KR20180077700A - Control method and apparatus for operating multilevel inverter - Google Patents

Abstract

Provided is a model predictive control method for operating a three-phase multilevel inverter using a cascade H-bridge (CHB). According to an embodiment of the present invention, the model predictive control method for operating a three-phase multilevel inverter using a CHB comprises the steps of: calculating a load current vector (i(k+1)) of a next step, based on a load of the CHB, and a two-phase load current vector (i(k)) and load voltage vector (v(k)) converted from a three-phase current and voltage of the load of the CHB; calculating a reference voltage vector (v^*(k+1)), based on a reference current vector (i^*(k)), a reference value which the i(k) value follows, the calculated i(k+1) value, and the load of the CHB; selecting a plurality of candidate voltage vectors adjacent to the v^*(k+1) value among a plurality of voltage vectors generated by the inverter, by a certain criterion; and determining a load voltage vector (v(k+1)) of a next step among the plurality of candidate voltage vectors, by using the v^*(k+1) value and a voltage of each of the plurality of candidate voltage vectors. Thus, improved dynamic response characteristics can be ensured.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to a control method and apparatus for operating a multilevel inverter,

The present invention relates to a model predictive control method and apparatus for operating a three-phase multilevel inverter using a cascaded H-bridge, and more particularly to a three-phase multilevel inverter that exhibits excellent dynamic response characteristics while reducing computational complexity And more particularly, to a model predictive control method and apparatus.

The model predictive control method (MPC) for operating a multi-level inverter requires a computational complexity proportional to the total number of voltage vectors generated by the inverter. Due to the characteristics of the MPC, as the level of the cascaded H-bridges (CHB) increases, the required amount of computation correspondingly increases explosively.

However, due to the time limitation such as the sampling time and the increase in the calculation amount, it is a reality that it is not easy to use all the voltage vectors generated by the inverter to find the optimal voltage vector.

In order to reduce the number of voltage vectors used for finding the optimal voltage vector, it is also possible to suppress the increase of the calculation amount by using only the 7 voltage vectors adjacent to the optimum voltage vector at the present stage, The MPC-adj method has been studied. However, the MPC-adj method has a drawback that the dynamic response characteristic is not excellent.

Therefore, there is a need for a model predictive control method and apparatus for operating a three-phase multilevel inverter so as to exhibit excellent dynamic response characteristics while reducing the amount of computation.

Related Prior Art Korean Patent Laid-Open No. 10-2016-0029914 (entitled H-bridge multilevel inverter control device and its operation method, published on Mar. 16, 2016) is known.

The present invention seeks to provide a model predictive control method and apparatus for operating a three-phase multilevel inverter using cascaded H-bridges to exhibit excellent dynamic response characteristics while reducing the amount of computation.

The problems to be solved by the present invention are not limited to the above-mentioned problem (s), and another problem (s) not mentioned can be clearly understood by those skilled in the art from the following description.

In order to achieve the above object, the present invention provides a model predictive control method for operating a three-phase multi-level inverter using cascaded H-bridges, Level inverter in a discrete time domain, the method comprising: converting a load of the CHB, a current and a voltage of three phases of the load of the CHB, Calculating a load current vector i (k + 1) of the next step based on the two-phase load current vector i (k) and the load voltage vector v (k); Based on the reference current vector i ^* (k) that is a reference to which i (k) is followed, the calculated i (k + 1), and the load of the CHB, the reference voltage vector v ^* +1)); Selecting a plurality of candidate voltage vectors adjacent to the v ^* (k + 1) by a predetermined reference among a plurality of voltage vectors generated by the inverter; And determining a load voltage vector v (k + 1) of the next step using the voltage of each of the plurality of candidate voltage vectors and v ^* (k + 1) among the plurality of candidate voltage vectors do.

Preferably, the load current vector and the load voltage vector of the two phases can be generated by alpha-beta transforming each of the three phases current and voltage.

Preferably, with respect to the V _α and V _β that the plurality of voltage vectors constituting the voltage vectors, on the αβ plane consisting of a second axis corresponding to the first axis and the V _β corresponding to the V _α, the inverter When a plurality of candidate voltage vectors are connected to adjacent voltage vectors among the plurality of voltage vectors, when v ^* (k + 1 (k + 1) ) Corresponding to the vertices constituting the smallest triangle.

Preferably, when dividing the cube into six equilateral triangles sharing the center of the cube as vertexes and dividing the six equilateral triangles into a first region to a sixth region, v ^* (k + 1) 2 region to the sixth region, it can be rotated with respect to the origin of the?? Plane so as to be located in the first region.

Preferably, when the load of the CHB includes the resistive component and the inductor component, the step of calculating the load current vector i (k + 1) of the next step may include calculating the resistance component and the inductor component using the resistive component and the inductor component, Deriving a first relationship of i (t) and v (t) in the region (t); (K), the i (k + 1) and the sampling time (T _s ) by approximating the differential component of the load current vector included in the first relational expression to the discrete time domain k To a second relationship in the second equation; And calculating the i (k + 1) using the second relation, the resistance component, the inductor component, the i (k), the v (k), and the t _s .

Preferably, i (k + 1) can be calculated by Equation (1).

[Equation 1]

Here, the i _αβ [k + 1] are the following, and the step load current vector (i (k + 1)) , i αβ [k] is the load current vector (i (k)), v αβ [k] Is the load voltage vector v (k), R is the resistive component, L is the inductor component, and T _s is the sampling time.

Preferably, the step of calculating the reference voltage vector v ^* (k + 1) of the next step uses the Lagrange extrapolation to calculate i ^* (k), i ^* (k-1) Sequentially calculating i ^* (k + 1) and i ^* (k + 2) from i ^* (k-2); And based on the resistance component (R) of the CHB load, the inductor component (L) of the CHB load, the i ^* (k + 2), wherein i (k + 1) and sampling time (T _s), And calculating the v ^* (k + 1).

Preferably, the step of determining the load voltage vector (v (k + 1)) of the next step includes calculating a value of a cost function indicating a difference between v ^* (k + 1) The voltage vector to be made smaller can be determined as v (k + 1).

Preferably, the cost function can be calculated by Equation (2).

&Quot; (2) "

Here, g _v is the cost ^{_{function, v α * [k + 1}} ] and v _β ^* [k + 1], respectively is two components constituting the ^{v * (k + 1),} v α [k + 1] and v [ _beta ] [k + 1] are the two components constituting the above v (k + 1).

In order to achieve the above object, the present invention provides a model predictive controller for operating a three-phase multilevel inverter using cascaded H-bridges, comprising: a three-phase multi-level inverter using cascaded H- (I (k)) and a load voltage vector (v (k)) transformed from the load of CHB, the current and voltage of the three phases of the load of the CHB, (k + 1)) of the next step on the basis of the reference current vector (i ^* (k)), the reference current vector i ^* (k) a calculation unit for calculating a reference voltage vector v ^* (k + 1) of the next step based on i (k + 1) and the load of CHB; A selector for selecting among a plurality of voltage vectors generated by the inverter a plurality of candidate voltage vectors adjacent to the v ^* (k + 1) by a predetermined reference; And a determination section for determining a load voltage vector v (k + 1) of the next step using the voltages of v ^* (k + 1) and each of the plurality of candidate voltage vectors, among the plurality of candidate voltage vectors do.

Preferably, the load current vector and the load voltage vector of the two phases may be generated by alpha-beta conversion of the current and voltage of the three phases, respectively.

Preferably, with respect to the V _α and V _β that the plurality of voltage vectors constituting the voltage vectors, on the αβ plane consisting of a second axis corresponding to the first axis and the V _β corresponding to the V _α, the inverter When a plurality of candidate voltage vectors are connected to adjacent voltage vectors among the plurality of voltage vectors, when v ^* (k + 1 (k + 1) ) Corresponding to the vertices constituting the smallest triangle.

Preferably, when dividing the cube into six equilateral triangles sharing the center of the cube as vertexes and dividing the six equilateral triangles into a first region to a sixth region, v ^* (k + 1) 2 region to the sixth region, it can be rotated with respect to the origin of the?? Plane so as to be located in the first region.

Preferably, when the load of the CHB includes a resistive component and an inductor component, the calculation unit calculates the i (t) and the v (t) in the continuous time domain (t) using the resistive component and the inductor component, (K), i (k + 1) and sampling time (T _s ) of the differential component of the load current vector included in the first relational expression, (K) by using the second relational expression, the resistance component, the inductor component, the i (k), the v (k) and the t _s in the discrete time domain (k) It is possible to calculate i (k + 1).

Preferably, i (k + 1) can be calculated by Equation (3).

&Quot; (3) "

Here, the i _αβ [k + 1] are the following, and the step load current vector (i (k + 1)) , i αβ [k] is the load current vector (i (k)), v αβ [k] Is the load voltage vector v (k), R is the resistive component, L is the inductor component, and T _s is the sampling time.

Preferably, the calculation unit using the Lagrange extrapolation, the ^{i * (k), i *} (k-1) and i ^* i from the ^{(k-2) * (k} + 1) and i ^* (k + 2) the resistance component of the CHB load (R calculated in sequence, and), the inductor element (L) of the CHB load, the i ^* (k + 2), wherein i (k + 1) and sampling time ( T _s ), the above v ^* (k + 1) can be calculated.

Preferably, the determination unit may determine, as v (k + 1), a voltage vector that minimizes a value of a cost function indicating a difference from v ^* (k + 1) among the plurality of candidate voltage vectors .

Preferably, the cost function can be calculated by Equation (4).

&Quot; (4) "

Here, g _v is the cost ^{_{function, v α * [k + 1}} ] and v _β ^* [k + 1], respectively is two components constituting the ^{v * (k + 1),} v α [k + 1] and v [ _beta ] [k + 1] are the two components constituting the above v (k + 1).

When a three-phase multilevel inverter using a cascaded H-bridge is operated by using the model predictive control method and apparatus according to an embodiment of the present invention, it is possible to secure an excellent dynamic response characteristic while reducing a calculation amount .

FIG. 1 is a flowchart for explaining a model predictive control method for operating a three-phase multilevel inverter using cascaded H-bridges according to an embodiment of the present invention.
Fig. 2 is a flowchart for explaining a method of calculating a load current vector i (k + 1) in the next step according to an embodiment of the present invention.
3 is a view for explaining a model predictive controller for operating a three-phase multi-level inverter using cascaded H-bridges according to an embodiment of the present invention.
4 is a diagram for explaining the operation of a model predictive controller for operating a three-phase multilevel inverter using cascaded H-bridges according to an embodiment of the present invention.
5 is a diagram for explaining a circuit configuration of a 3-phase multilevel inverter using CHB according to an embodiment of the present invention.
6 is a view for explaining a form of a voltage vector generated by a multi-level inverter according to an embodiment of the present invention.
7 is a diagram for explaining a method of rotating a reference voltage vector v ^* (k + 1) in the next step according to an embodiment of the present invention.
8 is a view for explaining the configuration and the structure of a first sector (sector 1) according to an embodiment of the present invention.
Fig. 9 is a diagram for explaining a corresponding region and a candidate voltage vector according to a component constituting the reference voltage vector v ^* (k + 1) in the next step according to an embodiment of the present invention.
10 is a diagram for explaining a candidate voltage vector corresponding to each sector according to an embodiment of the present invention.
11 is a diagram for explaining the dynamic response characteristic of the model predictive control method according to an embodiment of the present invention in comparison with the conventional technique (MPC-adj).

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals are used for like elements in describing each drawing.

The terms first, second, A, B, etc. may be used to describe various elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a flowchart for explaining a model predictive control method (MPC method) for operating a three-phase multilevel inverter using a cascaded H-bridge according to an embodiment of the present invention.

In step S110, the model predictive control apparatus determines whether or not the two-phase load current vector i (k) and the load voltage vector v (k) are converted from the load of CHB, the current and voltage of the three phases of the load of CHB, , The load current vector i (k + 1) of the next step is calculated.

For example, referring to Fig. 5, a circuit diagram of a 3-phase (a, b, c) multilevel inverter including j H-bridges connected in series is shown. At this time, each CHB is connected to a load in which a resistance component (R) and an inductor component (L) are connected in series. On the other hand, the model predictive controller controls four switches (e.g., _Sa11 , _Sa12 , _Sa13 , _Sa14 ) included in the total j cells for each CHB, It can be operated.

Here, letting a current flowing in the load in a phase be i _a , the voltage v _ao applied to the load in the continuous time region can be calculated by the following equation (1).

[Equation 1]

Where v _ao is the voltage across the load, R is the resistive component of the load, i _a is the current flowing through the load, and L is the inductor component of the load.

On the other hand, the details of calculating the load current vector i (k + 1) of the next step using the equation (1) by the model predictive controller will be described later in detail in the following embodiments and FIG. 3.

In another embodiment, the two-phase load current vector and the load voltage vector may be generated by alpha-beta transforming each of the three phase current and voltage.

Alpha Beta conversion is a mathematical conversion method used to simplify the analysis of three-phase circuits, conceptually similar to dq0 conversion. In particular, the alpha beta conversion is useful in that it can generate reference vectors that are used for the control of three-phase inverters.

For example, referring to FIG. 5, when the currents flowing through the respective three-phase loads a, b, and c are denoted by i _a , i _b , and i _c , the alpha beta transformed i _α and i _β denote the current vector i _αβ And can be calculated using the following equation (2).

&Quot; (2) "

Here, the 3x2 matrix is a simplified alpha beta transformation matrix.

Equation 1 can be rewritten with respect to the current vector and the voltage vector using Equation (3) using the alpha beta conversion result.

&Quot; (3) "

Here, i is an _αβ load consisting of i _α and i _β current vector, v _αβ v is a vector consisting of the load voltage and the _α v _β.

At this time, the model predictive controller sets the differential value in the continuous time domain for i _{alpha beta} in Equation (3) to the load current vector (i (k + 1)) in the next step in the discrete time domain, i (k)) and the sampling time (T _s ), the load current vector i (k + 1) of the next step can be predicted and calculated. Concrete contents related to this will be described later in detail with reference to FIG.

In step S120, based on the reference current vector i ^* (k), the calculated i (k + 1), and the load of CHB, which is a reference to which i (k) And calculates a vector v ^* (k + 1).

That is, the model predictive controller calculates the reference current vectors i ^* (k + 1) and i ^* (k + 1) of the next step and the next step using the extrapolation from the reference current vectors of the past step and the present step, 2) can be predicted. On the other hand, extrapolation is also called extrapolation, and is a method of estimating data on an unknown region using data in regions close to the unknown region. For example, Richardson extrapolation, Lagrangian extrapolation, and the like.

Then, the model predictive controller uses the reference current vector i ^* (k + 2) of the second next step, the load of i (k + 1) and CHB calculated in the preceding step S110, It is possible to calculate the voltage vector v ^* (k + 1). At this time, the model predictive controller can calculate v ^* (k + 1) using Equation (4) which is an approximate expression in the discrete time domain derived from Equation (3).

&Quot; (4) "

Here, v _αβ ^* a (k + 1) is a reference to the next step voltage vector, L is the inductor component of the CHB load, Ts is a sampling ^{_{time, i αβ * (k + 2}} ) will see the second following step a current vector is, _αβ i (k + 1) is the next step load current vector, R is the resistance component of the load of the CHB.

Equation (4) can be derived on the assumption that the reference voltage vector and the reference current vector are the same as the load voltage vector and the load current vector in the next step (k + 1) or the second next step (k + 2), respectively.

In another embodiment, when the model predictive controller calculates v ^* (k + 1), it can be calculated using Lagrange extrapolation.

The Lagrangian extrapolation method can predict the reference current vector of the next step by using the following equation (5), and the model predictive controller can calculate the resistance component (R) of the load of the CHB, the inductor component (L) It can be calculated the _αβ i (k + 1) and sampling time (T _s) applied to the equation _{^{4 v * αβ (k + 1}} ).

&Quot; (5) "

That is, the model predictive controller calculates the reference current vector i? _Beta ^* (k + 1) from the reference current vector of the past two steps (k-1, k- to predict: 1)) and, if repeated, it may then predict the second current reference vector ^{_{(i αβ * (k + 2}} )) of the step (k + 2).

In step S130, the model predictive controller selects a plurality of candidate voltage vectors adjacent to v ^* (k + 1) from the plurality of voltage vectors generated by the inverter by a predetermined criterion.

For example, referring to FIG. 6, there are shown results of a plurality of voltage vectors generated by a three-phase multilevel inverter using CHB. In the case of the 5th level, a plurality of voltage vectors corresponding to the intersections of the green cubes are generated. In the seventh level, a plurality of voltage vectors corresponding to the intersections of the blue cubes are generated. A plurality of voltage vectors are generated. At this time, the voltage vector in Fig. 6 is the result shown on the alpha beta plane for the first axis corresponding to alpha and the second axis corresponding to beta, for the voltage vector to which the alpha beta transform is applied.

Here, when the model predictive controller puts the previously calculated v ^* (k + 1) on the area where the plurality of voltage vectors are shown, the adjacent voltage vectors are divided into a plurality of candidate voltage vectors . Concrete contents related to a method of selecting a plurality of candidate voltage vectors will be described in detail below in the description of the embodiments.

In another embodiment, for a plurality of voltage vectors V _alpha and V _beta constituting the voltage vector, on an _? Beta plane constituted by a first axis corresponding to V _{? And} a second axis corresponding to V _? When a plurality of candidate voltage vectors are connected to adjacent voltage vectors among the plurality of voltage vectors, when v ^* (k + 1) is placed at a predetermined interval within the boundary and inside of a square of a size proportional to the level And may be three voltage vectors corresponding to the vertices constituting the smallest triangle included.

For example, referring to FIG. 6, when a plurality of voltage vectors are located at the boundary and inside of a cube, the model predictive controller connects the three voltage vectors to each of the plurality of voltage vectors, The triangles of At this time, the plurality of triangles are the minimum triangles generated as a result of connecting the most adjacent voltage vectors, and the model predictive controller can select three voltage vectors corresponding to the vertices of the minimum triangles as candidate voltage vectors have.

That is, FIG. 9, the model predictive control unit is V _α ^* (k + 1) and V _β ^* from three inequalities related to (k + 1), a triangle is the v ^* (k + 1) calculated previously in Region (e.g., T ₁ ). Further, the model predictive controller can select three voltage vectors (e.g., V ₀ , V ₁ , and V ₂ corresponding to T ₁ ) corresponding to the vertices of the triangular area as candidate voltage vectors.

Finally, in step S140, the model predictive controller calculates the load voltage vector v (k + 1) of the next step using the voltages of v ^* (k + 1) and the plurality of candidate voltage vectors from among the plurality of candidate voltage vectors, )).

That is, the model predictive controller uses the voltages of v ^* (k + 1) and the plurality of candidate voltage vectors calculated in the previous step, and calculates the load voltage vector v (k +1)).

For example, in the model predictive control method according to an embodiment of the present invention, the calculation is performed only on the three candidate voltage vectors corresponding to the vertexes of the triangular area. Therefore, the total calculation amount can be reduced, The dynamic response characteristic can be improved.

Concrete contents related to the method of determining v (k + 1) will be described in detail in the description of the embodiments below.

In another embodiment, when the model predictive controller determines v (k + 1), among the plurality of candidate voltage vectors, a voltage vector that minimizes the value of the cost function indicating the difference from v ^* (k + 1) Can be determined as v (k + 1).

At this time, the cost function may be a result of performing a vector operation on the difference between the candidate voltage vector and v ^* (k + 1). That is, the model predictive controller calculates a result of the cost function by performing a vector operation on the difference between v ^* (k + 1) for each of the plurality of candidate voltage vectors, and calculates a cost function (K + 1) can be determined as the candidate voltage vector having the smallest result value of v (k + 1).

8, assuming that the arrow starting from the origin indicates v ^* (k + 1), the model predictive control apparatus determines the position of the triangle T ₁₁ including v ^* (k + 1) The cost function can be calculated for three candidate voltage vectors V ₂₀ , V ₂₁ , and V ₃₉ corresponding to the three vertices (see FIG. 10), and v (k + 1) can be determined.

On the other hand, since the cost function may vary depending on the design of the function, depending on the design of the cost function, the model predictive controller may determine the voltage vector that maximizes the value as v (k + 1) Of course it is possible.

In yet another embodiment, the cost function may be calculated by:

&Quot; (5) "

Here, g _v is the cost function, v _α ^* [k + 1] and v _β ^* a [k + 1] Each of the two components that make up the v ^* (k + 1), v _α [k + 1] And v _? [K + 1] are the two components constituting v (k + 1).

That is, the cost function is a result of the vector operation on the difference between the candidate voltage vector v (k + 1) selected from the load voltage vectors that can be generated by the inverter and the calculated reference voltage vector v ^* (k + 1) . For example, the difference between the (a1, b1) vector and the (a2, b2) vector can be calculated as the size of the (a1-a2, b1-b2) vector.

In another embodiment, when a cube including a plurality of voltage vectors is divided into six equilateral triangles sharing the center of the cube as vertexes and the six equilateral triangles are divided into the first to sixth regions, v ^* ( k + 1) can be rotated with respect to the origin of the? beta plane so as to be located in the first region when the first to sixth regions are located in the second to sixth regions.

That is, the model predictive controller can hold the value of the voltage vector corresponding to each intersection for a cube containing a plurality of voltage vectors as a look-up table. However, in order to minimize the size of the look-up table, the size of the look-up table becomes larger as the number of levels of the inverter increases. The cube is divided into first to sixth regions, It may be considered to retain the lookup table only for the voltage vector.

Therefore, if the model predictive control apparatus rotates the v ^* (k + 1) on the basis of the origin of the? Β plane when v ^* (k + 1) is located in the second to sixth regions, . For example, referring to FIG. 7, the model predictive controller can rotate v ^* (k + 1) located in the third region (sector 3) to the first region (sector 1).

Referring to FIG. 4, the overall operation of the model predictive control method according to another embodiment of the present invention is shown. That is, the flow of each variable corresponding to each step of the flowchart of FIG. 1 is shown.

That is, i (k + 1) is calculated in the Predict future current corresponding to step S110, v ^* (k + 1) is calculated in the Predict future reference voltage corresponding to step S120, the candidate for v (k + 1) is selected from the voltage vector, and the switching to the CHB is performed by determining v (k + 1) finally in the Minimize voltage based cost function corresponding to step S140.

Referring to FIG. 11, there is shown a result of comparing the model predictive control method according to an embodiment of the present invention with the prior art (MPC-adj). That is, when comparing the dynamic response characteristics of the load current to the phase a, the response time is 253.4 ms and 77.4 ms, respectively, which indicates that the present invention exhibits much better dynamic response characteristics.

As described above, the model predictive control method for operating the 3-phase multilevel inverter using the cascaded H-bridge according to the embodiment of the present invention, when operating the 3-phase multilevel inverter using the skid H- And an excellent dynamic response characteristic can be ensured.

Fig. 2 is a flowchart for explaining a method of calculating a load current vector i (k + 1) in the next step according to an embodiment of the present invention.

In step S210, the model predictive controller derives a first relational expression of i (t) and v (t) in the continuous time domain (t) using the resistive component and the inductor component.

At this time, the model predictive controller can derive the first relational expression from Equation (3).

In step S220, the model predictive controller approximates the differential component of the load current vector included in the first relational expression by i (k), i (k + 1) and sampling time (T _s ) To the second relation in the discrete time domain (k).

At this time, the approximation of the differential component can be performed using Equation (6) below, and the accuracy of the approximation can be improved as the sampling time (T _s ) is sufficiently small.

&Quot; (6) "

Here, i _{alpha beta} (k + 1) is the load current vector of the next step, i _{alpha beta} (k) is the load current vector of the current step, and T _s is the sampling time.

In this case, the second relational expression for the discrete time domain obtained by applying Equation (6) to Equation (3) and calculating an approximate value is calculated as Equation (7) below.

&Quot; (7) "

Here, i and _αβ [k + 1] is the next step in load current vector, i _αβ [k] is the current step of the load current vector, v _αβ [k] are present and the step of load voltage vector, R is resistance minutes , and L is the inductor component, T _s is the sampling time.

On the other hand, Equation (4) can be calculated from Equation (7) by applying the above-described assumption.

Finally, in step S230, the model predictive controller calculates i (k + 1) using the second relational expression, the resistance component, the inductor component, i (k), v (k), and t _s .

That is, the model predictive control device to the resistance component in the equation (7), an inductor component, by substituting the i (k), v (k ) and t _s, calculating the next step of the load current vector, i (k + 1) have.

3 is a view for explaining a model predictive controller for operating a three-phase multi-level inverter using cascaded H-bridges according to an embodiment of the present invention.

3, a model predictive controller 300 for operating a three-phase multilevel inverter using cascaded H-bridges according to an embodiment of the present invention includes a calculator 310, a selector 320, And a determination unit 330.

On the other hand, the model predictive controller 300 is installed in a device such as a PC, a notebook computer, an industrial computer, a smart phone, or a tablet PC, and operates the three-phase multilevel inverter through the device.

The calculation unit 310 calculates the load voltage vector v (k) based on the two-phase load current vector i (k) and the load voltage vector v (k) that are converted from the CHB load, the three- (K + 1)) based on the load of the reference current vector i ^* (k), the calculated i (k + 1), and the CHB, which are references to which i (k) , The reference voltage vector v ^* (k + 1) of the next step is calculated.

In another embodiment, the two-phase load current vector and the load voltage vector may be generated by alpha-beta conversion of the three phase current and voltage, respectively.

In another embodiment, when the load of the CHB includes the resistive component and the inductor component, the calculation portion 310 calculates the difference between the i (t) and the i (t) in the continuous time region t using the resistive component and the inductor component, v deriving a first relation of (t), and by the approximation with the the finely divided component of the vector of the load current comprises a first relation i (k), i (k + 1) and sampling time (t _s), the first changing the first equation into the second equation of the discrete-time domain (k), and the second relation, the resistance component, the inductor component, i (k), v (k) and using the t _s, i (k + 1 ) Can be calculated.

In yet another embodiment, i (k + 1) may be calculated by:

&Quot; (8) "

Here, i _αβ [k + 1] is the next step in load current vector (i (k + 1)) , i αβ [k] is a load current vector (i (k)), v αβ [k] is the load and the voltage vector (v (k)), and R is resistance minutes, L is the inductor component, T _s is the sampling time.

In yet another embodiment, the calculation unit 310 using the Lagrange extrapolation, i ^* (k), i ^* (k-1) and i ^* i from the ^{(k-2) * (k} + 1) and i ^* (k + 2) for the calculation, and the CHB load resistance component (R) sequentially, an inductor component of the CHB load ^{(L), i * (k} + 2), i (k + 1) and sampling time ( T _s ), it is possible to calculate v ^* (k + 1).

The selector 320 selects a plurality of candidate voltage vectors adjacent to v ^* (k + 1) according to a predetermined criterion among a plurality of voltage vectors generated by the inverter.

In another embodiment, on the basis of V _alpha and V _beta constituting the voltage vector, the plurality of voltage vectors are arranged on the _{alpha beta} plane constituted by the first axis corresponding to V _alpha and the second axis corresponding to V _beta , when located in the border and at predetermined intervals in the interior of the cube of the size which is proportional to, the plurality of candidate voltage vectors when connecting the voltage vector between adjacent among the plurality of voltage vector, v ^* contains a (k + 1) And the three voltage vectors corresponding to the vertices constituting the smallest triangle.

The determining unit 330 determines the load voltage vector v (k + 1) of the next step using the voltages of v ^* (k + 1) and the plurality of candidate voltage vectors, among the plurality of candidate voltage vectors do.

In another embodiment, when dividing a cube containing a plurality of voltage vectors into six equilateral triangles sharing the center of the cube as vertices and dividing the six equilateral triangles into the first to sixth regions, v ^* (k +1) can be rotated with respect to the origin of the? Beta plane so as to be located in the first region, if it is located in the second to sixth regions.

In another embodiment, the determination unit 330 determines, among the plurality of candidate voltage vectors, a voltage vector that minimizes the value of the cost function indicating the difference from v ^* (k + 1) to v (k + 1) You can decide.

In yet another embodiment, the cost function may be calculated by:

&Quot; (9) "

Here, g _v is the cost function, v _α ^* [k + 1] and v _β ^* a [k + 1] Each of the two components that make up the v ^* (k + 1), v _α [k + 1] And v _? [K + 1] are the two components constituting v (k + 1).

The above-described embodiments of the present invention can be embodied in a general-purpose digital computer that can be embodied as a program that can be executed by a computer and operates the program using a computer-readable recording medium.

The computer readable recording medium includes a magnetic storage medium (e.g., ROM, floppy disk, hard disk, etc.), optical reading medium (e.g., CD ROM, DVD, etc.).

The present invention has been described with reference to the preferred embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

Claims (10)

A method of model predictive control (MPC) for operating a three-phase multi-level inverter using a cascaded H-bridge (CHB) in a discrete time domain,
Based on the two-phase load current vector i (k) and the load voltage vector v (k) converted from the load of the CHB, the three-phase current and the voltage of the load of the CHB, Calculating a current vector i (k + 1);
Based on the reference current vector i ^* (k) that is a reference to which i (k) is followed, the calculated i (k + 1), and the load of the CHB, the reference voltage vector v ^* +1));
Selecting a plurality of candidate voltage vectors adjacent to the v ^* (k + 1) by a predetermined reference among a plurality of voltage vectors generated by the inverter; And
Determining a next step load voltage vector v (k + 1) using the voltages of v ^* (k + 1) and each of the plurality of candidate voltage vectors from the plurality of candidate voltage vectors,
Phase multi-level inverter using a cascaded H-bridge. The method according to claim 1,
The two-phase load current vector and the load voltage vector are
Phase inverter is generated by alpha-beta transforming the current and voltage of each of the three phases. 3. The method of claim 2,
Wherein the plurality of voltage vectors are proportional to the level of the inverter on an?? Plane constituted by a first axis corresponding to V _? And a second axis corresponding to V _{? With} respect to V _? And V _? And when they are positioned at a predetermined interval in the boundary and inside of the cube,
The plurality of candidate voltage vectors
(K + 1), when the adjacent voltage vectors are connected to each other among the plurality of voltage vectors, the three voltage vectors corresponding to the vertices constituting the smallest triangle including v ^* (k + 1) Model Predictive Control Method for Operating 3 - Phase Multi - Level Inverter Using Bridge. The method of claim 3,
Dividing the cube into six equilateral triangles that share the center of the cube as a vertex, and dividing the six equilateral triangles into a first region to a sixth region,
The v ^* (k + 1)
And the third region is rotated based on the origin of the?? Plane so as to be located in the first region when the first region is located in the second region to the sixth region. Model predictive control method. The method according to claim 1,
When the load of the CHB includes a resistive component and an inductor component,
The step of calculating the load current vector i (k + 1) of the next step
Deriving a first relationship between i (t) and v (t) in the continuous time domain (t) using the resistive component and the inductor component;
(K), the i (k + 1) and the sampling time (T _s ) by approximating the differential component of the load current vector included in the first relational expression to the discrete time domain k To a second relationship in the second equation; And
Calculating the i (k + 1) using the second relation, the resistance component, the inductor component, the i (k), the v (k), and the t _s ,
Phase multi-level inverter using a cascaded H-bridge. 6. The method of claim 5,
The i (k + 1)
Phase inverter using a cascaded H-bridge, which is calculated according to Equation (1).
[Equation 1]

Here, the i _αβ [k + 1] are the following, and the step load current vector (i (k + 1)) , i αβ [k] is the load current vector (i (k)), v αβ [k] Is the load voltage vector v (k), R is the resistive component, L is the inductor component, and T _s is the sampling time. The method according to claim 1,
The step of calculating the reference voltage vector v ^* (k + 1) of the next step
Lagrange the extrapolation using (Lagrange extrapolation), the ^{i * (k), i *} (k-1) and ^{i * i * (k + 1} ) and i ^* (k + 2) from the (k-2) Sequentially calculating; And
Based on the resistance component (R) of the CHB load, the inductor component (L) of the CHB load, the i ^* (k + 2), wherein i (k + 1) and sampling time (T _s), wherein calculating v ^* (k + 1)
Phase multi-level inverter using a cascaded H-bridge. The method according to claim 1,
The step of determining the load voltage vector v (k + 1) of the next step
(K + 1) is determined as the voltage vector that minimizes the value of the cost function indicating the difference from v ^* (k + 1) among the plurality of candidate voltage vectors. A Model Predictive Control Method for Operating 3 - Phase Multi - level Inverter Using Bridge. 9. The method of claim 8,
The cost function
Wherein the three-phase multi-level inverter is calculated using Equation (2).
&Quot; (2) "

Here, g _v is the cost ^{_{function, v α * [k + 1}} ] and v _β ^* [k + 1], respectively is two components constituting the ^{v * (k + 1),} v α [k + 1] and v [ _beta ] [k + 1] are the two components constituting the above v (k + 1). A model predictive controller for operating a three-phase multilevel inverter using a cascaded H-bridge (CHB) in a discrete time domain,
The load current vector (V (k)) of the next step based on the two-phase load current vector i (k) and the load voltage vector v (k) converted from the CHB load, the three- (k + 1)), based on the reference current vector (i ^* (k)) which is a reference followed by i (k), the calculated i (k + A calculation unit for calculating a reference voltage vector v ^* (k + 1) of the next step;
A selector for selecting among a plurality of voltage vectors generated by the inverter a plurality of candidate voltage vectors adjacent to the v ^* (k + 1) by a predetermined reference; And
(K (k + 1)) of the next step using the voltage of each of the plurality of candidate voltage vectors and v ^* (k + 1) among the plurality of candidate voltage vectors,
Phase multi-level inverter using the cascaded H-bridge.

Images