Parallel distributed compensation method for grid-connected inverter based on double fuzziness

Abstract

The invention belongs to the technical field of new energy grid-connected control, and particularly relates to a grid-connected inverter parallel distributed compensation method based on double fuzziness. The double-fuzzy grid-connected synchronous control system DF-PDC performs nonlinear control on a nonlinear grid-connected synchronous control system, so that the grid-connected synchronous control system can be stable at any phase angle working point, and then a Gaussian membership function is designed according to a clustering result to fuzzy state feedback gain; at the moment, the state feedback gain of the linear subsystem is not considered to be constant, a plurality of feedback gains are associated to each subsystem according to a certain performance criterion, and the gains are subjected to fuzzy mixing to obtain the final gain of each subsystem, so that the self-adaptive capability of the grid-connected synchronous control system to cope with a power grid fault can be greatly improved, and the system reliability is improved. And the conservative property of the stability and the complexity of analysis are reduced by using double-fuzzy parallel distributed compensation, and the method has important significance for improving the stability of a grid-connected synchronous control system.

Description

Technical Field

[0001] This invention belongs to the field of new energy grid-connected control technology, specifically relating to a parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic. Background Technology

[0002] With the acceleration of energy transition, renewable energy (such as wind power and photovoltaics) is becoming the mainstay of my country's future power supply due to its clean characteristics. However, the large-scale grid connection of these energy sources relies on power electronic converters as the core of the grid connection interface. This will promote the deep power electronicsization of the power system and make the stability problem of grid-connected converters increasingly prominent. Their stable operation has a crucial impact on the grid connection of new energy sources. As the penetration rate of grid-connected converters in grid-connected systems continues to increase, related dynamic problems are gradually emerging in actual systems. The dynamic problems they cause (such as instability in new energy projects) have led to widespread attention and research on the stability of grid-connected converters. The transient synchronous instability of phase-locked loop (PLL) grid-connected converters involves the nonlinear dynamic process of the PLL. Current research mainly follows the classical transient stability analysis methods of synchronous machines, including the equal area criterion, the energy function method, and Lyapunov stability theory.

[0003] Current methods to enhance the transient stability of phase-locked loops (PLLs) mainly include: 1) Dynamically adjusting the active / reactive injection current after a fault: This method can change the current injection phase angle according to the grid impedance ratio, which helps to avoid power oscillations or recovery lag caused by traditional fixed limiting. However, this method requires real-time acquisition of accurate grid voltage phase / amplitude. In the event of asymmetrical faults or harmonic distortion, the performance of the PLL may degrade, potentially leading to control failure. 2) Increasing the damping of the PLL: This adjustment method is similar to the existing proportional control. Since the introduction of damping can avoid the reverse adjustment of the sine function within one and a half cycles, this method also depends on whether the system's equilibrium point exists during grid faults. 3) Adjusting the control characteristics of the PLL: This method temporarily freezes the output frequency / phase of the PLL to avoid the negative damping effect of the PLL from exacerbating system oscillations. Although it can improve transient stability, it loses phase tracking capability during the freeze period. 4) Traditional parallel distributed compensation: This method uses parallel distributed compensation to increase the system's damping to improve the stability of the system during transient instability. It not only does not ignore the system's damping term but also reduces the conservatism of the stability assessment results. However, this method has poor adaptability and its ability to cope with system instability caused by changes in system parameters is limited. The control methods mentioned above cannot restore the system's stability in a timely manner, and their ability to adaptively restore stability in the face of uncertain changes in the system is poor.

[0004] To address the transient synchronization instability problem in grid-connected synchronous control systems, a dual-fuzzy parallel distributed compensation method can be adopted. Dual-fuzzy parallel distributed compensation is a controller designed to improve the asymptotic stability of closed-loop control systems. Currently, fuzzy controllers are designed based on the traditional parallel distributed compensation (PDC) method. However, the traditional method cannot achieve adaptive stability recovery when faced with synchronization system instability, and traditional PDC also leads to excessive overshoot and slow phase-locking speed.

[0005] In view of this, the inventors hope to propose a dual-fuzzy parallel distributed compensation (DF-PDC) based on the traditional PDC design, to solve the problems of poor adaptability, excessive overshoot, and slow phase-locking speed of the traditional PDC, and improve the system's adaptability and stability in the face of instability. Summary of the Invention

[0006] The technical problem to be solved by this invention is: how to improve the shortcomings of the traditional PDC adaptive method when facing the transient instability of the synchronization system, and to provide a DF-PDC method, which associates multiple feedback gains with each subsystem and performs fuzzy mixing of these gains to obtain the final gain of each subsystem.

[0007] To achieve the above-mentioned technical objectives and effects, the present invention is implemented through the following technical solution:

[0008] This invention provides a parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic, characterized by comprising the following steps:

[0009] S1. Based on the phase-locked loop control structure of the grid-connected inverter, establish the nonlinear mathematical model of the system and perform TS fuzzy equivalence.

[0010] S2. Based on the control performance of the phase-locked loop, the feedback matrix of the subsystem is solved by using the pole placement method to obtain the state space matrix obtained by TS fuzzy logic.

[0011] S3. Use fuzzy C-means clustering to cluster the feedback matrix of the subsystem, construct a Gaussian membership function using the cluster centers, solve the membership degree corresponding to the system parameters based on the Gaussian membership function, and perform fuzzy weighted fusion on the feedback matrix of the subsystem accordingly.

[0012] S4. Inside the phase-locked loop, a dual-fuzzy controller is constructed to achieve distributed compensation control by weighting and fusing the subsystem feedback matrix based on the nonlinear term TS fuzzy and combining the Gaussian membership function to generate the total feedback gain.

[0013] Furthermore, the specific process of step S1 is as follows:

[0014] S101. Establish a second-order nonlinear model of the phase-locked loop:

[0015]

[0016] Where, k p，pll and k i，pll These are the proportional-integral (PI) and integral (I) coefficients in the phase-locked loop (PLL) PI controller, respectively. g For line inductance, I sd For grid-connected current, w g V is the angular velocity of the grid voltage. g Let d be the grid voltage amplitude, d be the difference between the phase angle of the phase-locked loop output phase angle and the grid phase angle, and a be the integrator output value in the phase-locked loop PI controller. Let d and a be used as state variables x1 and x2, respectively.

[0017] S102. Based on the state-space model, the state-space matrix A is determined as follows:

[0018]

[0019] Determine the fuzzy boundary of the fuzzy object sinx1 / x1, let its maximum and minimum values ​​be sina / a and 1 respectively, and then obtain the value of a according to the system of fuzzy matrix inequalities:

[0020]

[0021] The membership function is determined using the maximum and minimum values ​​of fuzzy objects as shown below:

[0022]

[0023] Where w1 + w2 = 1;

[0024] S103. The state-space matrix A contains only one nonlinear term, therefore there are two fuzzy rules:

[0025] Rule R1: If x1 ≠ 0, then the state space matrix A = A1;

[0026] Rule R2: If x1 = 0, then the state space matrix A = A2;

[0027] Based on the state space matrices A1 and A2, the membership function of the fuzzy object, and the fuzzy rules, the equivalent model of the phase-locked loop is obtained as follows:

[0028]

[0029] in,

[0030] Furthermore, the specific process of step S2 is as follows:

[0031] By inputting matrix B = [k p，pll ki，pll ] T The feedback matrix of the subsystem is solved using the pole placement method:

[0032] |λ i EA i +BK i |=0

[0033] Among them, K i Let be the feedback matrix, i be the eigenvalue to be solved, and E be the identity matrix.

[0034] Furthermore, the specific process of step S3 is as follows:

[0035] S301. Use the fuzzy C-means clustering method to solve the parameters in the Gaussian membership function. Use the pole placement formula to take the obtained subsystem feedback matrix as the C-means clustering input. Use the pole placement formula to solve the values ​​of reference current and grid impedance for the cluster centers generated by the clustering results. Use the obtained current and impedance values ​​as the center of the membership function curve, that is, use the center value of the result as the u value in the Gaussian membership function.

[0036] S302. Select the fuzzy range of reference current and grid impedance, divide the fuzzy range into segments, calculate the magnitude of A based on the range of each segment, and then use the maximum and minimum values ​​of the magnitude of subsystem A to solve the parameter of s in the Gaussian membership function. Based on the "3s criterion" of Gaussian function, extend the definition of the function to 6s.

[0037]

[0038] Based on the u-value and s-value obtained using the fuzzy C-means clustering method, the Gaussian membership function is derived, and its expression is:

[0039] m(x) = exp[-(xu)] 2 / 2*σ 2 ]*100%

[0040] The membership degree of each interval is calculated based on the Gaussian membership function of the system current and impedance values. The fuzzy weighting of the feedback matrices of multiple subsystems is then performed using this membership degree, and the result is the total feedback gain matrix.

[0041] Furthermore, the specific process of step S4 is as follows:

[0042] By acquiring the inverter output voltage signal V pcc Phase-locked control is performed to increase the total feedback gain K generated by the fuzzy subsystem. i x is added to the output voltage signal V pcc The q-axis component V obtained after coordinate transformation pccq Then, by inputting matrix B = [kp，pll k i，pll ] T This makes the new state matrix A i -BK i All eigenvalues ​​have negative real parts. Finally, the designed dual-fuzzy parallel distributed compensation is added to V. pccq Afterwards; at this point the system becomes:

[0043]

[0044] Where, m it K is the membership degree of the actual working condition corresponding to the t-th preset working condition. it Q is the state feedback gain of the i-th linear subsystem designed under the t-th preset operating condition. i The number of preset operating conditions.

[0045] The beneficial effects of this invention are:

[0046] To address the issue that traditional phase-locked loops (PLLs) can become unstable due to system divergence caused by operating point shifts during grid faults, this invention, the dual-fuzzy grid-connected synchronous control system DF-PDC, performs nonlinear control on the nonlinear grid-connected synchronous control system, enabling it to remain stable at any phase angle operating point. Secondly, based on clustering results, a Gaussian membership function is designed to fuzzify the state feedback gain. In this case, the state feedback gain of the linear subsystem is not considered constant but is associated with each subsystem according to certain performance criteria. These gains are then fuzzily mixed to obtain the final gain of each subsystem. This significantly improves the adaptive capability of the grid-connected synchronous control system to cope with grid faults. Furthermore, the use of dual-fuzzy parallel distributed compensation reduces the conservatism of stability and the complexity of analysis, which is of great significance for improving the stability of the grid-connected synchronous control system.

[0047] Of course, any product implementing this invention does not necessarily need to achieve all of the above advantages at the same time. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This is a schematic diagram of a new energy grid connection system in an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of a phase-locked loop system in an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the maximum stability region of the system in an embodiment of the present invention;

[0052] Figure 4 In the equivalent model of the embodiments of the present invention Schematic diagram of dynamic characteristic verification;

[0053] Figure 5 This is a schematic diagram of the membership function of the TS fuzzy model in an embodiment of the present invention;

[0054] Figure 6 K represents two working conditions in this embodiment of the invention. I K Z Schematic diagram of clustering and cluster centers;

[0055] Figure 7 This is a schematic diagram of the membership function under the reference current and grid impedance conditions in an embodiment of the present invention;

[0056] Figure 8 This is a schematic diagram of the DF-PDC control implementation of the phase-locked loop in an embodiment of the present invention;

[0057] Figure 9 This is a waveform diagram of a grid-connected system with two impedance jumps in an embodiment of the present invention;

[0058] Figure 10 The waveform diagrams of the PDC and DF-PDC systems are respectively added for the first transition in the embodiments of the present invention;

[0059] Figure 11 The waveform diagrams of the PDC and DF-PDC systems are respectively added to the second transition in the embodiment of the present invention. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] This embodiment provides a technical solution: based on the research of phase-locked loops (PLLs) in grid-connected systems, the PLL is modeled, and the nonlinearity in the model is processed to establish a fuzzy equivalent model of the new energy grid-connected system. A subsystem feedback matrix is ​​designed, and fuzzy C-means clustering is used to cluster the obtained subsystem feedback matrix. The maximum and minimum values ​​of the subsystem A-mode values ​​are used to solve for the parameter values ​​of 's' in the Gaussian membership function. The dual-fuzzy parallel distributed compensation can enable the grid-connected synchronous control system to adaptively recover stability when facing instability. Specifically, the solution includes the following steps:

[0062] S1. Based on the phase-locked loop control structure of the grid-connected inverter, establish the nonlinear mathematical model of the system and perform TS fuzzy equivalence.

[0063] S2. Based on the control performance of the phase-locked loop, the feedback matrix of the subsystem is solved by using the pole placement method to obtain the state space matrix obtained by TS fuzzy.

[0064] S3. Use fuzzy C-means clustering to cluster the feedback matrix of the subsystem, construct a Gaussian membership function using the cluster centers, solve for the membership degree corresponding to the system parameters based on the Gaussian membership function, and then perform fuzzy weighted fusion on the feedback matrix of the subsystem accordingly.

[0065] S4. Inside the phase-locked loop, a dual-fuzzy controller is constructed to achieve distributed compensation control by weighting and fusing the subsystem feedback matrix based on the nonlinear term TS fuzzy and combining the Gaussian membership function to generate the total feedback gain.

[0066] In this embodiment, the influence of other control loops on the phase-locked loop is not considered, and the grid-connected converter is equivalent to a current source, I s The reference current amplitude is pll, the phase angle of the phase-locked loop output is L. g For the mains inductance, V g denoted as , and g as , where g is the power grid amplitude and g is the power grid phase angle.

[0067] In this embodiment, the reference value I of the grid-connected current of the system sd The current is 1.5A, the reference current phase angle is the phase angle of the phase-locked loop output, and the magnitude of the grid voltage V. g The voltage is 220V, the mains frequency is 50Hz, and the mains parasitic inductance L is... g The value is 0.1mH, and the proportionality coefficient is k. p The integral coefficient k is 0.7. i It is 78.4.

[0068] Nonlinear model of traditional phase-locked loop:

[0069]

[0070] Where a is the integrator output in the phase-locked loop PI controller, and q pll The phase angle output by the phase-locked loop is q. g For the phase angle of the power grid, w g Let the angular velocity of the power grid rotation be d = q. pll -q g And by selecting state variables d and a as x1 and x2, the resulting state space matrix A is:

[0071]

[0072] Based on the fuzzy boundary in sinx1 / x1, the spatial matrix A yields the state subspaces A1 and A2. Substituting the obtained A1 and A2 into the linear matrix inequality:

[0073]

[0074] By continuously changing the domain of discourse, the positive definite matrix P is obtained by solving in the LMI toolbox. Finally, the domain of discourse with the largest state variable is obtained as x1∈(-0.78p, 0.78p). The common positive definite matrix obtained at this time is... Then we can obtain the Lyapunov function V(x) = x T Mx, where let The maximum stability region can be plotted, such as Figure 3 As shown.

[0075] The membership function is determined based on the maximum universe of discourse, as shown in the following formula:

[0076]

[0077] Membership function diagram as shown below Figure 4 As shown. Where w1 + w2 = 1.

[0078] The state-space matrix A contains only one nonlinear term, therefore there are two fuzzy rules:

[0079] Rule R1: If x1 ≠ 0, then the state space matrix A = A1;

[0080] Rule R2: If x1 = 0, then the state space matrix A = A2;

[0081] The final fuzzy equivalent model of the new energy grid-connected system is determined as follows:

[0082]

[0083] in,

[0084] Figure 5 To verify the dynamic relationship between the grid-connected system and the fuzzy equivalent model, two transitions were added at 0.2s and 0.35s. The fuzzy equivalent model was able to maintain the same dynamic characteristics as the grid-connected model, which proves that the phase-locked loop modeling is correct.

[0085] In step S2, according to the input matrix B = [k p，pll k i，pll ] T The feedback matrix of the subsystem is solved using the pole placement method of formula (7):

[0086] |λ i EA i +BKi |=0 (7)

[0087] When the reference current I sd When the grid impedance Lg experiences a jump, it can lead to instability in the synchronous system. Therefore, the feedback matrix of each subsystem is determined within a certain jump range. For the reference current, the feedback matrix of each subsystem can be designed at equal intervals (0.1A) within the ranges of 0 to 10A, 10A to 20A, and 20A to 30A. Similarly, for the grid impedance, the feedback matrix of each subsystem can be designed at equal intervals (0.2mH) within the ranges of 0 to 20mH, 20mH to 40mH, and 40mH to 60mH. Fuzzy C-means clustering is then used to cluster the obtained subsystem feedback matrices. After generating cluster centers, as shown... Figure 6 As shown. Based on the center value of the feedback matrix, the values ​​of the reference current and the grid impedance are solved using the pole placement formula and used as the center of the membership function curve. That is, the center value of the result is used as the u value in the Gaussian membership function. At this time, the cluster centers obtained by the reference current are c1 = 5.177, c2 = 15.044, and c3 = 25.128, respectively. Similarly, the cluster centers obtained by the grid impedance are d1 = 9.447, d2 = 32.40, and d3 = 51.072, respectively.

[0088] To solve for s in the Gaussian membership function, the maximum and minimum values ​​of the subsystem A modulus are solved using the range of the reference current and the grid impedance, as shown in formula (8). This is derived from the "3s criterion" of the Gaussian function, and 6s is used as the domain of the function.

[0089]

[0090] According to formula (19), the reference current s values ​​are s1 = 18.78, s2 = 19.75, and s3 = 20.79. Similarly, the grid impedance s values ​​are s4 = 16.04, s5 = 16.75, and s6 = 17.49. The Gaussian membership function expression is:

[0091] m(x) = exp[-(xu)] 2 / 2*σ 2 *100% (9)

[0092] Therefore, the membership functions of the reference current and the grid impedance are as follows:

[0093]

[0094] The membership function graph can be obtained from the above function, such as... Figure 7 As shown.

[0095] In step S3, based on the membership function obtained above, the number of fuzzy rules is 3. 2Therefore, a series of preset operating condition subsystem feedback gain K It K Zt The IF-THEN fuzzy rule can be described as follows:

[0096] Fuzzy rule i:

[0097] If I = S1 and Z = S2, then K i1 =K I1 K Z1 ...; If I = B1 and Z = B2, then K i9 =K I9 K Z9 ;

[0098]

[0099] Based on the pre-designed TS fuzzy model, the DF-PDC control fuzzy rule is defined as follows:

[0100] If x1 = 0, then the state matrix A = A1, K = K1;

[0101] If x1≠0, then the state matrix A=A2, K=K2;

[0102] Based on the membership function and fuzzy rules of the TS fuzzy model, the state feedback gains of different subsystems are fuzzily combined to construct the state feedback quantity of the entire system. Thus, the global fuzzy controller can be represented as follows:

[0103]

[0104] Wherein, K i The state feedback gain of the linear subsystem is generally expressed as:

[0105]

[0106] Under dual-fuzzy parallel distributed compensation control, the system dynamics can be described as follows:

[0107]

[0108] Where, m it K is the membership degree of the actual working condition corresponding to the t-th preset working condition. it Q is the state feedback gain of the i-th linear subsystem designed under the t-th preset operating condition. i The number of preset operating conditions is shown in the control implementation diagram of the dual fuzzy controller in the phase-locked loop, as follows. Figure 8 As shown.

[0109] The above embodiments overcome the shortcomings of traditional parallel distributed compensation, which cannot achieve self-adaptation. Figure 9The system becomes unstable when the grid impedance changes twice, and the second change is more severe. Figure 10 After the first jump, both traditional PDC control and DF-PDC control were added, and the system was able to recover stability. Figure 11 After the second transition, if traditional PDC control and DF-PDC control are added respectively, traditional PDC cannot restore the system to stability, while DF-PDC can still stabilize the system again. The comparison shows that DF-PDC can significantly improve the ability of the grid-connected synchronous control system to cope with grid faults, and in terms of system stability, it can improve the system's adaptive recovery stability capability.

[0110] The distributed compensation control method for grid-connected inverters of this invention, namely the fuzzy TS of the synchronous system and the fuzzy Gaussian membership function on the system parameters, enables the grid-connected synchronous control system to adaptively recover stability when transient instability occurs, improves the adaptive capability of the grid-connected synchronous control system to cope with grid faults, and reduces the conservatism of traditional parallel distributed compensation. Compared with a single fuzzy controller, this invention has better adaptive capability.

[0111] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic, characterized in that, Includes the following steps: S1. Based on the phase-locked loop control structure of the grid-connected inverter, establish the nonlinear mathematical model of the system and perform TS fuzzy equivalence. S2. Based on the control performance of the phase-locked loop, the feedback matrix of the subsystem is solved by using the pole placement method to obtain the state space matrix obtained by TS fuzzy logic. S3. Use fuzzy C-means clustering to cluster the feedback matrix of the subsystem, construct a Gaussian membership function using the cluster centers, solve the membership degree corresponding to the system parameters based on the Gaussian membership function, and perform fuzzy weighted fusion on the feedback matrix of the subsystem accordingly. S4. Inside the phase-locked loop, a dual-fuzzy controller is constructed to achieve distributed compensation control by weighting and fusing the subsystem feedback matrix based on the nonlinear term TS fuzzy and combining the Gaussian membership function to generate the total feedback gain.

2. The parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic as described in claim 1, characterized in that, The specific process of step S1 is as follows: S101. Establish a second-order nonlinear model of the phase-locked loop: Where, k p，pll and k i，pll These are the proportional-integral (PI) and integral (I) coefficients in the phase-locked loop (PLL) PI controller, respectively. g For line inductance, I sd For grid-connected current, w g V is the angular velocity of the grid voltage. g Let d be the grid voltage amplitude, d be the difference between the phase angle of the phase-locked loop output phase angle and the grid phase angle, and a be the integrator output value in the phase-locked loop PI controller. Let d and a be used as state variables x1 and x2, respectively. S102. Based on the state-space model, the state-space matrix A is determined as follows: Determine the fuzzy boundary of the fuzzy object sinx1 / x1, let its maximum and minimum values ​​be sina / a and 1 respectively, and then obtain the value of a according to the system of fuzzy matrix inequalities: The membership function is determined using the maximum and minimum values ​​of fuzzy objects as shown below: Where w1 + w2 = 1; S103. The state-space matrix A contains only one nonlinear term, therefore there are two fuzzy rules: Rule R1: If x1 ≠ 0, then the state space matrix A = A1; Rule R2: If x1 = 0, then the state space matrix A = A2; Based on the state space matrices A1 and A2, the membership function of the fuzzy object, and the fuzzy rules, the equivalent model of the phase-locked loop is obtained as follows: in, 3. The parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic as described in claim 1, characterized in that, The specific process of step S2 is as follows: By inputting matrix B = [k p，pll k i，pll ] T The feedback matrix of the subsystem is solved using the pole placement method: |l i EA i +BK i |=0 Among them, K i Let be the feedback matrix, i be the eigenvalue to be solved, and E be the identity matrix.

4. The parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic as described in claim 3, characterized in that, The specific process of step S3 is as follows: S301. Use the fuzzy C-means clustering method to solve the parameters in the Gaussian membership function. Use the pole placement formula to take the obtained subsystem feedback matrix as the C-means clustering input. Use the pole placement formula to solve the values ​​of reference current and grid impedance for the cluster centers generated by the clustering results. Use the obtained current and impedance values ​​as the center of the membership function curve, that is, use the center value of the result as the u value in the Gaussian membership function. S302. Select the fuzzy range of reference current and grid impedance, divide the fuzzy range into segments, calculate the magnitude of A based on the range of each segment, and then use the maximum and minimum values ​​of the magnitude of subsystem A to solve the parameter of s in the Gaussian membership function. Based on the "3s criterion" of Gaussian function, extend the definition of the function to 6s. Based on the u-value and s-value obtained using the fuzzy C-means clustering method, the Gaussian membership function is derived, and its expression is: m(x)=exp[-(x-u) 2 / 2*σ 2 ]*100% The membership degree of each interval is calculated based on the Gaussian membership function of the system current and impedance values. The fuzzy weighting of the feedback matrices of multiple subsystems is then performed using this membership degree, and the result is the total feedback gain matrix.

5. The parallel distributed compensation method for grid-connected inverters based on dual fuzzy logic as described in claim 4, characterized in that, The specific process of step S4 is as follows: By acquiring the inverter output voltage signal V pcc Phase-locked control is performed to increase the total feedback gain K generated by the fuzzy subsystem. i x is added to the output voltage signal V pcc The q-axis component V obtained after coordinate transformation pccq Then, by inputting matrix B = [k p，pll k i，pll ] T This makes the new state matrix A i -BK i All eigenvalues ​​have negative real parts. Finally, the designed dual-fuzzy parallel distributed compensation is added to V. pccq Afterwards; at this point the system becomes: Where, m it K is the membership degree of the actual working condition corresponding to the t-th preset working condition. it Q is the state feedback gain of the i-th linear subsystem designed under the t-th preset operating condition. i The number of preset operating conditions.

Images