Three-vector model predictive control method for T-type rectifier based on sliding mode control

Through sliding mode controller and three-vector model prediction control, the problems of large calculation volume and slow dynamic response of traditional three-level rectifiers are solved, efficient current ripple reduction and mid-point voltage balance are achieved, and the power quality of the rectifier is improved and Application capability.

CN114785166BActive Publication Date: 2024-11-19SOUTH CHINA UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210246329.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-14
Publication Date
2024-11-19
Estimated Expiration
2042-03-14

AI Technical Summary

Technical Problem

The traditional three-level rectifier model has large calculations, slow dynamic response, large current ripple, and low midpoint voltage balance capability, making it difficult to effectively apply in fast-paced applications.

Method used

Using the three-vector model prediction control method of T-type rectifier based on sliding mode control, the active power reference value is calculated by sliding mode controller, the active and reactive power of the sector center vector is predicted, and the functions of the optimal virtual switch vector and switch vector are determined. Sequentially, achieve precise control and mid-point voltage balance to reduce the calculation amount.

Benefits of technology

The dynamic response of the T-type three-level rectifier is improved, the current ripple is reduced, the power quality is improved, the midpoint voltage balancing capacity is enhanced, and the calculation burden is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114785166B_ABST
    Figure CN114785166B_ABST
Patent Text Reader

Abstract

The present invention discloses a three-vector model predictive control method for a T-type rectifier based on sliding mode control, comprising the following steps: calculating a power reference value P by sliding mode control ref , and at the same time, a preset method is used to predict the power at the next moment; the small sector where the optimal virtual switch vector is located is found according to the preset cost function; the action order of the switch vector and the action time of the corresponding vector are determined according to the cost function of the balanced midpoint voltage; finally, the various switch devices of the T-type three-level rectifier are controlled according to the found optimal switch state. The predictive control method disclosed in the present invention has at least the following beneficial effects: the cost function does not require a weight factor, which reduces the calculation burden; the sliding mode control speeds up the dynamic response speed of the system; selecting three vectors to control can reduce current ripple, reduce AC current harmonics, balance the midpoint voltage, improve the ability to maintain midpoint voltage balance, and effectively improve the power quality on the grid side.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of three-level PWM rectification, and in particular to a three-vector model predictive control method for a T-type rectifier based on sliding mode control. Background Art

[0002] With the rapid development of electric vehicles, large-scale charging piles have achieved large-scale development and application. At the same time, various large-scale energy storage devices have also developed rapidly, among which rectifiers are an indispensable link. The traditional two-level rectifier has high harmonic content in the input current and low power quality, which is not good for the power grid. T-type three-level rectifier can better solve the above problems.

[0003] There are many control strategies for T-type three-level rectifiers, but the most widely used and effective one is the model predictive control strategy. The model predictive control strategy is a nonlinear control strategy. The strategy determines the control quantity through the cost function according to the predicted value, and then controls the T-type three-level rectifier to achieve the control effect of tracking the given value. The traditional three-level rectifier model predictive control strategy has a large amount of calculation, consumes a lot of computing resources, and increases the hardware requirements. In addition, it only uses a single vector for control, the current ripple is large, the power quality is still not good enough, and the midpoint voltage balance requires the participation of weight factors, the balancing ability is low, and the effect is poor. At the same time, the model predictive control strategy used for three-level rectifiers currently uses PI controllers to achieve reference power control. The dynamic response of the PI controller is slow and the transition time is long, which is difficult to apply in fast-paced applications. Summary of the invention

[0004] The purpose of the present invention is to solve the above-mentioned defects in the prior art and provide a three-vector model predictive control method for a T-type rectifier based on sliding mode control. The method is combined with a sliding mode controller to improve the dynamic response of the T-type three-level rectifier. The three-vector segmented control can accurately track the optimal virtual switch vector, thereby reducing the current ripple and improving the power quality. The balance of the midpoint voltage is achieved by selecting the action sequence of the three vectors, eliminating the weight factor, improving the balancing ability, and greatly improving the balancing effect. The method of finding the center vector reduces the amount of calculation and reduces the calculation burden.

[0005] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0006] A three-vector model predictive control method for a T-type rectifier based on sliding mode control, the predictive control method comprising the following steps:

[0007] S1. Collect the current I on the DC side at a certain time interval dc , the two capacitor voltages U1 and U2 on the DC side, and the AC side voltage e abc and current iabc The active power reference value P is calculated through sliding mode control ref , and at the same time, the model prediction method is used to predict the active and reactive power corresponding to the central vector of each sector at the next moment;

[0008] S2. Find the small sector where the optimal virtual switch vector is located according to the cost function J of the sector center vector;

[0009] S3, according to the cost function J of the balanced midpoint voltage c Determine the action sequence and corresponding action time of the three switch vectors;

[0010] S4. Control each switch device of the T-type three-level rectifier according to the found optimal switch state.

[0011] Furthermore, in step S1, the output voltage is calculated using the two capacitor voltages U1 and U2 collected on the DC side, and the sliding surface is designed to track the given value of the output voltage, thereby calculating the active power reference value, wherein the sliding surface expression is as follows:

[0012]

[0013] Where s is the sliding surface, U ref is the given value of the output voltage, U dc is the value of the output voltage, k1 and k2 are the first and second sliding mode coefficients for realizing the function of the output voltage following the given value of the output voltage;

[0014] The active power reference value expression is as follows:

[0015]

[0016] Where P ref is the active power reference value, and C is the capacitance value on the DC side.

[0017] Further, in step S1, the active power and reactive power at the next moment are predicted by the model prediction method, that is, the active power and reactive power at moment k are respectively added to the power change rate of the active power and reactive power at moment k multiplied by the sampling time to obtain the active power and reactive power at moment k+1. Because the power change rate of the active power and reactive power corresponding to the center vector of each sector is different, the predicted values ​​of the active power and reactive power corresponding to the center vector of each sector are different. The expressions for active power prediction and reactive power prediction are as follows:

[0018]

[0019] Where P i (k+1), Qi (k+1) are the predicted values ​​of active power and reactive power of the center vector of the i-th sector, P(k) and Q(k) are the actual values ​​of active power and reactive power at time k, respectively. i ′(k),Q i ′(k) are the power change rates of active power and reactive power of the center vector of the i-th sector, T s is the sampling time.

[0020] Furthermore, the predicted values ​​of active power and reactive power of each sector center vector are compared and calculated with the active power reference value and the reactive power reference value, respectively, to obtain the cost function of the sector center vector. The cost function is used to determine which sector the optimal virtual switch vector is closest to, thereby determining its specific sector. The expression for calculating the cost function J of the sector center vector is as follows:

[0021]

[0022] In the formula, u iα 、u iβ is the component of the i-th sector center vector u on the α and β axes, u αref 、u βref is the voltage reference value U ref The components on the α and β axes, e α 、e β is the component of the grid voltage on the α and β axes, L is the value of the grid-side filter inductance, T s is the sampling time.

[0023] Furthermore, after finding the sector where the optimal virtual switch vector is located, the action times of the three switch vectors corresponding to the two types of switch vector sequences of the sector are respectively calculated. The calculation expressions of the action times corresponding to the three switch vectors are as follows:

[0024]

[0025] In the formula, t1, t2, and t3 are the action times of the three vectors, Q′ 1pre , Q′ 2pre , Q′ 3pre are the rates of change of reactive power of the three vectors at time k-1, P′ 1pre , P′ 2pre , P′ 3pre are the rates of change of the active power of the three vectors at time k-1, ΔP = P ref -P(k), ΔQ=Q ref -Q(k), P(k), Q(k) are the actual values ​​of active power and reactive power at time k, respectively. s is the sampling time, Pref , Q ref are the reference values ​​of active power and reactive power respectively.

[0026] Furthermore, in order to balance the midpoint voltage, that is, to make the difference between the upper and lower capacitor voltages on the DC side approach 0, the action times of the three switching vectors corresponding to the two types of switching vector sequences obtained above are substituted into the cost function J of the balanced midpoint voltage. c , the smaller the cost function value, the better the balancing effect, J c The expression is as follows:

[0027]

[0028] In the formula, S j =[S aj ,S bj ,S cj ] T is the state of each phase switch of each vector, S aj , S bj , S cj are the states of the switches of the three phases a, b, and c of the jth vector, i = [i a ,i b ,i c ] T is the grid side line current, i a 、i b 、i c are the line currents of phases a, b, and c, respectively, t j is the action time of each switching vector, C is the capacitance value of the DC side, ΔU(k) ​​is the difference between the upper and lower capacitor voltages of the DC side at time k, and ΔU(k)=U1-U2.

[0029] Furthermore, by comparing the cost function J of different switch vector sequences c The value of J c The switch vector sequence with the smallest value is used to determine the switch vector action sequence and its action time, and then control each switch tube to realize the function of the T-type rectifier system.

[0030] Compared with the prior art, the present invention has the following advantages and effects:

[0031] (1) The present invention uses three vectors of a specific small sector where the optimal virtual switch vector corresponding to the active power reference value is located for segmented control. The control time of each vector is calculated by a cost function, so that the vector acting on the switch tube can completely and accurately track the optimal virtual switch vector corresponding to the active power reference value, thereby reducing current ripple, reducing input harmonics, and improving power quality;

[0032] (2) The present invention selects the action order of the three vectors through the midpoint voltage balance cost function to achieve midpoint voltage balance, eliminating the weight factor, improving the balancing ability, and greatly improving the balancing effect;

[0033] (3) The present invention combines a sliding mode controller and designs a corresponding sliding mode surface to improve the dynamic response of a T-type three-level rectifier;

[0034] (4) The present invention first uses the distance between the center vector of each large sector and the optimal virtual switch vector to determine the large sector where the optimal virtual switch vector is located. Similarly, the distance between the center vector of each small sector in the large sector and the optimal virtual switch vector is used to determine the specific small sector where it is located. This search method reduces the amount of calculation of predictive control and alleviates the calculation burden. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0036] Figure 1 It is a control block diagram of a three-vector model predictive control method for a T-type three-level rectifier based on sliding mode control in the present invention;

[0037] Figure 2 It is a circuit topology diagram of a T-type three-level rectifier in the present invention;

[0038] Figure 3 It is a three-level space vector sector diagram in the present invention;

[0039] Figure 4 It is a grid-side current waveform diagram of the three-vector model predictive control method of a T-type three-level rectifier based on sliding mode control in the present invention;

[0040] Figure 5 It is a schematic diagram of output active power and reactive power of a three-vector model predictive control method for a T-type three-level rectifier based on sliding mode control in the present invention;

[0041] Figure 6 It is a schematic diagram of the transient response of the output voltage of the three-vector model predictive control method of the T-type three-level rectifier based on sliding mode control in the present invention when the load suddenly changes;

[0042] Figure 7 It is a schematic diagram of two capacitor voltages on the DC side of the three-vector model predictive control method for a T-type three-level rectifier based on sliding mode control in the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] Example

[0045] Figure 1 and Figure 2 It is a control block diagram and a rectifier circuit topology diagram of a three-vector model predictive control method for a T-type three-level rectifier based on sliding mode control disclosed in this embodiment. The predictive control method includes the following steps:

[0046] S1. Collect the current I on the DC side at a certain time interval dc , the two capacitor voltages U1 and U2 on the DC side, and the AC side voltage e abc and current i abc The active power reference value P is calculated through sliding mode control ref , and at the same time, the model prediction method is used to predict the active and reactive power corresponding to the central vector of each sector at the next moment;

[0047] S2. Find the small sector where the optimal virtual switch vector is located according to the cost function J of the sector center vector;

[0048] S3, according to the cost function J of the balanced midpoint voltage c Determine the action sequence and corresponding action time of the three switch vectors;

[0049] S4. Control each switch device of the T-type three-level rectifier according to the found optimal switch state.

[0050] The details are as follows:

[0051] like Figure 2 As shown, the rectifier AC side is directly connected to the grid through the input filter inductor L. The mathematical model of the AC side in the two-phase stationary αβ coordinate system is as follows:

[0052]

[0053] Among them, e α 、e β 、i α 、i β 、u α 、u β represents the grid phase voltage, line current and rectifier input voltage in the stationary two-phase αβ coordinate system, and L is the value of the grid-side filter inductance.

[0054] In a three-phase balanced system, there is the following relationship:

[0055]

[0056] Where ω is the angular frequency on the grid side.

[0057] According to the instantaneous power theory in the αβ coordinate system, the expressions of AC input active power and reactive power are as follows:

[0058] Combining the above formulas, in the stationary two-phase αβ coordinate system, the expressions of the active power and reactive power derivatives are as follows:

[0059]

[0060] Among them, P' represents the derivative of active power, and Q' represents the derivative of reactive power.

[0061] Then, the power at time k+1 is as follows:

[0062]

[0063] Among them, T s is the sampling period, active power P(k) and reactive power Q(k) are the active and reactive power at time k, active power P(k+1) and reactive power Q(k+1) are the predicted active and reactive power at time k+1.

[0064] like Figure 3 The three-level space vector sector diagram is shown in FIG. 1, and the switching vector of each sector is shown in Table 1. There are 3 zero vectors, 6 long vectors, 6 medium vectors and 12 short vectors. For the traditional model predictive control, 25 calculation cycles are required, while for the model predictive control of the present invention, only 14 calculation cycles are required.

[0065] Table 1. Switching vector table of sectors

[0066]

[0067]

[0068] The first step of the three-vector model predictive control method for T-type rectifier based on sliding mode control is to generate the active power reference value, such as Figure 2 As shown, the following expression can be obtained at point P:

[0069]

[0070] Among them, I inis the DC side input current, C is the DC side capacitance value, Udc is the DC side output voltage, I dc is the value of the DC side output current.

[0071] Multiply both sides by U dc ,get:

[0072]

[0073] Among them, P in is the DC side input active power.

[0074] According to the sliding mode controller, the expressions of sliding mode surface and reference power can be obtained as follows:

[0075]

[0076] Among them, s is the sliding surface, U ref is the given value of the output voltage, U dc is the value of the output voltage, I dc is the value of the output current, C is the capacitance value on the DC side, k1 and k2 are the first and second sliding mode coefficients for realizing the function that the output voltage follows the given value of the output voltage.

[0077] The next step is to determine the specific sector where the optimal switch virtual vector is located based on the active power reference value. When the optimal switch virtual vector is applied at time k, the active power and reactive power at time k+1 are the active power reference value and reactive power reference value, so the following expressions are available:

[0078]

[0079] Among them, u αref 、u βref is the value of the reference voltage vector in the α and β components, P ref , Q ref It is the reference value of active and reactive power.

[0080] The expression for calculating the center vector of each sector is as follows:

[0081] u ic =(u i1 +u i2 +u i3 ) / 3 (Formula 10)

[0082] Among them, u ic is the central voltage vector of the ith sector.

[0083] Substitute the central voltage vector of each sector into the corresponding predicted value, as follows:

[0084]

[0085] Among them, P i (k+1), Q i (k+1) is the predicted active and reactive power of the i-th center vector, P′ i (k), Q′ i (k) is the power change rate of active and reactive power of the ith center vector.

[0086] Then substitute the active power reference value, reactive power reference value and the active and reactive power prediction values ​​of the i-th central voltage vector into the following cost function:

[0087] J=[P ref -P i (k+1)] 2 +[Q ref -Q i (k+1)] 2 (Formula 12)

[0088] It turns out that:

[0089]

[0090] Among them, u iα 、u iβ is the voltage vector of the center vector of the i-th sector under the α and β axes.

[0091] The three-level space vector diagram can be divided into 6 large sectors, and each large sector can be divided into 4 small sectors. It can be seen from the above formula that the closer the center vector is to the optimal virtual switch vector, the smaller the J value is. The large sector where the center vector of the large sector with the smallest J value is located is the large sector where the optimal virtual switch vector is located. After determining the large sector, the J values ​​of the center vectors of the four small sectors in this large sector are calculated respectively. Similarly, the sector where the center vector of the small sector with the smallest J value is located is the small sector where the optimal virtual switch vector is located.

[0092] Next, after determining the specific small sector, it is necessary to determine the action sequence and corresponding action time of the three vectors of the small sector, so as to achieve midpoint voltage balance.

[0093] There are two types of vectors that have the same effect on rectifier power, one is called P-type vector and the other is called N-type vector, but their effects on midpoint voltage balance are completely opposite. Figure 2 As shown, assuming that the value of ia is positive, when POO is used, the upper capacitor is charged and the lower capacitor is discharged, while when ONN is used, the result is the opposite.

[0094] Each phase of the rectifier has three switching states, namely P, O, and N states. The corresponding switching functions are as follows:

[0095]

[0096] Among them, Sx is the switching state of each phase, and Sx1, Sx2, Sx3, and Sx4 are the switching states of four MOSFET tubes in each phase.

[0097] In the three-phase equilibrium state, Figure 2 As shown, the current flowing into point O is as follows:

[0098] i o =-(|S a |i a +|S b |i b +|S c |i c ) (Formula 15)

[0099] Among them, Sa, Sb, and Sc are the switching states of phases A, B, and C respectively.

[0100] At point O, according to Kirchhoff's current law, there is the following expression:

[0101]

[0102] Among them, U1 and U2 are the voltages of the upper and lower capacitors, and C is the capacitance value.

[0103] Combining the above two equations, the rate of change of the voltage difference between the upper and lower capacitors can be obtained as follows:

[0104]

[0105] Among them, ΔU=U1-U2 is the difference between the voltages of the upper and lower capacitors.

[0106] Discretizing the above formula yields:

[0107]

[0108] Wherein, ΔU(k+1) and ΔU(k) ​​are the differences between the upper and lower capacitor voltages on the DC side at time k+1 and k respectively.

[0109] According to the above formula, the cost function of setting the midpoint voltage balance is as follows:

[0110]

[0111] Among them, S j =[S aj ,S bj ,S cj ] T is the state of each phase switch of each vector, S aj , Sbj , S cj are the states of the switches of the three phases a, b, and c of the jth vector, i = [i a ,i b ,i c ] T is the grid side line current, i a 、i b 、i c are the line currents of phases a, b, and c, respectively, t j is the action time of each switching vector, C is the capacitance value of the DC side, ΔU(k) ​​is the difference between the upper and lower capacitor voltages of the DC side at time k, and ΔU(k)=U1-U2.

[0112] After determining the small sector where the optimal virtual switch vector is located, three switch vectors can be selected in the sector to form two types of switch vector sequences, as shown in Table 2, which is the switch vector sequence of the first large sector.

[0113] Table 2. Switching vector sequence table of the first large sector

[0114]

[0115]

[0116] Therefore, the Jc values ​​corresponding to the P-type switch vector sequence and the N-type switch vector sequence of the small sector where the optimal virtual switch vector is located are calculated respectively, so that the switch vector sequence type with the smallest JC value can achieve a better midpoint voltage balance effect, thereby determining the action order of the three switch vectors. According to the expression of Jc, it is known that before determining the action order of the three switch vectors, their action time needs to be calculated first.

[0117] Taking into account the calculation delay, the power at time k+1 is predicted at time k-1 to compensate for the delay. The specific action time calculation formula is as follows:

[0118]

[0119] Among them, P(k+1) and Q(k+1) are the predicted values ​​of active power and reactive power under the action of three vectors, P(k) and Q(k) are the actual values ​​of active power and reactive power at time k, respectively. ipre , Q′ ipre are the power change rates of the i-th vector of active power and reactive power at time (k-1), t ipre is the action time of the i-th vector at time (k-1).

[0120] Substitute the predicted values ​​of active and reactive power and the reference values ​​of active and reactive power into the following cost function:

[0121] G=[P ref -P(k+1)] 2 +[Q ref -Q(k+1)] 2 (Formula 21)

[0122] Because the predicted value will completely track the reference value, the derivative of the above formula with respect to t1 and t2 is equal to 0, as follows:

[0123]

[0124] Through the above calculation, the expression of action time can be obtained as follows:

[0125]

[0126] Among them, t1, t2, and t3 are the action times of the three vectors, Q′ 1pre , Q′ 2pre , Q′ 3pre are the rates of change of reactive power of the three vectors at time k-1, P′ 1pre , P′ 2pre , P′ 3pre are the rates of change of the active power of the three vectors at time k-1, ΔP = P ref -P(k), ΔQ=Q ref -Q(k).

[0127] As shown in the above formula, the action time of the two types of vector sequences is calculated and then substituted into the Jc expression, and the vector sequence that minimizes Jc is selected. In this way, the action order and action time of the three vectors are determined, and the balance of the midpoint voltage is well achieved.

[0128] Finally, according to the action sequence and time of the three optimal switching vectors, each power switch tube is controlled by DSP to achieve the preset effect.

[0129] exist Figure 1 In this paper, the control block diagram of the three-vector model predictive power control of a T-type three-level rectifier based on sliding mode control is given. The control algorithm is summarized as follows: First, the active power reference value P is generated by the sliding mode controller. refSecondly, the center vectors of the six large sectors are substituted into the objective function J to obtain the center vector that minimizes the objective function J and the optimal large sector. Then, the optimal small sector is obtained in the large sector using the same method. Then, the switching vector sequence in the optimal small sector is divided into two types, and their action times are calculated respectively. Then, they are substituted into the objective function Jc to obtain the optimal switching vector sequence and action time that minimizes the objective function Jc. Finally, the power switching devices of the rectifier are controlled according to the switching states corresponding to the optimal switching vector sequence and action time, so that the active power and reactive power track the active power reference value and the reactive power reference value, and the DC side midpoint potential can be well balanced. When the load suddenly changes, the system will also have a fast dynamic response, thereby quickly entering a new steady state.

[0130] The simulation software MATLAB / Simulink is used to build a three-vector model predictive power control system for T-type three-level rectifier based on sliding mode control for simulation verification, where the parameters are shown in Table 3.

[0131] Table 3. Simulation parameters

[0132] parameter value Filter inductor L 6mH DC link capacitor C 1000uF Grid phase voltage peak 60V Output voltage setpoint 120V Grid frequency f 50Hz Sampling frequency fs 10kHz <![CDATA[Sliding mode control parameter k1]]> 0.003 <![CDATA[Sliding mode control parameter k2]]> 0.03

[0133] exist Figure 4 In the paper, the grid-side current waveform of the three-vector model predictive control method of T-type rectifier based on sliding mode control is given; Figure 5 In , the output active power and reactive power of the three-vector model predictive control method of T-type rectifier based on sliding mode control are given; Figure 6 In this paper, the transient response of the output voltage of a T-type rectifier three-vector model predictive control method based on sliding mode control is given when the load suddenly changes. Figure 7 In the paper, the DC side capacitor voltage of the three-vector model predictive control method of T-type rectifier based on sliding mode control is given. From the simulation results above, it can be seen that the three-vector model predictive control method of T-type rectifier based on sliding mode control uses three vectors for control and does not require weight factors. It not only reduces the calculation cycle, but also improves the midpoint voltage balance ability a lot. At the same time, it can fully track the active power reference value and reactive power reference value, reduce the harmonic content and current ripple in the grid side current, and improve the power quality. At the same time, it has a fast dynamic response when the load changes suddenly, and can rise to the output voltage set value again in a very short time.

[0134] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A three-vector model predictive control method for a T-type rectifier based on sliding mode control, characterized in that: The predictive control method comprises the following steps: S1. Collect the current I on the DC side at a certain time interval dc , the two capacitor voltages U1 and U2 on the DC side, and the AC side voltage e abc and current i abc The active power reference value P is calculated through sliding mode control ref , and at the same time, the model prediction method is used to predict the active and reactive power corresponding to the central vector of each sector at the next moment; S2. Cost function based on the sector center vector Find the small sector where the optimal virtual switch vector is located; S3. Cost function based on balanced midpoint voltage Determine the action sequence and corresponding action time of the three switch vectors; S4, controlling each switch device of the T-type three-level rectifier according to the found optimal switch state; Wherein, in the step S1, the active power reference value is calculated by sliding mode control, wherein the sliding mode surface expression is as follows: Where s is the sliding surface, U ref is the given value of the output voltage, U dc is the value of the output voltage, k1 and k2 are the first and second sliding mode coefficients for realizing the function of the output voltage following the given value of the output voltage; The active power reference value expression is as follows: Among them, P ref is the active power reference value, C is the capacitance value on the DC side; The process of using the model prediction method to predict the active and reactive power corresponding to the central vector of each sector at the next moment in step S1 is as follows: The expressions for active power prediction and reactive power prediction are as follows: In the formula, , are the predicted values ​​of active power and reactive power of the center vector of the i-th sector, respectively. P(k) and Q(k) are the actual values ​​of active power and reactive power at time k, respectively. , are the power change rates of active power and reactive power of the center vector of the i-th sector, T s is the sampling time; The cost function of the sector center vector The expression is as follows: In the formula, , is the component of the i-th sector center vector u on the α and β axes, , is the voltage reference value U ref The components on the α and β axes, , is the component of the grid voltage on the α and β axes, is the value of the grid-side filter inductance, is the sampling time; The calculation expression of the action time corresponding to the switching vector is as follows: In the formula, t1, t2, and t3 are the action times of the three vectors respectively. , , are the rates of change of reactive power of the three vectors at time k-1, , , are the rates of change of the active power of the three vectors at time k-1, , , P(k) and Q(k) are the actual values ​​of active power and reactive power at time k, respectively. is the sampling time, P ref , Q ref are the reference values ​​of active power and reactive power respectively; The cost function of the balanced midpoint voltage is The expression is as follows: In the formula, For each vector, the state of each phase switch tube, , , are the states of the switches of the three phases a, b, and c of the jth vector, is the grid side line current, , , are the line currents of phases a, b, and c, respectively, t j is the action time of each switching vector, C is the capacitance value on the DC side, is the difference between the upper and lower capacitor voltages on the DC side at time k, .

2. The three-vector model predictive control method for T-type rectifier based on sliding mode control according to claim 1 is characterized in that: In step S3, by comparing the values ​​of cost functions of different switching sequences, the switching sequence that minimizes the cost function is selected, thereby determining the switching sequence and its action time.