阅读说明：本技术 矢量扇区定位方法、局部寻优模型预测控制方法及装置 (Vector sector positioning method, local optimization model prediction control method and device ) 是由 张晓� 李枝亮 史军伟 曾朝玮 韩笑笑 张辉 于 2020-05-25 设计创作，主要内容包括：本发明公开了矢量扇区定位方法、局部寻优模型预测控制方法及装置,通过建立详细的预测模型在实现控制输出电流的同时,可以实现直流侧中点电位以及各相悬浮电容电压的控制,同时针对传统模型预测控制进行全局寻优时存在计算量大,难以硬件实现的问题,本发明利用提出的矢量扇区定位方法提前判断参考电压在输出电压矢量图中所处的三角区域,在目标三角形区域中进行代价函数判断,将一个全局寻优问题化为一个在参考区域内的局部寻优问题。本发明在实现多个输出变量控制的前提下,大幅减小了系统计算负荷,节约计算资源,减小算法控制的延迟,该方法控制下的五电平逆变器具有良好的输出效果。(The invention discloses a vector sector positioning method, a local optimization model prediction control method and a device, which can realize control of a midpoint potential on a direct current side and the voltage of each phase of a suspended capacitor while realizing control of output current by establishing a detailed prediction model, and simultaneously solve the problems of large calculated amount and difficult hardware realization when the global optimization is carried out by the traditional model prediction control. On the premise of realizing the control of a plurality of output variables, the method greatly reduces the calculation load of the system, saves the calculation resources and reduces the delay of algorithm control, and the five-level inverter controlled by the method has good output effect.)

1. A vector sector positioning method is characterized by comprising the following specific steps:

obtaining a phase angle of a reference vector under a two-phase rotating coordinate system, and judging a large sector where the reference vector is located;

carrying out gh coordinate transformation on the components of the reference vector under the two-phase rotating coordinate system to determine a table boundary area or a diamond area in a large sector where the reference vector is located;

if the reference vector is in the table boundary area, directly obtaining a small triangular area of the reference vector in the vector diagram;

if the reference vector is in the diamond-shaped area, judging a small triangular area of the reference vector in the vector diagram by a translation positioning method of the reference vector;

the reference vector is a component of a three-phase reference voltage output by the inverter in a two-phase rotating coordinate system

2. The vector sector positioning method according to claim 1, wherein the method for determining the table boundary area or the diamond area in the large sector is:

and (3) rounding the unit length of the vector diagram by the component of the reference vector under a gh coordinate system to determine a diamond area or a table boundary area in the large sector where the reference vector is located, wherein the table boundary area refers to an independent small triangular area at the outermost periphery of the vector diagram.

3. The vector sector positioning method according to claim 1 or 2, wherein the translational positioning method of the reference vector is:

taking the component of the reference vector in a gh coordinate system for the unit length in the vector diagram, and translating the target diamond region to the central diamond region of the vector diagram;

and determining a small triangular area of the reference vector in the target diamond-shaped area according to the boundary conditions of the two triangles in the central diamond-shaped area.

4. A local optimization model prediction control method is suitable for a five-level grid-connected inverter and is characterized by comprising the following specific steps:

discretizing through establishing a mathematical model to obtain a prediction model of the output current of the inverter; the mathematical model is a model of a grid-connected process of a five-level active neutral point clamped inverter; the prediction model is used for predicting the output value of the controlled variable at the next sampling moment by using the sampling value of the variable at the current sampling moment;

constructing a dq vector according to a given reference value of the output current of the inverter, performing vector operation, and calculating a voltage reference vector output by the inverter;

determining a small triangular area of an inverter voltage reference vector in a vector diagram by using the vector sector positioning method of any one of claims 1 to 3 on a five-level inverter output voltage vector diagram;

sequentially substituting all switch states in the target small triangular area into a prediction model of the output current of the inverter to obtain a predicted value;

and substituting the predicted value into a cost function, and comparing the cost function with a reference value of the output current of the inverter, so that the switching state which enables the cost function to be minimum and meets the switching limiting condition acts on the system.

5. The local optimization model predictive control method of claim 4, further comprising constructing a multivariable cost function, and substituting the inverter output current, the DC side capacitance voltage difference and the predicted value of each phase bridge arm suspension capacitance voltage into the multivariable cost function;

the multivariable cost function is realized by adding the inverter output current control, the DC side midpoint potential control and the suspension capacitor voltage control of each phase bridge arm into one cost function;

and the output current of the inverter, the voltage difference of the direct current side capacitor and the predicted value of the voltage of the suspension capacitor of each phase bridge arm are respectively obtained through respective prediction models.

6. The local optimization model prediction control method according to claim 5, wherein the prediction models of the inverter output current, the dc-side capacitance voltage difference and the suspension capacitance voltage of each phase bridge arm are respectively:

wherein the content of the first and second substances,

7. The local optimization model predictive control method of claim 5, wherein the calculation of the voltage reference value of the inverter output is as follows:

will the network voltage e_a(k),e_b(k),e_c(k) Clark and park transformation to obtain e_d(k),e_q(k)；

Reference value of grid-connected currentSubjected to Clark and park transformation to obtain

e_dq(k)＝e_d(k)+je_q(k),

the inverter reference voltage dq vector is calculated as:

calculating the components of the inverter reference voltage dq vectorThe component of the inverter output voltage reference vector under the two-phase rotating coordinate system is obtained by carrying out reverse park conversion

8. The local optimization model predictive control method of claim 5, wherein the multivariate cost function is:

wherein the content of the first and second substances,

9. A local optimization model predictive control device is characterized in that: the device comprises a direct-current voltage source, two series capacitors, a main circuit module, a first voltage sensor, a second voltage sensor, a third voltage sensor, a fourth voltage sensor, a current sensor, a vector calculation module, a target reference sector judgment module and a prediction calculation module;

the direct-current voltage source supplies power to the two series capacitors;

the first sensor and the second sensor are used for respectively collecting voltages at two ends of a direct current side series capacitor;

the third sensor is used for collecting the voltage of the suspension capacitor of the bridge arm;

the fourth sensor is used for collecting the voltage of the network side;

the current sensor is used for collecting the output grid-connected current of the inverter;

the main circuit module is used for realizing electric energy conversion and grid connection;

the vector calculation module is used for constructing a dq vector and carrying out vector operation to calculate an inverter output voltage reference vector;

the reference sector judging module is used for judging a small triangular area where a reference vector is located in a vector diagram;

the prediction calculation module is used for predicting variable reference values by combining information acquired by the first sensor, the second sensor, the third sensor and the fourth sensor and the current sensor with switch states in a small triangular area, and carrying out optimization judgment by introducing a cost function to find out an optimal control target.

10. The local optimization model predictive control apparatus according to claim 9, characterized in that:

the reference sector judging module is used for judging that a small triangular area of a reference vector in a vector diagram is realized on a five-level inverter output voltage vector diagram by using a vector sector positioning method;

the vector sector positioning method specifically comprises the following steps:

obtaining a phase angle of a reference vector under a two-phase rotating coordinate system, and judging a large sector where the reference vector is located;

carrying out gh coordinate transformation on the components of the reference vector under the two-phase rotating coordinate system to determine a table boundary area or a diamond area in a large sector where the reference vector is located;

if the reference vector is in the table boundary area, directly obtaining a small triangular area of the reference vector in the vector diagram;

if the reference vector is in the diamond-shaped area, judging a small triangular area of the reference vector in the vector diagram by a translation positioning method of the reference vector;

the reference vector is a component of a three-phase reference voltage output by the inverter in a two-phase rotating coordinate system

Technical Field

The invention belongs to the technical field of multi-level inverter control, and particularly relates to a vector sector positioning method, a local optimization model prediction control method and a local optimization model prediction control device for a five-level grid-connected inverter.

Background

The inverter is a key component in the fields of power grid connection, motor variable frequency speed regulation and the like as an interface for energy conversion. However, with the development of power systems, the conventional two-level inverter gradually fails to meet the requirements of high capacity and high quality grid connection. The five-level inverter has smaller distortion rate of output voltage and current, limits dv/dt, needs smaller filter inductance, greatly reduces the weight and the volume of a system, and has great application value in the fields of new energy grid connection, rail transit, high-voltage power transmission and distribution and the like.

As a novel control method, model predictive control is used, a discrete mathematical model and a predictive model of a system are established, and a cost function is evaluated to enable a predicted quantity to track a reference quantity so as to achieve an expected control target. The model prediction control is simple, the response is rapid, the multi-objective optimization can be realized, and the method is widely applied to high-power converters at present. However, when the traditional model prediction control is applied to the control of the five-level active midpoint clamping type inverter, 8^3 ^ 512 times are needed to be calculated in each sampling period, a large amount of calculation resources are consumed, and the system delay is increased.

Disclosure of Invention

According to the defects of the prior art, the invention provides a vector sector positioning method, a local optimization model prediction control method and a local optimization model prediction control device, and the optimization area is reduced to 1/96 by judging a small triangular area of a reference voltage vector in an output voltage vector diagram in advance by providing the vector sector positioning method, so that the calculation amount caused by prediction control is greatly reduced, and the control delay is reduced.

The invention provides a vector sector positioning method, which comprises the following specific steps:

obtaining a phase angle of a reference vector under a two-phase rotating coordinate system, and judging a large sector where the reference vector is located;

carrying out gh coordinate transformation on the components of the reference vector under the two-phase rotating coordinate system to determine a table boundary area or a diamond area in a large sector where the reference vector is located;

if the reference vector is in the table boundary area, directly obtaining a small triangular area of the reference vector in the vector diagram;

if the reference vector is in the diamond-shaped area, judging a small triangular area of the reference vector in the vector diagram by a translation positioning method of the reference vector;

the reference vector is a component of a three-phase reference voltage output by the inverter in a two-phase rotating coordinate system

And (4) forming.

Preferably, the method for determining the table boundary area or the diamond area in the large sector is as follows:

and (3) rounding the unit length of the vector diagram by the component of the reference vector under a gh coordinate system to determine a diamond area or a table boundary area in the large sector where the reference vector is located, wherein the table boundary area refers to an independent small triangular area at the outermost periphery of the vector diagram.

Preferably, the method for translational positioning of the reference vector comprises:

taking the component of the reference vector in a gh coordinate system for the unit length in the vector diagram, and translating the target diamond region to the central diamond region of the vector diagram;

and determining a small triangular area of the reference vector in the target diamond-shaped area according to the boundary conditions of the two triangles in the central diamond-shaped area.

The invention also comprises a local optimization model prediction control method, which is suitable for a five-level grid-connected inverter and comprises the following specific steps:

discretizing through establishing a mathematical model to obtain a prediction model of the output current of the inverter; the mathematical model is a model of a grid-connected process of a five-level active neutral point clamped inverter; the prediction model is used for predicting the output value of the controlled variable at the next sampling moment by using the sampling value of the variable at the current sampling moment;

constructing a dq vector according to a given reference value of the output current of the inverter, performing vector operation, and calculating a voltage reference vector output by the inverter;

judging a small triangular area of an inverter voltage reference vector in a vector diagram by utilizing any one of the vector sector positioning methods on a five-level inverter output voltage vector diagram;

sequentially substituting all switch states in the target small triangular area into a prediction model of the output current of the inverter to obtain a predicted value;

and substituting the predicted value into a cost function, and comparing the cost function with a reference value of the output current of the inverter, so that the switching state which enables the cost function to be minimum and meets the switching limiting condition acts on the system.

Preferably, the method further comprises the steps of constructing a multivariable cost function, and substituting the output current of the inverter, the voltage difference of the capacitors on the direct current side and the predicted value of the voltage of the suspension capacitor of each phase of the bridge arm into the multivariable cost function;

the multivariable cost function is realized by adding the inverter output current control, the DC side midpoint potential control and the suspension capacitor voltage control of each phase bridge arm into one cost function;

and the output current of the inverter, the voltage difference of the direct current side capacitor and the predicted value of the voltage of the suspension capacitor of each phase bridge arm are respectively obtained through respective prediction models.

Preferably, the prediction models of the inverter output current, the dc-side capacitance voltage difference, and the suspension capacitance voltage of each phase bridge arm are respectively:

wherein the content of the first and second substances,the predicted value of the inverter output phase current at the moment of k + 1; i.e. i_x(k) Outputting phase current for the inverter at the moment k; e.g. of the type_x(k) The voltage of the power grid phase at the moment k; u. of_x(k) Outputting phase voltage for the inverter at the moment k; Δ u_dc ^P(k +1) is a predicted value of the voltage difference between the upper capacitor and the lower capacitor on the direct current side at the moment of k + 1; Δ u_dc(k) The voltage difference between the upper capacitor and the lower capacitor on the direct current side at the moment k;predicting the voltage of each bridge arm suspension capacitor at the moment k + 1; u. of_fx(k) The voltage of each bridge arm suspension capacitor at the time k; t is_sIs a sampling period; r, L is a grid-connected resistor and a filter inductor; c, C_fRespectively a direct current side capacitor and a suspension capacitor; h is_xIs the influence factor of the phase current on the midpoint potential; f. of_xThe influence factor of the phase current on the suspension capacitance is shown; and the value of the influence factor of the current on the suspension capacitor is determined according to the influence of the current on the midpoint potential and the voltage of the suspension capacitor under various switch states.

Preferably, the calculation process of the voltage reference value output by the inverter is as follows:

will the network voltage e_a(k),e_b(k),e_c(k) Clark and park transformation to obtain e_d(k),e_q(k)；

Reference value of grid-connected currentSubjected to Clark and park transformation to obtain

Constructing a dq vector:

the inverter reference voltage dq vector is calculated as:

calculating the components of the inverter reference voltage dq vectorThe component of the inverter output voltage reference vector under the two-phase rotating coordinate system is obtained by carrying out reverse park conversion

Preferably, the multivariate cost function is:

wherein the content of the first and second substances,respectively representing the reference value of the output current of the inverter, the reference value of the voltage difference of the capacitor on the DC side and the reference value of the voltage of the floating capacitor, lambda_iA weighting factor representing the index of the item.

The invention also comprises a local optimization model prediction control device which comprises a direct-current voltage source, two series capacitors, a main circuit module, a first voltage sensor, a second voltage sensor, a third voltage sensor, a fourth voltage sensor, a current sensor, a vector calculation module, a target reference sector judgment module and a prediction calculation module;

the direct-current voltage source supplies power to the two series capacitors;

the first sensor and the second sensor are used for respectively collecting voltages at two ends of a direct current side series capacitor;

the third sensor is used for collecting the voltage of the suspension capacitor of the bridge arm;

the fourth sensor is used for collecting the voltage of the network side;

the current sensor is used for collecting the output grid-connected current of the inverter;

the main circuit module is used for realizing electric energy conversion and grid connection;

the vector calculation module is used for constructing a dq vector and carrying out vector operation to calculate an inverter output voltage reference vector;

the reference sector judging module is used for judging a small triangular area where a reference vector is located in a vector diagram;

the prediction calculation module is used for predicting variable reference values by combining information acquired by the first sensor, the second sensor, the third sensor and the fourth sensor and the current sensor with switch states in a small triangular area, and carrying out optimization judgment by introducing a cost function to find out an optimal control target.

Preferably, the reference sector judging module is configured to judge that the small triangular region where the reference vector is located in the vector diagram is implemented on a five-level inverter output voltage vector diagram by using a vector sector positioning method;

the vector sector positioning method specifically comprises the following steps:

obtaining a phase angle of a reference vector under a two-phase rotating coordinate system, and judging a large sector where the reference vector is located;

carrying out gh coordinate transformation on the components of the reference vector under the two-phase rotating coordinate system to determine a table boundary area or a diamond area in a large sector where the reference vector is located;

if the reference vector is in the table boundary area, directly obtaining a small triangular area of the reference vector in the vector diagram;

if the reference vector is in the diamond-shaped area, judging a small triangular area of the reference vector in the vector diagram by a translation positioning method of the reference vector;

the reference vector is a component of a three-phase reference voltage output by the inverter in a two-phase rotating coordinate system

Forming;

the method for judging the table boundary area or the diamond area in the large sector comprises the following steps:

the method comprises the following steps of (1) rounding the unit length of a vector diagram by the components of a reference vector under a gh coordinate system to determine a diamond area or a table boundary area in a large sector where the reference vector is located, wherein the table boundary area refers to an independent small triangular area at the outermost periphery of the vector diagram;

the translation positioning method of the reference vector comprises the following steps:

taking the component of the reference vector in a gh coordinate system for the unit length in the vector diagram, and translating the target diamond region to the central diamond region of the vector diagram;

and determining a small triangular area of the reference vector in the target diamond-shaped area according to the boundary conditions of the two triangles in the central diamond-shaped area.

Compared with the prior art, the invention adopting the technical scheme has the following technical effects:

1. according to the method, the small triangular area where the reference vector is located can be determined only by judging the same boundary condition by using the provided translation positioning method of the vector sector and performing remainder translation on all target diamond areas in each large sector, and compared with the traditional method for judging the sector, the redundant and fussy boundary judgment is omitted;

2. compared with the traditional global optimization model prediction control, the local optimization model prediction control for judging the sector in advance provided by the invention reduces the optimization calculation amount by nearly 26 times, greatly reduces the system calculation load, improves the efficiency and is easy to realize by hardware;

3. the output voltage reference vector of the inverter is obtained by constructing dq vectors and carrying out vector operation, so that derivation operation in the operation process can be avoided;

4. the improved algorithm of the invention has good control effect on the DC side midpoint potential and the suspension capacitor voltage on the basis of realizing grid-connected current control, and has quick dynamic response;

5. the method provided by the invention is strong in popularization and can be applied to converters with other topologies.

Drawings

The accompanying drawings, which are included to provide a further understanding of the invention, are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention without limiting the invention to the right. It is obvious that the drawings in the following description are only some embodiments, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.

In the drawings:

fig. 1 is a flowchart of a vector sector positioning method according to an embodiment of the present invention;

FIG. 2 is a flowchart of a local optimization model predictive control method according to an embodiment of the present invention;

fig. 3 is a schematic diagram of a five-level active midpoint clamping type inverter topology and grid connection provided in an embodiment of the present invention;

fig. 4 is a vector diagram of an output voltage of a five-level active midpoint clamping type inverter provided by an embodiment of the invention;

fig. 5 is a schematic view illustrating a translation determination of a small triangular area where an inverter outputs a reference voltage according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a five-level active midpoint clamping type inverter local optimization model predictive control system according to an embodiment of the present invention;

FIG. 7 is a graph of inverter output current, DC side voltage, and floating capacitor voltage control effect and dynamic response provided by an embodiment of the present invention;

fig. 8 is a diagram of a frequency spectrum analysis of the inverter grid-connected current provided by the embodiment of the invention;

fig. 9 is a statistical chart of the number of times of optimizing calculation per sampling period in the conventional model predictive control and the local optimizing model predictive control according to the embodiment of the present invention.

It should be noted that the drawings and the description are not intended to limit the scope of the inventive concept in any way, but to illustrate it by a person skilled in the art with reference to specific embodiments.

Detailed description of the invention

In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and the following embodiments are used for illustrating the present invention and are not intended to limit the scope of the present invention.

In the description of the present invention, it should be noted that the terms "upper", "lower", "front", "rear", "left", "right", "vertical", "inner", "outer", etc., indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, and are only for convenience of description and simplicity of description, but do not indicate or imply that the device or element being referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus, should not be construed as limiting the present invention.

In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; may be directly connected or indirectly connected through an intermediate. The specific meanings of the above terms in the present invention can be understood in specific cases to those skilled in the art.

Fig. 1 shows a flowchart of a vector sector positioning method provided by the present invention, in order to determine a small triangular region where an inverter output voltage reference vector is located in an inverter output voltage vector diagram, the method specifically includes the following steps: a

Step 100: obtaining a phase angle of a reference vector under a two-phase rotating coordinate system, and judging a large sector where the reference vector is located;

step 101: carrying out gh coordinate transformation on the components of the reference vector under the two-phase rotating coordinate system to determine a table boundary area or a diamond area in a large sector where the reference vector is located;

step 102: if the reference vector is in the table boundary area, directly obtaining a small triangular area of the reference vector in the vector diagram;

step 103: if the reference vector is in the diamond-shaped area, judging the small triangular area of the reference vector in the vector diagram by using a translation positioning method of the reference vector.

Further, the reference vector is a component of a three-phase reference voltage output by the inverter in a two-phase rotating coordinate systemAnd (4) forming.

Further, the judgment of the diamond area or the table boundary area in the large sector is to integrate the component of the reference vector in the gh coordinate system with the unit length in the vector diagram, and determine the diamond area or the table boundary area in the large sector where the reference vector is located, wherein the table boundary area is an independent small triangular area at the outermost periphery of the vector diagram.

Further, the translational positioning method of the reference vector comprises the following steps: taking the component of the reference vector in a gh coordinate system for the unit length in the vector diagram, and translating the target diamond region to the central diamond region of the vector diagram;

and determining a small triangular area of the reference vector in the target diamond-shaped area according to the boundary conditions of the two triangles in the central diamond-shaped area.

Fig. 3 shows a structure diagram of a three-phase five-level midpoint clamping type grid-connected inverter, wherein a direct-current voltage source Udc supplies power to two series capacitors C1 and C2, and the stable voltage of each capacitor is 1/2 of the direct-current voltage; each phase of bridge arm of the inverter is composed of 12 power switching devices and 1 suspension capacitor, and the reference voltage of the suspension capacitor is 1/4 of direct-current voltage. The premise of normal operation of the five-level inverter is that the direct-current side capacitor voltage and the suspension capacitor voltage are stabilized at reference values.

As shown in table (1), each phase of the inverter has eight switching states, and five output levels-Udc/2, -Udc/4, 0, Ud/4 and Udc/2 are provided correspondingly, fig. 4 shows a voltage vector diagram corresponding to all three-phase output level combinations, wherein three levels-Udc/4, 0 and Ud/4 have two redundant states, and table 1 shows the level outputs in various switching states and the influences on the midpoint potential and the floating capacitor voltage. It can be seen from the table that although the output voltages of two redundant states of the same level have the same effect, the influence on the midpoint potential and the floating capacitor voltage is opposite.

TABLE 1 Effect of the switch states

According to fig. 3, a mathematical model of a five-level active midpoint clamped inverter is established:

I_np＝I_c1-I_c2＝H^TI (3)

wherein the content of the first and second substances,

wherein e_xRepresents a supply voltage; i.e. i_xRepresenting the grid-connected current; u. of_xoRepresenting an inverter output voltage; u. of_c1,u_c2，i_c1，i_c2Representing the voltage and the current of a direct-current side capacitor; u. of_fx，i_fxRepresenting the voltage and the current of the suspension capacitor; u. of_cmRepresents a common mode voltage; i.e. i_npRepresents the current flowing through the neutral point; l and R represent filter inductance and resistance; c, C_fRepresenting the direct current side capacitance and the suspension capacitance; h is_x,f_xRepresenting the influence factor of the phase current on the midpoint potential on the DC side and the voltage of the floating capacitor, where x ∈ { a, b, c }.

Discretizing the equations (1), (2) and (4) by applying the forward Euler equation, T_sFor a sampling time interval, k represents the current sampling moment, and the predicted values of the output current, the direct current side capacitance voltage difference and the suspension capacitance voltage at the next moment are deduced by using a mathematical model under a discrete domain as follows:

the above predicted values are obtained by bringing different switching states into the inverter, and it should be noted that, according to table (1), two redundant states of the same level act differently on the balanced midpoint potential and the floating capacitor voltage, so that eight states are brought into each phase instead of five for optimal control effect.

The reference value of the controlled variable at the sampling moment of K +1 is as follows:

the prediction of the reference value adopts a third-order Lagrange extrapolation method:

constructing a cost function by using the obtained predicted value and the reference value of the variable, and adding output current control, midpoint potential control and suspension capacitor voltage control into one cost function:

wherein λ is_iA weighting factor representing the overall effect of the control quantity.

Combining the analysis, a complete cost function is constructed according to a prediction model of the five-level active midpoint clamp type inverter, and the subsequent work is to bring all the switch state evaluation cost functions, find the switch state which enables the cost function to be minimum and serve as an execution target.

Because the effect of the redundant state of the same level on the control target is different, 8^3 ^ 512 states are required to be taken in the optimization process of each sampling period for calculation, huge calculation amount consumes a large amount of calculation resources, control delay is caused, and tracking error is increased.

In order to solve the above problems, the local optimization model prediction control method provided by the invention is suitable for a five-level grid-connected inverter, and simplifies global optimization into optimization of a target triangular region, reduces the optimization region to 1/96, and reduces the calculation amount, as shown in fig. 2. The method comprises the following specific steps:

step 200: discretizing through establishing a mathematical model to obtain a prediction model of the output current of the inverter; the mathematical model is a model of a grid-connected process of a five-level active neutral point clamped inverter; the prediction model is used for predicting the output value of the controlled variable at the next sampling moment by using the sampling value of the variable at the current sampling moment; in particular, the method comprises the following steps of,

step 201: constructing a dq vector according to a given reference value of the output current of the inverter, performing vector operation, and calculating a voltage reference vector output by the inverter;

step 202: judging a small triangular area of an inverter voltage reference vector in a vector diagram by utilizing a vector sector positioning method on an output voltage vector diagram of a five-level inverter;

step 203: sequentially substituting all switch states in the target small triangular area into a prediction model of the output current of the inverter to obtain a predicted value;

step 204: and substituting the predicted value into a cost function, and comparing the cost function with a reference value of the output current of the inverter, so that the switching state which enables the cost function to be minimum and meets the switching limiting condition acts on the system.

Specifically, the vector sector positioning method shown in fig. 1 is used to determine a small triangular region where the inverter voltage reference vector is located in the vector diagram. Firstly, calculating a reference vector of the output voltage of the inverter according to the reference current value, and converting the grid voltage e_a(k),e_b(k),e_c(k) Clark and park transformation to obtain e_d(k),e_q(k) (ii) a Will output a reference current

Subjected to Clark and park transformation to obtainConstructing a dq vector:

the inverter reference voltage dq vector is obtained by ignoring the common mode voltage according to equation (1):

calculating the components of the output reference voltage dq vector of the inverter

Obtaining the components of the reference vector of the output voltage by inverse park transformationThe inverter output voltage reference vector is obtained by constructing the dq vector and carrying out vector operation, so that derivation operation in the operation process can be avoided. As shown in fig. 4, based on the components of the voltage reference vectorThe phase angle ω of (c) can determine the large sector in which the voltage reference vector is located.

The reference vector of the output voltage of the inverter is obtained by carrying out gh coordinate transformationIn the gh coordinate system

And rounding the unit length in the vector diagram of the component pair to determine a diamond area or a table boundary area in the large sector where the component pair is located, wherein the table boundary area refers to an independent small triangular area at the outermost periphery of the vector diagram. If the reference vector is in the table boundary region, directly obtaining a small triangular region of the reference vector in the vector diagram; if the reference vector is in the diamond region, it will be in the gh coordinate system

The target diamond area can be translated to the central diamond area by taking the remainder of the unit length in the component pair vector diagram; and finally, determining which triangular area of the rhombic area the voltage reference vector is positioned in by one-step simple judgment. Taking the reference vector located in the large sector I as shown in FIG. 5 as an example, the reference vector is obtained by matchingRounding can easily determine that the diamond region A is located, and thenTaking the remainder of the unit length, the diamond area A can be translated to the central diamond area B, and then if u is judged_h＜1-u_gThe reference vector is located in the diamond-shaped area aTriangle area No. 1, otherwise, triangle area No. 2. Other large sectors can be switched to the sector I for the same judgment.

Through the judgment, the small triangular area where any reference vector is located can be determined. And then all preset switch states in the small triangular area are brought into the formulas (9), (10) and (11) to obtain predicted values, the predicted values are substituted into a cost function (14) for judgment, and the switch state which enables the cost function to be minimum and meets the switch switching limiting conditions is screened, namely the optimal control is obtained.

As a preferred scheme of the invention, the local optimization model predictive control method further comprises the steps of constructing a multivariable cost function, and substituting the output current of the inverter, the voltage difference of the direct-current side capacitor and the predicted value of each phase of the suspension capacitor voltage of the bridge arm into the multivariable cost function;

the multivariable cost function is realized by adding the inverter output current control, the DC side midpoint potential control and the suspension capacitor voltage control of each phase bridge arm into one cost function.

And the output current of the inverter, the voltage difference of the direct current side capacitor and the predicted value of the voltage of the suspension capacitor of each phase bridge arm are respectively obtained through respective prediction models.

The prediction models of the output current of the inverter, the difference of the direct current side capacitance voltage and the suspension capacitance voltage of each phase bridge arm are respectively as follows:

wherein the content of the first and second substances,the predicted value of the inverter output phase current at the moment of k + 1; i.e. i_x(k) For outputting phase power for inverter at time kA stream; e.g. of the type_x(k) The voltage of the power grid phase at the moment k; u. of_x(k) Outputting phase voltage for the inverter at the moment k; Δ u_dc ^P(k +1) is a predicted value of the voltage difference between the upper capacitor and the lower capacitor on the direct current side at the moment of k + 1; Δ u_dc(k) The voltage difference between the upper capacitor and the lower capacitor on the direct current side at the moment k;predicting the voltage of each bridge arm suspension capacitor at the moment k + 1; u. of_fx(k) The voltage of each bridge arm suspension capacitor at the time k; t is_sIs a sampling period; r, L is a grid-connected resistor and an inductor; c, C_fRespectively a direct current side capacitor and a suspension capacitor; h is_xIs the influence factor of the phase current on the midpoint potential; f. of_xThe values of the influence factors are determined according to different influences of the current on the midpoint potential and the voltage of the suspension capacitor under various switch states, and are shown in table 1, wherein x ∈ { a, b, c }.

The multivariate cost function is:

wherein the content of the first and second substances,

respectively representing an inverter output current reference value, a direct current side upper and lower capacitor voltage difference reference value and a suspension capacitor voltage reference value; lambda [ alpha ]_iIndicating the index. The value of the influence factor of the current on the suspension capacitor is determined according to the influence of the current on the midpoint potential and the voltage of the suspension capacitor under various switch states.

The local optimization model predictive control solves the global optimization problem into the local optimization problem in the reference area by judging the sector of the voltage reference vector in advance, thereby greatly reducing the calculation amount, improving the control efficiency and increasing the realizability of hardware. Besides controlling the grid-connected current of the inverter, the invention also obtains a prediction model by establishing a mathematical model of the midpoint potential and the suspension capacitor voltage on the direct current side and carrying out discretization treatment, constructs a multivariable cost function, adds the output current control, the midpoint potential control and the suspension capacitor voltage control into one cost function, and simultaneously plays a good control effect on the midpoint potential and the suspension capacitor voltage on the direct current side on the premise of realizing the grid-connected current control.

Fig. 6 is a schematic structure of a local optimization model prediction control system of a five-level active midpoint clamping inverter, where the device includes a dc voltage source, two series capacitors, a main circuit module, a first voltage sensor, a second voltage sensor, a third voltage sensor, a fourth voltage sensor, a current sensor, a vector calculation module, a target reference sector judgment module, and a prediction calculation module; the DC voltage source supplies power to the two series capacitors.

The first sensor and the second sensor are used for respectively collecting voltages at two ends of the direct current side series capacitor; the third sensor is used for collecting the voltage of the suspension capacitor of the bridge arm; the fourth sensor is used for collecting the voltage of the network side; the current sensor is used for collecting the grid-connected current output by the inverter; the main circuit module is used for realizing electric energy conversion and grid connection.

The vector calculation module is used for constructing a dq vector and carrying out vector operation to calculate an inverter output voltage reference vector; the reference sector judging module is used for judging a small triangular area where the reference vector is located in the vector diagram; the prediction calculation module is used for combining information collected by the first sensor, the second sensor, the third sensor, the fourth sensor and the current sensor with a switch state in a small triangular area to predict a variable reference value, and bringing the variable reference value into a cost function to perform optimization judgment so as to find an optimal control target.

The reference sector judgment mode is realized by using a vector sector positioning method on a five-level inverter output voltage vector diagram, and the specific steps of the vector sector positioning method are shown in figure 1.

And (3) utilizing an MATLAB/Simulink module to set up model prediction control simulation of the five-level active midpoint clamping type inverter, and verifying the feasibility of the proposed control method, wherein the simulation system parameters are shown in Table 1.

TABLE 1 simulation parameters

Parameter(s)	Size and breadth
		Voltage on the direct current side	7.5kV
Network phase voltage	4.16kV
		Frequency of the grid	50Hz
Sampling frequency	20kHz
		DC side capacitor	4700uF
Suspension capacitor	1500uF
		Filter inductor	2mH
Resistance (RC)	4Ω

Fig. 7 shows the output effect of the five-level inverter improved model predictive control, and it can be seen that the waveform of the grid-connected current is regular, fig. 8 performs spectrum analysis on the output current, the THD is only 0.81%, the distortion and the ripple are small, and simultaneously, the upper and lower capacitor voltages on the direct current side and the suspended capacitor voltage of each phase of bridge arm are all stabilized near the reference value. Meanwhile, as shown in fig. 7, when the set current is suddenly increased at 0.05S and decreased at 0.1S, the set current is suddenly decreased at 300A, and the dynamic response capability of the system is observed, it can be seen that the transition time when the reference current is suddenly increased is 0.4ms, the transition time when the reference current is suddenly decreased is 0.2ms, the output quickly follows the set when the set suddenly changes, and the dynamic response is quick. Fig. 9 shows the number of times of cost function optimization calculation performed in each sampling period to calculate the optimal control, and it can be seen that, while the conventional limited control set model prediction control needs to calculate 512 times in each sampling period, the improved local optimization model prediction control needs to calculate only 19.3 times in each sampling period on average, which reduces the calculation load by 26 times, saves the calculation resources, and reduces the system delay.

In the description provided herein, numerous specific details are set forth. It is understood, however, that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been shown in detail in order not to obscure an understanding of this description.

Furthermore, those skilled in the art will appreciate that while some embodiments described herein include some features included in other embodiments, rather than others, combinations of features of different embodiments are also meant to be within the scope of the invention and form different embodiments. For example, in the above embodiments, those skilled in the art can use the combination according to the known technical solutions and technical problems to be solved by the present application.

Although the present invention has been described with reference to a preferred embodiment, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.