CN115065294A - Switched reluctance motor model prediction torque control method based on multi-level power converter - Google Patents

Abstract

The present invention proposes a model prediction torque control method for a switched reluctance motor based on a multilevel power converter. First, the inductance characteristics and torque characteristics of the switched reluctance motor are obtained by offline measurement, and the flux linkage current torque of the motor is produced. Then, according to the current position, speed and current information, combined with the switch state, look up the table to predict the current and position information at the next moment; and in order to perform delay compensation, it is necessary to further predict the current and position information at the next two moments, and then check The table obtains the torque under each switching state and brings it into the cost function. Through the optimization of the cost function, the switching state of the optimal solution is obtained, and then the switching state of the switching tube of the three-phase power converter is corresponding to the decoupling of the switching signal. , so as to achieve the effect of torque ripple suppression.

Description

A Model Predictive Torque Control of Switched Reluctance Motor Based on Multilevel Power Converter manufacturing method

technical field

The invention belongs to the field of motor control, in particular to a model prediction torque control method of a switched reluctance motor based on a multi-level power converter.

Background technique

In terms of its own structure, both the stator and the rotor of the switched reluctance motor are salient pole structures, and there are no rotor windings and permanent magnets. The structure is simple and robust, and the cost is relatively low. It has high torque density and power density without cogging torque. Therefore, the structure of the reluctance motor is stronger than other motors, and it will last longer in harsh environments. In motor control, the switch reluctance motor has high controllability, and its electromagnetic torque direction is determined by the winding excitation sequence and has nothing to do with the current direction. Because switched reluctance motors have many inherent advantages such as high efficiency, high reliability, high starting torque and high fault tolerance, they are widely used in electric vehicles, household appliances, aerospace, industrial transmission and other fields. However, due to the highly nonlinear nature of their electromagnetic properties, switched reluctance motors suffer from disadvantages such as torque ripple and servo vibration, which limit their application fields. Therefore, in order to improve the performance of the switch reluctance motor speed control system, suppressing the torque ripple and vibration has become a research hotspot of the switch reluctance motor.

At present, the commonly used methods to reduce torque ripple mainly include torque distribution function, phase current PI controller, direct torque control and direct instantaneous torque control, etc. These methods have their own advantages and disadvantages. Model predictive control can achieve multi-objective optimization intuitively and conveniently by constructing a cost function, and has received more and more attention in switched reluctance motor control. By constructing the cost functions of the torque and radial force of the switched reluctance motor, model predictive control not only solves the problems of torque ripple and vibration at the same time, but also plays an important role in improving the applicability and speed regulation performance of the switched reluctance motor.

However, the power converters used by the current model predictive control methods are all three-level traditional asymmetric half-bridges. Although they also have the effect of reducing torque fluctuations, the three-level power converters are increasingly unable to meet the requirements of high-speed motors and various The demand of the application field has increased, so a more flexible and efficient multi-level power converter is required, which makes a requirement for the control method of the multi-level power converter.

SUMMARY OF THE INVENTION

Aiming at the problem that the traditional model prediction is the control of the asymmetric half-bridge three-level power converter, but the three-level can not well meet the requirements of the device and the level diversity, the present invention proposes a multi-level power converter based on This method is no longer aimed at the control of traditional asymmetric half-bridge power converters, but is aimed at five-level power converters.

This method does not simply predict all switch states when three levels are continued, but predicts through certain selections, and proposes an improved strategy for the commutation algorithm and vector optimization. The overall steps of the method are as follows: first, the inductance characteristics and torque characteristics of the switched reluctance motor are obtained by offline measurement, and the flux linkage current torque table of the motor is produced; then according to the current position, speed and current information, combined with the switch state, look up the table Predict the current and position information at the next moment; and in order to perform delay compensation, it is necessary to further predict the current and position information at the next two moments, and then look up the table to obtain the torque in each switching state and bring it into the cost function, through the cost The function optimization obtains the switch state of the optimal solution, and then through the decoupling of the switch signal, it corresponds to the switch state of the switch tube of the three-phase power converter, so as to achieve the effect of torque ripple suppression.

The technical scheme of the present invention is:

The method for predictive torque control of a switched reluctance motor model based on a multi-level power converter includes the following steps:

Step 1: Determine the reference torque T _ref ; obtain the inductance characteristics, flux linkage characteristics and torque characteristics of the switched reluctance motor, and construct data tables L _ph (I _ph ,θ), T _ph (I _ph ,θ) according to the above characteristics ; Among them, L _ph , T _ph , I _ph , and θ represent the switched reluctance motor phase inductance, phase torque, phase current and rotor position, respectively;

Step 2: Collect the rotor position θ(k), phase current i _ph (k), rotational speed ω(k), and phase voltage V _ph (k) and flux linkage ψ _ph (k) of the motor at time k;

Step 3: According to the measurement data obtained in Step 2, predict the rotor position θ(k+1) and flux linkage ψ _ph (k+1) at time k+1; and obtain the phase current i _ph (k+1) by looking up the table ;in

θ(k+1)=θ(k)+ω(k)Ts

ψ _ph (k+1)=[V _ph (k)-i _ph (k)R]Ts+ψ _ph (k)

where R is the load resistance, Ts is the sampling frequency, ω(k), θ(k), i _ph (k), and V _ph (k) are the rotational speed, rotor position, phase current, and phase voltage at time k, respectively, θ (k+1), i _ph (k+1) are the rotor position and phase current value at the time of k+1, respectively;

Step 4: Predict the rotor position θ(k+2) at time k+2 according to the formula θ(k+2)=2θ(k+1)-θ(k), where θ(k), θ(k+1) , θ(k+2) are the rotor positions at moments k, k+1, and k+2 respectively; according to the rotor position θ(k+2) at moment k+2, it is judged whether it is in the commutation zone, if it is in the commutation zone , then the torque prediction is performed for the 12 possible switching states at the time k+2, respectively, and if it is not in the commutation region, the torque prediction is performed for the 5 possible switching states at the time k+2 respectively;

The five switch states are:

The switch state variables of phase A, phase B and phase C are (1, -1, -1), (0.5, -1, -1), (0, -1, -1), (-0.5, -1 respectively , -1), (-1, -1, -1);

The 12 switch states are:

The switching state variables of phase A, phase B and phase C are (1, 1, -1), (1, 0.5, -1), (1, 0, -1), (1, -0.5, -1) , (1, -1, -1), (0.5, 0.5, -1), (0.5, 0, -1), (0.5, -0.5, -1), (0.5, -1, -1), ( 0, 0, -1), (0, -0.5, -1), (0, -1, -1);

The process of torque prediction for each switching state is:

Step 4.1: According to the three-phase switch state variable of the switch state and the relationship between the switch state variable and the phase voltage, the output voltage is obtained, which is the predicted phase voltage V _ph (k+1) at the time of k+1; and then use the formula

ψ _ph (k+2)=[V _ph (k+1)-i _ph (k+1)R]Ts+ψ _ph (k+1)

Predict the flux linkage ψ _ph (k+2) at time k+2, and then obtain the phase current i _ph (k+2) at time k+2 by looking up the table method;

Step 4.2: Combine the phase current and rotor position information at time k+2, predict the phase torque T _ph (k+2) at time k+2 by looking up the table T _ph (I _ph , θ), and then obtain the total value according to the formula. Torque:

In the formula, N _ph represents the number of phases of the switched reluctance motor, ph represents the number of phases, T _ph (k+2) represents the phase torque at the moment of k+2, and T(k+2) represents the total torque of the switched reluctance motor;

Step 4.3: Predict the total torque at time k+2 according to step 4.2, and obtain the cost function:

J=q _T *(|T _a -T _aref | ² +|T _b -T _bref | ² +|T _c -T _cref | ² )+q _I *(I _a ² +I _b ² +I _c ² )

where J is the cost function, q _T is the proportion of torque error, q _I is the proportion of phase current error, T _a , T _b , and T _c are the total torque T(k) of the switched reluctance motor obtained according to step 4.2 +2) Decomposed three-phase torque, I _a , I _b , I _c are three-phase phase currents, T _aref , T _bref , T _cref are three-phase reference torques;

Step 5: For the 5 switch states or 12 switch states described in Step 4, the corresponding cost function values are obtained respectively, and the switch state corresponding to the minimum cost function value is found, which is the optimal solution switch that minimizes the torque ripple. signal; through the obtained optimal solution, the three-phase switch state variable corresponding to the optimal solution is obtained, and then the switch state of each phase switch tube is obtained by decoupling the corresponding relationship between the switch state variable and the switch tube of each phase;

Step 6: On the basis of the completion of Step 5, return to Step 2, and repeat this cycle.

Further, the reference torque T _ref is determined by a torque distribution function, and a cosine function is used as the torque distribution function.

Further, in step 5, the proportion q _T of the torque error is 15, and the proportion q _I of the phase current error is 0.02.

Further, in step 5, the correspondence table between the switch state variable and the switch tube of each phase is:

Mode S1 S2 S3 S4 1 On On On On 0.5 On On On Off 0 On On Off Off -0.5 Off On Off Off -1 Off Off Off Off

.

Further, in step 4.1, the relationship between the switch state variable and the phase voltage is: when the switch state variable U _ph =1, the phase voltage V _ph =2U; when U _ph =0.5, the phase voltage V _ph =U; U _ph = When 0, the phase voltage V _ph =0; when U _ph =-0.5, the phase voltage V _ph =-U; when U _ph =-1, the phase voltage V _ph =-2U.

beneficial effect

The invention discloses a model prediction torque control method of a switched reluctance motor based on a multilevel power converter. At the same time, the torque distribution is combined with the model predictive control. In the torque distribution, the compensation of the model prediction error is considered, and the reference torque of each phase is obtained. In the model predictive control, the inductance characteristics and torque characteristics of the switched reluctance motor are obtained through offline measurement, and the current and position information at the next moment are predicted according to the current position, speed and current information combined with the switch state look-up table; Compensation, and further predicts the current and position information, and then looks up the table to obtain the torque under each switching state and brings it into the cost function, constructs a cost function including the phase reference torque tracking error, and seeks the optimal control signal for the cost function. , and then use the optimal control signal as the switch signal to control the switch in the power converter, so as to realize the torque fluctuation suppression of the switched reluctance motor taking into account the operating efficiency. Simulation results verify the effectiveness of the method. The method has simple control logic, obvious torque ripple effect and easy engineering implementation.

Additional aspects and advantages of the present invention will be set forth, in part, from the following description, and in part will be apparent from the following description, or may be learned by practice of the invention.

Description of drawings

The above and/or additional aspects and advantages of the present invention will become apparent and readily understood from the following description of embodiments taken in conjunction with the accompanying drawings, wherein:

Fig. 1 is the control block diagram of the torque ripple suppression method of switched reluctance motor based on model predictive control;

Figures 2 to 6 are schematic diagrams of five switching states of a single-phase bridge arm of a switched reluctance motor multilevel power converter; (dotted line is the current flow);

in:

Fig. 2 is the circuit diagram of the fast excitation state of the switched reluctance motor multilevel power converter;

Fig. 3 is the circuit diagram of the normal excitation state of the switched reluctance motor multilevel power converter;

FIG. 4 is a circuit diagram of a zero-voltage freewheeling state of a switched reluctance motor multilevel power converter;

Fig. 5 is the circuit diagram of the normal demagnetization state of the switched reluctance motor multilevel power converter;

6 is a circuit diagram of a fast demagnetization state of a switched reluctance motor multilevel power converter;

Fig. 7, Fig. 8 are optimized vector selection charts (the underlined ones are the deleted vectors, and the black ones are for the selection vector);

in:

Fig. 7 is a chart of 5 kinds of vector selection of single-phase conduction region;

Fig. 8 is the 12 vector selection chart of the commutation area;

9 is a flowchart of a method for suppressing torque ripple of a switched reluctance motor based on model predictive control;

Figure 10 and Figure 11 are the torque ripple comparison diagrams between angular position control and model predictive control when running at 2000rpm;

in:

Figure 10 is a diagram of the total torque ripple of angular position control;

FIG. 11 is a diagram of the model predictive control total torque ripple.

Detailed ways

The present invention is described below with reference to specific embodiments, which are exemplary and intended to explain the present invention, but should not be construed as limiting the present invention. The motor used in the example is a 1kW three-phase 12/8-pole switched reluctance motor.

This embodiment proposes a model predictive torque control method for a switched reluctance motor based on a multi-level power converter, and performs model predictive control for five levels, including the following steps:

Step 1: Given a reference torque T _ref , the reference torque is a given value that actually needs to be achieved by the motor stably. In the closed-loop system, T _ref can be obtained from the torque distribution function, which can help the model predictive control to commutate in the commutation region. Generally, the cosine function can be used as the torque distribution function. Its demagnetization phase reference torque T _dref and excitation phase T _eref can be expressed by the reference torque T _ref as

T _eref =T _ref *(0.5-0.5*cos(pi*(θ-θ _on )/θ _ov ))

T _dref =T _ref *(0.5+0.5*cos(pi*(θ-θ _off )/θ _ov ))

where θ is the rotor position, θ _on is the turn-on angle, θ _off is the turn-off angle, and θ _ov is the transition angle.

The inductance characteristics, flux linkage characteristics and torque characteristics of the switched reluctance motor are obtained by the rotor fixed clamping method, and the data tables L _ph (I _ph ,θ) and T _ph (I _ph ,θ) are constructed according to the above characteristics; where, L _ph , T _ph , I _ph , and θ represent the switched reluctance motor phase inductance, phase torque, phase current, and rotor position, respectively.

Step 2: Collect the rotor position θ(k), phase current i _ph (k), rotational speed ω(k), and phase voltage V _ph (k) and flux linkage ψ _ph (k) of the motor at time k;

Step 3: According to the measurement data obtained in Step 2, predict the rotor position θ(k+1) and flux linkage ψ _ph (k+1) at time k+1; and obtain the phase current i _ph (k+1) through the table look-up method );in

θ(k+1)=θ(k)+ω(k)Ts

ψ _ph (k+1)=[V _ph (k)-i _ph (k)R]Ts+ψ _ph (k)

where R is the load resistance, Ts is the sampling frequency, ω(k), θ(k), i _ph (k), and V _ph (k) are the rotational speed, rotor position, phase current, and phase voltage at time k, respectively, θ (k+1), i _ph (k+1) are the rotor position and phase current value at the time of k+1, respectively;

Step 4: According to the formula

θ(k+2)=2θ(k+1)-θ(k)

Predict the rotor position θ(k+2) at time k+2, where θ(k), θ(k+1), and θ(k+2) are the rotor positions at times k, k+1, and k+2, respectively; According to the rotor position θ(k+2) at the time of k+2, it is judged whether it is in the commutation region. If it is in the commutation region, the torque current prediction is performed for the 12 possible switching states at the time of k+2. In the commutation region, the torque current prediction is performed for the 5 possible switch states at the time k+2 respectively;

The five switch states are:

The switch state variables of phase A, phase B and phase C are (1, -1, -1), (0.5, -1, -1), (0, -1, -1), (-0.5, -1 respectively , -1), (-1, -1, -1);

The 12 switch states are:

The switching state variables of phase A, phase B and phase C are (1, 1, -1), (1, 0.5, -1), (1, 0, -1), (1, -0.5, -1) , (1, -1, -1), (0.5, 0.5, -1), (0.5, 0, -1), (0.5, -0.5, -1), (0.5, -1, -1), ( 0, 0, -1), (0, -0.5, -1), (0, -1, -1).

For a certain phase, the relationship between the switching state variable U _ph and the phase voltage V _ph is as follows:

Among them, U represents half of the bus voltage, U _ph represents the switch state variable, U _ph = 1 means that all four switches of the five-level power converter are turned on, and U _ph = 0.5 means that all three switches of the five-level power converter are turned on. Conduction, U _ph = 0 five-level power converter has only two switches on, U _ph = -0.5 means that only one switch is on in the five-level power converter, U _ph = -1 means four switches are closed.

The combination rule of switch states is: in the one-way conduction area, only the switch state of the current conduction phase is calculated, and the remaining phases are -1; in the commutation area, only the two-phase switch states that are being commutated are predicted.

Mode S1 S2 S3 S4 Uph 0 Off Off Off Off -1 1 On Off Off Off × 2 Off On Off Off -0.5 3 On On Off Off 0 4 Off Off On Off -0.5 5 On Off On Off × 6 Off On On Off 0 7 On On On Off 0.5 8 Off Off Off On × 9 On Off Off On × 10 Off On Off On × 11 On On Off On × 12 Off Off On On 0 13 On Off On On × 14 Off On On On 0.5 15 On On On On 1

The above table is a table of switch state variables corresponding to the state of the switch tubes, in which U _ph is the switch state variable, that is, the ratio of the output voltage at both ends of the bridge arm of each phase to the bus voltage, and S1, S2, S3, and S4 represent four switch tubes per phase , On stands for turn on, Off stands for turn off.

For a five-level power converter, there are 125 space vectors that can be selected, but there are many redundant vectors, so vector optimization is required.

The optimal selection of vectors is carried out according to the following rules:

In the single-phase conduction area, the zero-voltage freewheeling state should be used as far as possible when reducing the torque to ensure the smooth reduction of the torque and avoid the violent torque fluctuation caused by the use of the reverse voltage demagnetization state.

In the commutation area, the working states of the two-phase windings must not be switched at the same time, so as to avoid the overshoot which leads to the larger torque ripple and the power loss caused by the frequent action of the power tube.

In the commutation area, the turn-on phase voltage should always be greater than or equal to zero to quickly build up the required phase current, ensure sufficient torque is provided in the single-phase conduction area, and give priority to the turn-on phase excitation when the torque needs to be increased. Therefore, the value of the turn-on phase voltage less than zero is artificially deleted; the turn-off phase voltage should not be greater than the turn-on phase voltage, so as to avoid the phase current entering the negative torque to generate negative torque, which will affect the motor efficiency, and give priority to when the torque needs to be reduced. The off-phase is demagnetized, so the value of the off-phase voltage greater than the on-phase voltage is artificially deleted.

In the commutation zone, when reducing the torque, the demagnetization phase should be kept in the zero-voltage freewheeling state as much as possible to avoid severe torque fluctuation.

Finally, 12 optimized candidate vectors are obtained.

Since the position signal cannot correspond to these switch vector signals one by one, we need to first divide the position into regions, corresponding to 125 kinds of vector signals, and then encode these vector signals to select the switch signals we need. The three phases are coded as follows:

The A-phase coding formula is: y1=[-floor(u/25.9)+2]/2

In the formula, the floor function is a rounding function towards negative infinity.

The B-phase coding formula is: y2=-{mod[floor(u/5.01),5]+2}/2

In the formula, the mod function is the remainder function

The C-phase encoding formula is:

The obtained coded values corresponding to different switching signals of the three phases will be retrieved again after the calculation of the cost function of minimizing torque ripple is finally completed, and decoded into switching signals.

The process of torque prediction for each switching state is:

Step 4.1: For a predicted possible switching state, according to the three-phase switching state variables and the relationship between the switching state variables and the phase voltage, the output voltage is obtained, and the corresponding output voltage is the phase voltage. Obtain the predicted phase voltage V _ph (k+1) at time k+1;

Use the formula

ψ _ph (k+2)=[V _ph (k+1)-i _ph (k+1)R]Ts+ψ _ph (k+1)

Predict the flux linkage ψ _ph (k+2) at time k+2, and then obtain the phase current i _ph (k+2) by looking up the table method; where R is the load resistance, Ts is the sampling frequency, θ(k+1), i _ph (k+1) and V _ph (k+1) are the rotor position, phase current and phase voltage at time k+1 respectively, θ(k+2) and i _ph (k+2) are respectively k+2 rotor position and phase current value at time;

Step 4.2: Combine the phase current and rotor position information at time k+2, predict the phase torque T _ph (k+2) at time k+2 by looking up the table T _ph (I _ph , θ), and then obtain the total torque ;

In the formula, N _ph represents the number of phases of the switched reluctance motor, ph represents the number of phases, T _ph (k+2) represents the phase torque at the moment of k+2, and T(k+2) represents the total torque of the switched reluctance motor;

Step 4.3: Predict the total torque at time k+2 according to step 4.2, and solve the cost function; the cost function is as follows:

J=q _T *(|T _a -T _aref | ² +|T _b -T _bref | ² +|T _c -T _cref | ² )+q _I *(I _a ² +I _b ² +I _c ² )

Among them, J is the cost function, q _T is the proportion of torque error, q _I is the proportion of phase current error, q _T and q _I are artificially given values, here the torque proportion q _T is set to 15, and the current The specific gravity q _I is set to 0.02, which can be changed according to the required motor performance. T _a , T _b , and T _c are three-phase torques decomposed from the total torque T(k+2) of the switched reluctance motor obtained in step 4.2, and I _a , I _b , and I _c are three-phase currents. T _aref , T _bref , T _cref are three-phase reference torques, which are given by the torque distribution function from the reference torques.

Step 5: Use 5 or 12 switching states to obtain the corresponding cost function values respectively, and find the switching state corresponding to the minimum cost function value, which is the optimal solution switching signal that minimizes the torque ripple. Through the obtained optimal solution, the three-phase switch state variable corresponding to the optimal solution is obtained, and then the switch state of each phase switch tube is decoupled through the switch state variable and the corresponding table of each phase switch tube; the decoupling function table is as follows:

Mode S1 S2 S3 S4 1 On On On On 0.5 On On On Off 0 On On Off Off -0.5 Off On Off Off -1 Off Off Off Off

Step 6: On the basis of the completion of Step 5, return to Step 2, and repeat this cycle.

FIG. 9 is a flow chart of the control method proposed by the present invention, and FIGS. 10 and 11 are comparison diagrams (2000 rpm) of torque ripple suppression using angular position control and the method used in the present invention, respectively. When switching from the angular position control to the model-predicted torque ripple suppression control method, the torque ripple is reduced from 131% to 23%. It can be seen that the method for suppressing the torque ripple of the switched reluctance motor based on the model predictive control proposed by the present invention has an obvious effect on reducing the torque ripple.

Although the embodiments of the present invention have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and those of ordinary skill in the art will not depart from the principles and spirit of the present invention Variations, modifications, substitutions, and alterations to the above-described embodiments are possible within the scope of the present invention without departing from the scope of the present invention.

Claims (5)

1. A switched reluctance motor model prediction torque control method based on a multilevel power converter is characterized by comprising the following steps: the method comprises the following steps:

step 1: determining a reference torque T _ref (ii) a Acquiring inductance characteristic, flux linkage characteristic and torque characteristic of the switched reluctance motor, and constructing a data table L according to the characteristics _ph (I _ph ,θ)、T _ph (I _ph θ); wherein L is _ph 、T _ph 、I _ph Theta respectively represents the phase inductance, the phase torque, the phase current and the rotor position of the switched reluctance motor;

step 2: collecting motor at time kRotor position θ (k) and phase current i _ph (k) Rotational speed ω (k) and phase voltage V _ph (k) And flux linkage psi _ph (k) A value of (d);

and step 3: predicting the rotor position theta (k +1) and the flux linkage psi at the moment k +1 according to the measurement data obtained in the step 2 _ph (k + 1); and obtaining phase current i by looking up the table _ph (k + 1); wherein

θ(k+1)＝θ(k)+ω(k)Ts

ψ _ph (k+1)＝[V _ph (k)-i _ph (k)R]Ts+ψ _ph (k)

Wherein R is a load resistor, Ts is a sampling frequency, and ω (k), θ (k), i _ph (k)、V _ph (k) The rotational speed, rotor position, phase current, and phase voltage at time k are θ (k +1), i _ph (k +1) is the rotor position and phase current value at the moment of k +1, respectively;

and 4, step 4: predicting a rotor position theta (k +2) at the moment k +2 according to a formula theta (k +2) ═ 2 theta (k +1) -theta (k), wherein theta (k), theta (k +1) and theta (k +2) are rotor positions at the moments k, k +1 and k +2 respectively; judging whether the rotor is in a phase conversion area or not according to the rotor position theta (k +2) at the moment k +2, if so, respectively carrying out torque prediction on 12 possible switching states at the moment k +2, and if not, respectively carrying out torque prediction on 5 possible switching states at the moment k + 2;

the 5 switch states are respectively as follows:

the switching state variables of the A phase, the B phase and the C phase are respectively (1, -1, -1), (0.5, -1, -1), (0, -1, -1), (-0.5, -1, -1) and (-1, -1, -1);

the 12 switch states are respectively as follows:

the switching state variables of the a phase, the B phase and the C phase are (1, 1, -1), (1, 0.5, -1), (1, 0, -1), (1, -0.5, -1), (1, -1, -1), (0.5, 0.5, -1), (0.5, 0, -1), (0.5, -0.5, -1), (0.5, -1, -1), (0, 0, -1), (0, -0.5, -1), (0, -1, 1);

the process of torque prediction for each switch state is:

step 4.1: three-phase switch state variable according to the switch state, and switch state variable and phase voltageThe magnitude of the output voltage is obtained, namely the predicted phase voltage V at the moment of k +1 _ph (k + 1); and then use the formula

ψ _ph (k+2)＝[V _ph (k+1)-i _ph (k+1)R]Ts+ψ _ph (k+1)

Predicting k +2 time flux linkage psi _ph (k +2), and then obtaining the phase current i at the time of k +2 by a table lookup method _ph (k+2)；

And 4.2: combining the phase current at time k +2 and rotor position information by looking up table T _ph (I _ph θ) predicting the phase torque T at the time k +2 _ph (k +2), and further, the total torque is calculated according to the formula:

in the formula N _ph Representing the number of phases of the switched reluctance motor, ph representing the numerical value of the number of phases, T _ph (k +2) represents the phase torque at the time of k +2, and T (k +2) represents the total torque of the switched reluctance motor;

step 4.3: predicting the total torque at the moment k +2 according to the step 4.2, and solving a cost function:

J＝q _T *(|T _a -T _aref | ² +|T _b -T _bref | ² +|T _c -T _cref | ² )+q _I *(I _a ² +I _b ² +I _c ² )

where J is a cost function, q _T As a proportion of the torque error, q _I Is the proportion of phase current error, T _a 、T _b 、T _c Three-phase torque resolved for the total torque T (k +2) of the switched reluctance machine obtained according to step 4.2, I _a 、I _b 、I _c For three-phase current, T _aref 、T _bref 、T _cref Is a three-phase reference torque;

and 5: respectively solving corresponding cost function values for the 5 switch states or 12 switch states in the step 4, and finding out the switch state corresponding to the minimum cost function value, namely the optimal turn-off signal for minimizing the torque ripple; obtaining three-phase switch state variables corresponding to the optimal solution through the obtained optimal solution, and then decoupling through the corresponding relation between the switch state variables and the switch tubes of each phase to obtain the switch state of the switch tubes of each phase;

step 6: upon completion of step 5, the process returns to step 2 to circulate.

2. The method for model-based predictive torque control of a switched reluctance motor based on a multilevel power converter according to claim 1, wherein: the reference torque T _ref The torque distribution function is determined by adopting a cosine function as the torque distribution function.

3. The method for model-based predictive torque control of a switched reluctance motor based on a multilevel power converter according to claim 1, wherein: in step 5, the ratio q of the torque error _T 15, proportion q of phase current error _I 0.02 was taken.

4. The method for model-based predictive torque control of a switched reluctance motor based on a multilevel power converter according to claim 1, wherein: in step 5, the corresponding table of the switch state variable and the switch tube of each phase is

Mode S1 S2 S3 S4 1 On On On On 0.5 On On On Off 0 On On Off Off -0.5 Off On Off Off -1 Off Off Off Off

。

5. The method for model-based predictive torque control of a switched reluctance motor based on a multilevel power converter according to claim 1, wherein: step 4.1, switching off of the switching state variables and the phase voltagesThe method comprises the following steps: when the state variable U is switched on or off _ph Phase voltage V equal to 1 _ph ＝2U；U _ph At 0.5, phase voltage V _ph ＝U；U _ph At 0, phase voltage V _ph ＝0；U _ph At-0.5 phase voltage V _ph ＝-U；U _ph When equal to-1, the phase voltage V _ph ＝-2U。

Images