Self-adaptive impedance matching method for transient disturbance of grid voltage of grid-connected inverter

Abstract

The invention discloses a grid-connected inverter power grid voltage transient disturbance self-adaptive impedance matching method, relates to the technical field of new energy power generation, and effectively avoids a frequency spectrum aliasing phenomenon and ensures the accuracy of data by collecting power grid voltage and inverter output current in real time and performing accurate anti-aliasing filtering processing. Subsequently, a windowing FFT analysis technology is utilized to accurately extract fundamental wave and specific sub-harmonic components, and a reliable data basis is provided for subsequent impedance calculation and harmonic compensation. The power grid impedance is calculated in real time through the recursive least square method, and the virtual inductance at the inverter side is dynamically adjusted, so that effective compensation for the fundamental wave impedance mismatch is realized, and the stability of the system is remarkably improved. Finally, generation of PWM signals is optimized by adopting a model prediction control technology, rapid tracking of a compensated current reference value is realized, and transient response time is further reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy power generation technology, in particular to a grid-connected inverter grid voltage transient disturbance adaptive impedance matching method. BACKGROUND

[0002] The existing grid-connected inverter generally adopts an impedance matching method based on fixed parameters, and realizes grid impedance matching through preset LCL filter parameters. A typical scheme adopts a feedforward voltage compensation strategy, and adjusts active power output through a PI controller when grid voltage suddenly changes. However, such a method depends on a static model of grid impedance, and in the transient disturbance scenarios such as voltage surge / dip and harmonic distortion, due to the dynamic change of the equivalent impedance of the grid, it is easy to cause the harmonic content of the grid-connected current to exceed the standard, and even cause resonance instability problem. Therefore, the present application provides a grid-connected inverter grid voltage transient disturbance adaptive impedance matching method. SUMMARY

[0003] The present application aims to solve the problem that the feedforward voltage compensation strategy is adopted, and the active power output is adjusted through the PI controller when the grid voltage suddenly changes. However, such a method depends on a static model of grid impedance, and in the transient disturbance scenarios such as voltage surge / dip and harmonic distortion, due to the dynamic change of the equivalent impedance of the grid, it is easy to cause the harmonic content of the grid-connected current to exceed the standard, and even cause resonance instability problem. The present application provides a grid-connected inverter grid voltage transient disturbance adaptive impedance matching method.

[0004] In order to achieve the above-mentioned purpose, the present application specifically adopts the following technical scheme:

[0005] A grid-connected inverter grid voltage transient disturbance adaptive impedance matching method, comprising the following steps:

[0006] S1, synchronously collecting the grid voltage v s (t) and the inverter output current i g (t) at a sampling frequency f inv ≥20kHz, and obtaining discrete sequences v g [k] and i inv [k] after anti-aliasing filtering, real-time sampling (above 20kHz) ensures to capture the high-frequency characteristics of the grid transient disturbance, and the anti-aliasing filtering avoids spectrum aliasing.

[0007] S2, performing windowed FFT analysis on v g [k] and i inv [k], and extracting complex voltages and complex currents of fundamental wave and 6n±1th harmonic (n=1, 2, …, 13);

[0008] S3. Calculate the impedance of each harmonic grid based on the recursive least squares (RLS) method. The calculation formula is as follows:

[0009]

[0010] Where L para For the parasitic inductance of the line, f1 = 50Hz, the RLS dynamic impedance calculation updates the impedance parameters in real time through a recursive algorithm to adapt to the dynamic changes in the grid impedance.

[0011] S4. Based on the fundamental impedance Z grid (1) = R1 + jX1, dynamically adjust the virtual inductance on the inverter side:

[0012]

[0013] Virtual inductor adjustment directly compensates for fundamental impedance mismatch and improves system stability;

[0014] S5. Generate the harmonic admittance compensation coefficient matrix:

[0015]

[0016] Among them, Z base C is the rated impedance of the inverter. eq As the equivalent capacitance, the harmonic admittance compensation matrix independently adjusts the phase and amplitude for each harmonic to suppress resonance and harmonic amplification.

[0017] S6. Model predictive control (MPC) is used to generate PWM signals, with the objective function being:

[0018]

[0019] in, λ is the weighting coefficient. Model predictive control (MPC) optimizes PWM generation, quickly tracks the compensated current reference value, and reduces transient response time.

[0020] Furthermore, the anti-aliasing filter in step S1 is a fourth-order Butterworth low-pass filter with a cutoff frequency of Passband ripple ≤0.1dB, stopband attenuation ≥40dB, ensuring a flat passband and steep stopband attenuation to suppress high-frequency noise.

[0021] Furthermore, the windowed FFT analysis in step S2 includes applying a Blackman-Harris window to the N=1024 data points per period to reduce sidelobe amplitude and decrease spectral leakage of non-integer harmonics. The window function expression is as follows:

[0022]

[0023] Where a0 = 0.35875, a1 = 0.48829, a2 = 0.14128, and amplitude correction factors are used. Compensation for spectral leakage and amplitude attenuation caused by the compensation window function ensures that the calculation error of Vn / In is less than 0.5%.

[0024] Furthermore, the implementation of the recursive least squares method in step S3 includes:

[0025] An ARX model is established to accurately describe the dynamic relationship between grid voltage and current, covering the main transient processes:

[0026]

[0027] Among them, the model order p = 4, the forgetting factor μ = 0.98 balances the weights of historical data and new data, and adapts to impedance changes within 5ms;

[0028] Parameter update formula:

[0029]

[0030] Where, φ[k]=[i inv [k-1],…,i inv [kp]] T covariance matrix Covariance matrix updates ensure fast algorithm convergence (<2ms) and avoid parameter drift.

[0031] Furthermore, in step S5, the phase compensation angle θ comp The calculation of (n) satisfies the condition that the phase deviation introduced by the virtual inductance ΔL is corrected to ensure phase matching of each harmonic admittance. The phase compensation formula is as follows:

[0032]

[0033] Where, ΔL=L vir -L base and equivalent capacitance The equivalent capacitance is designed based on the fundamental frequency parameters, avoiding additional capacitance measurements and simplifying implementation.

[0034] Furthermore, the weighting coefficient λ of the model predictive control in step S6 satisfies:

[0035]

[0036] Among them, V dc f is the DC bus voltage. sw The switching frequency is used as the weighting formula to dynamically balance the current tracking accuracy and voltage change rate, avoiding oscillations near the switching frequency.

[0037] An adaptive impedance matching system for implementing the above method includes:

[0038] Voltage and current sampling module: equipped with AD7606 chip, 16-bit resolution, supports 8-channel synchronous sampling;

[0039] Impedance calculation module: Parallelized RLS algorithm implemented using FPGA, with a calculation delay ≤5μs;

[0040] Dynamic matching controller: integrates an STM32H7 processor, operates at a frequency of 480MHz, and stores a compensation coefficient lookup table;

[0041] PWM generation module: based on SiC MOSFET driver, switching frequency f sw =20kHz, dead time t d =100ns ° .

[0042] The beneficial effects of this invention are as follows:

[0043] 1. This invention effectively avoids spectral aliasing by real-time acquisition of grid voltage and inverter output current, followed by precise anti-aliasing filtering, ensuring data accuracy. Subsequently, windowed FFT analysis is used to accurately extract the fundamental and specific harmonic components, providing a reliable data foundation for subsequent impedance calculation and harmonic compensation. By calculating the grid impedance in real-time using the recursive least squares method and dynamically adjusting the virtual inductance on the inverter side, effective compensation for fundamental impedance mismatch is achieved, significantly improving system stability. Furthermore, this invention effectively suppresses resonance and harmonic amplification by generating a harmonic admittance compensation coefficient matrix and independently adjusting the phase and amplitude of each harmonic. Finally, model predictive control technology is used to optimize PWM signal generation, enabling rapid tracking of the compensated current reference value and further reducing transient response time. In summary, this invention, through a series of innovative technical means, achieves adaptive impedance matching of the grid-connected inverter under grid voltage transient disturbances, significantly improving system stability and performance.

[0044] 2. Through the coordinated operation of the above modules, the system can monitor transient disturbances in the grid voltage in real time and quickly adjust the inverter's output impedance to achieve precise matching with the grid impedance. The voltage and current sampling module accurately captures changes in voltage and current, providing basic data for impedance calculation. The impedance calculation module quickly and accurately calculates the current grid impedance, providing a decision-making basis for the dynamic matching controller. Based on the impedance calculation results, the dynamic matching controller adjusts the compensation coefficients in the lookup table in real time and precisely controls the inverter's output through the PWM generation module, thereby achieving adaptive impedance matching to transient disturbances in the grid voltage. This adaptive impedance matching system not only improves the stability and efficiency of the grid-connected inverter but also effectively reduces the impact of transient disturbances in the grid voltage on inverter operation. Detailed Implementation

[0045] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below.

[0046] This invention provides an adaptive impedance matching method for grid-connected inverters to withstand transient disturbances in grid voltage, comprising the following steps:

[0047] S1, with sampling frequency f s ≥20kHz synchronous acquisition of grid voltage v g (t) and inverter output current i inv (t), after anti-aliasing filtering, yields the discrete sequence v. g [k] and i inv [k] Real-time sampling (above 20kHz) ensures the capture of high-frequency characteristics of power grid transient disturbances, and anti-aliasing filtering avoids spectral aliasing.

[0048] The anti-aliasing filter is a fourth-order Butterworth low-pass filter with a cutoff frequency of With passband ripple ≤0.1dB and stopband attenuation ≥40dB, the system ensures a flat passband and steep stopband attenuation, suppressing high-frequency noise. Precise filter design addresses harmonic analysis errors caused by phase nonlinearity in traditional RC filters. By improving sampling accuracy and filtering performance, it ensures accurate measurement of grid voltage and inverter output current, providing a reliable data foundation for subsequent analysis.

[0049] S2, for v g [k] and i inv [k] Perform windowed FFT analysis to extract the complex voltages of the fundamental frequency and the 6n±1st harmonics (n=1,2,…,13). and complex current Windowed FFT analysis is used to extract the fundamental frequency and specific subharmonics, covering typical harmonic components of the power grid;

[0050] By accurately analyzing the harmonic components of the grid voltage and inverter output current, the dynamic changes in grid impedance are effectively identified, providing crucial information for subsequent adaptive impedance matching. The windowed FFT analysis employs the Blackman-Harris window function, significantly reducing spectral leakage, improving the accuracy of harmonic extraction, and ensuring the reliability of impedance calculation.

[0051] Windowed FFT analysis involves applying a Blackman-Harris window to N=1024 points of data per period to reduce sidelobe amplitude and decrease spectral leakage from non-integer harmonics. The window function expression is as follows:

[0052]

[0053] Where w[k] represents the k-th value of the window function, k is the sampling point index in the window function, a0, a1, and a2 represent the coefficients of the window function, a0 = 0.35875, a1 = 0.48829, a2 = 0.14128, k represents the sampling point index in the window function, and its value ranges from 0 to N-1, i.e., 0 to 1023. 2πk / N and 4πk / N represent the angle parameters in the window function, which are related to the sampling point index k and the window function length N. These angle parameters are used to calculate the shape of the window function, using an amplitude correction factor. Compensation for spectral leakage and amplitude attenuation caused by the compensation window function ensures that the calculation error of Vn / In is less than 0.5%.

[0054] Windowed FFT analysis effectively improves the accuracy of harmonic analysis. By accurately calculating the coefficients of the window function and applying it to each sampling point, spectral leakage can be significantly reduced, especially for non-integer harmonics. Furthermore, this method employs an amplitude correction factor to further compensate for the amplitude attenuation caused by the window function, thereby ensuring that the calculation error of the effective values ​​of voltage and current (Vn / In) is controlled within 0.5%. This innovation not only improves the response speed and accuracy of grid-connected inverters to grid voltage transient disturbances but also enhances their stability and reliability in complex grid environments.

[0055] S3. Calculate the impedance of each harmonic grid based on the recursive least squares (RLS) method. The calculation formula is as follows:

[0056]

[0057] Among them, Z grid (n) represents the power grid impedance under the nth harmonic. This represents the voltage amplitude under the nth harmonic. This represents the current amplitude under the nth harmonic. ω represents the square of the current amplitude under the nth harmonic, where j is the imaginary unit and satisfies j^2 = -1. nL represents the angular frequency of the nth harmonic. para Assuming the line parasitic inductance, f1 = 50Hz, the RLS dynamic impedance calculation updates impedance parameters in real time using a recursive algorithm, adapting to dynamic changes in grid impedance; this significantly improves the real-time performance and accuracy of impedance calculation. Through the recursive algorithm, this invention can quickly respond to changes in grid impedance, maintaining stable impedance matching even under grid transient disturbances, thereby improving the operating efficiency and power quality of the grid-connected inverter.

[0058] The implementation of recursive least squares includes:

[0059] An ARX model is established to accurately describe the dynamic relationship between grid voltage and current, covering the main transient processes:

[0060]

[0061] Among them, the model order p = 4, the forgetting factor μ = 0.98 balances the weights of historical data and new data, and adapts to impedance changes within 5ms;

[0062] v g [k] represents the grid voltage at time k. This represents the summation operation from m=1 to m=p, where p is the order of the model. p=4 means considering the current data at four time points prior to the current time. m Let i represent the coefficients of the m-th current term. These coefficients are the parameters that the ARX model needs to fit, reflecting the degree of influence of current at different historical moments on the grid voltage at the current moment. inv [km] represents the inverter current at time km. It is assumed that the current data is known or can be obtained through measurement. e[k] represents the error or noise term at time k, representing other influencing factors not captured by the model. Forgetting factor μ: In adaptive filtering or parameter estimation, the forgetting factor μ is used to balance the weights of historical and new data. The value of μ is between 0 and 1; the closer to 1, the greater the influence of historical data on the new estimate; the closer to 0, the greater the influence of new data. μ = 0.98 means that the model pays more attention to historical data, but can also adapt to new data changes to a certain extent, especially impedance mutations within 5ms, significantly improving the dynamic adaptability and robustness of grid impedance calculation. Through the recursive algorithm, this invention can adjust the impedance matching strategy in real time, effectively responding to grid transient disturbances and ensuring the stable operation of the grid-connected inverter under various operating conditions.

[0063] Parameter update formula:

[0064]

[0065] Where, φ[k]=[i inv[k-1],…,i inv [kp]] T covariance matrix Covariance matrix updates ensure fast algorithm convergence (<2ms) and avoid parameter drift.

[0066] This formula is a recursive parameter update formula used in online learning or Recursive Least Squares (RLS) algorithms. It is used to update the value of the parameter vector θ at time step k. The symbols in the formula have the following meanings: θ[k]: The parameter vector at time step k. θ[k-1]: The parameter vector at time step k-1. P[k-1]: The covariance matrix at time step k-1. φ[k]: The regression vector at time step k, composed of the values ​​of the inverse variables. μ: The forgetting factor, used to give higher weight to new data. g[k]: The gain vector at time step k. v: A constant, which may represent the variance of noise or other parameters used to adjust the algorithm. The superscript T in the formula indicates the transpose of a matrix or vector, and the superscript -1 indicates the inverse of a matrix.

[0067] This update formula ensures that the parameter θ can adapt quickly and stably as new data arrives. At the same time, the update of the covariance matrix P guarantees the convergence speed of the algorithm and avoids parameter drift, that is, avoids instability during the parameter update process.

[0068] S4. Based on the fundamental impedance Z grid (1) = R1 + jX1, dynamically adjust the virtual inductance on the inverter side:

[0069]

[0070] Virtual inductor adjustment directly compensates for fundamental impedance mismatch and improves system stability;

[0071] This formula is used to dynamically adjust the virtual inductance value on the inverter side to compensate for fundamental impedance mismatch and improve system stability. The meanings of the parameters in the formula are as follows: Z grid (1) represents the fundamental impedance of the power grid, expressed as a complex number. R1 represents the real part of the fundamental impedance of the power grid, representing the resistive component. X1 represents the imaginary part of the fundamental impedance of the power grid, representing the reactive component. L vir This represents the value of the virtual inductance on the inverter side, which needs to be dynamically adjusted. L base This represents the basic virtual inductance value on the inverter side. X grid (1) represents the desired fundamental reactance value of the power grid. f1 represents the fundamental frequency of the power grid.

[0072] This adjustment is based on the current fundamental impedance Z of the power grid. grid (1) The real and imaginary parts of the equation are related to the desired fundamental reactance of the power grid, X. grid(1) The difference between them. By adjusting the virtual inductance value, the system stability problem caused by grid impedance mismatch can be compensated, and the adaptability and operating efficiency of the grid-connected inverter to grid voltage transient disturbances can be improved. The dynamic adjustment strategy ensures a good match between the inverter output and grid demand under different operating conditions, thereby reducing harmonic pollution and improving power quality. In addition, this method effectively identifies and compensates for the dynamic changes in grid impedance through accurate harmonic analysis and impedance calculation, enhancing the robustness and stability of the system.

[0073] S5. Generate the harmonic admittance compensation coefficient matrix:

[0074]

[0075] Among them, Z base C is the rated impedance of the inverter. eq As the equivalent capacitance, the harmonic admittance compensation matrix independently adjusts the phase and amplitude for each harmonic to suppress resonance and harmonic amplification.

[0076] This formula is used to calculate the harmonic admittance compensation coefficient matrix K_comp(n), which is a complex matrix for the nth harmonic. It is used to independently adjust the phase and amplitude of each harmonic to suppress resonance and harmonic amplification. The meanings of the parameters in the formula are as follows: Z grid (n) represents the grid impedance at the nth harmonic. Z base This indicates the rated impedance of the inverter. Let represent the phase angle of the grid impedance at the nth harmonic. arctan(2πf) n ΔL / C eq This represents the phase angle increment related to frequency (f_n), line inductance increment (ΔL), and equivalent capacitance (C_eq). Using this formula, the required harmonic admittance compensation from the inverter under specific frequency and system parameters can be calculated to ensure stable system operation, reduce the impact of harmonics on the grid, optimize the output characteristics of the grid-connected inverter, and improve its adaptability in complex grid environments. The generation of the harmonic admittance compensation coefficient matrix fully considers the dynamic characteristics of the grid impedance and the frequency characteristics of the inverter output current, ensuring the accuracy and effectiveness of the compensation. Furthermore, this method achieves independent adjustment of harmonic phase and amplitude by accurately calculating the phase angle increment, further enhancing the system's stability and reliability.

[0077] Phase compensation angle θ comp The calculation of (n) satisfies the condition that the phase deviation introduced by the virtual inductance ΔL is corrected to ensure phase matching of each harmonic admittance. The phase compensation formula is as follows:

[0078]

[0079] Where, ΔL=L vir -L base and equivalent capacitance The equivalent capacitance is designed based on the fundamental frequency parameters, avoiding additional capacitance measurements and simplifying implementation.

[0080] Phase compensation angle θ comp The calculation of (n) is to correct the phase deviation introduced by the virtual inductance ΔL, ensuring phase matching of each harmonic admittance. The formula represents a calculated angle used to adjust the phase error in the system caused by the virtual inductance ΔL. Here, ΔL is the difference between the virtual inductance and the reference inductance, and C... eq This is an equivalent capacitance designed based on the fundamental frequency f1, used to simplify the implementation process and avoid additional capacitance measurements. The ultimate goal is to ensure that the admittance phases of each harmonic in the system are matched, thereby guaranteeing system performance.

[0081] S6. Model predictive control (MPC) is used to generate PWM signals, with the objective function being:

[0082]

[0083] in, λ is the weighting coefficient. Model predictive control (MPC) optimizes PWM generation, quickly tracks the compensated current reference value, and reduces transient response time.

[0084] This formula is the objective function used in Model Predictive Control (MPC) to generate the PWM signal. The objective function aims to minimize the weighted sum of the two parts:

[0085] The first part is the predicted inverter current i inv (k+1) and the current reference value i′ after fast tracking compensation ref The square of the difference between them. Here, i′ ref It is by using the current reference value i ref,n With compensation coefficient K comp The product of (n) and summation is obtained, where (n) ranges from 1 to 13, representing 13 possible different compensation coefficients.

[0086] The second part is the voltage V at the current moment. k Compared with the voltage V at the previous moment k-1 The square of the difference between them is multiplied by the weighting coefficient λ.

[0087] The objective function aims to optimize the PWM signal so that the inverter's current output quickly tracks and approaches the compensated current reference value as closely as possible, while minimizing voltage variations to reduce the system's transient response time. The weighting coefficient λ is used to balance the relative importance of current tracking and voltage variations.

[0088] The weighting coefficients λ of the model predictive control satisfy:

[0089]

[0090] Among them, L vir V represents the filter inductance. dc f is the DC bus voltage. sw The weighting formula dynamically balances current tracking accuracy and voltage change rate to avoid oscillations near the switching frequency, ensuring stable operation of the grid-connected inverter over a wide frequency range. Through model predictive control, this method can predict and adjust the PWM signal in advance to adapt to transient disturbances in the grid voltage, thereby effectively reducing harmonic generation and improving power quality. Furthermore, the weighting formula design also considers key parameters such as filter inductance, DC bus voltage, and switching frequency, ensuring the applicability and robustness of the control strategy under different operating conditions.

[0091] By acquiring grid voltage and inverter output current in real time and performing precise anti-aliasing filtering, spectral aliasing is effectively avoided, ensuring data accuracy. Subsequently, windowed FFT analysis is used to accurately extract the fundamental and specific harmonic components, providing a reliable data foundation for subsequent impedance calculation and harmonic compensation. The grid impedance is calculated in real time using the recursive least squares method, and the virtual inductance on the inverter side is dynamically adjusted to effectively compensate for fundamental impedance mismatch, significantly improving system stability. Furthermore, this invention generates a harmonic admittance compensation coefficient matrix and independently adjusts the phase and amplitude of each harmonic, effectively suppressing resonance and harmonic amplification. Finally, model predictive control technology is used to optimize PWM signal generation, enabling rapid tracking of the compensated current reference value and further reducing transient response time. In summary, this invention, through a series of innovative technical means, achieves adaptive impedance matching of the grid-connected inverter under grid voltage transient disturbances, significantly improving system stability and performance.

[0092] An adaptive impedance matching system for implementing the above method includes:

[0093] Voltage and current sampling module: This module is equipped with an AD7606 chip, which has a high resolution of 16 bits and can support synchronous sampling of up to 8 channels.

[0094] Impedance calculation module: This module uses an FPGA chip to implement a parallelized recursive least squares (RLS) algorithm, thereby reducing the delay time of impedance calculation to no more than 5 microseconds;

[0095] Dynamic matching controller: This controller integrates an STM32H7 processor with an operating frequency of up to 480MHz, and has a built-in lookup table for storing compensation coefficients to ensure high efficiency and accuracy of dynamic matching.

[0096] PWM generation module: Based on SiC MOSFET driver design, the switching frequency of the PWM signal is set to 20kHz, and a dead time of only 100 nanoseconds is set to optimize power conversion efficiency and reduce switching losses.

[0097] Through the coordinated operation of the aforementioned modules, the system can monitor transient disturbances in grid voltage in real time and quickly adjust the inverter's output impedance to achieve precise matching with the grid impedance. The voltage and current sampling module accurately captures changes in voltage and current, providing fundamental data for impedance calculation. The impedance calculation module quickly and accurately calculates the current grid impedance, providing a decision-making basis for the dynamic matching controller. Based on the impedance calculation results, the dynamic matching controller adjusts the compensation coefficients in the lookup table in real time and precisely controls the inverter's output through the PWM generation module, thereby achieving adaptive impedance matching to transient disturbances in grid voltage. This adaptive impedance matching system not only improves the stability and efficiency of the grid-connected inverter but also effectively reduces the impact of transient disturbances in grid voltage on inverter operation.

[0098] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An adaptive impedance matching method for grid-connected inverters to withstand transient disturbances in grid voltage, characterized in that: It includes the following steps: S1, with sampling frequency f s ≥20kHz synchronous acquisition of grid voltage v g (t) and inverter output current i inv (t), after anti-aliasing filtering, yields the discrete sequence v. g [k] and i inv [k] Real-time sampling (above 20kHz) ensures the capture of high-frequency characteristics of power grid transient disturbances, and anti-aliasing filtering avoids spectral aliasing. S2, for v g [k] and i inv [k] Perform windowed FFT analysis to extract the complex voltages of the fundamental frequency and the 6n±1st harmonics (n=1,2,…,13). and complex current Windowed FFT analysis is used to extract the fundamental frequency and specific subharmonics, covering typical harmonic components of the power grid; S3. Calculate the impedance of each harmonic grid based on the recursive least squares (RLS) method. The calculation formula is as follows: Where L para For the parasitic inductance of the line, f1 = 50Hz, the RLS dynamic impedance calculation updates the impedance parameters in real time through a recursive algorithm to adapt to the dynamic changes in the grid impedance. S4. Based on the fundamental impedance Z grid (1) = R1 + jX1, dynamically adjust the virtual inductance on the inverter side: Virtual inductor adjustment directly compensates for fundamental impedance mismatch and improves system stability; S5. Generate the harmonic admittance compensation coefficient matrix: Among them, Z base C is the rated impedance of the inverter. eq As the equivalent capacitance, the harmonic admittance compensation matrix independently adjusts the phase and amplitude for each harmonic to suppress resonance and harmonic amplification. S6. Model predictive control (MPC) is used to generate PWM signals, with the objective function being: in, λ is the weighting coefficient. Model predictive control (MPC) optimizes PWM generation, quickly tracks the compensated current reference value, and reduces transient response time.

2. The adaptive impedance matching method for grid-connected inverter grid voltage transient disturbances according to claim 1, characterized in that: In step S1, the anti-aliasing filter is a fourth-order Butterworth low-pass filter with a cutoff frequency of [missing information]. Passband ripple ≤0.1dB, stopband attenuation ≥40dB, ensuring a flat passband and steep stopband attenuation to suppress high-frequency noise.

3. The adaptive impedance matching method for grid-connected inverter grid voltage transient disturbances according to claim 1, characterized in that: The windowed FFT analysis in step S2 includes applying a Blackman-Harris window to the N=1024 data points per period to reduce sidelobe amplitude and decrease spectral leakage of non-integer harmonics. The window function expression is as follows: Where a0 = 0.35875, a1 = 0.48829, a2 = 0.14128, and amplitude correction factors are used. Compensation for spectral leakage and amplitude attenuation caused by the compensation window function ensures that the calculation error of Vn / In is less than 0.5%.

4. The adaptive impedance matching method for grid-connected inverter grid voltage transient disturbance according to claim 1, characterized in that: The implementation of the recursive least squares method in step S3 includes: An ARX model is established to accurately describe the dynamic relationship between grid voltage and current, covering the main transient processes: Among them, the model order p = 4, the forgetting factor μ = 0.98 balances the weights of historical data and new data, and adapts to impedance changes within 5ms; Parameter update formula: Where, φ[k]=[i inv [k-1],…,i inv [kp]] T covariance matrix Covariance matrix updates ensure fast algorithm convergence (<2ms) and avoid parameter drift.

5. The adaptive impedance matching method for grid-connected inverter grid voltage transient disturbance according to claim 1, characterized in that: In step S5, the phase compensation angle θ comp The calculation of (n) satisfies the condition that the phase deviation introduced by the virtual inductance ΔL is corrected to ensure phase matching of each harmonic admittance. The phase compensation formula is as follows: Where, ΔL=L vir -L base and equivalent capacitance The equivalent capacitance is designed based on the fundamental frequency parameters, avoiding additional capacitance measurements and simplifying implementation.

6. The adaptive impedance matching method for grid-connected inverter grid voltage transient disturbance according to claim 1, characterized in that: In step S6, the weight coefficient λ of the model predictive control satisfies: Among them, V dc f is the DC bus voltage. sw The switching frequency is used as the weighting formula to dynamically balance the current tracking accuracy and voltage change rate, avoiding oscillations near the switching frequency.

7. An adaptive impedance matching system implementing the method of any one of claims 1-6, characterized in that, include: Voltage and current sampling module: equipped with AD7606 chip, 16-bit resolution, supports 8-channel synchronous sampling; Impedance calculation module: Parallelized RLS algorithm implemented using FPGA, with a calculation delay ≤5μs; Dynamic matching controller: integrates an STM32H7 processor, operates at a frequency of 480MHz, and stores a compensation coefficient lookup table; PWM generation module: based on SiC MOSFET driver, switching frequency f sw =20kHz, dead time t d =100ns ° .

Images