Three-vector modulation model predictive control method and system for three-level inverter

Abstract

The invention relates to a three-vector modulation model prediction control method and system for a three-level inverter, and belongs to the technical field of three-level inverter control. The method comprises the following steps: firstly, predicting a reference voltage at a next moment according to a mathematical model of a three-level inverter topological structure and a reference current, and then judging a large sector and a small sector where the reference voltage is located to obtain three voltage vectors corresponding to the synthesized reference voltage; calculating the cost function values of the three voltage vectors to obtain the corresponding duty ratios, and finally using the duty ratios corresponding to the three voltage vectors to control the switching action of the inverter at the next moment. According to the invention, the problem of large calculation amount of model prediction control in the prediction and rolling optimization process in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to a three-vector modulation model predictive control method and system of a three-level inverter, and belongs to the technical field of three-level inverter control. BACKGROUND

[0002] The new energy power generation technology in China is developing rapidly. The T-type three-level inverter is widely used in the field of new energy power generation due to its high output voltage waveform quality and low harmonic content. Model predictive control, as one of the mainstream control strategies for grid-connected inverters, was first born in the field of industrial process control. It is a control algorithm that first constructs a predictive model according to mathematical modeling, and then uses the trend of the predicted changes to construct a value function to control the target and achieve the control purpose.

[0003] However, there are still some shortcomings in the application of model predictive control in three-phase T-type three-level inverters: for example, model predictive control requires a large amount of calculation in the process of prediction and rolling optimization. The main reason is that the three-phase three-level inverter can output three levels per phase, and after combination, there are 27 kinds of switching states in three phases, which need to be calculated 27 times in the prediction of current and cost function optimization, and it is also difficult to select the weight factor in the multi-objective control of grid-connected current and midpoint voltage balance. SUMMARY

[0004] The purpose of the present application is to provide a three-vector modulation model predictive control method and system of a three-level inverter to solve the problem of large amount of calculation in the process of prediction and rolling optimization of model predictive control in the prior art.

[0005] To achieve the above-mentioned purpose, the scheme of the present application includes:

[0006] The three-vector modulation model predictive control method of a three-level inverter of the present application includes the following steps:

[0007] 1) According to the mathematical model corresponding to the topology structure of the three-level inverter and the given reference current, the reference voltage output by the three-level inverter at the next moment is predicted;

[0008] 2) Determine the large sector where the reference voltage is located and the small sector under the large sector, and determine the three voltage vectors corresponding to the small sector required by the synthesized reference voltage v i , v j , v k , wherein v i is a redundant small vector;

[0009] 3) Calculate the cost function values g i , g j , g kAccording to the cost function values of the three voltage vectors, duty cycles d i , d j , d k are calculated

[0010] 4) The three-level inverter is controlled by using the obtained duty cycles of the three voltage vectors.

[0011] Further, when the neutral point voltage is unbalanced, the duty cycles of the redundant small vectors in the three voltage vectors are adjusted according to the designed unbalance coefficient, and the three-level inverter is controlled according to the adjusted duty cycles of the redundant small vectors and the obtained duty cycles of the other two voltage vectors.

[0012] Further, the duty cycle adjustment formula of the redundant small vector is:

[0013]

[0014] Wherein, x is the unbalance coefficient; d i is the duty cycle of the redundant small vector; d pi is the duty cycle of the positive redundant small vector; d ni is the duty cycle of the negative redundant small vector; Δu o = v c1 -v c2 ; v c1 is the upper capacitor voltage in the three-level inverter; v c2 is the lower capacitor voltage in the three-level inverter.

[0015] Further, the calculation formula of the reference voltage is:

[0016]

[0017] Wherein, represents the reference current of the α-axis at K+1 moment; i Lα (k) represents the grid-connected current of the α-axis at K moment; represents the reference voltage of the inverter output of the α-axis at K+1 moment; v gα (k) represents the grid-connected voltage of the α-axis at K moment; represents the reference current of the β-axis at K+1 moment; i Lβ (k) represents the grid-connected current of the β-axis at K moment; represents the reference voltage of the inverter output of the β-axis at K+1 moment; v gβ (k) represents the grid-connected voltage of the β-axis at K moment; T s is the sampling period; L represents the filter inductance.

[0018] Further, the large sector where the reference voltage is located is determined according to the value of E, and the calculation formula of E is:

[0019]

[0020] wherein, is the reference voltage under the α-axis; is the reference voltage under the β-axis.

[0021] Further, the small sector under the large sector where the reference voltage is located is determined according to the value of M, and the calculation formula of M is:

[0022]

[0023] wherein, is the reference voltage under the α-axis in the mapping relationship; is the reference voltage under the β-axis in the mapping relationship.

[0024] Further, the duty cycles d i , d j , d k of the three voltage vectors are respectively:

[0025]

[0026] wherein, is the reference voltage under the αβ-axis at the K+1 moment; v iαβ (k+1) is the voltage corresponding to the voltage vector v i (k+1) under the αβ-axis at the K+1 moment; v jαβ (k+1) is the voltage corresponding to the voltage vector v j (k+1) under the αβ-axis at the K+1 moment; v kαβ (k+1) is the voltage corresponding to the voltage vector v k (k+1) under the αβ-axis at the K+1 moment.

[0027] Further, the duty cycles d i , d j , d k of the three voltage vectors are respectively:

[0028]

[0029] wherein, g i , g j , g k are the cost function values of the voltage vectors v i , v j , v k .

[0030] Further, the method delivers the duty cycles of the three voltage vectors to the SVPWM module, the SVPWM module generates the action signals of the switch tubes of the three-level inverter, and the SVPWM module adopts the seven-segment form with the N-type small vector as the first vector.

[0031] A three-vector modulation model predictive control system of a three-level inverter comprises a processor configured to execute a computer program to implement the steps of the three-vector modulation model predictive control method of the three-level inverter according to any one of the above.

[0032] The method provided by the application has the advantages that: as an improved invention, the method predicts the reference voltage output by the inverter at the next moment according to the mathematical model corresponding to the topology structure of the three-level inverter and the given reference current, judges the large sector where the reference voltage is located and the small sector under the large sector, obtains the three voltage vectors corresponding to the reference voltage, calculates the cost function values corresponding to the three voltage vectors according to the three voltage vectors and the reference voltage, and finally obtains the duty cycles corresponding to the cost function values, so that when the inverter at the next moment is controlled, only the duty cycles of the three voltage vectors need to be used, without the need of a large amount of calculation, and thus the problem of large calculation amount in the process of prediction and rolling optimization of the model predictive control in the prior art is solved. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is the overall control flowchart provided by the application;

[0034] Figure 2 is the overall control block diagram provided by the application;

[0035] Figure 3 is the topology structure diagram of the three-phase T-type three-level inverter provided by the application;

[0036] Figure 4 is the space voltage vector sector division diagram provided by the application;

[0037] Figure 5 is the seven-segment modulation voltage vector action time allocation diagram provided by the application. DETAILED DESCRIPTION

[0038] In order to make the purpose, technical scheme and advantages of the application more clear, the application will be clearly and completely described in detail below with reference to the drawings and examples.

[0039] The concept of the present application is that firstly, the reference voltage of the next moment is predicted according to the topology structure of the three-level inverter and the reference current, then the three voltage vectors corresponding to the reference voltage are obtained after judging the large sector and the small sector where the reference voltage is located, and then the duty cycles corresponding to the three voltage vectors are obtained by calculating the cost function values of the three voltage vectors, and finally the switching action of the inverter of the next moment is controlled using the duty cycles corresponding to the three voltage vectors.

[0040] Three-vector modulation model predictive control method for three-level inverter

[0041] As Figure 1 The overall control flowchart provided by the present application is shown, mainly including the following steps:

[0042] (1) A mathematical model is established according to the topology structure of the three-phase T-type three-level inverter, and then the reference voltage of the three-level inverter of the next moment is predicted in combination with the given reference current;

[0043] (2) The large sector and the small sector under the large sector where the predicted reference voltage is located are judged, and the three voltage vectors corresponding to the small sector required by the synthesized reference voltage are determined;

[0044] (3) The cost function values and the duty cycles of the three voltage vectors are calculated according to the three voltage vectors and the reference voltage obtained above;

[0045] (4) The three-level inverter is controlled by using the duty cycles of the three voltage vectors obtained.

[0046] In addition, in order to realize the balance control of the midpoint voltage, when the midpoint voltage is unbalanced, the duty cycle of the redundant small vector in the three voltage vectors can be adjusted by the designed unbalance coefficient, so as to achieve the balance of the midpoint voltage.

[0047] The overall control block diagram is shown in Figure 2 The specific control steps are as follows:

[0048] (1) A mathematical model is established according to the topology structure of the three-phase T-type three-level inverter, and then the reference voltage is obtained by processing the inverter side inductance current state equation using the forward Euler method and the deadbeat control.

[0049] Specifically, the mathematical model established in the abc coordinate system according to the topology structure of the three-phase T-type three-level inverter shown in Figure 3 is as follows:

[0050]

[0051] Wherein, L represents the filter inductance, i L represents the grid-connected current, represents i LThe derivative of the function with respect to time t, v represents the inverter output voltage, v g represents the grid-connected voltage.

[0052] The mathematical model of the above-mentioned abc coordinate system is subjected to Clarke transformation to obtain a mathematical model of the αβ axis:

[0053]

[0054] The mathematical model is discretized by using the forward Euler method to obtain a predicted current expression at k+1 time:

[0055]

[0056] wherein X(k) and X(k+1) are variable values at k time and k+1 time respectively, T s is a sampling period.

[0057] The predicted current expression is rewritten to obtain an inverter output voltage expression as follows:

[0058]

[0059] Since the control target of the model predictive control is to make the predicted current track the given reference current and the error is as small as possible, the reference current can be approximately equal to the predicted current, that is,

[0060] According to the zero-error principle, the reference voltage can be obtained as follows:

[0061]

[0062] wherein i represents the reference current at k+1 time on the α axis; i Lα (k) represents the grid-connected current at k time on the α axis; represents the reference voltage output by the inverter at k+1 time on the α axis; v gα (k) represents the grid-connected voltage at k time on the α axis; represents the reference current at k+1 time on the β axis; i Lβ (k) represents the grid-connected current at k time on the β axis; represents the reference voltage output by the inverter at k+1 time on the β axis; v gβ (k) represents the grid-connected voltage at k time on the β axis; T s is a sampling period; and L represents a filter inductance.

[0063] (2) judging the large sector in which the reference voltage is located and the small sector under the large sector, and determining three voltage vectors corresponding to the small sector required to be used for synthesizing the reference voltage.

[0064] Large and small sectors are obtained by dividing the space voltage vector diagram. The space voltage vector diagram is divided according to the reference voltage under the α-axis and the reference voltage under the β-axis, as shown below. Figure 4 The six major sectors shown in Table 1 are defined by the following specific judgment formulas:

[0065]

[0066] Table 1

[0067]

[0068] in, This is the reference voltage along the α axis; This is the reference voltage along the β axis.

[0069] Assuming we determine A = 1, B = 1, and C = 1, then E = 7, meaning the sector containing the reference voltage is II. For example... Figure 4 As shown, sector I refers to the red hexagon in the figure; sector II refers to the green hexagon in the figure; sector III refers to the purple hexagon in the figure; sector IV refers to the yellow hexagon; sector V refers to the orange hexagon; and sector VI refers to the light orange hexagon.

[0070] Each large sector comprises six smaller triangular sectors. For example, the red I-th large sector is divided into 6 smaller sectors, such as... Figure 4 The six triangles are numbered 1 to 6. The numbers 1, -1, and 0 at the vertices of each triangle represent three switch states. In this embodiment, taking phase A as an example, 1 represents state P, corresponding to... Figure 3 S-type switching transistor a1 S a2 On, -1 represents N state, corresponding to Figure 3 S-type switching transistor a3 S a4 On, 0 represents the O state, corresponding to Figure 3 S-type switching transistor a2 S a3 Conduction. Based on the large sector determined above, the reference voltage is mapped, and then the small sector is determined according to the mapped reference voltage. The specific mapping relationship is shown in Table 2:

[0071] Table 2

[0072]

[0073] If the above determination indicates that the large sector containing the reference voltage is II, then according to the mapping relationship, the mapped α-axis voltage can be obtained. Mapped β-axis voltage

[0074] Substituting the α-axis reference voltage under the above mapping relationship and β-axis reference voltage The sub-sectors within the large sector containing the reference voltage are shown in Table 3. The specific determination formula is as follows:

[0075]

[0076] Table 3

[0077]

[0078] in, The reference voltage for the α-axis under the mapping relationship; This is the reference voltage for the β-axis under the mapping relationship.

[0079] Assuming we determine that X = 1, Y = 1, and Z = 0, then M = 6, meaning the large sector containing the reference voltage is II, and the small sector is 5.

[0080] Table 4

[0081]

[0082] The relationship between the switching state of a small sector and the voltage vector is expressed as follows:

[0083]

[0084] Among them, U dc This is the DC bus voltage. Therefore, the three voltage vectors v corresponding to the current reference voltage can be obtained using the voltage vector transformation relationship between the smaller sectors within the larger sector of the reference voltage. i v j v k , where v i The redundant small vectors are shown in Table 4.

[0085] (3) Substitute the three voltage vectors and reference voltage corresponding to the small sector obtained above into the cost function of the model predictive control to obtain the cost function values ​​of the three voltage vectors and their respective duty cycles.

[0086] Specifically, the cost function values ​​corresponding to the three voltage vectors are as follows:

[0087]

[0088] in, The reference voltage at time K+1 along the αβ axis; v iαβ (k+1) represents v at time K+1 under the αβ axis. i Voltage corresponding to the voltage vector; v jαβ (k+1) represents v at time K+1 under the αβ axis.j Voltage corresponding to voltage vector; v kαβ (k+1) is the voltage of K+1 time v k Voltage corresponding to voltage vector.

[0089] Since the voltage vector duty ratio is inversely proportional to the cost function value, the duty ratio d i , d j , d k of each voltage vector can be obtained

[0090]

[0091] Wherein, g i , g j , g k are the cost function values of voltage vectors v i , v j , v k .

[0092] After sorting, we can get:

[0093]

[0094] Wherein,

[0095] (4) Determine whether the midpoint voltage is balanced.

[0096] If the midpoint voltage is unbalanced, the duty ratio d i of the redundant small vector v i in the three voltage vectors is adjusted.

[0097] Specifically, since the current injected by the positive and negative small vectors at the DC side midpoint has opposite effects, only the duty ratio of the redundant small vector needs to be redistributed using a fixed imbalance coefficient, which can achieve the balance control of the midpoint potential and solve the problem of difficult selection of weight factors in the multi-objective control of grid-connected current and midpoint voltage balance.

[0098] The duty ratio adjustment formula of the redundant small vector is:

[0099]

[0100] Wherein, x is the imbalance coefficient, which is taken as 0.5 in the embodiment; d i is the duty ratio of the redundant small vector; d pi is the duty ratio of the positive redundant small vector; d ni is the duty ratio of the negative redundant small vector; Δu o =v c1 -v c2 ; v c1is the upper capacitor voltage in the three-level inverter; v c2 is the lower capacitor voltage in the three-level inverter.

[0101] (5) controlling the three-level inverter according to the duty ratios of the three voltage vectors obtained.

[0102] When the neutral point voltage is balanced, the duty ratios d i , d j , and d k of the three voltage vectors obtained in step (3) are delivered to the SVPWM module, and the three-level inverter is controlled by the SVPWM module.

[0103] When the neutral point voltage is unbalanced, the duty ratios d j , d k of the voltage vectors obtained in step (3) and the duty ratios d pi , d ni of the positive and negative redundant small vectors obtained in step (4) are delivered to the SVPWM module, and the three-level inverter is controlled by the SVPWM module.

[0104] In this embodiment, the SVPWM can adopt a seven-segment form with the N-type small vector as the first vector, and the time distribution of the seven-segment modulated voltage vector is shown in Figure 5 , wherein v in is the negative redundant small vector, and v ip is the positive redundant small vector. Assuming that the large sector where the reference voltage is located is II, and the small sector is 5, the switching state switching sequence used is: OON-OOO-POO-PPO-POO-OOO-OON, the negative redundant small vector switching state is OON, and the duty ratio thereof is d ni . The positive redundant small vector switching state is PPO, and the duty ratio thereof is d pi . The v j vector switching state is OOO, and the duty ratio thereof is d j . The v k vector switching state is POO, and the duty ratio thereof is d k .

[0105] Three-vector modulation model predictive control system embodiment of a three-level inverter:

[0106] The three-vector modulation model predictive control system of the three-level inverter of the present application comprises a processor for executing computer program instructions to realize the three-vector modulation model predictive control method of the three-level inverter introduced in the method embodiment, which has been sufficiently clear in the method embodiment and will not be repeated here.

Claims

1. A three-vector modulation model predictive control method for a three-level inverter, characterized in that, Includes the following steps: 1) Based on the mathematical model corresponding to the three-level inverter topology and the given reference current, predict the reference voltage output of the three-level inverter at the next moment; 2) Determine the large sector where the reference voltage is located and the small sectors within the large sector, and determine the three voltage vectors v corresponding to the small sectors required to synthesize the reference voltage. i v j v k , where v i It is a redundant small vector; 3) Calculate the cost function value g of the three voltage vectors based on the obtained three voltage vectors and the reference voltage. i g j g k Then, the duty cycle d of the three voltage vectors is calculated based on the cost function values ​​of the three voltage vectors. i d j d k ; 4) The duty cycle of the three voltage vectors is used to control the three-level inverter.

2. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, When the midpoint voltage is unbalanced, the duty cycle of the redundant small vector among the three voltage vectors is adjusted according to the designed unbalance coefficient, and the three-level inverter is controlled according to the adjusted duty cycle of the redundant small vector and the duty cycles of the other two voltage vectors.

3. The three-vector modulation model predictive control method for a three-level inverter according to claim 2, characterized in that, The duty cycle adjustment formula for the redundant small vector is: Where x is the imbalance coefficient; d i The duty cycle of the redundant small vector; d pi The duty cycle of a positive redundant small vector; d ni The duty cycle of the negative redundant small vector; Δu o =v c1 -v c2 ;v c1 This refers to the voltage across the upper capacitor in a three-level inverter; v c2 This refers to the voltage of the lower capacitor in a three-level inverter.

4. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, The formula used to calculate the reference voltage is as follows: in, This represents the reference current at time K+1 along the α axis; i Lα (k) represents the grid-connected current at time K under the α axis; This represents the reference voltage output by the inverter at time K+1 along the α axis; v gα (k) represents the grid-connected voltage at time K along the α axis; This represents the reference current at time K+1 along the β axis; i Lβ (k) represents the grid-connected current at time K under the β axis; This represents the reference voltage output by the inverter at time K+1 along the β axis; v gβ (k) represents the grid-connected voltage at time K along the β axis; T s The sampling period is L; L represents the filter inductance.

5. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, The sector containing the reference voltage is determined based on the E value, which is calculated using the following formula: in, This is the reference voltage along the α axis; This is the reference voltage along the β axis.

6. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, The smaller sector under the larger sector containing the reference voltage is determined based on the M value, which is calculated using the following formula: in, The reference voltage for the α-axis under the mapping relationship; This is the reference voltage for the β-axis under the mapping relationship.

7. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, The cost function value g of the three voltage vectors i g j g k The calculation formulas used are as follows: in, The reference voltage at time K+1 along the αβ axis; v iαβ (k+1) represents v at time K+1 under the αβ axis. i Voltage corresponding to the voltage vector; v jαβ (k+1) represents v at time K+1 under the αβ axis. j Voltage corresponding to the voltage vector; v kαβ (k+1) represents v at time K+1 under the αβ axis. k The voltage corresponding to the voltage vector.

8. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, The duty cycle d of the three voltage vectors i d j d k They are respectively: Among them, g i g j g k These are the voltage vectors v i v j v k The cost function value.

9. The three-vector modulation model predictive control method for a three-level inverter according to claim 1, characterized in that, This method transmits the duty cycle of three voltage vectors to the SVPWM module, which generates the operation signals of each switch of the three-level inverter. The SVPWM module adopts a seven-segment form with the N-type small vector as the first vector.

10. A three-vector modulation model predictive control system for a three-level inverter, comprising a processor, characterized in that, The processor is used to execute a computer program to implement the steps of a three-vector modulation model predictive control method for a three-level inverter as described in any one of claims 1 to 9.

Images