:

F _sw =|S _a (k)-S _a (k-1)|+|S _b (k)-S _b (k-1)|+|S _c (k)-S _c (k-1)|

Among them, S _a (k), S _b (k), S _c (k) are the switching states of the three-phase inverter at time k; S _a (k-1), S _b (k-1), S _c (k -1) is the switching state of the three-phase inverter at the last moment k-1.

As shown in Figure 2, the space vectors of the three-level inverter are 3 zero vectors, 12 positive and negative redundant small vectors, 6 medium vectors and 6 large vectors, of which the large vectors have no effect on the midpoint potential , so the space vector is divided into six sectors centered on the large vector.

Substitute the large vector in the three-level inverter space vector diagram as the candidate vector of the first optimal voltage vector into the first value function g, and select the voltage vector that minimizes the result as the first optimal voltage vector u _opt1 ; Include the non-zero vectors in the same sector as the first optimal voltage vector u _opt1 into the second optimal voltage vector candidate set, and there are three candidate voltage vectors in each sector; The zero vector of the voltage vector u _opt1 in the same sector is used as the third voltage vector.

Select three candidate voltage vectors in turn as the second optimal voltage vector, calculate the action time of the first optimal voltage vector u _opt1 , the second optimal voltage vector u _opt2 and the third voltage vector (zero vector) u ₀ , according to According to the deadbeat control principle of torque and flux linkage, after a sampling period ends, the predicted value of torque and flux linkage is equal to the given value, namely:

为转矩和磁通在第一最优电压矢量作用下的变化率；

为转矩和磁通在第二最优电压矢量作用下的变化率；

为转矩和磁通在零矢量作用下的变化率；

为第一最优电压矢量u_opt1作用下的转矩和磁链的预测值；T_e(k+1)_{opt2_n}、|ψ_s(k+1)_{opt2_n}|为第二最优电压矢量集合中第n个电压矢量u_{opt2_n}作用下的转矩和磁链的预测值；T_e(k+1)_{_0}、|ψ_S(k+1)_{_0}|为第三电压矢量(零矢量)作用下转矩和磁链的预测值。in

is the rate of change of torque and magnetic flux under the action of the first optimal voltage vector;

is the rate of change of torque and magnetic flux under the action of the second optimal voltage vector;

is the rate of change of torque and magnetic flux under the action of zero vector;

are the predicted values of torque and flux linkage under the action of the first optimal voltage vector u _opt1 ; T _e (k+1) _{opt2_n} , |ψ _s (k+1) _{opt2_n} | Predicted values of torque and flux linkage under the action of _n voltage vectors u _{opt2_n} ; Te (k+1) _{_0} , |ψ _S (k+1) _{_0} | are the torque under the action of the third voltage vector (zero vector) and the predicted value of the flux linkage.

The action times of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector are respectively t ₁ , t ₂ and t ₀ :

t ₀ =T _s -t ₁ -t ₂

S为一个量，且in,

S is a quantity, and

After calculating t ₁ , t ₂ , and t ₀ , in order to ensure that t ₁ , t ₂ , and t ₀ are within the range of 0 to T _s , the following processing is required:

(6.1) When any two of the action times t ₁ , t ₂ and t ₀ are less than 0, and the other action time is greater than T _s , then the action time less than 0 is equal to 0, and the action time greater than T _s is equal to T _s ;

(6.2) When the action time t ₁ +t ₂ ≤T _s , the action time of each vector remains unchanged;

(6.3) When the action time t ₁ +t ₂ ≥T _s , the action time of the three vectors needs to be substituted into the following formula for correction:

Wherein: t ₁ ', t' ₂ and t' ₀ respectively represent the action time of the corrected first optimal voltage vector, the second optimal voltage vector and the third voltage vector.

The duty cycles of the first optimal voltage vector u _opt1 , the second optimal voltage vector u _{opt2_n} and the zero vector u ₀ are η _opt1 , η _{opt2_n} , and η ₀ respectively, and a new voltage vector is modulated according to the duty cycle:

Among them, u _α2 and u _β2 are the α and β axis components of the new voltage vector, u _{opt1_α} and u _{opt1_β} are the α and β axis components of the first optimal voltage vector, and u _{opt2_n_α} and u _{opt2_n_β} are the second optimal voltage vector set The α and β axis components of the nth voltage vector in .

In a three-level inverter, the first optimal voltage vector, the second optimal voltage vector and the third voltage vector have different effects on the midpoint potential, so the two voltages on the DC side are re-established according to the action time of the three voltage vectors. The predicted voltage of the capacitor, the second capacitor voltage prediction model is:

Among them, v' _c1 (k+1), v' _c2 (k+1) are the two voltages on the DC side at time k+1 under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector Capacitor voltage, v' _c1 (k), v' _c2 (k) are the two capacitor voltages at time k under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, i _c11 (k) , _ic21 (k), _ic31 (k), _ic12 (k), _ic22 (k), _ic32 (k) are the first optimal voltage vector, the second optimal voltage vector and the third voltage vector respectively Output current value when active.

The second value function g _n constructed according to the second capacitor voltage prediction model is:

The three new voltage vectors modulated by sequentially selecting three candidate voltage vectors as the second optimal voltage vector are substituted into the second value function g _n , and the voltage vector with the smallest result is selected as the second optimal voltage vector, and finally the The selected switching sequences corresponding to the first optimal voltage vector, the second optimal voltage vector and the third voltage vector (zero vector) and their action time are input to the pulse modulator to control the action of the three-level inverter.

Claims (10)

1. a three-level permanent magnet synchronous motor three-vector model predictive torque control method, is characterized in that, comprises the following steps: (1) According to the state space model of the permanent magnet synchronous motor, the state equations of current and magnetic flux are constructed, and the prediction model of torque and magnetic flux is obtained according to the state equations of current and magnetic flux. According to the DC side capacitance of the three-level inverter obtaining a first capacitor voltage prediction model from the voltage model; constructing a first value function according to the first capacitor voltage prediction model; (2) Sampling to obtain the current at time k; (3) Substitute the data obtained by sampling into the prediction model of torque and magnetic flux, and obtain the torque and flux linkage at the next moment k+1; (4) Substitute the large vector in the three-level inverter space vector diagram into the first value function as the candidate vector of the first optimal voltage vector, and select the first optimal voltage vector; (5) Three non-zero vectors in the sector corresponding to the first optimal voltage vector are used as candidate vectors for the second optimal voltage vector, and the zero vector is used as the third voltage vector; (6) Select three candidate vectors in turn as the second optimal voltage vector, calculate the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, and obtain the duty ratio. The empty ratio modulates a new voltage vector; (7) establishing a second capacitance voltage prediction model according to the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, and building a second cost function according to the second capacitance voltage prediction model; (8) Substitute the three new voltage vectors modulated in step 6 into the second value function to select the optimal second optimal voltage vector; (9) Compare the switching sequence corresponding to the first optimal voltage vector selected in step (4), the second optimal voltage vector selected in step (8) and the third voltage vector selected in step (5) and the corresponding The action time input pulse modulator controls the action of the three-level inverter. 2. The three-vector model predictive torque control method according to claim 1, wherein the state equation of the current and the magnetic flux in the step 1 is:

Among them, i _α and i _β represent the components of the stator current vector on the α and β axes respectively; L _s represents the stator inductance; u _α and u _β represent the components of the stator voltage vector on the α and β axes respectively; R _s represents the stator resistance ; ω _r represents the rotor electrical angular velocity of the motor; ψ _α , ψ _β represent the components of the motor flux linkage vector under the α and β axes, respectively. 3. The three-vector model predictive torque control method according to claim 2, wherein in the step 1, the predictive models of torque and magnetic flux are:

ψ _s (k+1)=ψ _s (k)+T _s [u _s (k)-R _s i _s (k)]

Among them, T _s is the sampling period; _{i s} (k+1), ψ _s (k+1), and Te (k+1) are the stator current vector, stator flux linkage and torque vector at time k+1, respectively; i _s (k), ψ _s (k), u _s (k), and ω _r (k) are the stator current vector, stator flux linkage, stator voltage vector and the electrical angular velocity of the rotor at time k, respectively; j is an imaginary unit, and p is Number of pole pairs,

is the cross product symbol. 4. The three-vector model predicted torque control method according to claim 3, wherein in the step 1, the capacitor voltage prediction model is:

Among them, v _c1 (k) and v _c2 (k) are the capacitor voltages sampled at time k; v _c1 (k+1) and v _c2 (k+1) are the predicted capacitor voltages at time k+1; i _c1 (k ) and i _c2 (k) are defined values according to the switching state and output current of the inverter at time k, C ₁ and C ₂ are the two capacitors on the DC side. 5. The three-vector model predictive torque control method according to claim 4, wherein in the step 4, the first value function is:

in,

is the torque given value,

is the given value of the stator magnetic flux, λ ₁ , λ ₂ , λ ₃ are the weight coefficients, and F _sw is the switching times of switching on and off in a control cycle. 6 . The three-vector model predictive torque control method according to claim 5 , wherein in the step 6, the action times of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector are respectively 6 . for t ₁ , t ₂ and t ₀ : t ₀ =T _s -t ₁ -t ₂

in,

S is a quantity, and

is the rate of change of torque and magnetic flux under the action of the first optimal voltage vector;

is the rate of change of torque and magnetic flux under the action of the second optimal voltage vector;

is the rate of change of torque and flux under the action of zero vector. 7 . The three-vector model predictive torque control method according to claim 6 , wherein t ₁ , t ₂ , and t ₀ are within the range of 0 to T _s , and in order to ensure that t ₁ , t ₂ , and t ₀ are within the range of t 1 , t 2 , and t 0 . In the range of 0～T _s , the following processing is required: (6.1) When any two of the action times t ₁ , t ₂ and t ₀ are less than 0, and the other action time is greater than T _s , then the action time less than 0 is equal to 0, and the action time greater than T _s is equal to T _s ; (6.2) When the action time t ₁ +t ₂ ≤T _s , the action time of each vector remains unchanged; (6.3) When the action time t ₁ +t ₂ ≥T _s , the action time of the three vectors needs to be substituted into the following formula for correction:

Wherein: t' ₁ , t' ₂ and t' ₀ respectively represent the action time of the corrected first optimal voltage vector, the second optimal voltage vector and the third voltage vector. 8. The three-vector model predictive torque control method according to claim 7, wherein in the step 6, the modulation formula of the new voltage vector:

Among them, u _α2 and u _β2 are the α and β axis components of the new voltage vector, u _{opt1_α} and u _{opt1_β} are the α and β axis components of the first optimal voltage vector, and u _{opt2_n_α} and u _{opt2_n_β} are the second optimal voltage vector set In the α and β axis components of the nth voltage vector, the duty ratios of the first optimal voltage vector u _opt1 , the second optimal voltage vector u _{opt2_n} and the zero vector u ₀ are η _opt1 , η _{opt2_n} , and η ₀ , respectively. 9. The three-vector model predicted torque control method according to claim 8, wherein the second capacitor voltage prediction model in the step 7 is:

Among them, v' _c1 (k+1), v' _c2 (k+1) are the two voltages on the DC side at time k+1 under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector Capacitor voltage, v' _c1 (k), v' _c2 (k) are the two capacitor voltages at time k under the action time of the first optimal voltage vector, the second optimal voltage vector and the third voltage vector, i _c11 (k) , _ic21 (k), _ic31 (k), _ic12 (k), _ic22 (k), _ic32 (k) are the first optimal voltage vector, the second optimal voltage vector and the third voltage vector respectively The output current value of the two capacitors on the DC side when acting. 10. The three-vector model predictive torque control method according to claim 9, wherein the second value function in the step 7 is:

Among them, g _n is the second value function.

Images