Three-level rectifier model predictive power control method based on dead-beat control

Abstract

The invention discloses a deadbeat control-based low-complexity three-level rectifier model predictive power control method, and provides improved model predictive control for adjusting instantaneous active power and reactive power and balancing neutral-point voltage for a three-phase three-level rectifier. And meanwhile, dead-beat control is adopted to replace traditional PI control to adjust the output voltage. According to the model predictive power control method, an active power reference value is calculated through dead-beat control, then a sector where an optimal virtual switching vector is located is found, then an optimal switching vector is found in the sector, and finally, each switching device of the three-level rectifier is controlled according to the found optimal switching state. Compared with traditional model prediction power control combined with PI, the method has the advantages that the calculation amount of an algorithm can be reduced, the dynamic response speed of the three-phase three-level rectifier is greatly increased, and the performance of the three-phase three-level rectifier is good.

Description

Technical Field

[0001] The present invention relates to the technical field of three-level PWM rectification, in particular to the technical field of model prediction power control combined with a deadbeat controller, and specifically to a model prediction power control method for a three-phase three-level rectifier based on deadbeat control. Background Art

[0002] With the popularization of new energy, a large number of power electronic devices are constantly pouring into people's lives. On the user side, electric vehicles occupy an increasingly larger share in the market, so the demand for large charging piles is also growing; on the grid side, the proportion of new energy sources such as photovoltaic power generation, wind power generation, and tidal power generation is increasing, and the requirements for high-power energy storage devices are also getting higher and higher. Among them, the rectifier is an important link. Compared with the traditional two-level rectifier, the three-level rectifier has low input current harmonic content, less impact on the grid, and more stable output voltage.

[0003] Model predictive control is a new type of nonlinear predictive control strategy that uses a dynamic model to predict the future behavior of an unknown system and optimizes the system based on the predictive model. Model predictive control mainly consists of three parts: prediction model, rolling optimization, and feedback correction. For the model predictive control of traditional three-level rectifiers, the amount of calculation is large, requiring 27 calculation cycles to obtain the optimal switching vector, consuming a lot of computing resources and increasing hardware costs. At present, the model predictive power control applied to rectifiers basically uses PI controllers to generate reference power, which has a slow dynamic response speed and a long transition time, making it difficult to adapt to scenarios with large load changes. Summary of the invention

[0004] The purpose of the present invention is to solve the above defects in the prior art and provide a low-complexity three-level rectifier model prediction power control method based on deadbeat control

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

[0006] A low-complexity three-level rectifier model predictive power control method based on deadbeat control comprises the following steps:

[0007] S1. Write a quadratic equation about the output voltage according to the mathematical model of the DC side of the three-phase three-level rectifier, and then use deadbeat control to obtain a preliminary active power reference value, and add a correction function on the basis of the preliminary active power reference value to obtain a final active power reference value;

[0008] S2. Design a first objective function for finding the best sector. The best sector refers to the sector where the best virtual vector is located. The best virtual switch vector refers to the virtual vector that can achieve the best control effect in the three-level space voltage vector diagram. Substitute the center vector of each sector into the first objective function, and take the sector corresponding to the center vector that minimizes the value of the first objective function as the best sector.

[0009] S3, design a second objective function for finding the optimal switching vector, substitute all the switching vectors contained in the optimal sector obtained in step S2 into the second objective function, and take the switching vector that minimizes the value of the second objective function as the optimal switching vector of the three-phase three-level rectifier;

[0010] S4. Outputting the switching state corresponding to the optimal switching vector of the three-phase three-level rectifier obtained in step S3 to the power switch device through the microcontroller, thereby controlling the power switch device.

[0011] Further, the three-phase three-level rectifier includes 27 switching vectors, including 3 zero vectors, namely PPP, OOO and NNN, which belong to the common part of all sectors and therefore exist in all sectors; the remaining 24 switching vectors are 6 long vectors, 6 medium vectors and 12 short vectors; the switching vectors of large sector I include: OOO, PNN, PPN; the switching vectors of large sector II include: OOO, PPN, NPN; the switching vectors of large sector III include: OOO, NPN, NPP; the switching vectors of large sector IV include: OOO, NPP, NNP; the switching vectors of large sector V include: OOO, NNP, PNP; the switching vectors of large sector VI include: OOO, PNP, PNN; the switching vectors of large sector I include: The switching vectors of small sectors ① include: OOO, POO, ONN, PPO, OON; the switching vectors of small sectors ② in large sector I include: POO, ONN, PNN, PON; the switching vectors of small sectors ③ in large sector I include: POO, ONN, PON, PPO, OON; the switching vectors of small sectors ④ in large sector I include: PON, PPN, PPO, OON; the switching vectors of small sectors ① in large sector II include: OOO, PPO, OON, OPO, NON; the switching vectors of small sectors ② in large sector II include: PPO, OON, PPN, OPN; the switching vectors of small sectors ③ in large sector II include: PPO, OON, OPN, OPO, NON; large sector The switching vectors of small sectors ④ in large sector II include: OPO, NON, OPN, NPN; the switching vectors of small sectors ① in large sector III include: OOO, OPO, NON, OPP, NOO; the switching vectors of small sectors ② in large sector III include: OPO, NON, NPN, NPO; the switching vectors of small sectors ③ in large sector III include: OPO, NON, NPO, OPP, NOO; the switching vectors of small sectors ④ in large sector III include: NPO, NPP, OPP, NOO; the switching vectors of small sectors ① in large sector IV include: OOO, OPP, NOO, OOP, NNO; the switching vectors of small sectors ② in large sector IV include: OPP, NOO, NPP, NOP ; The switching vectors of the small sector ③ in the large sector IV include: OPP, NOO, NOP, OOP, NNO; the switching vectors of the small sector ④ in the large sector IV include: OOP, NNO, NOP, NNP; the switching vectors of the small sector ① in the large sector V include: OOO, OOP, NNO, POP, ONO; the switching vectors of the small sector ② in the large sector V include: OOP, NNO, NNP, ONP; the switching vectors of the small sector ③ in the large sector V include: OOP, NNO, ONP, POP, ONO; the switching vectors of the small sector ④ in the large sector V include: POP, ONO, ONP, PNP; the switching vectors of the small sector ① in the large sector VI include: OOO, POO, ONN, POP, ONO;The switching vectors of small sector ② in large sector VI include: POP, ONO, PNP, PNO; the switching vectors of small sector ③ in large sector VI include: POP, ONO, PNO, POO, ONN; the switching vectors of small sector ④ in large sector VI include: POO, ONN, PNN, PNO. By distinguishing sectors in this way, the optimal vector selection range can be narrowed and the calculation amount of the entire control process operation can be reduced. ;

[0012] Furthermore, in step S1, the expressions of the quadratic equation and the active power reference value obtained are as follows:

[0013] Among them, P ref (k) represents the reference power at the kth moment, T s is the sampling time, U dc (t) is the continuous expression of the output voltage, U ref is the reference value of the output voltage, U dc (k) is the value of the output voltage at the kth moment, U dc (k+1) is the value of the output voltage at the k+1th moment, I dc (k) is the value of the output current at the kth moment, C is the capacitance value on the DC side, R is the load resistance, y c is the correction function, y c (k-1) is the correction function value at the k-1th moment, λ is the positive parameter value to be set, and dt represents the correction function value for (U ref -U dc (t)) is integrated.

[0014] Furthermore, in step S2, the formula for finding the sector where the optimal virtual switch vector is located is as follows:

[0015] Among them, J represents the first objective function, Q ref is the reactive power reference value, P ref is the reference value of active power, T s represents the sampling time, P i (k) and P i (k+1) is the active power when the i-th switch vector is used at the k-th moment and the k+1-th moment, Q i (k) and Q i (k+1) are the reactive powers when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. P(k) and Q(k) are the active power and reactive power of the three-level rectifier at the k-th moment. i ′ and Q i′ are the derivatives of active power and reactive power when the i-th switching vector is used, sector(h) indicates that the h-th sector is the best sector, and argmin(J) indicates the minimum value in the set of all objective function values. Through this sector selection method, the largest and optimal sector can be selected to narrow the selection range of the vector.

[0016] In the stationary two-phase αβ coordinate system, the expressions of the active power and reactive power derivatives of the three-level rectifier are as follows:

[0017] Among them, e α 、e β 、i α 、i β represents the α component and β component of the grid phase voltage and the α component and β component of the line current in the stationary two-phase αβ coordinate system, L is the value of the grid side filter inductance, ω represents the angular velocity, P′ represents the derivative of the active power of the three-level rectifier, Q′ represents the derivative of the reactive power of the three-level rectifier; when the switching state of the rectifier is v i When the output voltage is u in the stationary two-phase αβ coordinate system, i (u αi ,u βi ), the above formula is expressed as follows:

[0018]

[0019] u α and u β represents the α and β components of the switch vector in the αβ coordinate system, u αi and u βi represents the α component and β component of the i-th switch vector in the αβ coordinate system;

[0020] At time k+1, the active power and reactive power of the three-level rectifier are written as:

[0021]

[0022] The three-level space vector diagram is divided into 6 large sectors and 4 small sectors. The switching vector in the middle of each sector is taken, which is called the central vector of the sector. The active power P of the three-level rectifier at time k+1 is calculated using the central vector of the large sector. i (k+1) and reactive power Q i (k+1), when finding the sector where the optimal virtual switch vector is located, the midpoint voltage balance problem is not considered first, and the objective function is defined as:

[0023] J=[P ref -P i (k+1)] 2+[Q ref -Q i (k+1)] 2

[0024] Find the central vector corresponding to the minimum value of the first objective function J. The large sector where the central vector is located is the large sector where the optimal virtual switch vector is located. Then use the same method to obtain the small sector where the optimal virtual switch vector is located in the large sector. This small sector is the sector where the optimal switch vector is located.

[0025] Furthermore, the process of step S3 is as follows:

[0026] After finding the small sector where the optimal switching vector is located, substitute the switching vector in the sector into the second objective function G, so that the switching vector with the smallest second objective function G is the optimal switching vector. The formula for finding the optimal switching vector is as follows:

[0027]

[0028] Among them, G is the second objective function for finding the optimal switching vector, u o (k) and u o (k+1) is the midpoint voltage of the output terminal of the three-level rectifier at the kth moment and the k+1th moment, and the expressions are u o (k) = u c1 (k)-u c2 (k),u o (k+1)=u c1 (k+1)-u c2 (k+1),u c1 (k) and u c2 (k) are the voltage values ​​of the upper and lower capacitors at the output end at time k, u c1 (k+1) and u c2 (k+1) are the voltage values ​​of the upper and lower capacitors at the output end at time k+1, C is the capacitance value on the DC side, S abc is the switching state of each phase, and |S abc |={|S a |,|S b |,|S c |}, S a , S b , S c The switching states i of phase A, phase B, and phase C are abc is the line current on the grid side, which is expressed as i abc ={i a ,i b ,i c}, i a 、i band i c are the line currents of phase A, phase B and phase C respectively, λ m is the weight factor of the midpoint potential, vector(m) indicates that the mth switch vector is the optimal switch vector, and argmin(G) indicates the minimum value in the set of all objective function values. By using this expression to select the optimal small sector from the previously found optimal large sector, the selection range of the optimal vector can be further narrowed, thereby achieving the purpose of reducing the amount of calculation.

[0029] Furthermore, according to multiple simulation experiments, we select m The value of is 0.25-0.35. Under this value, the midpoint potential can be well balanced, and the impact on the output power is also small.

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

[0031] (1) Based on the traditional model predictive power control, the present invention adopts a control method of first determining the optimal sector and then determining the optimal switching vector. Compared with the traditional three-level rectifier model predictive control, this control method does not need to calculate all 27 switching vectors, which greatly reduces the calculation amount of the control algorithm and saves the computing resources of the control chip.

[0032] (2) The present invention adds a weight factor of the midpoint voltage to the objective function. This method can not only effectively balance the midpoint voltage, but also reduce the harmonic content of the grid-side current, so that the output active power and reactive power of the three-phase three-level rectifier can track the given reference values ​​of the active power and reactive power.

[0033] (3) The present invention uses deadbeat control to generate a reference value of active power. Compared with the traditional PI controller, the deadbeat control has a fast response speed. After adding the correction function, it not only greatly improves the dynamic response speed of the three-phase three-level rectifier, but also eliminates the steady-state error caused by the oscillation of the midpoint potential and the error of the system model parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] 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:

[0035] Figure 1 is a circuit topology diagram of a T-type three-phase three-level rectifier in an embodiment of the present invention;

[0036] Figure 2 is a three-level space voltage vector diagram in an embodiment of the present invention;

[0037] Figure 3 is a three-level space vector sector diagram in an embodiment of the present invention;

[0038] Figure 4 It is a control block diagram of an improved model predictive power control of a low-complexity three-level rectifier based on deadbeat control in an embodiment of the present invention;

[0039] Figure 5 It is a grid-side current waveform diagram of the improved model predictive power control of the low-complexity three-level rectifier based on deadbeat control in Example 1 of the present invention;

[0040] Figure 6 Schematic diagram of output active power and reactive power of improved model predictive power control of a low-complexity three-level rectifier based on deadbeat control in Embodiment 1 of the present invention;

[0041] Figure 7 It is a schematic diagram of transient response of output voltage of the improved model prediction power control of the low-complexity three-level rectifier based on deadbeat control in embodiment 1 of the present invention when the load changes suddenly;

[0042] Figure 8 It is a schematic diagram of the DC side capacitor voltage of the improved model predictive power control of the low-complexity three-level rectifier based on deadbeat control in Example 1 of the present invention.

[0043] Fig. 9 It is a grid-side current, voltage and output voltage waveform diagram of the improved model predictive power control of the low-complexity three-level rectifier based on deadbeat control in Example 2 of the present invention;

[0044] Fig.10 Schematic diagram of output active power and reactive power of improved model predictive power control of low-complexity three-level rectifier based on deadbeat control in Embodiment 2 of the present invention;

[0045] Fig.11 It is a schematic diagram of the DC side capacitor voltage of the improved model predictive power control of the low-complexity three-level rectifier based on deadbeat control in Example 2 of the present invention. DETAILED DESCRIPTION

[0046] 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.

[0047] Example 1

[0048] like Figure 1 As shown in FIG. 1 , a topological structure diagram of a T-type three-phase three-level rectifier is given. The three-phase three-level rectifier is directly connected to the power grid through a filter inductor L. Figure 2 As shown, the space vector diagram of 27 switching vectors of the three-level converter is given, and each switching vector corresponds to the voltage vector of the output of the rectifier.

[0049] When the switching vector of the rectifier is v i When the output voltage vector of the three-level rectifier is u i (u αi ,u βi ), the output voltage vector of the three-level rectifier can be obtained as follows:

[0050]

[0051] In formula (1), u i is the output voltage when the i-th switch vector acts; L is the AC filter inductance; e is the grid voltage; i represents the grid line voltage; after Clark transformation, the equation can be transformed from the three-phase coordinate system to the two-phase stationary αβ coordinate system:

[0052]

[0053] Among them, e α 、e β 、i α 、i β Represents the α component and β component of the grid phase voltage and the α component and β component of the line current in the static two-phase αβ coordinate system.

[0054] Taking the three-phase grid voltage as an example, the Clark equal amplitude transformation formula is:

[0055]

[0056] Among them, e a 、e b and e c is the grid voltage of phase A, phase B and phase C, e α and e β There are the following relations:

[0057]

[0058] Where ω is the grid voltage angular frequency.

[0059] The formula describing instantaneous active power and reactive power in the αβ coordinate system is:

[0060]

[0061] Among them, P and Q represent the active power and reactive power of the three-level rectifier respectively. Combining the above formulas, it can be obtained that the derivatives of the active power and reactive power of the three-level rectifier in the αβ coordinate system are:

[0062]

[0063] Among them, P' and Q' represent the derivatives of the active power and reactive power of the three-level rectifier. The above formula can also be expressed as:

[0064]

[0065] Where P i ′ and Q i ′ represents the derivative of the active power and reactive power of the three-level rectifier when the i-th switching vector acts; u α and u β represents the α component and β component of the switch vector in the αβ coordinate system; u αi and u βi represents the α component and β component of the output voltage of the i-th switching vector in the αβ coordinate system;

[0066] At time k+1, the active power and reactive power can be written as:

[0067]

[0068] Among them, T s is the sampling period, the active power P(k) and the reactive power Q(k) are the active power and reactive power of the three-level rectifier at time k, and the active power P(k+1) and the reactive power Q(k+1) are the predicted values ​​of the active power and reactive power of the three-level rectifier at time k+1.

[0069] exist Figure 3 In the scheme, the three-level space vector diagram is divided into sectors, and the switching vectors of each sector are 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 improved model predictive control, only 14 calculation cycles are required.

[0070] Table 1. Sector switching vector table

[0071]

[0072] The first step of the model predictive power control of a low-complexity three-level rectifier based on deadbeat control is to generate a reference active power. According to the deadbeat control, the expression of the reference power can be obtained as follows:

[0073]

[0074] Among them, P ref (k) represents the reference power at the kth moment, T s is the sampling time, U ref (t) is the given value of the output voltage, U dc (k) is the value of the output voltage at the kth moment, U dc (k+1) is the value of the output voltage at the k+1th moment, I dc (k) is the value of the output current at the kth moment, C is the capacitance value on the DC side, R is the load resistance, y c is the correction function, y c (k-1) is the correction function value at the k-1th moment, λ is the positive parameter value to be set, and dt represents the correction function value for (U ref -U ref (t)) is integrated.

[0075] The next step is to determine the large sector where the optimal virtual switch vector is located based on the reference power generated by equation (9). The objective function for determining the sector is:

[0076] Among them, J represents the first objective function, Q ref is the reactive power reference value of the three-level rectifier, T s represents the sampling time, P i (k) and P i (k+1) are the active powers of the three-level rectifier when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. i (k) and Q i (k+1) is the reactive power of the three-level rectifier when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. P(k) and Q(k) are the active power and reactive power of the three-level rectifier at the k-th moment, respectively. P i ′ and Q i ′ are the derivatives of the active power and reactive power of the three-level rectifier when the i-th switching vector is used, sector(h) indicates that the h-th sector is the optimal sector, and argmin(J) indicates the minimum value in the set of all objective function values.

[0077] Find the central vector corresponding to the minimum value of the first objective function J. The large sector where the central vector is located is the large sector where the optimal virtual switch vector is located. Then use the same method to obtain the small sector where the optimal virtual switch vector is located in the large sector. This small sector is the sector where the optimal switch vector is located.

[0078] After finding the small sector where the optimal switching vector is located, substitute the switching vector in the sector into the second objective function G, so that the switching vector with the smallest second objective function G is the optimal switching vector; the formula for finding the optimal switching vector is as follows:

[0079]

[0080] Among them, G is the second objective function for finding the optimal switching vector, u o (k) and u o (k+1) is the midpoint voltage of the output terminal of the three-level rectifier at the kth moment and the k+1th moment, and the expressions are u o (k) = u c1 (k)-u c2 (k),u o (k+1)=u c1 (k+1)-u c2 (k+1),u c1 (k) and u c2 (k) are the voltage values ​​of the upper and lower capacitors at the output end at time k, respectively, u c1 (k+1) and u c2 (k+1) are the voltage values ​​of the upper and lower capacitors at the output end at time k+1, C is the capacitance value on the DC side, S abc is the switching state of each phase, and |S abc |={|S a |,|S b |,|S c |}, S a , S b , S c The switching states i of phase A, phase B, and phase C are abc is the line current on the grid side, which is expressed as i abc ={i a ,i b ,i c}, i a 、i b and i c are the line currents of phase A, phase B and phase C respectively, λ m is the weight factor of the midpoint potential, vector(m) indicates that the mth switching vector is the optimal switching vector, and argmin(G) indicates the minimum value in the set of all objective function values.

[0081] After the optimal switching vector is obtained, the microprocessor controls the power switching device according to the switching state corresponding to the optimal switching vector.

[0082] exist Figure 4In this paper, a control block diagram of model predictive power control of a low-complexity three-level rectifier based on deadbeat control is given. The control algorithm is summarized as follows: First, the reference active power P is generated by the deadbeat controller. ref Secondly, substitute the center vectors of the six large sectors into the objective function J to obtain the center vector that minimizes the objective function J and obtain its optimal large sector. Then, use the same method in the large sector to obtain the optimal small sector. Next, substitute the switch vector in the optimal small sector into the objective function G to obtain the optimal switch vector that minimizes the objective function G. Finally, control the power switch device of the rectifier according to the switch state corresponding to the optimal switch vector, so that its active power and reactive power track the reference power and can balance the midpoint potential of the DC side. When the load changes suddenly, the system will quickly enter a new steady state.

[0083] The simulation software MATLAB / Simulink is used to build a model predictive power control system of a T-type three-phase three-level rectifier based on deadbeat control for simulation verification, where the parameters are shown in Table 2.

[0084] Table 2. Simulation parameters

[0085] parameter value Filter inductor L 6mH DC link capacitor C 1000uF Grid phase voltage peak 100V <![CDATA[Output voltage set value V ref > 400V <![CDATA[Active power setpoint P ref > 2kW <![CDATA[Reactive power setpoint Q ref > 0Var Grid frequency f 50Hz Sampling frequency fs 10kHz <![CDATA[Midpoint potential weighting factor λ m > 0.25 Deadbeat control parameter λ 0.15

[0086] exist Figure 5 In the paper, the grid-side current waveform of the model predictive power control of the T-type three-phase three-level rectifier based on the deadbeat control is given. It can be seen from the figure that the waveform quality of the grid-side current is relatively good and the degree of sinusoidalization is high; Figure 6 In the figure, the output active power and reactive power of the model predictive power control of the T-type three-phase three-level rectifier based on the deadbeat control are given. It can be seen that after using the model predictive power control of the T-type three-phase three-level rectifier, the output active power and reactive power can track the given value of the reference power well; Figure 7 In the paper, a model prediction of the output voltage transient response of the power control based on deadbeat control of a T-type three-phase three-level rectifier is given when the load suddenly changes. It can be seen from the figure that when the load suddenly changes, the output voltage changes little and can rise to the given output voltage value in a very short time. Figure 8In the paper, the DC side capacitor voltage of the model predictive power control of the T-type three-phase three-level rectifier based on the beatless control is given. As can be seen from the figure, the midpoint voltage is balanced, and the fluctuation does not exceed 3V. From the above simulation results, it can be seen that the model predictive power control of the T-type three-phase three-level rectifier based on the beatless control not only reduces the calculation cycle, but also can better track the given reference active power and reactive power, reduce the harmonic content in the grid-side current, balance the midpoint voltage, and at the same time, the output voltage changes less when the load changes suddenly, and can rise to the given output voltage value again in a very short time.

[0087] Example 2

[0088] like Figure 1 As shown in FIG. 1 , a topological structure diagram of a T-type three-phase three-level rectifier is given. The three-phase three-level rectifier is directly connected to the power grid through a filter inductor L. Figure 2 As shown, the space vector diagram of 27 switching vectors of the three-level converter is given, and each switching vector corresponds to the voltage vector of the output of the rectifier.

[0089] When the switching vector of the rectifier is v i When the output voltage vector of the three-level rectifier is u i (u αi ,u βi ), the output voltage vector of the three-level rectifier can be obtained as follows:

[0090]

[0091] In formula (12), u i is the output voltage when the i-th switch vector acts; L is the AC filter inductance; e is the grid voltage; i represents the grid line voltage; after Clark transformation, the equation can be transformed from the three-phase coordinate system to the two-phase stationary αβ coordinate system:

[0092]

[0093] Among them, e α 、e β 、i α 、i β Represents the α component and β component of the grid phase voltage and the α component and β component of the line current in the static two-phase αβ coordinate system.

[0094] Taking the three-phase grid voltage as an example, the Clark equal amplitude transformation formula is:

[0095]

[0096] Among them, e a 、e b and ec is the grid voltage of phase A, phase B and phase C, e α and e β There are the following relations:

[0097]

[0098] Where ω is the grid voltage angular frequency.

[0099] The formula describing instantaneous active power and reactive power in the αβ coordinate system is:

[0100]

[0101] Among them, P and Q represent the active power and reactive power of the three-level rectifier respectively. Combining the above formulas, it can be obtained that the derivatives of the active power and reactive power of the three-level rectifier in the αβ coordinate system are:

[0102]

[0103] Among them, P' and Q' represent the derivatives of the active power and reactive power of the three-level rectifier. The above formula can also be expressed as:

[0104]

[0105] Where P i ′ and Q i ′ represents the derivative of the active power and reactive power of the three-level rectifier when the i-th switching vector acts; u α and u β represents the α component and β component of the switch vector in the αβ coordinate system; u αi and u βi represents the α component and β component of the output voltage of the i-th switching vector in the αβ coordinate system;

[0106] At time k+1, the active power and reactive power can be written as:

[0107]

[0108] Among them, T s is the sampling period, the active power P(k) and the reactive power Q(k) are the active power and reactive power of the three-level rectifier at time k, and the active power P(k+1) and the reactive power Q(k+1) are the predicted values ​​of the active power and reactive power of the three-level rectifier at time k+1.

[0109] exist Figure 3In the scheme, the three-level space vector diagram is divided into sectors, and the switching vectors of each sector are 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 improved model predictive control, only 14 calculation cycles are required.

[0110] Table 3. Switching vector table of sectors

[0111]

[0112] The first step of the model predictive power control of a low-complexity three-level rectifier based on deadbeat control is to generate a reference active power. According to the deadbeat control, the expression of the reference power can be obtained as follows:

[0113]

[0114] Among them, P ref (k) represents the reference power at the kth moment, T s is the sampling time, U ref (t) is the given value of the output voltage, U dc (k) is the value of the output voltage at the kth moment, U dc (k+1) is the value of the output voltage at the k+1th moment, I dc (k) is the value of the output current at the kth moment, C is the capacitance value on the DC side, R is the load resistance, y c is the correction function, y c (k-1) is the correction function value at the k-1th moment, λ is the positive parameter value to be set, and dt represents the correction function value for (U ref -U ref (t)) is integrated.

[0115] The next step is to determine the large sector where the optimal virtual switch vector is located based on the reference power generated by equation (9). The objective function for determining the sector is:

[0116] Among them, J represents the first objective function, Q ref is the reactive power reference value of the three-level rectifier, T s represents the sampling time, P i (k) and P i (k+1) are the active powers of the three-level rectifier when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. i (k) and Q i (k+1) is the reactive power of the three-level rectifier when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. P(k) and Q(k) are the active power and reactive power of the three-level rectifier at the k-th moment, respectively. Pi ′ and Q i ′ are the derivatives of the active power and reactive power of the three-level rectifier when the i-th switching vector is used, sector(h) indicates that the h-th sector is the optimal sector, and argmin(J) indicates the minimum value in the set of all objective function values.

[0117] Find the central vector corresponding to the minimum value of the first objective function J. The large sector where the central vector is located is the large sector where the optimal virtual switch vector is located. Then use the same method to obtain the small sector where the optimal virtual switch vector is located in the large sector. This small sector is the sector where the optimal switch vector is located.

[0118] After finding the small sector where the optimal switching vector is located, substitute the switching vector in the sector into the second objective function G, so that the switching vector with the smallest second objective function G is the optimal switching vector; the formula for finding the optimal switching vector is as follows:

[0119]

[0120] Among them, G is the second objective function for finding the optimal switching vector, u o (k) and u o (k+1) is the midpoint voltage of the output terminal of the three-level rectifier at the kth moment and the k+1th moment, and the expressions are u o (k) = u c1 (k)-u c2 (k),u o (k+1)=u c1 (k+1)-u c2 (k+1),u c1 (k) and u c2 (k) are the voltage values ​​of the upper and lower capacitors at the output end at time k, respectively, u c1 (k+1) and u c2 (k+1) are the voltage values ​​of the upper and lower capacitors at the output end at time k+1, C is the capacitance value on the DC side, S abc is the switching state of each phase, and |S abc |={|S a |,|S b |,|S c |}, S a , S b , S c The switching states i of phase A, phase B, and phase C are abc is the line current on the grid side, which is expressed as i abc ={i a ,i b ,i c}, ia 、i b and i c are the line currents of phase A, phase B and phase C respectively, λ m is the weight factor of the midpoint potential, vector(m) indicates that the mth switching vector is the optimal switching vector, and argmin(G) indicates the minimum value in the set of all objective function values.

[0121] After the optimal switching vector is obtained, the microprocessor controls the power switching device according to the switching state corresponding to the optimal switching vector.

[0122] exist Figure 4 In this paper, a control block diagram of model predictive power control of a low-complexity three-level rectifier based on deadbeat control is given. The control algorithm is summarized as follows: First, the reference active power P is generated by the deadbeat controller. ref Secondly, substitute the center vectors of the six large sectors into the objective function J to obtain the center vector that minimizes the objective function J and obtain its optimal large sector. Then, use the same method in the large sector to obtain the optimal small sector. Next, substitute the switch vector in the optimal small sector into the objective function G to obtain the optimal switch vector that minimizes the objective function G. Finally, control the power switch device of the rectifier according to the switch state corresponding to the optimal switch vector, so that its active power and reactive power track the reference power and can balance the midpoint potential of the DC side. When the load changes suddenly, the system will quickly enter a new steady state.

[0123] The simulation software MATLAB / Simulink is used to build a model predictive power control system of a T-type three-phase three-level rectifier based on deadbeat control for simulation verification, where the parameters are shown in Table 4.

[0124] Table 4. Simulation parameters

[0125] parameter value Filter inductor L 6mH DC link capacitor C 1000uF Grid phase voltage peak 70V <![CDATA[Output voltage set value V ref > 120V Grid frequency f 50Hz Sampling frequency fs 10kHz <![CDATA[Midpoint potential weighting factor λ m > 0.1 Deadbeat control parameter λ 0.2

[0126] exist Fig. 9 In the paper, the grid-side current, voltage and output voltage waveforms of the model predictive power control of the T-type three-phase three-level rectifier based on deadbeat control are given. It can be seen from the figure that the waveform quality of the grid-side current is relatively good, the degree of sinusoidalization is high, and it meets the unity power factor operation. Fig.10 In the paper, the output active power and reactive power of the model predictive power control of the T-type three-phase three-level rectifier based on the deadbeat control are given. It can be seen from the figure that after using the model predictive power control of the T-type three-phase three-level rectifier, the output active power and reactive power can be relatively stable, and the purpose of controlling the output voltage is achieved by controlling the active power; Fig.11In the figure, the DC side capacitor voltage of the model predictive power control of the T-type three-phase three-level rectifier based on the deadbeat control is given. As can be seen from the figure, the midpoint voltage is balanced, and the fluctuation does not exceed 1V. From the above simulation results, it can be seen that the model predictive power control of the T-type three-phase three-level rectifier based on the deadbeat control not only reduces the calculation cycle, but also can better track the given reference output voltage, reduce the harmonic content in the grid-side current, and balance the midpoint voltage.

[0127] 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-level rectifier model prediction power control method based on deadbeat control, characterized in that: The model prediction power control method comprises the following steps: S1. Write a quadratic equation about the output voltage according to the mathematical model of the DC side of the three-phase three-level rectifier, and then use deadbeat control to obtain a preliminary active power reference value, and add a correction function on the basis of the preliminary active power reference value to obtain a final active power reference value; S2. Design a first objective function for finding the best sector. The best sector refers to the sector where the best virtual vector is located. The best virtual switch vector refers to the virtual vector that can achieve the best control effect in the three-level space voltage vector diagram. Substitute the center vector of each sector into the first objective function, and take the sector corresponding to the center vector that minimizes the value of the first objective function as the best sector. S3, design a second objective function for finding the optimal switching vector, substitute all the switching vectors contained in the optimal sector obtained in step S2 into the second objective function, and take the switching vector that minimizes the value of the second objective function as the optimal switching vector of the three-phase three-level rectifier; S4. Outputting the switching state corresponding to the optimal switching vector of the three-phase three-level rectifier obtained in step S3 to the power switch device through the microcontroller to complete the control of the power switch device.

2. The three-level rectifier model prediction power control method based on deadbeat control according to claim 1, characterized in that: The three-phase three-level rectifier is a T-type three-level rectifier, including 27 switching vectors, of which 3 are zero vectors, namely PPP, OOO and NNN, and the remaining 24 switching vectors are respectively 6 long vectors, 6 medium vectors and 12 short vectors; The switching vectors of large sector I include: OOO, PNN, PPN; the switching vectors of large sector II include: OOO, PPN, NPN; the switching vectors of large sector III include: OOO, NPN, NPP; the switching vectors of large sector IV include: OOO, NPP, NNP; the switching vectors of large sector V include: OOO, NNP, PNP; the switching vectors of large sector VI include: OOO, PNP, PNN; the switching vectors of small sector ① within large sector I include: OOO, POO, ONN, PPO, OON; the switching vectors of small sector ② within large sector I include: POO, ONN, PNN, PON; the switching vectors of small sector ③ within large sector I include: POO, ONN, PON, PPO, OON; large sector I The switching vectors of small sector ④ in large sector II include: PON, PPN, PPO, OON; the switching vectors of small sector ① in large sector II include: OOO, PPO, OON, OPO, NON; the switching vectors of small sector ② in large sector II include: PPO, OON, PPN, OPN; the switching vectors of small sector ③ in large sector II include: PPO, OON, OPN, OPO, NON; the switching vectors of small sector ④ in large sector II include: OPO, NON, OPN, NPN; the switching vectors of small sector ① in large sector III include: OOO, OPO, NON, OPP, NOO; the switching vectors of small sector ② in large sector III include: OPO, NON, NPN, NPO; The switching vectors of small sectors ③ in sector III include: OPO, NON, NPO, OPP, NOO; the switching vectors of small sectors ④ in large sector III include: NPO, NPP, OPP, NOO; the switching vectors of small sectors ① in large sector IV include: OOO, OPP, NOO, OOP, NNO; the switching vectors of small sectors ② in large sector IV include: OPP, NOO, NPP, NOP; the switching vectors of small sectors ③ in large sector IV include: OPP, NOO, NOP, OOP, NNO; the switching vectors of small sectors ④ in large sector IV include: OOP, NNO, NOP, NNP; the switching vectors of small sectors ① in large sector V include: OOO, OOP, NNO, POP, ONO; the switching vectors of the small sector ② within the large sector V include: OOP, NNO, NNP, ONP; the switching vectors of the small sector ③ within the large sector V include: OOP, NNO, ONP, POP, ONO; the switching vectors of the small sector ④ within the large sector V include: POP, ONO, ONP, PNP; the switching vectors of the small sector ① within the large sector VI include: OOO, POO, ONN, POP, ONO; the switching vectors of the small sector ② within the large sector VI include: POP, ONO, PNP, PNO; the switching vectors of the small sector ③ within the large sector VI include: POP, ONO, PNO, POO, ONN; the switching vectors of the small sector ④ within the large sector VI include: POO, ONN, PNN, PNO.

3. The three-level rectifier model predictive power control method based on deadbeat control according to claim 2, characterized in that: The quadratic equation of the output voltage in step S1 is as follows: After discretization, the active power reference value P ref The expression of (k) is as follows: Among them, P ref (k) represents the reference power at the kth moment, T s is the sampling time, U dc (t) is the continuous expression of the output voltage, U ref is the reference value of the output voltage, U dc (k) is the value of the output voltage at the kth moment, U dc (k+1) is the value of the output voltage at the k+1th moment, I dc (k) is the value of the output current at the kth moment, C is the capacitance value on the DC side, R is the load resistance, y c is the correction function, y c (k-1) is the correction function value at the k-1th moment, λ is the positive parameter value to be set, and dt represents the correction function value for (U ref -U dc (t)) is integrated.

4. The three-level rectifier model prediction power control method based on deadbeat control according to claim 2, characterized in that: In step S2, the formula for finding the sector corresponding to the central vector that minimizes the value of the first objective function is as follows: Among them, J represents the first objective function, Q ref is the reactive power reference value, P ref is the reference value of active power, T s represents the sampling time, P i (k) and P i (k+1) is the active power when the i-th switch vector is used at the k-th moment and the k+1-th moment, Q i (k) and Q i (k+1) are the reactive powers when the i-th switching vector is used at the k-th moment and the k+1-th moment, respectively. P(k) and Q(k) are the active power and reactive power of the three-level rectifier at the k-th moment. P i ′ and Q i ′ are the derivatives of active power and reactive power when the i-th switching vector is used, sector(h) indicates that the h-th sector is the best sector, and argmin(J) indicates the minimum value in the set of all objective function values; In the stationary two-phase αβ coordinate system, the expressions of active power and reactive power derivatives are as follows: Among them, e α 、e β 、i α 、i β represents the α component and β component of the grid phase voltage and the α component and β component of the line current in the stationary two-phase αβ coordinate system, L is the value of the grid side filter inductance, ω represents the angular velocity, P′ represents the derivative of the active power of the three-level rectifier, Q′ represents the derivative of the reactive power of the three-level rectifier; when the switching state of the rectifier is v i When the three-level rectifier outputs voltage u in the stationary two-phase αβ coordinate system i (u αi ,u βi ), the above formula is expressed as follows: u α and u β represents the α and β components of the switch vector in the αβ coordinate system, u αi and u βi represents the α component and β component of the i-th switch vector in the αβ coordinate system; At time k+1, the active power and reactive power of the three-level rectifier are written as: The three-level space vector diagram is divided into 6 large sectors and 4 small sectors. The switching vector in the middle of each sector is taken, which is called the central vector of the sector. The active power P of the three-level rectifier at time k+1 is calculated using the central vector of the large sector. i (k+1) and reactive power Q i (k+1), when finding the sector where the optimal virtual switch vector is located, the midpoint voltage balance problem is not considered first, and the first objective function is defined as: J=[P ref -P i (k+1)] 2 +[Q ref -Q i (k+1)] 2 Find the central vector corresponding to the minimum value of the first objective function J. The large sector where the central vector is located is the large sector where the optimal virtual switch vector is located. Then use the same method to obtain the small sector where the optimal virtual switch vector is located in the large sector. This small sector is the sector where the optimal switch vector is located.

5. The three-level rectifier model predictive power control method based on deadbeat control according to claim 2, characterized in that: The process of step S3 is as follows: After finding the small sector where the optimal switching vector is located, substitute the switching vector in the sector into the second objective function G, so that the switching vector with the smallest second objective function G is the optimal switching vector. The formula for finding the optimal switching vector is as follows: Among them, u o (k) and u o (k+1) is the midpoint voltage of the output terminal of the three-level rectifier at the kth moment and the k+1th moment, and the expressions are u o (k) = u c1 (k)-u c2 (k),u o (k+1)=u c1 (k+1)-u c2 (k+1),u c1 (k),u c2 (k) are the voltage values ​​of the upper and lower capacitors at the output end at time k, respectively, u c1 (k+1) and u c2 (k+1) are the voltage values ​​of the upper and lower capacitors at the output end at time k+1, C is the capacitance value on the DC side, S abc is the switching state of each phase, and |S abc |={|S a |,|S b |,|S c |}, S a , S b , S c The switching states i of phase A, phase B, and phase C are abc is the line current on the grid side, which is expressed as i abc ={i a ,i b ,i c }, i a 、i b and i c are the line currents of phase A, phase B and phase C respectively, λ m is the weight factor of the midpoint potential, vector(m) indicates that the mth switching vector is the optimal switching vector, and argmin(G) indicates the minimum value in the set of all objective function values.

6. The three-level rectifier model predictive power control method based on deadbeat control according to claim 5, characterized in that: Weight factor λ of midpoint potential m The value range is [0.25,0.35].

Images