A grid-connected converter adaptive oscillation suppression method and device based on online identification of grid impedance, an electronic device, and a storage medium

Abstract

The application discloses a grid-connected converter adaptive oscillation suppression method and device based on grid impedance online identification, electronic equipment and a storage medium, and relates to the technical field of power electronics and new energy grid connection control. The method synchronously collects grid-connected point voltage, current and control state quantity, identifies grid impedance through amplitude-limited multi-frequency point perturbation and / or background harmonic and natural disturbance response in grid-connected operation, and obtains a wideband grid impedance model with stable poles through fitting; based on the model and closed-loop output admittance under current control parameters, a stability risk index is calculated, and phase-locked loop bandwidth, active damping coefficient and virtual resistance parameters are generated according to the resonance frequency band, which are written into the controller after amplitude limiting, rate constraint, hysteresis switching and rollback. The method can improve the oscillation suppression capability of the grid-connected converter under the condition of weak grid and time-varying impedance.

Description

Technical Field

[0001] This invention belongs to the field of power electronics and new energy grid-connected control technology, specifically a method, device, electronic device and storage medium for adaptive oscillation suppression of grid-connected converters based on online identification of grid impedance. Background Technology

[0002] As the penetration rate of new energy power generation in the power system continues to increase, the power grid structure is gradually shifting from a rigid grid dominated by traditional synchronous generators to a flexible grid with highly integrated power electronic equipment. Compared with the traditional strong power grid, new energy power plants have longer transmission lines, lower short-circuit ratios, and multiple converters operating in parallel. Under these conditions, the equivalent impedance at the grid connection point exhibits strong time-varying and wide-bandgap uncertainties. When the control loop, phase-locked loop, and filter resonance characteristics of the grid-connected converter couple with the grid impedance, problems such as subsynchronous oscillation, mid-to-high frequency resonance, phase-locked loop synchronism loss, and amplified distortion of grid-connected voltage and current can easily occur, affecting the safe and stable operation of the new energy grid-connected system.

[0003] Existing grid-connected converters typically employ offline modeling and fixed-parameter tuning. During the design phase, control parameters such as phase-locked loop bandwidth, current loop parameters, active damping coefficient, and virtual impedance are preset based on typical grid impedance conditions. When the actual grid impedance is close to the design nominal value, this type of fixed-parameter scheme can meet the basic stable operation requirements; however, in weak grids or scenarios with rapidly changing grid impedance, fixed-parameter control struggles to balance dynamic performance and stability margin, easily leading to overcompensation, undercompensation, or instability risks.

[0004] Some existing solutions propose online identification or adaptive control of grid impedance, but they still have the following shortcomings: most solutions only achieve online impedance identification and do not convert the identification results into quantitative indicators of stability risk and executable control parameter adjustment commands; some adaptive adjustment solutions lack a complete closed-loop link of "online impedance sensing - stability risk assessment - adaptive reconfiguration of control parameters - smooth parameter writing"; some solutions adopt a single active injection or a single passive identification method, which makes it difficult to simultaneously take into account identification accuracy, identification speed and grid-connected power quality; parameter updates lack limiting, rate constraints, hysteresis switching and risk removal backoff mechanisms, and parameter mutations may trigger secondary transient oscillations.

[0005] Therefore, it is necessary to propose an adaptive oscillation suppression method for grid-connected converters that can identify broadband grid impedance online, quantify grid-connected stability risks, adaptively generate oscillation suppression parameters according to frequency band, and smoothly write the parameters into the controller, so as to improve the operational stability of grid-connected converters in weak grids and time-varying impedance grids. Summary of the Invention

[0006] Purpose of the invention

[0007] The purpose of this invention is to provide an adaptive oscillation suppression method, device, electronic equipment, and storage medium for grid-connected converters based on online grid impedance identification. This addresses the problems of existing grid-connected converters where fixed parameter control struggles to adapt to time-varying grid impedance, online impedance identification is disconnected from stable control, and parameter updates easily trigger secondary transient shocks. This solution is not simply a data processing method; rather, it directly affects the grid-connected converter controller through steps such as grid connection point voltage and current acquisition, control loop perturbation injection, impedance model construction, risk assessment, control parameter writing, and rollback, thereby improving the actual grid-connected current and voltage oscillation state.

[0008] Technical solution

[0009] To achieve the above objectives, this invention provides an adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance. The method is applied to grid-connected converters in three-phase balanced systems and includes the following steps:

[0010] S1, synchronously collect the grid connection point voltage, grid connection current and control status variables of the grid-connected converter, and obtain the current control parameters of the grid-connected converter based on the control status variables.

[0011] S2, when the grid-connected converter is in grid-connected operation, the voltage and current response related to the grid impedance is obtained by adopting active identification mode and / or passive identification mode; wherein, the active identification mode is to inject multi-frequency micro-perturbations with limited amplitude into the control loop of the grid-connected converter, and the passive identification mode is to extract the background harmonics and natural disturbance response of the grid connection point.

[0012] S3. Based on the voltage and current response, identify the grid impedance online at multiple identification frequency points to obtain discrete frequency grid impedance; perform continuous fitting on the discrete frequency grid impedance to obtain a continuous frequency domain broadband grid impedance model with a pole real part less than zero, and extract the amplitude, phase angle, resonant frequency and resonant peak value of the grid impedance within a preset frequency band covering the grid-connected converter control bandwidth, phase-locked loop bandwidth and filter resonant frequency.

[0013] S4. Based on the broadband grid impedance model and the closed-loop output admittance of the grid-connected converter under the current control parameters, calculate the stability risk index after coupling the grid-connected converter with the grid impedance; wherein, the stability risk index is obtained by weighting one or more of the impedance ratio, phase margin, sensitivity function peak value, and resonance peak value.

[0014] S5, when the stability risk index is greater than or equal to the preset entry threshold, it is determined that the preset triggering condition is met, and an oscillation suppression parameter is generated according to the stability risk index and the frequency band where the resonant frequency is located; wherein, the oscillation suppression parameter includes at least one of the phase-locked loop bandwidth, active damping coefficient and virtual resistance parameter.

[0015] S6, after performing amplitude limiting, rate constraint and hysteresis switching on the oscillation suppression parameter, write it into the grid-connected converter controller; continuously monitor the grid impedance and the stability risk index, and cyclically update the oscillation suppression parameter; when the stability risk index is less than or equal to the preset exit threshold and continues to reach the preset hold time, gradually back the oscillation suppression parameter to the default parameter according to the preset back-off rate.

[0016] Further, in step S1, the control state quantities include phase-locked loop state quantities, current loop state quantities, controller integral state, modulation signal, DC side voltage for calculating closed-loop output admittance, and one or more of active power, reactive power, and grid frequency for auxiliary operation status monitoring; the control state quantities are collected synchronously with grid connection point voltage and grid connection current, and the sampling frequency is not less than twice the highest frequency of the identification frequency point.

[0017] Furthermore, in step S2, the multi-frequency perturbation used in the active identification mode is injected into at least one of the current reference channel, voltage reference channel, virtual impedance compensation channel, or active damping compensation channel of the control loop, and the injection position of the multi-frequency perturbation satisfies that the response signal of the corresponding frequency point can be detected in the grid connection point voltage and grid connection current after injection; the multi-frequency perturbation is one or more of the following: multi-sine wave signal, single-frequency sweep signal, pseudo-random sequence signal, or discrete pulse sequence signal.

[0018] Furthermore, the amplitude of the multi-frequency perturbation satisfies the single-frequency perturbation amplitude constraint and the total perturbation energy constraint; the single-frequency perturbation amplitude constraint is that the single-frequency perturbation amplitude does not exceed 0.5% of the rated current of the grid-connected converter, and the total perturbation energy constraint is that the total harmonic distortion rate of all frequency perturbations does not exceed the limit specified in the grid-connected power quality standard; the single-frequency perturbation amplitude constraint and the total perturbation energy constraint are determined based on the rated current, rated power fluctuation limit of the grid-connected converter, or grid-connected power quality limit.

[0019] Further, in step S3, the grid connection point voltage response and grid connection current response are synchronously detected at each identification frequency point to obtain the complex voltage response and complex current response at the corresponding identification frequency point; the grid impedance at the corresponding identification frequency point is calculated based on the ratio of the complex voltage response and complex current response; the calculated grid impedance is subjected to one or more of the following processing: residual verification, amplitude mutation constraint, and phase angle mutation constraint, in order to eliminate invalid identification results or suppress abnormal mutations; when using the passive identification mode, the passive identification quantity is constructed using the voltage and current cross power spectrum and the current self power spectrum, and the grid impedance at the corresponding frequency is obtained based on the passive identification quantity.

[0020] Further, in step S3, the grid impedance identified at multiple identification frequency points is used as the discrete frequency point grid impedance, and rational function fitting, vector fitting, least squares fitting or recursive least squares fitting is performed on the discrete frequency point grid impedance. During the fitting process, a forced constraint condition is that the real part of the pole is less than zero to obtain a stable continuous frequency domain broadband grid impedance model; the impedance peak point is searched traversally within a preset frequency band according to the broadband grid impedance model, and the frequency corresponding to the impedance peak point is used as the resonance frequency, and the impedance amplitude corresponding to the impedance peak point is used as the resonance peak.

[0021] Further, in steps S4 to S6, the weighting coefficients of the stability risk index are adaptively adjusted according to the operating conditions of the grid-connected converter and the grid short-circuit ratio; the weights of the impedance ratio and the resonance peak are increased under weak grid conditions, and the weights of the phase margin and the sensitivity function peak are increased under strong grid conditions; the generation rules of the oscillation suppression parameters include: when the stability risk in the low frequency band increases, the PLL bandwidth is reduced; when the resonance risk in the medium and high frequency bands increases, the active damping coefficient is increased; when the impedance ratio near the dominant resonance frequency exceeds the upper bound of the target impedance ratio, the virtual resistance parameter is increased; wherein, the low frequency band is the frequency band lower than the PLL bandwidth, and the medium and high frequency bands are the frequency bands higher than the PLL bandwidth and covering the filter resonance frequency; the preset entry threshold is greater than the preset exit threshold, and the difference between the two is not less than 20% of the preset entry threshold, and the preset holding time is not less than 3 complete identification and adaptive control cycles; the adjustment amplitude of a single oscillation suppression parameter within the same adaptive cycle does not exceed 50% of its preset maximum adjustment amount.

[0022] The present invention also provides a grid-connected converter adaptive oscillation suppression device, including a data acquisition module, a disturbance injection and response extraction module, an impedance identification module, a risk assessment module, a parameter generation module and a parameter update module; each module is respectively used to perform the corresponding data acquisition, response acquisition, impedance identification, risk assessment, parameter generation and parameter update functions in the above method.

[0023] The present invention also provides an electronic device, including a processor and a memory, and a computer program is stored in the memory. When the computer program is executed by the processor, the above grid-connected converter adaptive oscillation suppression method is implemented; the electronic device is deployed in a digital signal processor, a microcontroller, a field programmable logic gate array or a main control board of the grid-connected converter.

[0024] The present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above grid-connected converter adaptive oscillation suppression method is implemented; the computer-readable storage medium is a non-volatile computer-readable storage medium supporting the grid-connected converter controller.

[0025] Beneficial effects

[0026] Compared with the prior art, the present invention provides an adaptive oscillation suppression method, device, electronic device and storage medium for grid-connected converters based on online identification of grid impedance, which has the following beneficial effects:

[0027] 1. This invention combines online identification of grid impedance with grid stability risk assessment, enabling real-time tracking of broadband impedance changes under weak grid and time-varying impedance conditions, and further converting impedance changes into a basis for updating control parameters.

[0028] 2. The present invention adopts an active identification mode and / or a passive identification mode, which can improve the identification accuracy through multi-frequency perturbation with limited amplitude, and reduce the impact of active injection on power quality by utilizing background harmonics and natural disturbances.

[0029] 3. This invention constructs stability risk indicators based on indicators such as impedance ratio, phase margin, peak sensitivity function, and resonant peak value, which can transform impedance identification results into a basis for actionable control parameter adjustment.

[0030] 4. This invention adaptively adjusts oscillation suppression parameters such as phase-locked loop bandwidth, active damping coefficient, and virtual resistance parameters according to the risk frequency band, which can differentiate the suppression of low-frequency stability risks and mid-to-high-frequency resonance risks.

[0031] 5. The present invention sets amplitude limiting, rate constraint, hysteresis switching and risk removal rollback logic during the parameter writing process, which can avoid secondary transient oscillations caused by sudden changes in control parameters, and make online identification, risk assessment and control execution form a closed loop. Attached Figure Description

[0032] Figure 1 This is a flowchart of the adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance, as described in this invention.

[0033] Figure 2 This is a schematic diagram of the grid-connected converter, the equivalent impedance of the power grid, and the online identification adaptive vibration suppression control structure in an embodiment of the present invention.

[0034] Figure 3 This is a time-varying tracking diagram of the online identification results of grid impedance during the strong grid, weak grid, and strong grid restoration processes in an embodiment of the present invention.

[0035] Figure 4 This is a graph showing the stability risk index and adaptive parameter update curves based on online impedance identification in an embodiment of the present invention.

[0036] Figure 5 This is a comparison chart of grid-connected current, grid-connected point voltage, oscillation envelope, and risk indicators for fixed parameter control and adaptive control in embodiments of the present invention.

[0037] Figure 6 This is a snapshot of impedance identification under strong network, weak network, and strong network recovery conditions in an embodiment of the present invention. Detailed Implementation

[0038] The specific embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the following embodiments are used to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0039] Example 1: Overall Method Flow

[0040] Reference Appendix Figure 1 and attached Figure 2 The grid-connected converter in this embodiment includes a DC-side power supply, a grid-connected converter, an LCL filter, a grid connection point, a grid equivalent impedance, a grid equivalent power supply, and a grid-connected controller. The grid connection point is the voltage and current detection point where the converter connects to the grid.

[0041] In this embodiment, the phase-locked loop bandwidth, active damping coefficient and virtual resistance parameter are combined to form the control parameter vector to be adjusted, as shown in Equation (1).

[0042] (1);

[0043] In equation (1), For the phase-locked loop bandwidth, This is the active damping coefficient. For virtual resistance parameters, This is the matrix transpose symbol.

[0044] 1. Synchronous data acquisition and current control parameter acquisition

[0045] In step S1, the controller synchronously acquires the grid connection point voltage, grid connection current, and control status quantities of the grid-connected converter. The control status quantities include phase-locked loop status quantities, current loop status quantities, controller integral status, modulation signals, DC side voltage, and one or more of active power, reactive power, and grid frequency.

[0046] The sampling frequency satisfies equation (2):

[0047] (2);

[0048] In equation (2), Sampling frequency, This is the highest frequency among the identified frequency points. This condition ensures that the sampling frequency meets the basic sampling requirements for identifying the frequency points.

[0049] 2. Active recognition mode and passive recognition mode

[0050] In step S2, the controller uses an active identification mode and / or a passive identification mode to obtain the voltage and current response.

[0051] In active identification mode, amplitude-limited multi-frequency perturbations are injected into the control loop. These perturbations can be injected into at least one of the current reference channel, voltage reference channel, virtual impedance compensation channel, or active damping compensation channel. (See attached document.) Figure 2 After the perturbation is injected, the voltage and current response is formed by the converter controller, power conversion unit, filter and grid impedance, and is collected at the grid connection point.

[0052] Multi-frequency perturbation can be expressed as equation (3):

[0053] (3);

[0054] In equation (3), For the injected multi-frequency perturbation signal, To identify the number of frequency points, For the first The disturbance amplitude at each identification frequency point For the first The perturbation angular frequency of the identification frequency point Let K be the initial phase of the disturbance at the Kth identified frequency point, where K is the frequency point number.

[0055] To avoid the impact of active injection on grid-connected power quality, the disturbance amplitude must satisfy equations (4) and (5):

[0056] (4);

[0057] (5);

[0058] In equations (4) and (5), This represents the upper limit of the maximum injection amplitude per frequency. The total harmonic distortion (THD) corresponds to all injected frequencies. The total harmonic distortion (THD) limit is specified in the grid-connected power quality standards. Preferably, the amplitude of a single-frequency disturbance does not exceed 0.5% of the rated current of the grid-connected converter.

[0059] In passive identification mode, no disturbances are actively injected. Instead, the voltage and current responses related to grid impedance are extracted by utilizing the background harmonics at the grid connection point, load disturbances, natural grid fluctuations, or natural disturbance responses caused by changes in operating states. This mode is suitable for operating scenarios with higher requirements for power quality disturbances or where there is sufficient background excitation.

[0060] 3. Discrete Frequency Point Power Grid Impedance Identification

[0061] In step S3, for the active identification mode, the controller synchronously detects the grid connection point voltage response and grid connection current response at each identification frequency. For the first... By identifying a frequency point, the complex voltage response and complex current response can be obtained, as shown in equations (6) and (7):

[0062] (6);

[0063] (7);

[0064] In equations (6) and (7), The grid connection point voltage represents the voltage signal at the connection point between the converter and the power grid. The complex voltage response at the Kth identification frequency point is... The complex current response at the Kth identification frequency point is... To identify the window length, For window functions, For integration time variable, Grid-connected current represents the current signal flowing from the grid-connected converter into the grid side. The imaginary unit, It is the base of the natural exponential function.

[0065] No. The estimated value of the grid impedance at each identified frequency point can be expressed as Equation (8):

[0066] (8);

[0067] In equation (8), This is the estimated value of the grid impedance at the Kth identification frequency point. These are the corresponding frequency domain points.

[0068] In the passive identification mode, the power spectrum estimation method can be used to construct the grid impedance identification quantity, as shown in equation (9):

[0069] (9);

[0070] In equation (9), Angular frequency Estimated grid impedance at the location, This is the cross-power spectrum of the grid connection point voltage and the grid connection current. This is the self-power spectrum of the grid-connected current.

[0071] To improve identification robustness, the effectiveness of discrete frequency point grid impedances can be verified. The residual can be expressed as equation (10): (10);

[0072] In equation (10), Let be the normalized identification residual of the Kth identification frequency point. When the residual exceeds a preset threshold, the corresponding identification result can be deemed invalid or the weight of that frequency point can be reduced. Simultaneously, amplitude and phase angle constraints can be set to suppress impedance estimation abrupt changes caused by noise, transient impacts, or abnormal sampling.

[0073] 4. Continuous Fitting of Wideband Power Grid Impedance Model

[0074] Based on the discrete impedance estimates at multiple identified frequency points, a continuous frequency domain broadband power grid impedance model is constructed. This model can be expressed in rational function form, as shown in equation (11):

[0075] (11);

[0076] In equation (11), This is a continuous frequency domain broadband power grid impedance model. For direct access items, For model order, For the first Residual number, For the Mth pole, This is the model item number.

[0077] During the fitting process, extreme point stability is used as a constraint, as shown in equation (12):

[0078] (12);

[0079] In equation (12), extreme point The real part of the constraint is used to ensure that the fitted broadband power grid impedance model is a stable model.

[0080] Based on a broadband power grid impedance model, within a preset frequency band The impedance peak point is searched by traversing the inner loop to obtain the resonant frequency and resonant peak value, as shown in equations (13) and (14):

[0081] (13);

[0082] (14);

[0083] In equations (13) and (14), For preset frequency band, For frequency, The resonant frequency, This is the resonant peak value. This is pi (π). The preset frequency band covers the grid-connected converter control bandwidth, phase-locked loop bandwidth, and filter resonant frequency. (See attached reference.)Figure 3 and attached Figure 6 It can track the impedance amplitude and phase angle changes during strong network, weak network and strong network recovery process through broadband identification results.

[0084] 5. Calculation of stability risk indicators

[0085] In step S4, the stability risk index is calculated based on the broadband power grid impedance model and the converter closed-loop output admittance.

[0086] Under current control parameters (Including phase-locked loop bandwidth, active damping coefficient, and virtual resistance parameters), the closed-loop output admittance of the grid-connected converter can be obtained through the controller model or the linearized model. If the linearized state-space model is as shown in equations (15) and (16), then the closed-loop output admittance is as shown in equation (17):

[0087] (15);

[0088] (16);

[0089] (17);

[0090] In equations (15) to (17), Let be the system state vector. Let be the derivative of the state vector with respect to time. The state matrix, For the input matrix, For the output matrix, For a through matrix, For the small signal voltage at the grid connection point, It is the identity matrix. The closed-loop output admittance represents the closed-loop transmission relationship between the grid connection point voltage disturbance and the grid connection current response under the action of control parameters.

[0091] The small-loop gain, sensitivity function, and complementary sensitivity function are defined as shown in equations (18) to (20), respectively:

[0092] (18);

[0093] (19);

[0094] (20);

[0095] In equations (18) to (20), This refers to the small-loop gain formed by the coupling between the grid impedance and the converter's closed-loop output admittance. For sensitivity function, This is a complementary sensitivity function.

[0096] The impedance ratio, peak value of the sensitivity function, peak value of the complementary sensitivity function, and phase margin are shown in equations (21) to (24), respectively:

[0097] (twenty one);

[0098] (twenty two);

[0099] (twenty three);

[0100] (twenty four);

[0101] In equations (21) to (24), The impedance ratio is an indicator. The peak value of the sensitivity function. The peak value of the complementary sensitivity function. For phase margin, This is the gain crossover frequency.

[0102] Based on the above indicators, the stability risk index can be expressed as equation (25):

[0103] (25);

[0104] In equation (25), Let t be the stability risk indicator at time t. For the impedance ratio normalized risk term, For the phase margin normalized risk term, This is the risk term for the normalization of the peak value of the sensitivity function. This is the risk term for the normalized resonant peak value. , , , These are the corresponding weighting coefficients, and they satisfy... .

[0105] Under weak grid conditions, the weights of impedance ratio and resonant peak value can be increased; under strong grid conditions, the weights of phase margin and sensitivity peak value can be increased. (See attached reference.) Figure 4 The stability risk index can rise when the power grid switches from a strong grid to a weak grid, and trigger the generation of subsequent adaptive vibration suppression parameters.

[0106] 6. Generation of Oscillation Suppression Parameters

[0107] In step S5, when the stability risk index is greater than or equal to the preset entry threshold, the preset triggering condition is met and the system enters the adaptive vibration suppression state. The oscillation suppression parameters can be generated through one or more of the following methods: rule mapping, lookup table mapping, projection gradient optimization, model prediction optimization, or reinforcement learning-assisted optimization.

[0108] Under the rule-based mapping method, parameter generation follows these principles: when the stability risk in the low-frequency band increases, the PLL bandwidth is reduced; when the resonance risk in the mid-to-high frequency band increases, the active damping coefficient is increased; when the impedance ratio near the dominant resonant frequency exceeds the upper limit of the target impedance ratio, the virtual resistance parameter is increased. The low-frequency band is the frequency band below the PLL bandwidth, and the mid-to-high frequency band is the frequency band above the PLL bandwidth and covers the filter's resonant frequency.

[0109] In one implementation, parameter instructions can be generated based on the degree of risk exceeding the limit, as shown in equations (26) to (28):

[0110] (26);

[0111] (27);

[0112] (28);

[0113] The positive part function, i.e., in equations (26) to (28), This is the phase-locked loop bandwidth command value. A preset entry threshold is used to determine whether to enter the adaptive vibration damping state. This is the commanded value for the active damping coefficient. This is the command value for the virtual resistance parameter. , , These are the corresponding default values. , , These are the corresponding adjustment coefficients. The positive part function indicates that the larger value between the value inside the parentheses and 0 is taken. This is the upper limit of the target impedance ratio.

[0114] Under the optimization approach, the optimization problem shown in equation (29) can be constructed, and the constraints shown in equation (30) can be applied:

[0115] (29);

[0116] (30);

[0117] In equations (29) and (30), Let be the objective function. To control the feasible region of parameters, , , , The weights are the objective function weights. Let be the current parameter vector for the nth adaptive cycle. Let the lower bound vector of parameters be . The upper limit vector of parameters, This represents the upper limit of the peak sensitivity. This represents the lower limit of the phase margin.

[0118] When using projection gradient update, it can be expressed as equation (31):

[0119] (31);

[0120] In equation (31), For the first A parameter vector of adaptive periods, Towards the feasible region The projection operator of projection, Step size, For the objective function in The gradient at that point.

[0121] 7. Amplitude limiting, rate constraint, hysteresis switching and backoff

[0122] In step S6, to avoid secondary transient oscillations caused by sudden parameter changes, amplitude and rate constraints are applied to the oscillation suppression parameters. Let... For the first The parameter instructions generated in each adaptive cycle are actually written into the controller as shown in equation (32):

[0123] (32);

[0124] In equation (32), For the upper and lower bound projection operators of parameters, The rate of change constraint function, This represents the maximum allowable parameter change in a single adaptive cycle.

[0125] The hysteresis switching logic is as follows: when the stability risk index is greater than or equal to a preset entry threshold, the system enters an adaptive vibration suppression state; when the stability risk index is less than or equal to a preset exit threshold and remains for a preset holding time, the system exits the adaptive vibration suppression state. The preset entry threshold is greater than the preset exit threshold, and the difference between the two is not less than 20% of the preset entry threshold. The preset holding time is not less than three complete identification and adaptive control cycles.

[0126] After the risk is eliminated, the parameters are gradually rolled back to the default parameters at a preset rollback rate, as shown in equation (33):

[0127] (33);

[0128] In equation (33), For the default parameter vector, This is the backoff coefficient, and 0 < This formula ensures that the control parameters do not abruptly revert to their default values, but rather roll back smoothly. (See attached reference.) Figure 4 After switching into a weak grid, if the risk indicators exceed the entry threshold, the system enters an adaptive vibration suppression state and smoothly adjusts the phase-locked loop bandwidth, active damping coefficient, and virtual resistance parameters. After the grid recovers to a strong grid and the risk indicators are below the exit threshold and remain at the preset time, the control parameters are gradually restored according to the backoff rate.

[0129] Example 2: Device Implementation

[0130] Reference Appendix Figure 2 This invention also provides an adaptive oscillation suppression device for grid-connected converters. This device can be integrated into the local controller of the grid-connected converter or deployed on a centralized control platform at the plant level. The device includes a data acquisition module, a disturbance injection and response extraction module, an impedance identification module, a risk assessment module, a parameter generation module, and a parameter update module.

[0131] The data acquisition module is used to synchronously acquire grid connection point voltage, grid connection current, and control status variables of the grid-connected converter, and obtain the current control parameters of the grid-connected converter based on the control status variables.

[0132] The disturbance injection and response extraction module is used to obtain the voltage and current response related to grid impedance using active identification mode and / or passive identification mode. In active identification mode, the module injects amplitude-limited multi-frequency micro-perturbations into at least one of the current reference channel, voltage reference channel, virtual impedance compensation channel, or active damping compensation channel; in passive identification mode, the module extracts the background harmonics and natural disturbance responses at the grid connection point.

[0133] The impedance identification module is used to calculate the discrete frequency grid impedance based on the voltage and current response at multiple identification frequency points, and obtain a broadband grid impedance model with a pole real part less than zero through rational function fitting, vector fitting, least squares fitting or recursive least squares fitting; at the same time, it extracts the impedance amplitude, phase angle, resonant frequency and resonant peak value.

[0134] The risk assessment module is used to calculate the stability risk index based on the broadband grid impedance model and the closed-loop output admittance of the grid-connected converter under the current control parameters.

[0135] The parameter generation module is used to generate oscillation suppression parameters according to the risk index and the frequency band where the resonant frequency is located when the stability risk index is greater than or equal to the preset entry threshold.

[0136] The parameter update module is used to perform amplitude limiting, rate constraint, and hysteresis switching on the oscillation suppression parameters and then write them into the grid-connected converter controller, and when the risk is eliminated, the oscillation suppression parameters are retreated to the default parameters at a preset retreat rate. Each of the above modules can be implemented by a software program, or by hardware logic, or by a combination of software and hardware.

[0137] Embodiment 3: Electronic device and storage medium

[0138] This embodiment provides an electronic device for implementing the above grid-connected converter adaptive oscillation suppression method based on online identification of grid impedance. The electronic device includes at least one processor and a memory communicatively connected to the processor. A computer program is stored in the memory, and when the computer program is executed by the processor, the processor implements the method described in Embodiment 1. The electronic device can be deployed in a digital signal processor, a microcontroller, a field programmable gate array, or a main control board of a grid-connected converter.

[0139] This embodiment also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the above grid-connected converter adaptive oscillation suppression method based on online identification of grid impedance is implemented. The computer-readable storage medium can be used as a non-volatile storage medium supporting a grid-connected converter controller.

[0140] Embodiment 4: Simulation verification

[0141] To verify the feasibility and technical effects of the method of the present invention, refer to Appendix Figure 3 to Appendix Figure 6 , in one embodiment, verification is performed based on a three-phase , , grid-connected converter simulation model. The model uses a filter, and the identified frequency points include , , , , , , , and the injection amplitude at each frequency point is about 0.35% of the rated current. The entry threshold of the risk index is set to 0.6, the exit threshold is set to 0.4, the low-risk holding time is set to 3 adaptive cycles, and each adaptive cycle is .

[0142] to The system operates stably under strong grid conditions to verify the impedance identification accuracy and disturbance-free operation characteristics under strong grid conditions.

[0143] At any given moment, the power grid switches from a strong grid operating condition to a weak grid operating condition to verify its online tracking, risk triggering, and adaptive vibration suppression capabilities under sudden changes in grid impedance.

[0144] to The system runs continuously under weak network conditions to verify the convergence and closed-loop stability of parameter adaptive reconstruction.

[0145] At any given moment, the power grid is restored from a weak grid condition to a strong grid condition to verify the parameter smooth rollback capability after the risk is eliminated.

[0146] Reference Appendix Figure 3 During the processes of strong grid, weak grid, and strong grid recovery, the grid impedance amplitude and phase angle at multiple identification frequency points can be tracked online. After a sudden change in grid impedance, the identification results can quickly follow the actual impedance change; After the strong network is restored, the identification results can fall back to near the true impedance.

[0147] Reference Appendix Figure 4 After switching to weak grid operation mode, the stability risk index rises and exceeds the entry threshold, triggering adaptive vibration suppression mode. Subsequently, the system reduces the phase-locked loop bandwidth, increases the active damping coefficient, and increases the virtual resistance parameter; after the grid returns to strong grid operation, the risk index decreases and meets the exit conditions, and all control parameters smoothly return to their default values ​​at a preset rate.

[0148] Reference Appendix Figure 5 Compared with fixed parameter control, the adaptive control of this invention can reduce the grid current oscillation envelope after weak grid switching and shorten the oscillation convergence time. Since the control parameters are protected by amplitude limiting, rate constraint and hysteresis switching during writing and rollback, parameter changes do not introduce obvious secondary transient shocks.

[0149] Reference Appendix Figure 6 Under three typical operating conditions—strong grid, weak grid, and restored strong grid—the identified grid impedance and the theoretical impedance maintain good consistency within a preset wide bandwidth range, indicating that the wideband impedance identification results of this invention can provide an effective basis for stability risk assessment and adaptive reconfiguration of control parameters.

[0150] The above verification shows that the present invention can achieve closed-loop control of "online broadband impedance identification - stability risk quantification assessment - frequency band control parameter adaptive reconstruction - parameter smooth writing and rollback" under typical new energy grid connection conditions such as time-varying grid impedance and strong and weak grid switching, effectively improving the operation stability of grid-connected converters under weak grid and time-varying impedance grids.

[0151] Other implementation methods

[0152] Without departing from the core concept of this invention, the perturbation signal in the active identification mode can be, in addition to multiple sinusoidal signals, a single-frequency sweep signal, a pseudo-random sequence signal, or a discrete pulse sequence signal. Regardless of the form of the perturbation signal used, it should satisfy the single-frequency perturbation amplitude constraint and the total perturbation energy constraint.

[0153] In passive identification mode, in addition to power spectrum estimation, sliding window Fourier transform, subspace identification and other methods can be used to extract impedance information in background harmonics and natural disturbance responses.

[0154] In stability risk assessment, in addition to impedance ratio, phase margin, peak sensitivity function and resonant peak value, indicators such as peak complementary sensitivity function and resonant frequency drift can also be introduced to participate in the weighted calculation, as long as the risk quantification is still based on the coupling relationship between grid impedance and converter output admittance.

[0155] In three-phase imbalance, positive and negative sequence coupling or In scenarios with significant axis coupling, the scalar grid impedance can be extended to The impedance matrix in the coordinate system is used, and the stability risk index is calculated based on the maximum singular value of the product of the converter output admittance matrix and the grid impedance matrix, as shown in equations (34) and (35):

[0156] (34);

[0157] (35);

[0158] In equations (34) and (35), for The power grid impedance matrix in the coordinate system for Converter closed-loop output admittance matrix in coordinate system For small loop gain in matrix form, For the maximum singular value, It is a stability risk indicator in matrix form.

[0159] This invention can be applied to grid-connected interface converters in photovoltaic grid-connected converters, wind power grid-connected converters, energy storage converters, charge-discharge converters, grid-type converters, or parallel multi-converter systems. For grid-type converters, the oscillation suppression parameters may further include virtual inertia and virtual damping coefficients; for parallel multi-converter systems, the broadband grid impedance model may include the coupling impedance between parallel converters.

[0160] Any equivalent substitutions, parameter adjustments, module combinations, or application scenario extensions made to the above embodiments within the scope of the technical concept disclosed in this invention shall fall within the protection scope of this invention.

Claims

1. An adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance, characterized in that, The method is applied to grid-connected converters and includes the following steps: S1 synchronously collects the grid connection point voltage, grid connection current, and control status variables of the grid-connected converter, and obtains the current control parameters of the grid-connected converter; S2, when the grid-connected converter is in grid-connected operation, the voltage and current response related to the grid impedance is obtained by adopting active identification mode and / or passive identification mode. S3. Based on the voltage and current response, identify the grid impedance online, construct a broadband grid impedance model, and extract grid impedance characteristics; S4. Based on the broadband grid impedance model and the closed-loop output admittance of the grid-connected converter under the current control parameters, calculate the stability risk index after the grid-connected converter is coupled with the grid impedance. S5, when the stability risk index is greater than or equal to the preset entry threshold and the preset triggering condition is met, an oscillation suppression parameter is generated based on the stability risk index and the grid impedance characteristics; wherein, the oscillation suppression parameter includes at least one of the phase-locked loop bandwidth, active damping coefficient and virtual resistance parameter; S6. After performing amplitude limiting, rate constraint and hysteresis processing on the oscillation suppression parameter, it is written into the grid-connected converter controller. After the stability risk index is less than or equal to the preset exit threshold and continues to reach the preset hold time, the oscillation suppression parameter is gradually rolled back to the default parameter.

2. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 1, characterized in that, In step S1, the control state quantities include phase-locked loop state quantities, current loop state quantities, controller integral state, modulation signal, DC side voltage for calculating closed-loop output admittance, and one or more of active power, reactive power, and grid frequency for auxiliary operation status monitoring; the control state quantities are collected synchronously with grid connection point voltage and grid connection current, and the sampling frequency is not less than twice the highest frequency of the identification frequency point.

3. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 1, characterized in that, In step S2, the multi-frequency perturbation used in the active identification mode is injected into at least one of the current reference channel, voltage reference channel, virtual impedance compensation channel, or active damping compensation channel of the control loop, and the injection position of the multi-frequency perturbation satisfies the requirement that the response signal of the corresponding frequency point can be detected in the grid connection point voltage and grid connection current after injection; the multi-frequency perturbation is one or more of the following: multi-sine wave signal, single-frequency sweep signal, pseudo-random sequence signal, or discrete pulse sequence signal.

4. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 3, characterized in that, The amplitude of the multi-frequency perturbation satisfies the single-frequency perturbation amplitude constraint and the total perturbation energy constraint; the single-frequency perturbation amplitude constraint is that the single-frequency perturbation amplitude does not exceed 0.5% of the rated current of the grid-connected converter, and the total perturbation energy constraint is that the total harmonic distortion rate of all frequency perturbations does not exceed the limit specified in the grid-connected power quality standard; the single-frequency perturbation amplitude constraint and the total perturbation energy constraint are determined based on the rated current and rated power fluctuation limit of the grid-connected converter or the grid-connected power quality limit.

5. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 1, characterized in that, In step S3, the grid connection point voltage response and grid connection current response are synchronously detected at each identified frequency point to obtain the complex voltage response and complex current response at the corresponding identified frequency point. The grid impedance at the corresponding identification frequency point is calculated based on the ratio of the complex voltage response and the complex current response; the calculated grid impedance is subjected to one or more of the following processing: residual verification, amplitude mutation constraint, and phase angle mutation constraint, in order to eliminate invalid identification results or suppress abnormal mutations. When using the passive identification mode, the passive identification quantity is constructed by the voltage-current cross power spectrum and the current self power spectrum, and the grid impedance at the corresponding frequency is obtained based on the passive identification quantity.

6. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 1, characterized in that, In step S3, the grid impedances identified at multiple identification frequency points are used as discrete frequency grid impedances. Rational function fitting, vector fitting, least squares fitting, or recursive least squares fitting are performed on the discrete frequency grid impedances. During the fitting process, the real part of the poles being less than zero is used as a mandatory constraint to obtain a stable continuous frequency domain broadband grid impedance model. Based on the broadband grid impedance model, the impedance peak points are searched traversally within a preset frequency band. The frequency corresponding to the impedance peak point is used as the resonant frequency, and the impedance amplitude corresponding to the impedance peak point is used as the resonant peak value.

7. The adaptive oscillation suppression method for grid-connected converters based on online identification of grid impedance as described in claim 1, characterized in that, In steps S4 to S6, the weighting coefficients of the stability risk index are adaptively adjusted according to the operating conditions of the grid-connected converter and the grid short-circuit ratio; the weights of impedance ratio and resonant peak value are increased under weak grid conditions, and the weights of phase margin and sensitivity function peak value are increased under strong grid conditions. The generation rules for the oscillation suppression parameters include: reducing the phase-locked loop bandwidth when the stability risk in the low-frequency band increases; increasing the active damping coefficient when the resonance risk in the mid-to-high-frequency band increases; and increasing the virtual resistance parameter when the impedance ratio near the dominant resonant frequency exceeds the upper limit of the target impedance ratio. The low-frequency band refers to the frequency band below the phase-locked loop bandwidth, and the mid-to-high-frequency band refers to the frequency band above the phase-locked loop bandwidth and covering the filter resonant frequency. The preset entry threshold is greater than the preset exit threshold, and the difference between the two is not less than 20% of the preset entry threshold. The preset hold time is not less than three complete identification and adaptive control cycles. Within the same adaptive cycle, the adjustment amplitude of a single oscillation suppression parameter does not exceed 50% of its preset maximum adjustment.

8. An adaptive oscillation suppression device for a grid-connected converter, characterized in that, include: The data acquisition module is used to synchronously acquire grid connection point voltage, grid connection current and control status variables of the grid-connected converter, and obtain the current control parameters of the grid-connected converter based on the control status variables; The disturbance injection and response extraction module is used to obtain the voltage and current response related to the grid impedance by adopting an active identification mode and / or a passive identification mode when the grid-connected converter is in grid-connected operation state; wherein, the active identification mode is to inject multi-frequency micro-disturbances with limited amplitude into the control loop, and the passive identification mode is to extract the background harmonics and natural disturbance response at the grid connection point. The impedance identification module is used to identify the grid impedance online at multiple identification frequency points based on the voltage and current response, and obtain the discrete frequency grid impedance; perform continuous fitting on the discrete frequency grid impedance to obtain a continuous frequency domain broadband grid impedance model with a pole real part less than zero, and extract the amplitude, phase angle, resonant frequency and resonant peak value of the grid impedance. A risk assessment module, configured to calculate a stability risk index after the coupling of the grid-connected converter and the grid impedance based on the broadband grid impedance model and the closed-loop output admittance of the grid-connected converter under the current control parameters; A parameter generation module, configured to generate oscillation suppression parameters according to the stability risk index and the frequency band where the resonant frequency is located when the stability risk index is greater than or equal to a preset entry threshold; A parameter update module, configured to write the oscillation suppression parameters into the grid-connected converter controller after performing amplitude limiting, rate constraint and hysteresis switching on the oscillation suppression parameters, and gradually retreat the oscillation suppression parameters to the default parameters at a preset retreat rate after the risk is eliminated.

9. An electronic device, characterized in that, It includes a processor and a memory. A computer program is stored in the memory. When the computer program is executed by the processor, it implements an adaptive oscillation suppression method for a grid-connected converter based on online identification of grid impedance according to any one of claims 1 to 7; the electronic device is deployed in a digital signal processor, a microcontroller, a field programmable gate array or a main control board of the grid-connected converter.

10. A computer-readable storage medium, characterized in that, A computer program is stored thereon. When the computer program is executed by a processor, it implements an adaptive oscillation suppression method for a grid-connected converter based on online identification of grid impedance according to any one of claims 1 to 7; The computer-readable storage medium is a non-volatile computer-readable storage medium supporting the grid-connected converter controller.

Images