SVG-based photovoltaic grid-connected reactive power compensation method and system

Abstract

The invention discloses an SVG-based photovoltaic grid-connected reactive power compensation method and system, and relates to the technical field of photovoltaic grid-connected compensation, and the method comprises the steps: monitoring the output power fluctuation of a photovoltaic array, and building a first electrical parameter feature vector; photovoltaic grid-connected impedance fluctuation characteristics are monitored, and a second electrical parameter feature vector is established; extracting a dynamic harmonic component and a reactive demand spectrum; generating a dynamic adjustment instruction set of SVG trigger pulse phase-amplitude according to the dynamic harmonic component and the reactive demand spectrum; and issuing the dynamic adjustment instruction set to a carrier phase shift controller of the SVG power unit for grid-connected interleaving compensation modulation. The technical problems that in the prior art, due to output power fluctuation and impedance change, harmonic waves of a photovoltaic grid-connected system are difficult to restrain, and then the electric energy quality is poor, and the system operation stability is insufficient are solved, and the technical effects that optimal modulation of SVG trigger pulses is achieved, and the electric energy quality and the operation stability of the photovoltaic grid-connected system are improved are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic grid-connected compensation, and in particular to a photovoltaic grid-connected reactive power compensation method and system based on SVG. Background Art

[0002] With the rapid development of the photovoltaic industry, the scale of photovoltaic grid-connected systems continues to expand. However, factors such as the volatility and randomness of photovoltaic array output power, as well as the inherent characteristics of inverters, can lead to problems in grid-connected systems such as high harmonic content, low power factor, and three-phase voltage imbalance. Traditional reactive power compensation methods struggle to adapt quickly to the dynamic changes of photovoltaic systems and cannot effectively address the problems of degraded power quality and insufficient system stability caused by the coupling of harmonics and reactive power.

[0003] The existing technology has the technical problem that photovoltaic grid-connected systems are difficult to suppress harmonics due to output power fluctuations and impedance changes, which in turn causes poor power quality and insufficient system operation stability. Summary of the Invention

[0004] The present application provides a photovoltaic grid-connected reactive power compensation method and system based on SVG, which is used to solve the technical problem in the prior art that photovoltaic grid-connected systems are difficult to suppress harmonics due to output power fluctuations and impedance changes, thereby causing poor power quality and insufficient system operation stability.

[0005] In view of the above problems, the present application provides a photovoltaic grid-connected reactive power compensation method and system based on SVG.

[0006] In a first aspect of the present application, a method for reactive power compensation of photovoltaic grid-connected devices based on SVG is provided, the method comprising:

[0007] Monitor the output power fluctuation of the photovoltaic array, synchronously obtain the voltage ripple characteristics and the dynamic change of the power factor from the DC side of the inverter, and establish a first electrical parameter characteristic vector; monitor the fluctuation characteristics of the photovoltaic grid-connected impedance, synchronously obtain the harmonic distortion rate and the three-phase voltage imbalance from the AC side of the inverter, and establish a second electrical parameter characteristic vector; based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector, perform cross-time scale harmonic and reactive coupling characteristic analysis, and extract dynamic harmonic components and reactive power demand spectrum; based on the dynamic harmonic components and reactive power demand spectrum, combine the weight distribution mechanism to generate a dynamic adjustment instruction set of the SVG trigger pulse phase-amplitude, and the dynamic adjustment instruction set performs nonlinear compensation priority sorting based on the improved quasi-proportional resonant control; the dynamic adjustment instruction set is sent to the carrier phase shift controller of the SVG power unit for grid-connected interleaved compensation modulation.

[0008] A second aspect of the present application provides a photovoltaic grid-connected reactive power compensation system based on SVG, the system comprising:

[0009] A first electrical parameter characteristic vector establishment module is used to monitor the fluctuation of photovoltaic array output power, synchronously obtain voltage ripple characteristics and dynamic changes in power factor from the DC side of the inverter, and establish a first electrical parameter characteristic vector; a second electrical parameter characteristic vector establishment module is used to monitor the fluctuation characteristics of photovoltaic grid-connected impedance, synchronously obtain harmonic distortion rate and three-phase voltage imbalance from the AC side of the inverter, and establish a second electrical parameter characteristic vector; a reactive power demand spectrum extraction module is used to perform cross-time scale harmonic and reactive power coupling characteristic analysis based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector, and extract dynamic harmonic components and reactive power demand spectrum; a dynamic adjustment instruction set generation module is used to generate a dynamic adjustment instruction set of SVG trigger pulse phase-amplitude based on the dynamic harmonic components and reactive power demand spectrum in combination with a weight distribution mechanism, and the dynamic adjustment instruction set is based on improved quasi-proportional resonant control to perform nonlinear compensation priority sorting; a grid-connected interleaved compensation modulation module is used to send the dynamic adjustment instruction set to the carrier phase shift controller of the SVG power unit for grid-connected interleaved compensation modulation.

[0010] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0011] The system monitors the output power fluctuations of the photovoltaic array, synchronously obtains the voltage ripple characteristics and dynamic changes in power factor from the DC side of the inverter, and establishes a first electrical parameter characteristic vector. The system also monitors the fluctuation characteristics of the photovoltaic grid-connected impedance, synchronously obtains the harmonic distortion rate and three-phase voltage imbalance from the AC side of the inverter, and establishes a second electrical parameter characteristic vector. It also conducts cross-timescale harmonic and reactive coupling analysis to extract dynamic harmonic components and reactive power demand spectra. It also generates a dynamic adjustment instruction set for the phase-amplitude of the SVG trigger pulse, which prioritizes nonlinear compensation based on improved quasi-proportional resonant control. The dynamic adjustment instruction set is then sent to the carrier phase-shift controller of the SVG power unit for grid-connected interleaved compensation modulation. This achieves the technical effect of optimizing the modulation of the SVG trigger pulse and improving the power quality and operational stability of the photovoltaic grid-connected system. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0013] Figure 1 A schematic flow chart of a photovoltaic grid-connected reactive power compensation method based on SVG provided in an embodiment of the present application;

[0014] Figure 2 A schematic diagram of the structure of a photovoltaic grid-connected reactive power compensation system based on SVG provided in an embodiment of the present application.

[0015] Description of the reference numerals: first electrical parameter characteristic vector establishing module 10 , second electrical parameter characteristic vector establishing module 20 , reactive power demand spectrum extracting module 30 , dynamic adjustment instruction set generating module 40 , grid-connected interleaved compensation modulating module 50 . DETAILED DESCRIPTION

[0016] This application provides a photovoltaic grid-connected reactive power compensation method and system based on SVG, which is used to solve the technical problem in the prior art that photovoltaic grid-connected systems are difficult to suppress harmonics due to output power fluctuations and impedance changes, thereby causing poor power quality and insufficient system operation stability.

[0017] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making any creative work are within the scope of protection of this application.

[0018] Example 1, as Figure 1 As shown, the present application provides a photovoltaic grid-connected reactive power compensation method based on SVG, the method comprising:

[0019] Step S100: monitoring the output power fluctuation of the photovoltaic array, synchronously obtaining the voltage ripple characteristics and the dynamic change of the power factor from the DC side of the inverter, and establishing a first electrical parameter characteristic vector.

[0020] Specifically, high-precision power sensors are used to monitor the output power fluctuations of the photovoltaic array in real time, capturing information such as the amplitude and frequency of power changes. Simultaneously, voltage ripple monitoring equipment and power factor meters are used to synchronously obtain voltage ripple characteristics and dynamic changes in power factor from the DC side of the inverter. Voltage ripple characteristics, including ripple amplitude and frequency, reflect the stability of the DC voltage; dynamic changes in power factor demonstrate the dynamic changes in the effective utilization of power on the DC side. This data is integrated and arranged according to a specific order and rules to construct the first electrical parameter feature vector.

[0021] Step S200: monitor the impedance fluctuation characteristics of the photovoltaic grid, synchronously obtain the harmonic distortion rate and three-phase voltage imbalance from the AC side of the inverter, and establish a second electrical parameter feature vector.

[0022] Specifically, a second electrical parameter feature vector is constructed through multi-dimensional electrical parameter monitoring to provide AC-side state representation for subsequent harmonic and reactive coupling analysis. First, an impedance spectrum analyzer is used to track the fluctuation characteristics of the photovoltaic grid-connected impedance in real time, covering a wide frequency range of 0.1Hz to 10kHz, capturing dynamic impedance changes caused by factors such as sudden changes in sunlight and changes in inverter switching frequency. Simultaneously, a high-precision power quality analyzer is used to collect harmonic distortion (THD) from the inverter AC side, decompose it into 50 harmonic components, and calculate the three-phase voltage imbalance (including the percentage of negative-sequence voltage components and the percentage of zero-sequence voltage components). After aligning these parameters in time series, a feature vector with 54 dimensions is constructed, including the real and imaginary parts of the grid-connected impedance, the harmonic distortion rates of each order (THD1-THD50), the negative-sequence voltage percentage, and the zero-sequence voltage percentage. To enhance the vector's ability to characterize the system's dynamic characteristics, each parameter is described using three statistical measures: mean, standard deviation, and rate of change within a sliding window (100ms time window, 10ms sliding step). This ultimately forms a 162-dimensional second electrical parameter feature vector. This vector is transmitted to the SVG control unit in real time via the OPC UA protocol, providing a precise data foundation for subsequent cross-timescale coupling characteristic analysis.

[0023] Step S300: performing cross-time-scale harmonic and reactive coupling characteristic analysis based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector to extract dynamic harmonic components and reactive power demand spectrum.

[0024] Specifically, based on the first electrical parameter feature vector (constructed by monitoring the PV array output power fluctuations and simultaneously obtaining voltage ripple characteristics and power factor dynamic changes from the inverter's DC side) and the second electrical parameter feature vector (constructed by monitoring the PV grid impedance fluctuation characteristics and simultaneously obtaining harmonic distortion and three-phase voltage imbalance from the inverter's AC side), a sliding time window grey correlation analysis method is used to accurately determine the coupling strength coefficient between the dynamic change of the DC-side power factor and the three-phase voltage imbalance on the AC side. This coefficient is the key basis for feature fusion. The timestamp-frequency band alignment table generated based on it has a one-to-one mapping relationship with the subsequent dynamic adjustment instruction set, providing the basis for subsequent precise control. Next, the information contained in these two vectors is further explored to extract the implicit coupling characteristics under grid impedance fluctuations. Based on these implicit coupling characteristics, a comprehensive assessment of the DC-side voltage ripple energy mutation events and the AC-side subharmonic amplitude growth is conducted, thereby configuring the harmonic coupling strength matrix. Using the harmonic coupling strength matrix, the overlapping energy spectral density of the DC-side voltage ripple energy mutation frequency band and the AC-side subharmonics is extracted to generate an initial set of resonance risk frequency bands. Subsequently, a timestamp-band alignment table is used to screen this initial set of resonance risk frequency bands, removing bands that do not meet the requirements and obtaining candidate resonance risk frequency bands. Taking into account the time delay characteristics of switches in actual circuits, the candidate resonance risk frequency bands are then aligned with these bands to accurately determine the harmonic resonance risk frequency bands and their corresponding dynamic harmonic components, while simultaneously extracting the reactive power demand spectrum.

[0025] Step S400: generating a dynamic adjustment instruction set of the SVG trigger pulse phase-amplitude according to the dynamic harmonic components and reactive power demand spectrum in combination with a weight distribution mechanism, wherein the dynamic adjustment instruction set performs nonlinear compensation priority sorting based on improved quasi-proportional resonant control.

[0026] Specifically, a weight allocation model is constructed based on multiple factors, including harmonic frequency, amplitude, and the urgency of reactive power demand. For example, higher weights are assigned to low-order harmonics with large amplitudes and a significant impact on power quality; the weight of rapidly changing reactive power demand is also increased accordingly. This model is used to calculate the weights of each harmonic component and reactive power demand. These weights are then incorporated into the command generation process. Based on the principles of improved quasi-proportional resonant control, this control adds adaptive parameter adjustment to traditional quasi-proportional resonant control, enabling more accurate tracking and compensation of harmonics and reactive power. Based on the weights of different harmonic components and reactive power demands, their compensation priorities are nonlinearly ranked. For example, compensation of harmonic components and reactive power demands with higher weights is prioritized to maximize system power quality. After the ranking is completed, the required phase and amplitude of the SVG trigger pulse are calculated based on the operation rules of the improved quasi-proportional resonant control and the weight allocation results. These phase and amplitude parameters are integrated into a dynamic adjustment instruction set. Each instruction in the instruction set specifies the trigger pulse phase and amplitude that the SVG should output at a specific moment. This enables precise control of the SVG, ensuring that it can efficiently perform compensation operations based on the system's real-time harmonic and reactive power conditions, thereby optimizing the operating performance of the PV grid-connected system.

[0027] Step S500: Send the dynamic adjustment instruction set to the carrier phase shift controller of the SVG power unit to perform grid-connected interleaved compensation modulation.

[0028] Specifically, a dynamic adjustment instruction set is transmitted to the carrier phase-shift controller of the SVG power unit. During transmission, a high-speed communication bus ensures accurate and timely delivery of the instructions, minimizing signal transmission delays and interference. After receiving the instruction set, the carrier phase-shift controller performs grid-connected interleaving compensation modulation on the SVG power unit based on the trigger pulse phase-amplitude information specified in the instruction. First, a phase margin analysis is performed on the harmonic resonance risk frequency band at the grid-connected interleaving point. If the phase margin falls below a set threshold, it indicates a potential harmonic resonance risk. Frequency bands that may cause impedance mismatch need to be filtered out to prevent system performance degradation. The selected impedance mismatch frequency bands are then processed using an adaptive notch filter. The adaptive notch filter automatically adjusts the transfer function's notch frequency and bandwidth parameters based on the characteristics of the impedance mismatch frequency bands, thereby deploying a cascaded complex notch filter group. These notch filter groups dissipate the impedance values ​​corresponding to multiple impedance mismatch frequency bands in the equivalent impedance trajectory, dynamically correcting the equivalent impedance trajectory and ensuring more stable system operation. During the modulation process, a compensation current component consisting of fundamental reactive and high-frequency harmonic components is distributed to the H-bridge power unit. The fundamental reactive component utilizes multi-carrier phase-shifting technology to achieve current sharing among power devices, improving their lifespan and operating efficiency. The high-frequency harmonic component uses dynamic carrier recombination technology to optimize switching frequency, reducing the impact of high-frequency harmonics on the system. Through this series of operations, the SVG power unit can accurately output compensation current based on the system's real-time harmonic and reactive power requirements, achieving efficient grid-connected interleaved compensation modulation and improving the power quality and stability of the photovoltaic grid-connected system.

[0029] In one possible implementation, step S100 further includes:

[0030] Step S110: extracting a subsynchronous oscillation component in the voltage ripple based on the voltage ripple characteristics.

[0031] Step S120: constructing a dynamic impedance matching network according to the subsynchronous oscillation component in the voltage ripple, wherein the dynamic impedance matching network adjusts the equivalent impedance trajectory.

[0032] Specifically, based on the acquired voltage ripple characteristics, multi-resolution signal decomposition technology is used to extract the subsynchronous oscillation components. The voltage ripple signal is decomposed into a specific frequency band (for example, 5-25Hz) through wavelet packet transform, and the number of decomposition layers is set to 5, each layer corresponding to a different frequency interval. The Hilbert transform is performed on the decomposed frequency band signals to calculate their instantaneous frequency and amplitude, and the oscillation components with a frequency lower than the power frequency (50 / 60Hz) and an amplitude exceeding the threshold (for example, 0.5% of the rated voltage) are identified as subsynchronous oscillation components. At the same time, in order to improve the extraction accuracy, the Kalman filter is used to dynamically track the identified subsynchronous oscillation components, and the state transfer matrix is ​​set to the diagonal matrix diag (0.95, 0.98). The observation matrix is ​​dynamically adjusted according to the signal characteristics.

[0033] A dynamic impedance matching network is constructed based on the extracted subsynchronous oscillation component. First, the equivalent admittance matrix in the dq coordinate system is calculated based on the frequency and amplitude characteristics of the subsynchronous oscillation component. Next, a dynamic impedance matching network consisting of an IGBT module and an LC resonant branch is designed. The IGBT module adopts an SVPWM modulation strategy with a switching frequency set to 2-4kHz. The parameters of the LC resonant branch are dynamically adjusted according to the subsynchronous oscillation frequency. This network adjusts the equivalent impedance trajectory to maintain the system's impedance angle within a ±30° range at the subsynchronous frequency and the impedance amplitude within a range of 0.8-1.2 times the rated impedance. By monitoring the subsynchronous oscillation component in real time, the IGBT trigger pulse phase and duty cycle, as well as the capacitance and inductance of the LC resonant branch, are dynamically adjusted to ensure that the equivalent impedance trajectory remains stable. This process is performed on an FPGA real-time processing platform with a processing cycle of 100μs to meet the requirements of fast dynamic response.

[0034] In one possible implementation, step S120 further includes:

[0035] Step S121: extracting a current-impedance coupling coefficient according to a subsynchronous oscillation component in the voltage ripple.

[0036] Step S122: Based on the dynamic impedance matching network, combined with the current-impedance coupling coefficient and the compensation current component, reverse coupling adjustment is performed on the equivalent impedance trajectory.

[0037] Specifically, the circuit is accurately modeled based on the subsynchronous oscillation component extracted from the voltage ripple. Since the subsynchronous oscillation component's frequency ranges from 5 to 15 Hz, the circuit is transformed within this specific frequency range to simplify the complex circuit into an equivalent circuit model consisting of a power supply, resistors, inductors, and capacitors. In this model, the subsynchronous oscillation component serves as the excitation source. Fundamental circuit laws, such as Ohm's law and Kirchhoff's laws, are used to establish the mathematical relationship between voltage, current, and impedance. Multiple monitoring points are set in the circuit. High-precision current and voltage sensors are used to synchronously collect current and voltage data at each monitoring point under the influence of the subsynchronous oscillation component. The sampling frequency is set to 1 kHz to accurately capture the changes in the subsynchronous oscillation component. The collected data are then fitted and analyzed using the least squares method. A fitting equation is constructed, using current data as the independent variable and voltage data after equivalent impedance calculation as the dependent variable. The fitting parameters are continuously adjusted to minimize the sum of squared errors between the fitted curve and the actual data. After multiple iterative calculations, the linear relationship coefficient between current and equivalent impedance at the subsynchronous oscillation frequency is obtained. This coefficient is the current-impedance coupling coefficient. It reflects the degree of influence of current change on equivalent impedance under the influence of subsynchronous oscillation components, and provides key parameter basis for subsequent equivalent impedance trajectory adjustment based on dynamic impedance matching network.

[0038] Relying on the constructed dynamic impedance matching network, the obtained current-impedance coupling coefficient, combined with the compensation current component, performs reverse coupling adjustment of the equivalent impedance trajectory. This adjustment results in an inverse tracking characteristic between the equivalent impedance and the grid impedance in the subsynchronous frequency band (5-15Hz). Based on the current-impedance coupling coefficient, the influence of current changes on the equivalent impedance is determined. This principle is incorporated into the control algorithm, which adjusts the parameters of the dynamic impedance matching network based on the real-time changes in the compensation current component. The dynamic impedance matching network typically consists of components such as inductors, capacitors, and controllable switches. The control algorithm adjusts the inductor and capacitor connection structure by varying the on-time and frequency of the controllable switches, thereby changing the equivalent impedance. In the subsynchronous frequency band (5-15Hz), when the grid impedance amplitude is detected to increase, the control algorithm, based on the reverse tracking characteristic, drives the dynamic impedance matching network to reduce the equivalent impedance amplitude. Conversely, when the grid impedance amplitude decreases, the equivalent impedance amplitude is increased. The equivalent impedance trajectory and the grid impedance amplitude are adjusted inversely within the subsynchronous frequency band to optimize the system's operating state under subsynchronous oscillation and improve system stability.

[0039] In one possible implementation, step S122 further includes:

[0040] Step S1221: allocating a compensation current component including a fundamental wave reactive component and a high frequency harmonic component to the H-bridge power unit.

[0041] Step S1222: wherein the fundamental reactive component is subjected to current sharing of power devices by multi-carrier phase shifting, and the high-frequency harmonic component is subjected to switching frequency optimization by dynamic carrier recombining.

[0042] Specifically, to effectively compensate for system harmonics and reactive power, the compensation current components, including fundamental reactive power and high-frequency harmonics, must be precisely allocated to the H-bridge power units. First, the total system compensation current is analyzed and calculated, and the fundamental reactive power component is separated using instantaneous reactive power theory. The magnitude and direction of this fundamental reactive power component are determined based on the grid's real-time power factor and reactive power demand. Simultaneously, spectrum analysis techniques such as Fourier transforms are used to extract high-frequency harmonic components from the total current, determining the amplitude and phase information of each high-frequency harmonic. Next, a reasonable allocation strategy is developed based on the performance parameters of each power device in the H-bridge power unit, its current operating status, and the overall system operation. For example, the fundamental reactive power and high-frequency harmonic components are allocated based on the heat dissipation and load-bearing capacity of each H-bridge power unit, ensuring that each H-bridge power unit operates efficiently and stably, collaboratively completing the system compensation task.

[0043] Different carrier modulation strategies are employed for compensation current components with different characteristics to improve system performance. For the fundamental reactive component, multi-carrier phase shifting technology is used to achieve current sharing among power devices. Multiple carriers are configured for each H-bridge power unit, and these carriers are phase-shifted by a specific angle. For example, assuming there are N H-bridge power units, the carrier phases of each unit are shifted by 360° / N. This allows the switching operations of different H-bridge power units to be staggered in time during compensation of the fundamental reactive component. Because the switching losses of each unit are evenly distributed over time, the current flowing through each power device is balanced, preventing overheating of some devices due to excessive current, extending the service life of the power devices, and improving system reliability. For high-frequency harmonic components, dynamic carrier reconfiguration is used to optimize the switching frequency. The frequency and amplitude changes of the high-frequency harmonics are monitored in real time, and the carrier combination and frequency are dynamically adjusted based on these changes. For example, when a high-frequency harmonic amplitude is detected, the carrier frequency of the corresponding H-bridge power unit is increased to enhance the ability to track and compensate for the harmonic. At the same time, the phase relationship between the carriers is properly adjusted to maintain an appropriate ratio between the switching frequency and the high-frequency harmonic frequency, effectively compensating for the harmonics while reducing switching losses. This dynamic adjustment not only ensures effective compensation for high-frequency harmonics, but also optimizes the switching frequency, reduces overall system losses, and improves the operating efficiency of the photovoltaic grid-connected system.

[0044] In one possible implementation, step S300 further includes:

[0045] Step S310: performing a sliding time window grey relational analysis based on the first electrical parameter eigenvector and the second electrical parameter eigenvector to determine a coupling strength coefficient between the dynamic change of the DC side power factor and the three-phase voltage imbalance on the AC side.

[0046] Step S320: performing feature fusion based on the dynamic change of the DC side power factor and the coupling strength coefficient of the AC side three-phase voltage imbalance, generating a timestamp-frequency band alignment table, wherein the timestamp-frequency band alignment table is mapped one-to-one with the dynamic adjustment instruction set.

[0047] Specifically, a sliding time window grey correlation analysis is performed based on the first and second electrical parameter eigenvectors to determine the coupling strength coefficient between the dynamic change in the DC power factor and the three-phase voltage imbalance on the AC side. First, a sliding time window of appropriate length is constructed, for example, set to 200ms with a step size of 20ms. As time progresses, the sliding time window sequentially sweeps through the time series data of the two eigenvectors. Within each time window, data related to the dynamic change in the DC power factor and the three-phase voltage imbalance on the AC side are accurately extracted. The dynamic change in the DC power factor is regarded as a reference series, and the three-phase voltage imbalance on the AC side is regarded as a comparison series. By calculating the absolute difference between the reference series and the comparison series at each time point, the maximum difference and the minimum difference between the two levels are determined. A resolution coefficient (usually around 0.5) is introduced, and the correlation coefficient of the two series at different time points is calculated according to the grey correlation calculation formula. The correlation coefficients within each time window are arithmetic averaged, and the coupling strength coefficient between the dynamic change of the DC side power factor and the three-phase voltage imbalance on the AC side within the time window is finally obtained. This coefficient can effectively quantify the degree of coupling between the two at different time points, providing a key basis for subsequent feature fusion and analysis.

[0048] Based on the obtained coupling strength coefficient between the dynamic change in the DC power factor and the three-phase voltage imbalance on the AC side, feature fusion is performed to generate a timestamp-frequency band alignment table, which is then mapped one-to-one with the dynamic adjustment instruction set. First, the two features, the dynamic change in the DC power factor and the three-phase voltage imbalance on the AC side, are weighted and fused, using the coupling strength coefficient as a weight. For example, if the coupling strength coefficient is 0.8 at a certain moment, it indicates a strong correlation between the two features at that moment, and a higher weight is given during fusion. Then, the fused time domain features are converted to the frequency domain using a discrete Fourier transform (DFT) to decompose the components into different frequency bands. To ensure accurate correspondence between the time and frequency dimensions, a fixed time interval (e.g., 10ms) is set as the timestamp, and the frequency domain is divided into multiple frequency bands (e.g., 0-50Hz, 50-100Hz, etc.). The energy distribution of each frequency band at each timestamp is calculated to construct an initial timestamp-frequency band matrix. To eliminate noise and smooth data fluctuations, the matrix is ​​filtered using a Savitzky-Golay filter with a window length of 5 and a polynomial order of 3. The filtered matrix is ​​then thresholded to remove frequency bands with energy contributions below 2% to reduce redundant information. The processed timestamp-frequency band matrix is ​​then mapped to a predefined dynamic adjustment instruction set. Based on system control requirements, each frequency band range is assigned a specific adjustment instruction type (such as reactive power compensation or harmonic suppression). The execution priority and parameter values ​​of the instructions are determined based on the energy intensity of the frequency band at each timestamp. To ensure the uniqueness and traceability of the mapping, each timestamp-frequency band combination is assigned a unique identifier, and a mapping comparison table is established. A hash algorithm is used to verify the consistency of the mapping, ensuring that each instruction set corresponds to a unique timestamp and frequency band combination. The resulting timestamp-frequency band alignment table contains fields such as timestamp, frequency band range, energy intensity, corresponding instruction type, and parameter value. This achieves a one-to-one mapping with the dynamic adjustment instruction set, providing accurate data support for subsequent intelligent control.

[0049] In one possible implementation, step S300 further includes:

[0050] Step S330: extracting implicit coupling characteristics under grid impedance fluctuation based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector.

[0051] Step S340: Based on the implicit coupling characteristics under grid impedance fluctuations, the DC side voltage ripple energy mutation event and the AC side subharmonic amplitude growth are evaluated, and a harmonic coupling strength matrix is ​​configured.

[0052] Step S350: determining the harmonic resonance risk frequency band and the corresponding dynamic harmonic component through the harmonic coupling strength matrix and the timestamp-frequency band alignment table.

[0053] Specifically, the implicit coupling characteristics of power grid impedance fluctuations are extracted using the first and second electrical parameter eigenvectors as data. First, these two sets of vectors are preprocessed, using the Kalman filter algorithm to remove noise and ensure data accuracy and stability. The Kalman filter's state transition matrix and observation matrix are dynamically adjusted based on the characteristics of the electrical parameters. Next, the processed vectors are subjected to multi-scale decomposition using wavelet transform technology. The db4 wavelet, which exhibits excellent time-frequency localization, is selected as the wavelet basis, and the decomposition level is set to six, enabling detailed information to be obtained across different time scales and frequency ranges. Among the numerous frequency band coefficients obtained from the decomposition, key frequency bands are identified by analyzing the correlation between electrical parameter variations and power grid impedance fluctuations within different frequency bands. The Pearson correlation coefficient between the elements of the first and second electrical parameter eigenvectors within each frequency band is calculated. Frequency bands with absolute correlation coefficients greater than 0.6 are considered key frequency bands closely related to power grid impedance fluctuations. Within these key frequency bands, parameters that reflect implicit coupling characteristics are further extracted. For example, parameters such as the phase difference, amplitude change rate, and energy distribution of voltage and current within the frequency band are calculated. These parameters are combined in a specific order to form a eigenvector, which represents the implicit coupling characteristics under grid impedance fluctuations. This approach can extract implicit coupling information from complex electrical parameters that is difficult to directly observe but has a significant impact on system performance, providing a key basis for subsequent assessment of system stability and configuration of the harmonic coupling strength matrix.

[0054] First, based on the implicit coupling characteristics of grid impedance fluctuations, we assess DC-side voltage ripple energy mutation events and AC-side subharmonic amplitude growth. For DC-side voltage ripple, we set a standard value for energy measurement. When the voltage ripple energy suddenly increases within a short period of time, exceeding this standard value by a certain percentage (e.g., 1.5 times), and this significant increase persists for a certain duration (e.g., 30 milliseconds), we determine a voltage ripple energy mutation event. For AC-side subharmonic amplitude growth, we first determine the frequency range within which the subharmonics reside (e.g., 5-15 Hz). We then continuously monitor changes in harmonic amplitude within this frequency range. When the amplitude of a particular subharmonic increases consistently over multiple consecutive sampling periods (e.g., five consecutive sampling periods) and the increase exceeds a certain percentage (e.g., 5%) of the fundamental amplitude, we determine an AC-side subharmonic amplitude growth event. Next, we use parameters such as phase difference and energy distribution from the implicit coupling characteristics of grid impedance fluctuations to calculate the correlation between DC-side voltage ripple energy mutation events and AC-side subharmonic amplitude growth. A copula function is used to establish their joint probability distribution, quantifying the nonlinear correlation between the two. A confidence level (e.g., 95%) is set to ensure the reliability of the calculated results. Finally, a harmonic coupling strength matrix is ​​constructed based on the correlation analysis results. The rows of this matrix represent different frequency bands on the DC side (for example, 0-50 Hz is divided into five sub-bands), and the columns represent different frequency bands of the AC subharmonics (assuming 5-15 Hz is divided into three sub-bands). Each element in the matrix represents the coupling strength between the corresponding DC-side frequency band and the AC-side subharmonic frequency band. This is calculated by dividing the covariance of the corresponding events in these two frequency bands by the product of their respective standard deviations, and then multiplying it by an exponential function related to the phase difference in the implicit coupling feature. This matrix is ​​then reduced in dimensionality using the singular value decomposition method, retaining only the top few singular values ​​with the greatest overall impact and a contribution exceeding 90%. This results in a simplified harmonic coupling strength matrix.

[0055] First, using the harmonic coupling strength matrix, we identify overlaps between the DC-side voltage ripple energy mutation frequency band and the AC-side subharmonic frequency band. We then calculate the energy spectral density of these overlapping regions and collect frequency bands with energy spectral densities exceeding a certain threshold to form an initial set of resonance risk frequency bands. These bands are preliminarily identified as those that may cause resonance based on harmonic coupling relationships. Next, we use a timestamp-band alignment table to screen this initial set of resonance risk frequency bands. Based on information such as the energy distribution and frequency characteristics of each frequency band at different time points, we eliminate frequency bands with a low probability of occurrence in actual operation and minimal impact on the system. This allows us to identify candidate resonance risk frequency bands and narrow the scope of risk bands. Finally, considering the time delay characteristics of the switch during operation, the candidate resonance risk frequency bands are matched with the switch delay characteristics for analysis, the degree of resonance risk of each candidate frequency band under the influence of the switch action is evaluated, the frequency band range and risk level are corrected, and finally the harmonic resonance risk frequency bands are determined. The dynamic harmonic components corresponding to these frequency bands are obtained through monitoring and calculation, including parameters such as the harmonic amplitude, phase and frequency, providing an accurate basis for subsequent system adjustments.

[0056] In one possible implementation, step S350 further includes:

[0057] Step S351: extracting the overlapping energy spectrum density of the DC side voltage ripple energy mutation frequency band and the AC side subharmonics through the harmonic coupling strength matrix to generate an initial resonance risk frequency band set.

[0058] Step S352: using the timestamp-frequency band alignment table, screening the initial resonance risk frequency band set to determine candidate resonance risk frequency bands.

[0059] Step S353: Based on the candidate resonance risk frequency band, a matching correction is performed with the delay characteristics of the switch to determine the harmonic resonance risk frequency band and the corresponding dynamic harmonic component.

[0060] Specifically, based on the harmonic coupling strength matrix, an in-depth analysis is conducted on the coupling relationship between the frequency band of the DC side voltage ripple energy mutation and the AC side subharmonic frequency band. First, the frequency band combinations whose coupling strength exceeds the set threshold (such as 0.6) are screened out from the matrix. These frequency band combinations indicate that there is a strong correlation between the DC side and the AC side in the corresponding frequency band. Then, for these strongly correlated frequency band combinations, the energy spectrum density of their overlapping areas is calculated, and the power spectrum density of the DC side voltage ripple in the corresponding frequency band is multiplied and integrated with the power spectrum density of the AC side subharmonic frequency band. By setting an energy spectrum density threshold (for example, 1.5 times the average energy spectrum density of the system), the frequency band combinations that exceed the threshold are marked as potential resonance risk bands. Finally, all potential resonance risk bands that meet the conditions are summarized and integrated to generate an initial resonance risk band set, which provides a preliminary screening basis for further determining the actual resonance risk bands.

[0061] With the help of the timestamp-band alignment table, the obtained initial resonance risk frequency band set is screened to determine the candidate resonance risk frequency bands. First, based on the energy distribution of each frequency band at different time points in the timestamp-band alignment table, the occurrence frequency and duration of each frequency band in the initial resonance risk frequency band set are counted. If a frequency band appears less than 5 times in multiple consecutive timestamps (such as 15 timestamps), or the duration of a single occurrence is less than 30ms, it means that the energy fluctuation of the frequency band may be caused by accidental factors and is not a stable source of resonance risk, and it is removed from the set. At the same time, combined with the mapping relationship between the timestamp-band alignment table and the dynamic adjustment instruction set, check whether the frequency bands in the initial resonance risk frequency band set can correspond to effective adjustment instructions. If a frequency band lacks matching instructions, or the execution success rate of the corresponding instructions in actual operation is less than 70%, it means that it is difficult for the system to effectively control the potential risks of the frequency band, and it is also excluded. Through this series of screening operations based on time characteristics and instruction matching, frequency bands with more practical risk significance are screened out from the initial set of resonance risk frequency bands and determined as candidate resonance risk frequency bands, thereby further narrowing the range of risk frequency bands.

[0062] Based on the candidate resonance risk frequency band, the delay characteristics of the power electronic switch (generally 50-100 microseconds) are taken into account and corrected to determine the final harmonic resonance risk frequency band and the corresponding dynamic harmonic component. Due to the time delay of the switching action, the harmonics will produce phase shifts during transmission and action, which in turn affects the actual occurrence of harmonic resonance. The phase difference between the harmonic frequency and the switching action frequency in the candidate resonance risk frequency band is calculated. If the phase difference is close to ±90°, it means that the switching action will intensify the interaction of harmonics, significantly increasing the resonance risk of this frequency band. In this case, the risk level of this frequency band is increased; conversely, when the phase difference is close to 0° or 180°, the switching action has a suppressive effect on the harmonics, correspondingly reducing the risk level of this frequency band. After such a matching correction, the frequency band with a higher risk level is screened out and determined as the harmonic resonance risk frequency band. At the same time, using the real-time data in the timestamp-frequency band alignment table and through algorithms such as Kalman filtering, the dynamic harmonic components such as the amplitude, phase, and frequency change rate of the harmonics in these risk frequency bands are tracked and calculated to obtain accurate dynamic harmonic parameters, providing a key basis for subsequent targeted adjustment of system parameters and suppression of harmonic resonance.

[0063] In one possible implementation, step S500 further includes:

[0064] Step S510: Analyze the phase margin of the harmonic resonance risk frequency band at the grid-connected crossover point, and filter the impedance mismatch frequency band.

[0065] Step S520: embedding the impedance mismatch frequency band into an adaptive notch filter to dynamically correct the equivalent impedance trajectory.

[0066] Specifically, a phase margin analysis is performed on identified harmonic resonance risk bands at the grid-connection intersection to identify impedance mismatch bands. First, a harmonic power flow calculation method is used to obtain the voltage and current phasor information for each harmonic resonance risk band at the grid-connection intersection. The initial phase margin for each band is calculated by calculating the phase difference between the two. Next, a phase margin judgment threshold is set, such as 45°. When the phase margin in a frequency band is less than this threshold, it indicates that the system is unstable in this frequency band and there is a potential impedance mismatch risk. Furthermore, the impedance-frequency characteristic curve is combined to analyze the impedance amplitude variation within this frequency band. If the impedance amplitude ratio within the band exceeds a reasonable range (e.g., 1:2), the band is identified as an impedance mismatch band. Through this analysis and judgment process, all frequency bands with impedance mismatch issues are screened out from the harmonic resonance risk bands, completing the filtering of the impedance mismatch bands and providing a clear target for subsequent targeted system adjustments.

[0067] The determined impedance mismatch frequency band parameters are embedded in an adaptive notch filter, and the equivalent impedance trajectory is corrected in real time by dynamically adjusting the filter parameters. The adaptive notch filter's transfer function is initialized based on the center frequency (e.g., 13.5 Hz) and bandwidth (e.g., 1 Hz) of the impedance mismatch frequency band, with a quality factor (Q) set to 30 to ensure sufficient frequency selectivity. The filter adopts an IIR structure and updates its weight coefficients in real time using the LMS algorithm. The step size is set to 0.01 to balance convergence speed and stability. Within each control cycle (e.g., 1 ms), the filter samples the grid-connected current in real time and extracts the harmonic components of the impedance mismatch frequency band using a fast Fourier transform (FFT). Based on the extracted harmonic data, the deviation vector between the current equivalent impedance trajectory and the ideal trajectory is calculated. This vector includes amplitude and phase deviations. When the amplitude deviation exceeds 0.5 Ω or the phase deviation exceeds 15°, the adaptive adjustment mechanism is triggered. During the adjustment process, the filter's center frequency, bandwidth, and attenuation factor are dynamically modified based on the deviation vector. For example, when an impedance mismatch is detected in the 13-14 Hz frequency band, the center frequency is fine-tuned to 13.5 Hz, and the bandwidth is expanded to 1.2 Hz to cover the entire mismatch band. At the same time, the attenuation factor in this frequency band is increased (up to -40 dB) to suppress harmonic amplification. To avoid overcorrection, a forgetting factor λ = 0.98 is introduced, causing the filter's weighting of historical data to decay exponentially over time. The impedance trajectory monitoring module continuously evaluates the correction effect. When the norm of the deviation vector is less than 0.2 Ω for 10 consecutive cycles, the correction is considered to have reached steady state, and the filter parameters are temporarily frozen. If the system operating state changes (such as a sudden load change) and the deviation increases again, adaptive adjustment is reactivated. Through this closed-loop feedback mechanism, the adaptive notch filter compensates for impedance mismatch in real time, allowing the equivalent impedance trajectory to approach the ideal curve across the entire frequency band, effectively suppressing harmonic resonance and improving grid-connected system stability.

[0068] In one possible implementation, step S520 further includes:

[0069] Step S521: according to the impedance mismatch frequency band, an adaptive notch filter conforming to the complex poles is set to determine the notch frequency and bandwidth parameters of the transfer function.

[0070] Step S522: deploying a cascaded complex notch filter group according to the notch frequency and bandwidth parameters of the transfer function, wherein the cascaded complex notch filter group is used to directionally dissipate the impedance values ​​corresponding to the multiple impedance mismatch frequency bands in the equivalent impedance trajectory.

[0071] Specifically, based on the determined impedance mismatch frequency band, an adaptive notch filter that meets the complex poles is set, and the notch frequency and bandwidth parameters of its transfer function are determined. First, for each impedance mismatch frequency band (for example, 13-14Hz), its center frequency is calculated, and the center frequency is set as the notch frequency of the notch filter, which is used as the key frequency point for the filter to suppress harmonics. Then, according to the width of the frequency band itself, the bandwidth parameters of the filter are determined. The setting of the bandwidth must ensure that it can effectively cover the entire impedance mismatch frequency band and ensure that all harmonics in the frequency band can be processed. In order to construct an adaptive notch filter, a biquad IIR filter structure is adopted, and deep suppression of specific frequency harmonics is achieved through complex pole configuration. In the complex domain, the poles are set within the unit circle and close to the circumference. The angle of the pole corresponds to the notch frequency, and its radius is determined according to the quality factor (Q value) required by the filter and the sampling frequency. The quality factor is determined by the ratio of the notch frequency to the bandwidth. Subsequently, the transfer function in the analog domain is converted to the digital domain using bilinear transformation. During the conversion process, parameters such as the quality factor are dynamically adjusted according to the real-time operating status and the requirements for filter performance to optimize the frequency response characteristics of the filter, ensuring that it can form sufficiently deep attenuation in the impedance mismatch frequency band while maintaining good signal transmission performance in other frequency bands, thereby achieving effective suppression of harmonics and optimization of system impedance.

[0072] Based on the determined transfer function notch frequency and bandwidth parameters, a cascaded complex notch filter bank, consisting of multiple complex notch filters connected in series, is deployed to precisely correct the impedance values ​​of multiple impedance mismatch bands in the equivalent impedance trace. For each identified impedance mismatch band, a separate complex notch filter is designed. This filter achieves extremely high attenuation at the specific notch frequency, typically reaching -40dB or even lower, while maintaining high signal transparency in other frequency regions. When multiple such filters are cascaded, a comprehensive processing system for multiple impedance mismatch bands is formed. During real-time grid current monitoring, the cascaded complex notch filter bank selectively suppresses harmonic currents in specific frequency bands. By dissipating the harmonic energy in these bands, the system's equivalent impedance in the corresponding bands gradually approaches the ideal design value. Simultaneously, the correction results are continuously evaluated and the parameters of each notch filter are dynamically adjusted to ensure precise dissipation of multiple impedance mismatch bands under different operating conditions, effectively suppressing harmonic amplification and improving system stability and power quality.

[0073] The second embodiment is based on the same inventive concept as the photovoltaic grid-connected reactive power compensation method based on SVG in the above embodiment. Figure 2 As shown, the present application provides a photovoltaic grid-connected reactive power compensation system based on SVG. The system and method embodiments in the present application are based on the same inventive concept. The system includes:

[0074] The first electrical parameter characteristic vector establishing module 10 is used to monitor the output power fluctuation of the photovoltaic array, synchronously obtain the voltage ripple characteristics and the dynamic change of the power factor from the DC side of the inverter, and establish the first electrical parameter characteristic vector.

[0075] The second electrical parameter characteristic vector establishing module 20 is used to monitor the photovoltaic grid impedance fluctuation characteristics, synchronously obtain the harmonic distortion rate and three-phase voltage imbalance from the inverter AC side, and establish the second electrical parameter characteristic vector.

[0076] The reactive power demand spectrum extraction module 30 is configured to perform cross-time-scale harmonic and reactive power coupling characteristic analysis based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector, and extract dynamic harmonic components and reactive power demand spectrum.

[0077] The dynamic adjustment instruction set generation module 40 is used to generate a dynamic adjustment instruction set of the SVG trigger pulse phase-amplitude based on the dynamic harmonic components and reactive power demand spectrum in combination with a weight distribution mechanism. The dynamic adjustment instruction set is based on the improved quasi-proportional resonant control to perform nonlinear compensation priority sorting.

[0078] The grid-connected staggered compensation modulation module 50 is configured to send the dynamic adjustment instruction set to the carrier phase shift controller of the SVG power unit to perform grid-connected staggered compensation modulation.

[0079] Furthermore, the system is also used for the following functions:

[0080] Based on the voltage ripple characteristics, a subsynchronous oscillation component in the voltage ripple is extracted; and according to the subsynchronous oscillation component in the voltage ripple, a dynamic impedance matching network is constructed, and the dynamic impedance matching network adjusts the equivalent impedance trajectory.

[0081] Furthermore, the system is also used for the following functions:

[0082] According to the subsynchronous oscillation component in the voltage ripple, a current-impedance coupling coefficient is extracted; based on the dynamic impedance matching network, combined with the current-impedance coupling coefficient and the compensation current component, the equivalent impedance trajectory is reversely coupled and adjusted.

[0083] Furthermore, the system is also used for the following functions:

[0084] A compensation current component including a fundamental reactive component and a high-frequency harmonic component is distributed to the H-bridge power unit; wherein the fundamental reactive component is current-balanced among power devices by multi-carrier phase shifting, and the high-frequency harmonic component is optimized for switching frequency by dynamic carrier reorganization.

[0085] Furthermore, the system is also used for the following functions:

[0086] A sliding time window grey correlation analysis is performed based on the first electrical parameter eigenvector and the second electrical parameter eigenvector to determine the coupling strength coefficient between the dynamic change of the DC side power factor and the three-phase voltage imbalance on the AC side; feature fusion is performed based on the coupling strength coefficient between the dynamic change of the DC side power factor and the three-phase voltage imbalance on the AC side to generate a timestamp-frequency band alignment table, and the timestamp-frequency band alignment table is mapped one-to-one with the dynamic adjustment instruction set.

[0087] Furthermore, the system is also used for the following functions:

[0088] Based on the first electrical parameter eigenvector and the second electrical parameter eigenvector, implicit coupling characteristics under grid impedance fluctuations are extracted. Based on the implicit coupling characteristics under grid impedance fluctuations, the DC side voltage ripple energy mutation event and the AC side subharmonic amplitude growth are evaluated, and a harmonic coupling strength matrix is ​​configured. The harmonic resonance risk frequency band and the corresponding dynamic harmonic component are determined using the harmonic coupling strength matrix and the timestamp-frequency band alignment table.

[0089] Furthermore, the system is also used for the following functions:

[0090] The harmonic coupling strength matrix is ​​used to extract the overlapping energy spectral density of the DC-side voltage ripple energy mutation frequency band and the AC-side subharmonic, generating an initial resonance risk frequency band set. The timestamp-frequency band alignment table is used to screen the initial resonance risk frequency band set to determine candidate resonance risk frequency bands. Based on the candidate resonance risk frequency bands, a matching correction is performed with the time delay characteristics of the switch to determine the harmonic resonance risk frequency band and the corresponding dynamic harmonic component.

[0091] Furthermore, the system is also used for the following functions:

[0092] For the grid-connected crossover point, the phase margin of the harmonic resonance risk frequency band is analyzed, and the impedance mismatch frequency band is filtered; the impedance mismatch frequency band is embedded in an adaptive notch filter, and the equivalent impedance trajectory is dynamically corrected.

[0093] Furthermore, the system is also used for the following functions:

[0094] According to the impedance mismatch frequency band, an adaptive notch filter that conforms to the complex pole is set to determine the notch frequency and bandwidth parameters of the transfer function; based on the notch frequency and bandwidth parameters of the transfer function, a cascaded complex notch filter group is deployed, and the cascaded complex notch filter group is used to directionally dissipate the impedance values ​​corresponding to multiple impedance mismatch frequency bands in the equivalent impedance trajectory.

[0095] It should be noted that the order in which the embodiments of the present application are presented is for illustrative purposes only and does not necessarily represent the superiority or inferiority of the embodiments. Furthermore, the foregoing descriptions of specific embodiments of this specification are provided. Furthermore, the processes depicted in the accompanying drawings do not necessarily require the specific order or sequential sequence shown to achieve the desired results. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0096] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

[0097] This specification and drawings are merely illustrative of the present application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Obviously, those skilled in the art may make various modifications and variations to this application without departing from the scope of this application. Thus, this application is intended to include such modifications and variations as fall within the scope of this application and its equivalents.

Claims

1. A photovoltaic grid-connected reactive power compensation method based on SVG, characterized in that: The method comprises: Monitor the output power fluctuations of the photovoltaic array, synchronously obtain the voltage ripple characteristics and power factor dynamic changes from the DC side of the inverter, and establish the first electrical parameter characteristic vector; Monitor the impedance fluctuation characteristics of the photovoltaic grid, synchronously obtain the harmonic distortion rate and three-phase voltage imbalance from the AC side of the inverter, and establish the second electrical parameter characteristic vector; Performing cross-time-scale harmonic and reactive coupling characteristic analysis based on the first electrical parameter eigenvector and the second electrical parameter eigenvector to extract dynamic harmonic components and reactive demand spectrum; According to the dynamic harmonic components and reactive power demand spectrum, a dynamic adjustment instruction set of the SVG trigger pulse phase-amplitude is generated in combination with a weight distribution mechanism, wherein the dynamic adjustment instruction set is based on improved quasi-proportional resonant control to perform nonlinear compensation priority sorting; The dynamic adjustment instruction set is sent to the carrier phase shift controller of the SVG power unit to perform grid-connected interleaved compensation modulation.

2. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 1, characterized in that: Synchronously obtaining voltage ripple characteristics and power factor dynamic change from the DC side of the inverter, the method further comprising: extracting a subsynchronous oscillation component in the voltage ripple based on the voltage ripple characteristics; A dynamic impedance matching network is constructed according to the subsynchronous oscillation component in the voltage ripple, and the dynamic impedance matching network adjusts the equivalent impedance trajectory.

3. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 2, characterized in that: The dynamic impedance matching network adjusts the equivalent impedance trajectory, and the method includes: extracting a current-impedance coupling coefficient according to a subsynchronous oscillation component in the voltage ripple; Based on the dynamic impedance matching network, combined with the current-impedance coupling coefficient and the compensation current component, the equivalent impedance trajectory is reversely coupled and adjusted.

4. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 3, characterized in that: Allocating a compensation current component including a fundamental reactive component and a high-frequency harmonic component to the H-bridge power unit; The fundamental reactive component uses multi-carrier phase shifting to balance the current of power devices, and the high-frequency harmonic component uses dynamic carrier recombining to optimize the switching frequency.

5. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 1, characterized in that: Performing reactive coupling characteristic analysis across time scales based on the first electrical parameter eigenvector and the second electrical parameter eigenvector, the method comprising: performing a sliding time window grey relational analysis based on the first electrical parameter eigenvector and the second electrical parameter eigenvector to determine a coupling strength coefficient between a dynamic change in the DC side power factor and an imbalance in the AC side three-phase voltage; Feature fusion is performed based on the dynamic change of the DC side power factor and the coupling strength coefficient of the AC side three-phase voltage imbalance to generate a timestamp-frequency band alignment table, which is mapped one-to-one with the dynamic adjustment instruction set.

6. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 5, characterized in that: Performing cross-time-scale harmonic coupling characteristic analysis based on the first electrical parameter eigenvector and the second electrical parameter eigenvector, the method comprising: Extracting implicit coupling characteristics under grid impedance fluctuation based on the first electrical parameter characteristic vector and the second electrical parameter characteristic vector; Based on the implicit coupling characteristics under grid impedance fluctuations, the DC side voltage ripple energy mutation events and the AC side subharmonic amplitude growth are evaluated, and the harmonic coupling strength matrix is ​​configured; The harmonic resonance risk frequency band and the corresponding dynamic harmonic component are determined by the harmonic coupling strength matrix and the timestamp-frequency band alignment table.

7. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 6, characterized in that: Determining harmonic resonance risk frequency bands and corresponding dynamic harmonic components by using the harmonic coupling strength matrix and the timestamp-frequency band alignment table, the method includes: The overlapping energy spectrum density of the DC side voltage ripple energy mutation frequency band and the AC side subharmonic is extracted through the harmonic coupling strength matrix to generate an initial resonance risk frequency band set; Using the timestamp-frequency band alignment table, screening the initial resonance risk frequency band set to determine candidate resonance risk frequency bands; Based on the candidate resonance risk frequency band, a matching correction is performed with the time delay characteristics of the switch to determine the harmonic resonance risk frequency band and the corresponding dynamic harmonic component.

8. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 3, characterized in that: Sending the dynamic adjustment instruction set to a carrier phase shift controller of an SVG power unit to perform grid-connected interleaved compensation modulation, the method further comprising: For the grid-connected crossover point, analyze the phase margin of the harmonic resonance risk frequency band and filter the impedance mismatch frequency band; The impedance mismatch frequency band is embedded in an adaptive notch filter to dynamically correct the equivalent impedance trajectory.

9. The photovoltaic grid-connected reactive power compensation method based on SVG according to claim 8, characterized in that: Dynamically correcting the equivalent impedance trajectory, the method comprising: According to the impedance mismatch frequency band, an adaptive notch filter conforming to the complex poles is set to determine the notch frequency and bandwidth parameters of the transfer function; A cascaded complex notch filter group is deployed by transferring the notch frequency and bandwidth parameters of the function. The cascaded complex notch filter group is used to directionally dissipate the impedance values ​​corresponding to the multiple impedance mismatch frequency bands in the equivalent impedance trajectory.

10. A photovoltaic grid-connected reactive power compensation system based on SVG, characterized in that: The system is used to implement the photovoltaic grid-connected reactive power compensation method based on SVG according to any one of claims 1 to 9, and the system includes: A first electrical parameter characteristic vector establishment module is used to monitor the output power fluctuation of the photovoltaic array, synchronously obtain the voltage ripple characteristics and the dynamic change of the power factor from the DC side of the inverter, and establish the first electrical parameter characteristic vector; The second electrical parameter characteristic vector establishment module is used to monitor the photovoltaic grid impedance fluctuation characteristics, synchronously obtain the harmonic distortion rate and three-phase voltage imbalance from the inverter AC side, and establish the second electrical parameter characteristic vector; a reactive power demand spectrum extraction module, configured to perform harmonic and reactive power coupling characteristic analysis across time scales based on the first electrical parameter eigenvector and the second electrical parameter eigenvector, and extract dynamic harmonic components and reactive power demand spectrum; a dynamic adjustment instruction set generation module, configured to generate a dynamic adjustment instruction set for the phase-amplitude of the SVG trigger pulse based on the dynamic harmonic components and reactive power demand spectrum in combination with a weight distribution mechanism, wherein the dynamic adjustment instruction set is based on improved quasi-proportional resonant control to perform nonlinear compensation priority sorting; The grid-connected staggered compensation modulation module is used to send the dynamic adjustment instruction set to the carrier phase shift controller of the SVG power unit to perform grid-connected staggered compensation modulation.

Images