Two-degree-of-freedom prediction control method of LCL filtering type three-phase grid-connected inverter system

Abstract

The invention discloses a two-degree-of-freedom prediction control method of an LCL filtering type three-phase grid-connected inverter system, and belongs to the field of power electronic converter control. A prediction control reference instruction is obtained by sampling parameters of inverter side current, capacitor voltage, network side current and power grid voltage at the moment k, and a continuous and discrete model of an LCL filtering type three-phase grid-connected inverter system is established; designing a control function based on a grid-side current reference tracking error and a gradient thereof, and establishing a prediction model; constructing a prediction cost function based on a control function, and obtaining a prediction control reference instruction by using a minimum value of the prediction cost function; and determining an optimal voltage vector and an action duty ratio according to the prediction control reference instruction, generating a control pulse signal according to a symmetrical five-section pulse mode, and acting on each switching tube in the inverter to realize prediction control of the LCL filtering type three-phase grid-connected inverter system. The method is simple in step, convenient to use, constant in switching frequency and good in stability.

Description

Technical Field

[0001] The invention relates to a two-degree-of-freedom predictive control method for an LCL filtering type three-phase grid-connected inverter system, and belongs to the field of power electronic converter control. Background Art

[0002] As a bridge between new energy power generation devices and three-phase AC power grids, the three-phase grid-connected inverter system plays an important role in the process of power transmission. Compared with L filters, LCL filters have smaller size and lighter weight, so they are often connected between the inverter and the grid to filter out grid current harmonics and improve power quality. However, the introduction of LCL filters will not only increase the control order of the system, but also cause resonance problems on the grid-side current, which seriously threatens the stability and reliability of the photovoltaic grid-connected system. Therefore, various damping methods need to be introduced to solve this resonance problem.

[0003] The quality of the control algorithm is closely related to the quality of the grid-side current. Therefore, the control method of the grid-connected inverter has been widely studied. The traditional control method of the LCL filter type grid-connected inverter is mainly the traditional multi-objective single vector predictive control. For the traditional multi-objective single vector predictive control method, although there are few parameters to be adjusted and no additional damping is required, it is limited by the single vector predictive control method, the steady-state performance is still poor and the switching frequency is not fixed, which will be unfavorable for the design of the LCL filter. Summary of the invention

[0004] In view of the shortcomings of the prior art, a two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system is provided. By designing a control function based on the grid-side current reference tracking error and its gradient, multi-state variable control of capacitor voltage and grid-side current is inherently realized, and inherent resonance damping is achieved without the need for additional damping. In addition, the setting is simple by only introducing a weight factor m. The method implements a three-vector control method according to a predictive control reference instruction, which can improve the steady-state performance of the grid-side current, and adopts a symmetrical 5-segment pulse mode in the final wave generation stage to generate a control pulse signal and act on each switch tube of the inverter to achieve a fixed switching frequency.

[0005] To achieve the above technical objectives, the present invention discloses a two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system, comprising the following steps:

[0006] Step 1: Sample the state variables of the LCL filter type three-phase grid-connected inverter system. The state variable sampling includes the inverter side current i at time k 1,k , capacitor voltage v c,k , three-phase AC grid side current i 2,k and grid voltage v g,k, transform the sampled state variables from the abc coordinate system to the αβ coordinate system, where the grid voltage v in the abc coordinate system g,abc The voltage phase θ of the three-phase AC grid is obtained by combining the phase-locked loop, and the continuous domain model of the LCL filter type three-phase grid-connected inverter system in the αβ coordinate system is established. The LCL filter type three-phase grid-connected inverter system is discretized using a zero-order holder to obtain the system discrete model x of the LCL filter type three-phase grid-connected inverter system at time k+1. k+1 ;

[0007] Step 2: Design a control function r based on the grid-side current reference tracking error and its gradient, and calculate the control function r based on the system discrete model x. k+1 Establish a two-degree-of-freedom prediction model r at time k+1 k+1 , to reflect the error between the predicted value of the three-phase AC grid-side current and its gradient and the corresponding reference value;

[0008] Step 3: Design a prediction model based on two degrees of freedom k+1 The prediction cost function g is used as the two-degree-of-freedom prediction model r k+1 αβ axis tracking error;

[0009] Step 4: Based on the principle of solving the minimum value of the function, let the prediction cost function g be the inverter voltage v iαβ,k The partial derivative of is equal to 0, and the predictive control reference instruction that can minimize the prediction cost function g is obtained.

[0010] Step 5: Based on the predictive control reference instruction Axis components in the αβ coordinate system and Get the predictive control reference instruction The spatial angle position of the control unit and the prediction control reference command are determined. The voltage vector sector where the predictive control reference command is selected Two adjacent non-zero voltage vectors v in the voltage vector sector m 、v n , and the zero vector v7 at the origin of the voltage vector sector as the synthetic predictive control reference instruction Three optimal voltage vectors;

[0011] Step 6: Calculate the three optimal voltage vectors v m 、v n and V7 at their respective duty cycles, and compare the synthesized voltage with the predicted control reference instruction The square sum of the tracking errors is used to calculate the partial derivative of each duty cycle, and the three duty cycles d corresponding to the three optimal voltage vectors are calculated.m d n and d0;

[0012] Step 7: According to the three optimal voltage vectors v m 、v n and v7 and three action duty cycle d m d n And d0, generate a control pulse signal S according to the symmetrical 5-segment pulse mode, and apply the control pulse signal S to each switch tube in the inverter to realize the predictive control of the LCL filter type three-phase grid-connected inverter system.

[0013] Further, the step 1 specifically includes: establishing a continuous domain model of the LCL filter type three-phase grid-connected inverter system in the αβ coordinate system:

[0014]

[0015] In the formula, x=[i 1αβ v cαβ i 2αβ ] is the state variable of the LCL filter type three-phase grid-connected inverter system, u=[v iαβ v gαβ ] is the control input of the LCL filter type three-phase grid-connected inverter system, and the coefficient matrices A and B are expressed as follows:

[0016]

[0017] In the formula, i 1αβ 、v cαβ 、i 2αβ They represent the inverter side current, LCL capacitor voltage, and grid side current of the three-phase AC grid of the LCL filter type three-phase grid-connected inverter system under the α and β axes, respectively. iαβ represents the inverter voltage under αβ axis, v gαβ represents the αβ-axis grid voltage; L1 and R1 represent the filter inductance and parasitic resistance on the output side of the LCL filter type three-phase grid-connected inverter system, L2 and R2 represent the filter inductance and parasitic resistance on the grid side, and C represents the filter capacitor;

[0018] Assume that the sampling period of the LCL filter type three-phase grid-connected inverter system is T s , the continuous domain model is accurately discretized using a zero-order holder, and the obtained discrete model of the LCL filter type three-phase grid-connected inverter system is expressed as:

[0019]

[0020] Wherein, the subscripts "k" and "k+1" represent the k and k+1 moments respectively. k =[i 1αβ,k vcαβ,k i 2αβ,k ] T represents the actual values ​​of the inverter side current, capacitor voltage and grid side current sampled at time k, x k+1 =[i 1αβ,k+1 v cαβ,k+1 i 2αβ,k+1 ] T is the predicted value at time k+1, u k =[v iαβ,k v gαβ,k ] T represents the inverter voltage and grid voltage at time k, A d , B d Coefficient matrix computed for the zero-order holder method: B d =A -1 (A d -I6)B, where e is the base of the natural logarithm, I6 is the sixth-order unit matrix, and a and b represent the coefficient matrix A respectively. d and B d The elements in , the subscripts represent their row and column position information in the coefficient matrix.

[0021] Furthermore, the step 2 specifically includes: designing a control function r based on the grid-side current reference tracking error and its gradient tracking, which is expressed as follows:

[0022]

[0023] In the formula, r = [r α r β ] represents the control function matrix, i 2αβ =[i 2α i 2β ]、 are the grid-side currents of α and β axes and their reference matrices, v cαβ =[v cα v cβ ]、v gαβ =[v gα v gβ ] are the αβ axis capacitor voltage matrix and the grid voltage matrix respectively, and m is the weight factor;

[0024] Based on the control function r and the system discrete model, the two-degree-of-freedom prediction model at time k+1 is constructed using the state variable sampling information at time k as follows:

[0025]

[0026] In the formula, i 2αβ，k+1 、v cαβ，k+1 and v gαβ，k+1represents the predicted value of the current and capacitor voltage on the αβ axis grid side at time k+1, from the system discrete model x at time k+1 k+1 Directly obtain, k+1 The control degrees of freedom include the grid-side current tracking error at time k+1 and the gradient of the grid-side current tracking error.

[0027] Furthermore, the step 3 specifically includes: in order to achieve accurate predictive control, it is necessary to control the αβ axis tracking errors of the two-degree-of-freedom prediction model to be 0, that is, assuming that the control reference value is and The prediction cost function g based on the two-degree-of-freedom prediction model is expressed as follows:

[0028]

[0029] In the formula, r α,k+1 and r β,k+1 The two-degree-of-freedom prediction model r k+1 The αβ axis component.

[0030] Further, the step 4 specifically includes: based on the function minimum solution principle, let the prediction cost function g be the inverter voltage v iαβ,k The partial derivative of is equal to 0, and the predictive control reference command that can minimize g at time k is obtained by the following formula

[0031]

[0032] Further, the step 5 specifically includes: according to the predictive control reference instruction The αβ axis component and And the inverse tangent function gets the spatial angle position angle: judge The voltage vector sector where it is located and selects two adjacent non-zero voltage vectors v in the voltage vector sector m =[v mα v mβ ],v n =[v nα v nβ ] and zero vector v7 =

[00] as the synthesis Three optimal voltage vectors.

[0033] Further, the step 6 specifically includes: defining three optimal voltage vectors v m 、v n The duty ratios corresponding to v7 are d m d n and d0, the component form of the synthetic voltage of the three optimal voltage vectors under these three action duty ratios in the αβ coordinate system is expressed as:

[0034]

[0035] In the formula, v sα and v sβ The composite voltage v sαβ The α and β axis components of v mα 、v mβ and v nα 、v nβ are the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β is the αβ axis component of the zero voltage vector v7;

[0036] Based on the principle of solving the minimum value of the function, the synthetic voltage v sαβ With predictive control reference instructions The sum of squares of the αβ axis tracking errors acts on the duty cycle d m d n By calculating the partial derivatives respectively, the three duty cycles corresponding to the three optimal voltage vectors can be obtained as follows:

[0037]

[0038] Solving the above equation yields:

[0039]

[0040] Where K = v 0α v mβ -v 0α v nβ -v 0β v mα +v 0β v nα +v mα v nβ -v mβ v nα , and Represents the predictive control reference instruction The αβ axis component; v mα 、v mβ and v nα 、v nβ Respectively represent the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β They represent the α and β axis components of the zero voltage vector v7 respectively; d m d nand d0 represent the three optimal voltage vectors v m 、v n The duty cycle corresponding to v7.

[0041] Further, the step 7 specifically includes: in order to achieve a fixed switching frequency, the following symmetrical 5-segment pulse mode is used to generate a control pulse signal of the inverter switch tube, and the specific expression is as follows:

[0042]

[0043] Finally, the pulse signal S is applied to the inverter switch tube to realize the two-degree-of-freedom predictive control of the LCL type filter three-phase grid-connected inverter system.

[0044] A computer device comprises a processor and a memory, wherein the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used to execute a two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention designs a control function based on the grid-side current reference tracking error and its gradient, inherently realizes multivariable control of the grid-side current and capacitor voltage, and has inherent resonant damping characteristics without the need for additional damping measures, effectively solving the resonance problem of the LCL type three-phase grid-connected inverter system and simplifying the control structure.

[0047] 2. The method provided by the present invention only includes one design parameter weight factor m, and there is no repeated traversal optimization process. The parameter setting is simple and convenient, and the implementation complexity is reduced.

[0048] 3. The method provided by the present invention adopts three optimal voltage vectors and a symmetrical 5-segment pulse generation strategy, which can achieve excellent steady-state performance and constant switching frequency, effectively improve the quality of grid-connected current, and the filter design is also simpler. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a structural block diagram of a two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system of the present invention;

[0050] Figure 2 The basic voltage vector diagram of the inverter output used in the embodiment of the present invention;

[0051] Figure 3(a) shows the steady-state waveforms of grid voltage and grid current of the traditional multi-objective single vector predictive control method.

[0052] FIG3( b ) is a diagram of the grid-side current harmonic spectrum of the traditional multi-objective single vector predictive control method;

[0053] FIG3(c) is a steady-state waveform diagram of the grid voltage and grid current of the two-degree-of-freedom predictive control method of the LCL filter type three-phase grid-connected inverter system of the present invention.

[0054] FIG3( d ) is a schematic diagram of a grid-side current harmonic spectrum of a two-degree-of-freedom predictive control method for an LCL filter-type three-phase grid-connected inverter system according to the present invention;

[0055] FIG4( a ) is a comparison diagram of grid voltage and grid current transient waveforms of the traditional multi-objective single vector predictive control method;

[0056] FIG. 4( b ) is a comparison diagram of grid voltage and grid current transient waveforms of the two-degree-of-freedom predictive control method of the LCL filter type three-phase grid-connected inverter system of the present invention. DETAILED DESCRIPTION

[0057] The embodiments of the present invention are further described below in conjunction with the accompanying drawings:

[0058] Figure 1 The present invention is an LCL filter type three-phase grid-connected inverter system, and a flowchart corresponding to the two-degree-of-freedom predictive control method for the LCL filter type three-phase grid-connected inverter system. The DC bus voltage is converted into an AC voltage square wave signal through a three-phase voltage source inverter, and then connected to the three-phase AC power grid through an output LCL filter. The inverter side current, three-phase capacitor voltage, three-phase grid side current and three-phase power grid voltage of the LCL filter type three-phase grid-connected inverter system are sampled. In the figure, V dc represents the DC bus voltage; L1 and R1 represent the filter inductance and parasitic resistance on the inverter output side, L2 and R2 represent the filter inductance and parasitic resistance on the grid side, and C represents the filter capacitor. 1,abc Indicates the current on the three-phase filter inverter side; v c,abc Represents the three-phase capacitor voltage; i 2,abc Indicates the grid-side current; v g,abc Indicates the grid voltage; is the grid current given value; i 2,k 、v g,k 、v c,k and i 1,k They represent the grid-side current, grid voltage, capacitor voltage and filter inverter-side current in the αβ coordinate system at time k respectively; the sampled grid voltage is sent to the phase-locked loop to obtain the grid voltage phase θ; is the predictive control reference instruction; S represents a symmetrical 5-segment pulse signal.

[0059] Figure 2 It is the basic voltage vector diagram of the inverter output in the αβ coordinate system; it includes 6 sectors, corresponding to 6 voltages v1-v6, and the 0 vector is v7;

[0060] A two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system comprises the following steps:

[0061] Step 1: Establish a continuous domain model of the LCL filter type three-phase grid-connected inverter system in the αβ coordinate system:

[0062]

[0063] In the formula, x=[i 1αβ v cαβ i 2αβ ] is the state variable of the LCL filter type three-phase grid-connected inverter system, u=[v iαβ v gαβ ] is the control input of the LCL filter type three-phase grid-connected inverter system, and the coefficient matrices A and B are:

[0064]

[0065] In the formula, i 1αβ 、v cαβ 、i 2αβ They represent the inverter side current, LCL capacitor voltage, and grid side current of the three-phase AC grid of the LCL filter type three-phase grid-connected inverter system under the α and β axes, respectively. iαβ represents the inverter voltage under αβ axis, v gαβ represents the αβ-axis grid voltage; L1 and R1 represent the filter inductance and parasitic resistance on the output side of the LCL filter type three-phase grid-connected inverter system, L2 and R2 are the filter inductance and parasitic resistance on the grid side, and C is the filter capacitor.

[0066] Furthermore, let the sampling period of the LCL filter type three-phase grid-connected inverter system be T s , the continuous domain model is accurately discretized using a zero-order holder, and the discrete model of the system is obtained as follows:

[0067]

[0068] Wherein, the subscripts "k" and "k+1" represent the k and k+1 moments respectively. k =[i 1αβ,k v cαβ,k i 2αβ,k ] T represents the actual values ​​of the inverter side current, capacitor voltage and grid side current sampled at time k, x k+1 =[i 1αβ,k+1 v cαβ,k+1 i 2αβ,k+1 ] T is the predicted value at time k+1, u k =[v iαβ,k vgαβ,k ] T represents the inverter voltage and grid voltage at time k, A d , B d Coefficient matrix computed for the zero-order holder method: B d =A -1 (A d -I6)B, where e is the base of the natural logarithm, I6 is the sixth-order unit matrix, and a and b represent the coefficient matrix A respectively. d and B d The elements in , the subscripts represent their row and column position information in the coefficient matrix.

[0069] Step 2: Design the following control function based on the grid-side current reference tracking error and its gradient tracking:

[0070]

[0071] In the formula, r = [r α r β ] is the control function matrix, i 2αβ =[i 2α i 2β ]、 are the grid-side currents of α and β axes and their reference matrices, v cαβ =[v cα v cβ ]、v gαβ =[v gα v gβ ] are the αβ axis capacitor voltage matrix and the grid voltage matrix respectively, and m is the weight factor; through the control function r based on the grid-side current reference tracking error and its gradient of the three-phase AC grid, multivariable control of the grid-side current and capacitor voltage is realized. Due to the constraint of Kirchhoff's current law at the connection point of L1 and C, the inherent resonant damping characteristics can be achieved without additional damping measures. This can not only effectively solve the resonance problem of the LCL filter type three-phase grid-connected inverter system, but also simplify the control structure.

[0072] Secondly, based on the control function and the system discrete model x in step 1 k+1 , the prediction model for time k+1 is constructed as follows:

[0073]

[0074] In the formula, i 2αβ，k+1 、v cαβ，k+1 and v gαβ，k+1 represents the predicted value of the grid-side current and capacitor voltage at time k+1 under the αβ axis, which can be obtained from the system discrete model x k+1 Directly obtain; k+1The two degrees of freedom are the grid-side current tracking error at time k+1 and the gradient of the grid-side current tracking error.

[0075] Step 3: In order to achieve accurate predictive control, the tracking errors of the α and β axes of the two-degree-of-freedom prediction model need to be set to 0, that is, the control reference value is and And the prediction cost function based on the two-degree-of-freedom prediction model is as follows:

[0076]

[0077] In the formula, r α,k+1 and r β,k+1 They are the α and β axis components of the two-degree-of-freedom prediction model, respectively. As can be seen from the above formula, the prediction cost function designed based on control function reference tracking realizes the multi-state variable control of grid-connected current and capacitor voltage, and there is a constraint of Kirchhoff's current law at the node of L1 and C, so the resonant energy in the LCL filter element can be effectively suppressed without additional damping to achieve resonant damping.

[0078] Step 4: Based on the principle of solving the minimum value of the function, that is, let the prediction cost function g be the inverter voltage v iαβ,k The partial derivative of is equal to 0, and the predicted control reference command that can minimize g at time k is obtained

[0079]

[0080] Step 5: Based on the predictive control reference instruction The αβ axis component and And the inverse tangent function to get its spatial angle position angle: judge The sector where the voltage is located is selected, and two adjacent non-zero voltage vectors v in the sector are selected. m =[v mα v mβ ],v n =[v nα v nβ ] and zero vector v7 =

[00] as the synthesis Three optimal voltage vectors.

[0081] Step 6: Solve for the optimal synthetic predictive control reference instruction The duty ratios of the three optimal voltage vectors are defined as follows: m d n and d0, the component form of the synthetic voltage of the three optimal voltage vectors under these three action duty ratios in the αβ coordinate system can be expressed as:

[0082]

[0083] In the formula, v sα and v sβ The composite voltage v sαβ The α and β axis components of v mα 、v mβ and v nα 、v nβ are the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β is the αβ axis component of the zero voltage vector v7;

[0084] Based on the principle of solving the minimum value of the function, the synthetic voltage v sαβ With predictive control reference instructions The sum of squares of the αβ axis tracking errors acts on the duty cycle d m d n By calculating the partial derivatives respectively, we can obtain the three action duty ratios corresponding to the three optimal voltage vectors as follows:

[0085]

[0086] Solved

[0087]

[0088] Where K = v 0α v mβ -v 0α v nβ -v 0β v mα +v 0β v nα +v mα v nβ -v mβ v nα , and Reference instructions for predictive control The αβ axis component; v mα 、v mβ and v nα 、v nβ are the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β is the αβ axis component of the zero voltage vector v7; d m d n and d0 are the three optimal voltage vectors v m 、vn The duty cycle corresponding to v7.

[0089] Step 7: In order to achieve a fixed switching frequency, the following symmetrical 5-segment pulse mode is used to generate the control pulse signal of the inverter switch tube. The specific expression is as follows:

[0090]

[0091] Each voltage vector represents the switching state of the inverter upper bridge arm switch tube. Figure 2 As shown in the figure (1 means on, 0 means off), the switch state of the lower bridge arm switch is opposite to that of the upper bridge arm. Finally, the 5-segment pulse signal S is applied to the inverter switch to realize the two-degree-of-freedom predictive control of the LCL filter type three-phase grid-connected inverter system.

[0092] In order to verify the two-degree-of-freedom predictive control method of an LCL filter type three-phase grid-connected inverter system provided by the present invention, the method provided by the present invention is applied to the LCL filter type three-phase grid-connected inverter system, and the system parameters are given in Table 1.

[0093] Table 1

[0094]

[0095] Fig. 3(a), Fig. 3(b), Fig. 3(c) and Fig. 3(d) are respectively the steady-state waveforms of the grid voltage and grid-side current of the traditional multi-objective single vector model predictive control method and the method of the present invention when the grid-side current reference value is set to 20A.

[0096] From Figure 3, we can see that the grid-side current amplitude under both control methods can track the reference value, and the grid-side current phase can also keep in sync with the grid voltage. However, compared with the traditional multi-objective single vector model predictive control method, the grid-side current ripple under the method of the present invention is reduced, the total harmonic distortion rate is reduced, and the grid-side current quality is improved; at the same time, by observing the grid-side current harmonic spectrum, we can see that compared with the traditional multi-objective single vector model predictive control method, the harmonic distribution frequency range of the method of the present invention is small, the harmonic content is lower, and the grid-side current harmonics are concentrated at the switching frequency f s A fixed switching frequency is achieved, which makes the steady-state performance of the proposed method better and is more convenient for the design of LCL filters.

[0097] Fig. 4(a) and Fig. 4(b) are the transient waveforms of the grid voltage and grid-side current of the conventional multi-objective single vector model predictive control method and the method of the present invention, respectively. The grid current reference value is set to suddenly change from 10A to 30A at 0.205s and suddenly change from 30A to 10A at 0.255s.

[0098] From Figure 4, we can see that the grid-side current under the two control methods can track the reference value, and the grid-side current phase can also keep synchronized with the grid voltage, and the entire dynamic process transitions smoothly without resonance. In addition, in the transient process, the grid-side voltage and grid-side current can still keep in phase. The two methods have good dynamic performance and meet the control requirements of the LCL filter type three-phase grid-connected inverter system.

Claims

1. A two-degree-of-freedom predictive control method for an LCL filter type three-phase grid-connected inverter system, characterized in that: The steps include: Step 1: Sample the state variables of the LCL filter type three-phase grid-connected inverter system. The state variable sampling includes the inverter side current i at time k 1,k , capacitor voltage v c,k , three-phase AC grid side current i 2,k and grid voltage v g,k , transform the sampled state variables from the abc coordinate system to the αβ coordinate system, where the grid voltage v in the abc coordinate system g,abc The voltage phase θ of the three-phase AC grid is obtained by combining the phase-locked loop, and the continuous domain model of the LCL filter type three-phase grid-connected inverter system in the αβ coordinate system is established. The LCL filter type three-phase grid-connected inverter system is discretized using a zero-order holder to obtain the system discrete model x of the LCL filter type three-phase grid-connected inverter system at time k+1. k+1 ; Step 2: Design a control function r based on the grid-side current reference tracking error and its gradient, and calculate the control function r based on the system discrete model x. k+1 Establish a two-degree-of-freedom prediction model r at time k+1 k+1 , to reflect the error between the predicted value of the three-phase AC grid-side current and its gradient and the corresponding reference value; Step 3: Design a prediction model based on two degrees of freedom k+1 The prediction cost function g is used as the two-degree-of-freedom prediction model r k+1 αβ axis tracking error; Step 4: Based on the principle of solving the minimum value of the function, let the prediction cost function g be the inverter voltage v iαβ,k The partial derivative of is equal to 0, and the predictive control reference instruction that can minimize the prediction cost function g is obtained. Step 5: Based on the predictive control reference instruction Axis components in the αβ coordinate system and Get the predictive control reference instruction The spatial angle position of the control unit and the prediction control reference command are determined. The voltage vector sector where the predictive control reference command is selected Two adjacent non-zero voltage vectors v in the voltage vector sector m 、v n , and the zero vector v7 at the origin of the voltage vector sector as the synthetic predictive control reference instruction Three optimal voltage vectors; Step 6: Calculate the three optimal voltage vectors v m 、v n and V7 at their respective duty cycles, and compare the synthesized voltage with the predicted control reference instruction The square sum of the tracking errors is used to calculate the partial derivative of each duty cycle, and the three duty cycles d corresponding to the three optimal voltage vectors are calculated. m d n and d0; Step 7: According to the three optimal voltage vectors v m 、v n and v7 and three action duty cycle d m d n And d0, generate a control pulse signal S according to the symmetrical 5-segment pulse mode, and apply the control pulse signal S to each switch tube in the inverter to realize the predictive control of the LCL filter type three-phase grid-connected inverter system.

2. The two-degree-of-freedom predictive control method of a LCL filter type three-phase grid-connected inverter system according to claim 1, characterized in that: The step 1 specifically includes: The continuous domain model of the LCL filter type three-phase grid-connected inverter system in the αβ coordinate system is established: In the formula, x=[i 1αβ v cαβ i 2αβ ] is the state variable of the LCL filter type three-phase grid-connected inverter system, u=[v iαβ v gαβ ] is the control input of the LCL filter type three-phase grid-connected inverter system, and the coefficient matrices A and B are expressed as follows: In the formula, i 1αβ 、v cαβ 、i 2αβ They represent the inverter side current, LCL capacitor voltage, and grid side current of the three-phase AC grid of the LCL filter type three-phase grid-connected inverter system under the α and β axes, respectively. iαβ represents the inverter voltage under αβ axis, v gαβ represents the αβ-axis grid voltage; L1 and R1 represent the filter inductance and parasitic resistance on the output side of the LCL filter type three-phase grid-connected inverter system, L2 and R2 represent the filter inductance and parasitic resistance on the grid side, and C represents the filter capacitor; Assume that the sampling period of the LCL filter type three-phase grid-connected inverter system is T s , the continuous domain model is accurately discretized using a zero-order holder, and the obtained discrete model of the LCL filter type three-phase grid-connected inverter system is expressed as: In the formula, the subscripts "k" and "k+1" represent the k and k+1 moments respectively. k =[i 1αβ,k v cαβ,k i 2αβ,k ] T represents the actual values ​​of the inverter side current, capacitor voltage and grid side current sampled at time k, x k+1 =[i 1αβ,k+1 v cαβ,k+1 i 2αβ,k+1 ] T is the predicted value at time k+1, u k =[v iαβ,k v gαβ,k ] T represents the inverter voltage and grid voltage at time k, A d , B d Coefficient matrix computed for the zero-order holder method: B d =A -1 (A d -I6)B, where e is the base of the natural logarithm, I6 is the sixth-order unit matrix, and a and b represent the coefficient matrix A respectively. d and B d The elements in , the subscripts represent their row and column position information in the coefficient matrix.

3. The two-degree-of-freedom predictive control method of a LCL filter type three-phase grid-connected inverter system according to claim 1, characterized in that: The step 2 specifically includes: Design a control function r based on the grid-side current reference tracking error and its gradient tracking, which is expressed as follows: In the formula, r = [r α r β ] represents the control function matrix, i 2αβ =[i 2α i 2β ]、i2 * αβ =[i2 * α i2 * β ] are the grid-side currents of α and β axes and their reference matrices, v cαβ =[v cα v cβ ]、v gαβ =[v gα v gβ ] are the αβ axis capacitor voltage matrix and the grid voltage matrix respectively, and m is the weight factor; Based on the control function r and the system discrete model, the two-degree-of-freedom prediction model at time k+1 is constructed using the state variable sampling information at time k as follows: In the formula, i 2αβ，k+1 、v cαβ，k+1 and v gαβ，k+1 represents the predicted value of the current and capacitor voltage on the αβ axis grid side at time k+1, from the system discrete model x at time k+1 k+1 Directly obtain, k+1 The control degrees of freedom include the grid-side current tracking error at time k+1 and the gradient of the grid-side current tracking error.

4. The two-degree-of-freedom predictive control method of a LCL filter type three-phase grid-connected inverter system according to claim 1, characterized in that: The step 3 specifically includes: In order to achieve accurate predictive control, it is necessary to control the αβ axis tracking errors of the two-degree-of-freedom prediction model to be 0, that is, the control reference value is and The prediction cost function g based on the two-degree-of-freedom prediction model is expressed as follows: In the formula, r α,k+1 and r β,k+1 The two-degree-of-freedom prediction model r k+1 The αβ axis component.

5. The two-degree-of-freedom predictive control method of a LCL filter type three-phase grid-connected inverter system according to claim 4, characterized in that: The step 4 specifically includes: Based on the principle of solving the minimum value of the function, let the prediction cost function g be the inverter voltage v iαβ,k The partial derivative of is equal to 0, and the predictive control reference command that can minimize g at time k is obtained by the following formula 6. The two-degree-of-freedom predictive control method of a LCL filter type three-phase grid-connected inverter system according to claim 5, characterized in that: The step 5 specifically includes: According to the predictive control reference command The αβ axis component and And the inverse tangent function gets the spatial angle position angle: judge The voltage vector sector where it is located and selects two adjacent non-zero voltage vectors v in the voltage vector sector m =[v mα v mβ ],v n =[v nα v nβ ] and zero vector v7 = [00] as the synthesis Three optimal voltage vectors.

7. The two-degree-of-freedom predictive control method for a LCL filter type three-phase grid-connected inverter system according to claim 6, characterized in that: The step 6 specifically includes: Define three optimal voltage vectors v m 、v n The duty ratios corresponding to v7 are d m d n and d0, the component form of the synthetic voltage of the three optimal voltage vectors under these three action duty ratios in the αβ coordinate system is expressed as: In the formula, v sα and v sβ The composite voltage v sαβ The α and β axis components of v mα 、v mβ and v nα 、v nβ are the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β is the αβ axis component of the zero voltage vector v7; Based on the principle of solving the minimum value of the function, the synthetic voltage v sαβ With predictive control reference instructions The sum of squares of the αβ axis tracking errors acts on the duty cycle d m d n By calculating the partial derivatives respectively, the three duty cycles corresponding to the three optimal voltage vectors can be obtained as follows: Solving the above equation yields: Where K = v 0α v mβ -v 0α v nβ -v 0β v mα +v 0β v nα +v mα v nβ -v mβ v nα , and Represents the predictive control reference instruction The αβ axis component; v mα 、v mβ and v nα 、v nβ Respectively represent the two selected non-zero voltage vectors v m With v n The αβ axis component; v 0α 、v 0β They represent the α and β axis components of the zero voltage vector v7 respectively; d m d n and d0 represent the three optimal voltage vectors v m 、v n The duty cycle corresponding to v7.

8. The two-degree-of-freedom predictive control method for a LCL filter type three-phase grid-connected inverter system according to claim 7, characterized in that: The step 7 specifically includes: In order to achieve a fixed switching frequency, the following symmetrical 5-segment pulse mode is used to generate the control pulse signal of the inverter switch tube. The specific expression is as follows: Finally, the pulse signal S is applied to the inverter switch tube to realize the two-degree-of-freedom predictive control of the LCL type filter three-phase grid-connected inverter system.

9. A computer device, characterized in that: It includes a processor and a memory, the processor is electrically connected to the memory, the memory is used to store instructions and data, and the processor is used to execute the two-degree-of-freedom predictive control method of the LCL filter type three-phase grid-connected inverter system described in claims 1-8.

Images