Active damping control LC resonance suppression method and device

Abstract

The invention relates to the technical field of resonance suppression, and discloses an active damping control LC resonance suppression method and device. The method comprises the following steps: performing parameter identification processing based on bus voltage to obtain an inductance value, a capacitance value and a resonance frequency of an LC circuit; extracting a resonant current component of the current signal based on the resonant frequency; calculating damping power according to the resonance current component, and converting the damping power into a voltage compensation instruction under an alpha-beta coordinate system; generating a PWM control signal according to the voltage compensation instruction and a voltage reference instruction; and performing on-off control on the inverter based on the PWM control signal, and injecting power opposite to the resonant current component into the LC circuit to generate a virtual damping effect. Virtual impedance characteristics are seamlessly integrated into inverter control, rapid response of the whole process from resonance detection to power injection is ensured, the resonance peak value is effectively suppressed, and the current quality is improved.

Description

Technical Field

[0001] This invention relates to the field of resonance suppression technology, and in particular to an active damping control method and apparatus for suppressing LC resonance. Background Technology

[0002] In electrolytic capacitor-free systems, LC filter circuits are prone to resonance under operating conditions, leading to system current distortion, increased electromagnetic interference, and severely impacting equipment stability and power quality. This LC resonance problem is particularly pronounced under conditions such as high-frequency switching, sudden load changes, and grid disturbances. Traditional LC resonance suppression methods mainly rely on passive damping components, such as connecting physical resistors or inductors in series or parallel in the LC circuit to dissipate resonance energy. However, these methods struggle to accurately identify LC parameter changes and resonant frequency drift and lack targeted frequency selectivity. Summary of the Invention

[0003] The main objective of this invention is to provide an LC resonance suppression method and device with active damping control. This invention seamlessly integrates virtual impedance characteristics into inverter control, ensuring a fast response throughout the entire process from resonance detection to power injection, effectively suppressing resonance peaks and improving current quality.

[0004] To achieve the above objectives, the present invention provides an active damping control method for suppressing LC resonance, comprising the following steps: Based on the bus voltage, parameter identification processing is performed to obtain the inductance, capacitance and resonant frequency of the LC circuit; The resonant current component of the current signal is extracted based on the resonant frequency; The damping power is calculated based on the resonant current components, and the damping power is converted into a voltage compensation command in the αβ coordinate system; A PWM control signal is generated based on the voltage compensation command and the voltage reference command; The inverter is switched on and off based on the PWM control signal, and power opposite to the resonant current component is injected into the LC circuit to produce a virtual damping effect.

[0005] Optionally, in a first implementation of the first aspect of the present invention, the parameter identification processing based on the bus voltage to obtain the inductance, capacitance, and resonant frequency of the LC circuit includes: The state equation of the first-order linear extended state observer is established based on the state matrix containing the LC natural frequency and quality factor. The bus voltage is input into the first-order linear extended state observer to obtain the first-order observation result. The bus voltage ripple component, voltage rate of change component and resonant disturbance component in the first-order observation results are used as state variables for linear tracking to obtain the second-order observation results. The second-order observation results are input into the third-order nonlinear extended state observer for nonlinear compensation in the resonant frequency band to obtain the third-order observation results; The third-order observation results are subjected to state separation to obtain the bus ripple signal and the resonant disturbance component; Based on the bus ripple signal and the resonant disturbance component, parameter identification processing is performed to obtain the inductance value, capacitance value and resonant frequency of the LC circuit.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the step of linearly tracking the bus voltage ripple component, voltage rate of change component, and resonant disturbance component in the first-order observation result as state variables to obtain the second-order observation result includes: Extract the bus voltage ripple component, voltage rate of change component, and resonant disturbance component from the first-order observation results; The second-order extended state vector is constructed by combining the bus voltage ripple component, the voltage change rate component, and the resonant disturbance component. Based on the second-order extended state vector, the linear relationship between each state variable is described by the state transition matrix and recursively calculated to obtain the state tracking equation. Solve the state tracking equation, update the second-order extended state vector, and output the second-order observation result.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the step of performing parameter identification processing based on the bus ripple signal and the resonant disturbance component to obtain the inductance value, capacitance value, and resonant frequency of the LC circuit includes: The bus ripple signal is used as the observation output vector, the resonant disturbance component is constructed as a regression vector, and a parameter vector containing inductance, capacitance and impedance values ​​is established. The parameter vector is corrected using the covariance matrix and forgetting factor to obtain the updated parameter identification results. Based on the updated parameter identification results, the forgetting factor is dynamically adjusted and the covariance matrix is ​​updated synchronously using a variable forgetting factor adjustment mechanism to obtain the target parameters. The inductance and capacitance values ​​of the LC circuit are extracted from the target parameters, and the resonant frequency is obtained by taking the square root of the inductance and capacitance components.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the extraction of the resonant current component of the current signal based on the resonant frequency includes: The center frequency of the bandpass filter is set according to the resonant frequency, and the damping coefficient is set at the same time. A resonant frequency selection filter is constructed based on the center frequency and the damping coefficient. The current signal is input into the resonant frequency selection filter for frequency domain separation to obtain the filtered current in the resonant frequency band; The amplitude and phase of the filtered current in the resonant frequency band are detected to obtain the amplitude and phase angle of the resonant current. The phase delay of the filter is calculated based on the transfer function of the resonant frequency selective filter. The phase angle of the resonant current and the phase delay of the filter are phase compensated. The resonant current component is generated by reconstructing the phase angle and the amplitude of the resonant current using a cosine function.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, the step of calculating the damping power based on the resonant current component and converting the damping power into a voltage compensation command in the αβ coordinate system includes: The amplitude of the resonant current component is extracted and multiplied by the resonant current component to obtain the square value of the resonant current. Then, the square value of the resonant current is multiplied by the virtual impedance to obtain the damping power. The three-phase current signal is transformed from the abc three-phase coordinate system to the αβ stationary coordinate system by the Clarke transformation matrix, and the α-axis current component and the β-axis current component are obtained. Based on the resonant frequency, the α-axis current component and the β-axis current component are extracted by bandpass filtering to obtain the α-axis resonant current component and the β-axis resonant current component. The sum of the squares of the α-axis resonant current component and the β-axis resonant current component is calculated as the denominator value. When the denominator value is less than a preset minimum threshold, the preset minimum threshold is used as the denominator protection value. The damping power is multiplied by the α-axis resonant current component and the β-axis resonant current component respectively, and then divided by the denominator protection value to obtain the voltage compensation command in the αβ coordinate system.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, generating the PWM control signal based on the voltage compensation command and the voltage reference command includes: The α-axis voltage compensation command and the α-axis voltage reference command are added together to obtain the α-axis total voltage command. At the same time, the β-axis voltage compensation command and the β-axis voltage reference command are added together to obtain the β-axis total voltage command. The total voltage command of the α-axis and the total voltage command of the β-axis are superimposed to obtain the first voltage command; Obtain the voltage amplitude of the first voltage command; when the voltage amplitude exceeds the maximum output voltage of the inverter, scale the α-axis total voltage command and the β-axis total voltage command proportionally to obtain the second voltage command. Based on the second voltage command, determine the sector where the voltage vector is located and calculate the duration of action of adjacent basic vectors; Based on the stated duration, the zero vector is evenly distributed across both ends of the switching cycle, and the switching sequence of the basic vector is arranged to obtain the PWM control signal.

[0011] Optionally, in a seventh implementation of the first aspect of the present invention, the step of uniformly distributing the zero vector to both ends of the switching cycle and arranging the switching sequence of the basic vector according to the action time to obtain the PWM control signal includes: Calculate the total action time of the zero vector based on the aforementioned action time; The total action time of the zero vector is symmetrically allocated to obtain the initial zero vector time and the final zero vector time; The basic vector switching order is determined based on the sector where the current voltage vector is located. The timing sequence of the first and second basic vectors is arranged according to the principle of minimum switching between adjacent vectors, thus obtaining the vector switching sequence. A PWM control signal is generated based on the start zero vector time, the vector switching sequence, and the end zero vector time.

[0012] Optionally, in an eighth implementation of the first aspect of the present invention, the step of controlling the switching of the inverter based on the PWM control signal and injecting power in the LC circuit that is opposite to the resonant current component to generate a virtual damping effect includes: The PWM control signal is input into the inverter's drive circuit to control the power transistor switching, thereby obtaining the switching state. Based on the switching state, the DC voltage is converted into an inverter output voltage containing virtual impedance characteristics; The inverter output voltage is applied to the LC circuit, causing the compensation component in the inverter output voltage to exchange power with the resonant current component in opposite phase, thus obtaining reverse injected power. The reverse injected power generates a virtual damping effect in the LC circuit.

[0013] The present invention also provides an LC resonance suppression device with active damping control, comprising: The parameter identification and processing module is used to perform parameter identification and processing based on the bus voltage to obtain the inductance value, capacitance value and resonant frequency of the LC circuit. An extraction module is used to extract the resonant current component of the current signal based on the resonant frequency; The calculation module is used to calculate the damping power based on the resonant current component and convert the damping power into a voltage compensation command in the αβ coordinate system; The generation module is used to generate a PWM control signal based on the voltage compensation command and the voltage reference command; The switching control module is used to control the switching of the inverter based on the PWM control signal and inject power in the LC circuit that is opposite to the resonant current component to generate a virtual damping effect.

[0014] In summary, the technical solution provided by this invention significantly improves the detection accuracy and anti-interference capability of LC resonance through a three-stage cascaded observer structure. Combined with the online parameter identification mechanism of recursive least squares, it can accurately track changes in LC circuit parameters in real time, ensuring the stability of the identification results. A two-stage design using an adaptive notch filter and virtual impedance achieves frequency-selective damping, precisely suppressing specific resonant frequencies while ensuring fundamental wave transmission performance. Through a power distribution strategy in the αβ coordinate system and SVPWM modulation optimization, the virtual impedance characteristics are seamlessly integrated into the inverter control, achieving precise energy dissipation. This invention can effectively suppress resonant peak values ​​and improve current quality. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the steps of an active damping control method for suppressing LC resonance in one embodiment of the present invention; Figure 2 This is a block diagram of an LC resonance suppression device with active damping control in one embodiment of the present invention.

[0016] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] Reference Figure 1 This embodiment provides an active damping control method for suppressing LC resonance, including the following steps: S1, based on the bus voltage, parameter identification processing is performed to obtain the inductance value, capacitance value and resonant frequency of the LC circuit; This involves constructing a hierarchical extended state observation system and embedding it into the main controller flow to achieve online dynamic estimation of the electrical parameters of the LC system. In the initial stage, a first-order linear extended state observer is built by establishing a state-space model containing the inherent frequency and quality factor of the LC circuit. The state matrix explicitly embeds the resonant frequency expression and the calculation form of the quality factor. The input signal is set to the actual acquired bus voltage waveform. The first-order observation results are obtained through the linear observer, which include the amplitude change representing the bus voltage ripple, the voltage change rate representing the ripple slope, and the resonant disturbance term caused by load changes or drive disturbances. These three characteristics from the first-order observation results are input as new state variables into the second-order observer. The observer maintains a linear structure and continuously tracks the voltage disturbance-related components to eliminate the time delay effects caused by sampling fluctuations or system interference, outputting the second-order observation results. Considering the significant frequency concentration characteristics of resonant disturbances, the second-order observation results are input into a frequency-selective third-order nonlinear extended state observer. The observer incorporates a bandpass gain function and nonlinear suppression features, exhibiting a high-gain response for signals near the resonant frequency and a low-gain or suppressed state for signals in other frequency bands, generating a third-order observation result containing high signal-to-noise ratio resonant information. A frequency bandwidth filter structure is used to perform state separation on the third-order observation results, extracting the bus voltage ripple signal and the resonant disturbance component. These components are then input into a recursive least squares algorithm for online parameter identification. The recursive least squares algorithm uses the analytical form of the resonant frequency as the dominant structure and constructs a regression vector. The ripple response and disturbance changes are used as the observation outputs, and the equivalent inductance, capacitance, and derived current resonant frequency of the LC circuit are continuously estimated through iterative correction. Specifically, the bus ripple signal is used as the observation output vector, the resonant disturbance component is constructed as a regression vector, and a parameter vector containing inductance, capacitance, and impedance values ​​is established. The parameter vector is iteratively updated using the recursive least squares method. The gain matrix is ​​calculated by multiplying the covariance matrix and the regression vector. The prediction error is calculated using the difference between the observation output vector and the product of the regression vector and the parameter vector. The gain matrix is ​​multiplied by the prediction error and then added to the current parameter vector to obtain the updated parameter identification result. Based on the updated parameter identification result, a variable forgetting factor adjustment mechanism is applied to dynamically adjust the forgetting factor and synchronously update the covariance matrix to obtain the target parameters. The inductance and capacitance values ​​of the LC circuit are extracted from the target parameters, and the resonant frequency is obtained by calculating the reciprocal square root of the inductance and capacitance components.

[0019] S2, extracting the resonant current component of the current signal based on the resonant frequency; Specifically, based on the resonant frequency value, the resonant frequency is set as the center frequency of the bandpass filter. The damping coefficient of the filter is then set according to the frequency response requirements of the actual system, constructing a frequency-selective resonant band filter. This resonant band filter employs a second-order bandpass structure, featuring adjustable bandwidth and smooth edge suppression to ensure high-gain transmission only for current components near the target resonant frequency, while effectively suppressing non-target frequency regions, especially the fundamental and high-frequency harmonic components. The current signal is input to the resonant frequency selective filter, and the current signal components near the resonant frequency are extracted through frequency domain separation, resulting in the filtered current waveform of the resonant band, which exhibits a single-frequency sinusoidal characteristic. Amplitude and phase detection operations are performed on the filtered current signal. Amplitude detection uses envelope demodulation or square averaging to obtain the steady-state amplitude, while phase detection uses a phase-locked loop, Hilbert transform, or phasor method to extract the phase angle of the band sinusoidal signal relative to the standard reference frequency. Based on the amplitude and phase angle of the resonant current, a cosine function is reconstructed. That is, the resonant frequency is used as the angular frequency, the extracted amplitude is used as the amplitude, and the detected phase angle is used as the initial phase. The time variation function of the resonant current component is constructed, and the obtained resonant current component can express the dynamic disturbance component of the system at the resonant frequency.

[0020] S3, calculate the damping power based on the resonant current component, and convert the damping power into a voltage compensation command in the αβ coordinate system; It should be noted that the instantaneous amplitude is extracted from the resonant current component, and multiplied by the original resonant current component to obtain the square of the resonant current. The square of the resonant current is then multiplied by a set virtual impedance factor to obtain the damping power representing the required dissipation of resonant energy. This damping power expresses the control objective of injecting reverse power into the system to achieve virtual energy consumption. The three-phase current signals of the system are mapped from the traditional abc three-phase rotating coordinate system to the αβ stationary coordinate system using the Clarke transform matrix, obtaining the instantaneous current components in the α-axis and β-axis directions. A bandpass filter is constructed, and the α-axis and β-axis current components are filtered separately using the resonant frequency as the center frequency to obtain the corresponding α-axis and β-axis resonant current components. This constitutes a vector representation of the resonant disturbance in the αβ plane, which is considered as the distribution of resonant energy in spatial coordinates. The damping power is multiplied by the resonant current components of the α-axis and β-axis respectively to construct the voltage control terms in the corresponding directions. The above results are then divided by the sum of the squares of the α-axis resonant current components and the squares of the β-axis resonant current components to ensure that the output voltage compensation command is consistent with the damping power in terms of total energy and orthogonal to the resonant current. This enables an active damping control compensation mechanism in the αβ coordinate system with fast dynamic response, strong suppression effect, and accurate power injection.

[0021] S4 generates a PWM control signal based on the voltage compensation command and the voltage reference command; Specifically, the α-axis voltage compensation component in the voltage compensation command is added to the α-axis voltage reference command provided by the main controller to construct the superimposed α-axis total voltage command. Simultaneously, the β-axis voltage compensation component is added to the corresponding β-axis voltage reference command to obtain the β-axis total voltage command. These two components together form a composite voltage vector that already includes the active damping target. The composite voltage vector is then synthesized to obtain the first voltage command, which is essentially the spatial composite amplitude of the two-dimensional vector formed by the α-axis and β-axis total voltage commands. The voltage amplitude of the first voltage command is used to assess whether the currently superimposed voltage exceeds the inverter's maximum output capacity. If the amplitude of the first voltage command exceeds the inverter's maximum output voltage limit, the total voltage command in the α-axis and β-axis directions is scaled proportionally. The scaling factor is determined based on the maximum amplitude constraint to ensure that the power output does not exceed the inverter's voltage boundary, resulting in the second voltage command. Based on the position of the second voltage command in the αβ coordinate system, its sector affiliation in space vector modulation is determined. The SVPWM sector number is determined by comparing the positional relationship between the vector angle and the regions divided by the six basic voltage vectors. Based on this, the duration of action of the two adjacent basic vectors corresponding to the voltage vector in the sector is calculated, depending on the control cycle, the DC bus voltage, and the projection ratio of the second voltage command onto the vector coordinates. According to the duration of action of the two basic vectors and the standard control logic of space vector modulation, the zero vector is symmetrically distributed throughout the switching cycle, i.e., evenly distributed at the beginning and end of the cycle, with the switching sequence of the two basic vectors arranged in the middle segment. The generated PWM signal achieves optimal configuration in terms of spectral symmetry, minimization of switching losses, and harmonic suppression, thereby outputting the PWM control signal.

[0022] S5 controls the switching of the inverter based on the PWM control signal and injects power in the LC circuit that is opposite to the resonant current component to produce a virtual damping effect.

[0023] The inverter's drive circuit receives a PWM control signal, which is then used by a gate drive module to logically control the power switches, resulting in a defined sequence of switching states at the hardware level. The switching state changes occur in cycles, and the turn-on and turn-off times of each half-bridge are scheduled under the instruction of a space vector modulation algorithm, forming a set of inverter output voltage waveforms that, on average, correspond to the voltage commands in the αβ coordinate system. Based on the output voltage, the inverter converts the DC bus voltage into an AC voltage with virtual impedance characteristics. The voltage compensation component is derived from the amplitude and direction of the resonant current, exhibiting a clear phase reversal characteristic, enabling spatial energy cancellation of the resonant components. Furthermore, the inverter output voltage is directly applied to the LC resonant circuit port through the power interface, superimposed on the existing system current. This results in the injected compensation voltage component and the resonant current component exhibiting the same frequency but a 180-degree phase difference in the frequency domain, forming an energy channel with opposite power exchange directions. Since the product of the current and voltage vectors is opposite in direction, the energy exchange process manifests as reverse power injection, thereby achieving dynamic dissipation and continuous attenuation of the resonant current component. Throughout the cycle, the reverse power injection mechanism constructs an energy absorption path equivalent to a resistive element. There is no real resistive element physically, but through the synergy of control logic and power electronics, an adjustable, trackable, and adaptable frequency-selective virtual damping effect is constructed.

[0024] In one example, parameter identification is performed based on the bus voltage to obtain the inductance, capacitance, and resonant frequency of the LC circuit, including: The state equation of the first-order linear extended state observer is established based on the state matrix containing the LC natural frequency and quality factor. The bus voltage is input into the first-order linear extended state observer to obtain the first-order observation result. The bus voltage ripple component, voltage rate of change component and resonant disturbance component in the first-order observation results are used as state variables for linear tracking to obtain the second-order observation results. The second-order observation results are input into the third-order nonlinear extended state observer for nonlinear compensation in the resonant frequency band to obtain the third-order observation results. State separation is performed on the third-order observation results to obtain the bus ripple signal and the resonant disturbance component; Based on the bus ripple signal and the resonant disturbance component, parameter identification processing is performed to obtain the inductance, capacitance and resonant frequency of the LC circuit.

[0025] In this example, a state matrix containing LC intrinsic frequency and quality factor information is established based on the bus voltage. A first-order linear extended state observer state equation is constructed, using the real-time acquired bus voltage as input to form a closed-loop estimate. This enables the observer to respond sensitively to bus voltage ripple, voltage rate of change, and resonance-related disturbances, outputting a first-order observation result that combines amplitude and dynamic variation characteristics. The bus voltage ripple component, voltage rate of change component, and resonance disturbance component from the first-order observation result are used as new state variables and continuously tracked linearly within the same sampling period. By configuring the observation gain and bandwidth under constraints, the resonance sensitivity and time-domain convergence speed are improved without introducing excessive amplification noise, resulting in a second-order observation result. This ensures that the energy related to resonance is stably captured on the time axis and that jitter caused by random noise is suppressed. The second-order observation results are input into a third-order nonlinear extended state observer. A selective nonlinear compensation mechanism is configured for the resonant frequency band, increasing the observation gain within the target frequency band and suppressing it within the non-target frequency band. This improves the signal-to-noise ratio of the resonant signal, reduces the intrusion of power frequency and high-frequency interference, and obtains the third-order observation results with a smaller overshoot. This achieves focused extraction of resonant disturbances and effective isolation of non-resonant components. State separation is performed on the third-order observation results. The bus ripple signal is extracted via a low-frequency path to describe DC-side quality and slow dynamics, while the resonant disturbance components are separated via a narrow-band path to describe the equivalent resonant energy and phase relationship, maintaining amplitude and phase consistency during the separation process. Online parameter identification is performed based on bus ripple signal and resonant disturbance component. A time-varying estimation strategy of recursive least squares is adopted. Inductance and capacitance are used as parameters to be estimated, and resonant frequency is used as a derived quantity for sample-by-sample updating. At the same time, a variable forgetting factor is used to balance fast tracking and steady-state accuracy. When a sudden change or drift exceeds the limit is detected, the covariance and constraint boundary are reset. In this way, stable, reliable inductance, capacitance and resonant frequency values ​​that can be updated adaptively with the operating conditions are obtained in the real operating environment.

[0026] In one example, the bus voltage ripple component, voltage rate of change component, and resonant disturbance component from the first-order observation results are linearly tracked as state variables to obtain the second-order observation results, including: Extract the bus voltage ripple component, voltage rate of change component, and resonant disturbance component from the first-order observation results; The second-order extended state vector is constructed by combining the bus voltage ripple component, voltage rate of change component, and resonant disturbance component. Based on the second-order extended state vector, the linear relationship between each state variable is described by the state transition matrix and recursively calculated to obtain the state tracking equation. Solve the state tracking equation, update the second-order extended state vector, and output the second-order observation results.

[0027] In this example, the bus voltage ripple component, voltage rate of change component, and resonant disturbance component are separated from the first-order observation results. Anti-aliasing preprocessing, DC component subtraction, and sideband suppression are performed within the same sampling cycle, employing a consistent timestamp alignment strategy to avoid cross-cycle error accumulation. Based on a linear observation model incorporating LC intrinsic frequency and quality factor information, the three components are recalibrated as standardized quantities within the observer, with amplitude ranges constrained and unit consistency checks performed. This processing is based on a pre-stage linear extended state observer, whose state matrix explicitly embeds the physical characteristics of LC intrinsic frequency and quality factor. This allows the bus voltage ripple and resonant disturbance to be stably amplified to a identifiable level and maintains a sensitive response to the target frequency band during dynamic processes, forming a reliable first-order observation output containing the three components, which serves as the input reference for second-order tracking. A second-order extended state vector is constructed by combining the bus voltage ripple component, voltage rate of change component, and resonant disturbance component in a fixed field order. Each field is configured with initial confidence level, allowable rate of change, soft limiting, and anomaly shielding flags. Simultaneously, an equivalent noise description related to the sensing path, sampling jitter, and temperature drift is established. A state transition matrix is ​​constructed based on the second-order extended state vector to describe the continuity constraint between voltage ripple and voltage rate of change, the excitation effect of resonant disturbance on both, and the suppression effect of observation gain on high-frequency noise. The structure of the state transition matrix follows the approximate relationship derived from the LC physical model and the numerical stability requirements of the sampling period setting. Combined with bandwidth adaptation and forgetting factor strategies, the rising edge, step disturbance, and slowly varying drift under different operating conditions are partitioned to obtain a set of recursively computeable state tracking equations. During the recursion process, frequency band selection weights are introduced for the resonant-related quantities, and residual thresholds, anomaly detection, and backoff logic are combined to suppress the interference of occasional spikes and artifacts on state evolution. The state tracking equation is solved sequentially by time update and measurement update. A prediction model is used to provide the state prior and uncertainty for the next time step. Corrections are then made using the bus voltage ripple component, voltage rate of change component, and resonance disturbance component from the first-order observations. The observation gain and weight allocation are dynamically adjusted, and the observation bandwidth is slightly reconfigured to balance convergence speed and steady-state jitter. When a concentrated increase in resonance energy is detected within the target frequency band, the observation sensitivity associated with the resonance disturbance component is appropriately increased. Conversely, when power frequency and high-frequency spurious emissions increase, the frequency band gain is reduced to avoid false tracking. The second-order extended state vector is continuously updated through this recursive solution, outputting second-order observation results containing three types of quantities: voltage ripple, voltage rate of change, and resonance disturbance.

[0028] In one example, parameter identification processing is performed based on the bus ripple signal and resonant disturbance components to obtain the inductance, capacitance, and resonant frequency of the LC circuit, including: The bus ripple signal is used as the observation output vector, the resonant disturbance component is constructed as a regression vector, and a parameter vector containing inductance, capacitance and impedance values ​​is established. The parameter vector is corrected using the covariance matrix and forgetting factor to obtain the updated parameter identification results; Based on the updated parameter identification results, the forgetting factor is dynamically adjusted and the covariance matrix is ​​updated synchronously using a variable forgetting factor adjustment mechanism to obtain the target parameters. The inductance and capacitance values ​​of the LC circuit are extracted from the target parameters, and the resonant frequency is obtained by taking the square root of the inductance and capacitance components.

[0029] In this example, the bus ripple signal is used as the observed output vector, and a regression vector is constructed using the resonant disturbance component. Simultaneously, a parameter vector containing inductance, capacitance, and equivalent impedance values ​​is established, creating a recursive linear mapping between the observed output and the regression quantity. During online operation, DC removal and bandwidth limiting of the observed output are performed, followed by amplitude normalization and correlation checks on the regression vector to reduce the impact of input dimension differences and collinearity on numerical stability. Initial parameter values, initial values ​​of the symmetric positive definite covariance matrix, and an initial forgetting factor are set. Soft limiting and anomaly masking are implemented on the parameter vector using residual thresholds and physical boundaries to ensure a controllable convergence region during the identification initiation phase. In the recursive phase, the correction gain is calculated using the covariance matrix and forgetting factor in each sampling period. A prediction-correction approach is used to update the parameter vector, ensuring that the new parameter vector can maximally explain the observed changes in the current bus ripple, while controlling the covariance contraction rate to balance responsiveness and steady-state accuracy. When a short-term surge occurs in the observation residual, the resonant center shifts, or a sudden change in dynamic characteristics caused by the load is detected, a covariance reset or covariance increase adjustment strategy is triggered to quickly restore the adaptability to the new operating conditions. To maintain identification stability under all operating conditions, a variable forgetting factor adjustment mechanism is set up: when the residual is consistently small and the parameter changes smoothly, the forgetting factor is gradually increased to enhance the weight of historical information and reduce jitter; when the residual increases or frequency drift is detected, the forgetting factor is decreased to enhance the weight of new data and improve the tracking speed, and the covariance matrix is ​​updated simultaneously according to the symmetric positive definite constraint. If necessary, square root form or Josephus form update is used to suppress numerical accumulation error, obtaining the target parameters that best match the observation statistical characteristics at the current moment. Physical consistency checks are performed on the target parameters, including ensuring that the inductance and capacitance values ​​are within preset boundaries, the equivalent impedance value remains non-negative and does not violate energy conservation, and then time continuity constraints and differential smoothing are applied to eliminate instantaneous fluctuations caused by single-point outliers. The inductance and capacitance values ​​are extracted from the target parameters through the above process. The resonant frequency is calculated by multiplying them, taking the reciprocal, and finally taking the square root, thus obtaining a frequency result that is synchronized with the current operating conditions.

[0030] In one example, the resonant current component of the current signal is extracted based on the resonant frequency, including: The center frequency of the bandpass filter is set according to the resonant frequency, and the damping coefficient is set at the same time. A resonant frequency selection filter is constructed based on the center frequency and the damping coefficient. The current signal is input into the resonant frequency selection filter for frequency domain separation to obtain the filtered current in the resonant frequency band; The amplitude and phase of the filter current in the resonant frequency band are detected to obtain the amplitude and phase angle of the resonant current. The phase delay of the filter is calculated based on the transfer function of the resonant frequency selective filter. The phase angle of the resonant current and the phase delay of the filter are phase compensated. The resonant current component is generated by reconstructing the phase angle and the amplitude of the resonant current using a cosine function.

[0031] In this example, the resonant frequency is used as the center frequency of the bandpass filter. The damping coefficient is set based on the target vibration suppression intensity, allowable overshoot, and noise environment to determine the filter's bandwidth and roll-off slope. The filter employs a second-order or equivalent second-order narrowband structure, and the center frequency is updated by sliding based on the small drift of the resonant frequency within the control cycle. Simultaneously, upper and lower limit constraints and rate of change constraints are superimposed on the damping coefficient to avoid phase jitter caused by excessively rapid parameter tuning. The filter is implemented in digital form obtained through bilinear transformation, combined with predistortion compensation to reduce frequency offset caused by discretization. The sampling rate should be higher than the resonant frequency. The filter bandwidth is temporarily increased during the rising edge to accelerate locking, and then the bandwidth is slowly narrowed during the steady-state phase to suppress adjacent spurious signals. The real-time current signal is input to the resonant frequency selection filter, and the full-frequency current is separated in the frequency domain, with the output retaining only the resonant frequency band component. To mitigate the impact at startup, the filter input is configured with soft-start evanescence, and the output is configured with differential prior to suppress DC drift. In overload scenarios, a bypass is triggered to ensure main loop stability. To obtain the amplitude and phase information of the resonant current, the filtered current enters the amplitude detection channel, where envelope demodulation or sliding RMS calculation is employed, and instantaneous jitter is smoothed using a first- or second-order low-pass filter. A minimum measurable threshold and quantization dead zone are incorporated into the amplitude detection to prevent spurious activations near the noise floor. The phase detection channel uses a phase-locked loop (PLL) or Hilbert transform analytical signal method. The PLL achieves robust phase locking through phase comparison and loop filtering, making it suitable for embedded fixed-point implementations. The Hilbert transform provides higher phase resolution when computational resources allow. To avoid phase measurement delay affecting subsequent compensation, fixed group delay compensation is applied to the phase output, and it is timestamped with the amplitude channel. After amplitude and phase synchronization, the controller reconstructs a cosine function based on the resonant frequency, resonant current amplitude, and resonant current phase angle, generating a resonant current component in the time domain that matches the resonant frequency, has a controlled phase, and adjustable amplitude. To suppress high-frequency glitches caused by jumps during reconstruction, a slope-limiting strategy is adopted for amplitude and phase, and continuous expansion is used when the phase crosses quadrants. Freezing logic is also added for extremely small amplitudes to prevent numerical amplification due to excessively small denominators. To improve anti-interference capability, the consistency between the reconstructed output and the original filter current is checked. If the correlation between the two drops below the threshold within a short window, the reconstruction weight is reduced and the parameters are rolled back to the previous time step. When external disturbances cause a short-term shift in the resonant frequency, the center frequency adopts gradual tracking instead of instantaneous jumps, and the deviation during the frequency locking process is offset by a temporary relaxation of the damping coefficient.

[0032] In one example, the damping power is calculated based on the resonant current component and then converted into voltage compensation commands in the αβ coordinate system, including: The amplitude of the resonant current component is extracted and multiplied with the resonant current component to obtain the square value of the resonant current. Then, the square value of the resonant current is multiplied with the virtual impedance to obtain the damping power. The three-phase current signal is transformed from the abc three-phase coordinate system to the αβ stationary coordinate system by the Clarke transformation matrix, and the α-axis current component and the β-axis current component are obtained. Based on the resonant frequency, the α-axis current component and the β-axis current component are extracted by bandpass filtering to obtain the α-axis resonant current component and the β-axis resonant current component. The sum of the squares of the α-axis resonant current component and the β-axis resonant current component is calculated as the denominator value. When the denominator value is less than the preset minimum threshold, the preset minimum threshold is used as the denominator protection value. The damping power is multiplied by the α-axis resonant current component and the β-axis resonant current component respectively, and then divided by the denominator protection value to obtain the voltage compensation command in the αβ coordinate system.

[0033] In this example, amplitude extraction is performed on the resonant current component within the control cycle. The resonant current component is multiplied by the amplitude result to obtain a square-shaped metric reflecting the energy level, followed by overflow prevention and normalization. The square-shaped metric is multiplied and mapped to the virtual impedance parameter to obtain the damping power target value, representing the energy intensity that needs to be actively dissipated within the resonant frequency band. The damping power target value is limited by a first- or second-order smoothing element, while upper limits and rising slope constraints are set to avoid overcompensation during grid disturbances or load abrupt changes. To project the damping power onto spatial coordinates, the controller transforms the three-phase current signals from the abc three-phase coordinate system to the αβ stationary coordinate system using a standard orthogonal matrix. During this process, the current samples are timestamped and aligned, and zero-sequence component suppression, proportional coefficient calibration, and DC bias subtraction are performed to obtain the α-axis and β-axis current components. The α-axis and β-axis current components pass through narrowband bandpass filters centered on the resonant frequency. The filter bandwidth and damping coefficient are adaptively adjusted according to the frequency drift and noise environment identified online. During the frequency locking phase, the bandwidth is appropriately widened to improve locking speed, and during the steady-state phase, the bandwidth is narrowed to improve selectivity. The output α-axis and β-axis resonant current components together constitute the vector expression of the resonant current in the αβ plane. Based on the principle of power conservation, the controller multiplies the target damping power value by the α-axis and β-axis resonant current components respectively, forming unnormalized voltage distributions in two directions. The sum of the squares of the two resonant current components is used as the normalization reference to achieve vector-proportional distribution of damping power in space. To prevent the denominator from approaching zero and causing numerical anomalies, a small protection amount and threshold freezing strategy are added to the normalization reference. When the overall resonant components are too small or the signal-to-noise ratio is low, the voltage compensation command is proportionally attenuated or temporarily frozen to a safe bias value. After normalization, the voltage compensation command in the αβ coordinate system is obtained. The voltage compensation command is consistent with the resonant current component in direction and matches the damping power target value in terms of power. It can construct a frequency-selective energy absorption path at the inverter output level.

[0034] In one example, a PWM control signal is generated based on a voltage compensation command and a voltage reference command, including: The α-axis voltage compensation instruction and the α-axis voltage reference instruction are added together to obtain the α-axis total voltage instruction. At the same time, the β-axis voltage compensation instruction and the β-axis voltage reference instruction are added together to obtain the β-axis total voltage instruction. The total voltage command on the α-axis and the total voltage command on the β-axis are superimposed to obtain the first voltage command; Obtain the voltage amplitude of the first voltage command. When the voltage amplitude exceeds the maximum output voltage of the inverter, scale the α-axis total voltage command and the β-axis total voltage command proportionally to obtain the second voltage command. The sector where the voltage vector is located is determined based on the second voltage command, and the duration of action of adjacent basic vectors is calculated. Based on the duration of action, the zero vector is evenly distributed across both ends of the switching cycle, and the switching sequence of the basic vector is arranged to obtain the PWM control signal.

[0035] In this example, co-domain fusion is performed within the αβ static coordinate domain. The α-axis voltage compensation command is added to the α-axis voltage reference command to obtain the α-axis total voltage command, and the β-axis voltage compensation command is added to the β-axis voltage reference command to obtain the β-axis total voltage command. This allows the compensation amount to be directly superimposed on the voltage reference layer rather than the current loop or power loop, thereby shortening the control link and reducing phase error, forming a voltage reference pair. The first voltage command, obtained by superimposing the α-axis and β-axis total voltage commands, is calculated in a vector manner, and the voltage amplitude and direction information are extracted. When the amplitude exceeds the inverter's maximum allowable output capacity, the α and β-axis total voltage commands are compressed simultaneously according to the principle of proportional amplitude scaling. This keeps the voltage vector direction unchanged while the amplitude falls within the hardware safety boundary, avoiding distortion and saturation caused by overmodulation, while preserving the correct spatial orientation of the compensation effect, resulting in a second voltage command that satisfies hardware constraints. The sector where the voltage vector is located is determined based on the distribution of the second voltage command in the αβ plane, and two adjacent basic voltage vectors are selected as the synthesis object accordingly. The effective time of the two fundamental vectors within one switching cycle and the remaining time allowance of the zero vector are calculated. The calculation results reflect the amplitude and phase requirements of the voltage reference. To balance DC bus voltage utilization and switching losses, the zero vector is placed symmetrically at both ends of the switching cycle, while the fundamental vectors are symmetrically arranged according to a seven-segment timing sequence, resulting in the duty cycle sequence and gate triggering sequence. This symmetrical arrangement reduces the sideband components of the synthesized voltage trajectory and improves harmonic quality, while keeping current ripple and electromagnetic noise within a controllable range.

[0036] In one example, based on the action time, the zero vector is evenly distributed across the two ends of the switching cycle, and the switching sequence of the basic vector is arranged to obtain the PWM control signal, including: Calculate the total action time of the zero vector based on the action time; By symmetrically distributing the total action time of the zero vector, the initial zero vector time and the final zero vector time are obtained; The basic vector switching order is determined based on the sector where the current voltage vector is located. The timing sequence of the first and second basic vectors is arranged according to the principle of minimum switching between adjacent vectors, thus obtaining the vector switching sequence. The PWM control signal is generated based on the start zero vector time, the vector switching sequence, and the end zero vector time.

[0037] In this example, a time budget is established based on the action time of two adjacent basic voltage vectors and the switching cycle. The total action time of the zero vector is obtained by subtracting the action time of the two basic vectors from the switching cycle. A non-negativity check and upper limit constraint are applied to the total action time of the zero vector to prevent negative duration or over-cycle accumulation caused by numerical jitter. The total action time of the zero vector is symmetrically allocated, dividing the total duration equally into the start and end zero vector times. During the equalization process, minimum conduction time and slope limits are applied simultaneously, ensuring that the zero vector segments are mirror-image distributed at both ends of the time axis. This symmetrical arrangement can cancel low-order sideband components in the frequency spectrum and provides a stable boundary. After completing the symmetrical allocation, the vector switching sequence is determined. Two adjacent basic vectors are selected based on the sector where the current voltage vector is located, and the timing of the first and second basic vectors is arranged according to the principle of minimum switching between adjacent vectors, so that each state transition only changes the conduction state of one bridge arm, thereby reducing switching losses and electromagnetic interference. In the switching sequence generation stage, DC bus voltage utilization, modulation linearity, and sector crossing continuity are considered to avoid unnecessary state jumps caused by excessively rapid sector crossings. In terms of specific timing construction, a switching cycle is used as the time axis. An idle segment is placed at the beginning of the cycle, with the initial zero vector time serving as the starting point. The first and second basic vector action times are then inserted sequentially, maintaining a minimum interval between the two basic vector segments to satisfy the turn-off buffer of the drive layer. The ending zero vector time is placed at the end of the time axis to form a symmetrical layout. When the operating point approaches the upper limit of linear modulation or bus voltage fluctuations occur, the duty cycle feasibility and waveform continuity are maintained by fine-tuning the order of the two basic vectors and the zero vector allocation ratio. To enhance stability under extreme conditions, the timing generation algorithm performs a "feasibility check" in each cycle. This check includes verifying whether the sum of the two basic vector action times and the zero vector time equals the switching cycle, whether each time slice is not less than the minimum conduction time, whether the dead-zone insertion window is satisfied between adjacent time slices, and whether the sector change is consistent with the angular velocity. If any condition is violated, a backoff strategy is executed, proportionally pushing back the excess time components to both ends of the zero vector while maintaining the voltage vector direction from jumping. After the time slices are arranged, the gate signals of the three-phase bridge arms are generated based on the start zero vector time, vector switching sequence, and end zero vector time. An independent timer comparison channel is established for each bridge arm, switching the conduction state of the high and low arms according to the segment boundaries on the time axis. All bridge arms are inserted with a fixed dead zone to eliminate the risk of cross-conduction. To reduce jitter and quantization errors, the gate duty cycle is implemented using a high-resolution counter in the digital domain, with constant rising and falling edges superimposed in the analog driver stage. The generated PWM control signal undergoes a consistency verification, which includes verifying whether the voltage vector synthesis is consistent with the second voltage command amplitude, whether the zero-sequence component remains within the expected range, and whether the state at the end of the previous cycle is continuous with the state at the beginning of the current cycle.When a sudden change in amplitude or phase is detected, the zero vector proportion is temporarily increased. The distribution strategy is then restored after the action time is recalculated in the next cycle to block abnormal energy injection. If the system detects a sudden change in load or an increase in harmonic background, the timing generation stage should allow a brief bandwidth relaxation. This is achieved by extending the zero vector time and reducing the edge slope of the basic vector action time to decrease the peak current, thereby ensuring that the inverter and LC circuit remain under control during the transition phase.

[0038] In one example, the inverter is switched on and off based on a PWM control signal, and power opposite to the resonant current component is injected into the LC circuit to produce a virtual damping effect, including: The PWM control signal is input to the inverter's drive circuit to control the power transistor switching and obtain the switching state. Based on the switching state, the DC voltage is converted into an inverter output voltage containing virtual impedance characteristics; The inverter output voltage is applied to the LC circuit, causing the compensation component in the inverter output voltage to exchange power with the resonant current component in opposite phase, thus obtaining reverse injected power. The reverse-injected power in the LC circuit generates a virtual damping effect.

[0039] In this example, the PWM control signal is input to the gate drive module of the inverter drive circuit. The gate drive module converts the logic level signal into drive pulses suitable for power transistors (such as IGBTs or MOSFETs), and controls the conduction state of the upper and lower transistors of each bridge arm through turn-on and turn-off timing, forming a dynamic switching state sequence of the inverter. During this process, the drive circuit must ensure that there is always a dead time between the upper and lower transistors to avoid shoot-through, while also meeting the minimum conduction time and temperature rise protection mechanism to ensure that the switching devices operate within the safe boundary for a long time. When the power transistor enters a specific switching state combination according to the PWM control signal, the DC bus voltage is modulated into an equivalent AC voltage vector through the inverter bridge. The average value of the AC voltage vector in one switching cycle is equivalent to the reference voltage issued by the controller. Since the reference voltage includes a voltage compensation component calculated through the αβ coordinate system, and the voltage compensation component carries the characteristics of virtual impedance, the inverter output voltage not only has the original fundamental component, but also includes a compensation voltage component that is opposite in direction to the resonant current component. This virtual impedance, achieved through modulation, is adjustable and real-time. It can flexibly adjust the compensation strength and phase relationship according to parameter identification results and dynamic changes in the resonant frequency, achieving a more efficient and precise damping effect than physical damping elements. The inverter output voltage is directly applied to the LC circuit via the output port, coupling with the inductor current and capacitor voltage. Near the resonant frequency, the impedance of the LC circuit itself exhibits a high frequency dependence, causing the amplitude of the resonant current component to amplify and deviate from the fundamental voltage in phase. When the compensation component is added, it is out of phase with the resonant current component, resulting in power exchange between them. That is, the inverter actively injects a portion of reverse power to offset the energy accumulation of the resonant current, causing the resonant amplitude of the LC circuit to decay rapidly. In other words, the inverter constructs an energy dissipation path in the electrical space through a modulation strategy. This energy dissipation path functions similarly to a physical resistor, but without additional heat loss or hardware volume. The reverse-injected power gradually dissipates the resonant energy in the LC circuit, manifested as a weakening of the resonant peak and a rapid decay of the resonant waveform, while simultaneously reducing the total harmonic distortion of the system. During dynamic operation, when the load or bus conditions change, the parameter identification module updates the inductance and capacitance values ​​and calculates the new resonant frequency. The filter and voltage compensation modules adjust the amplitude and phase of the compensation components in real time. The inverter then injects the updated compensation into the LC circuit through PWM modulation, thereby ensuring that the virtual damping effect remains stable and effective under all operating conditions.

[0040] Reference Figure 2 This embodiment provides an LC resonance suppression device with active damping control, comprising: The parameter identification and processing module 021 is used to perform parameter identification and processing based on the bus voltage to obtain the inductance value, capacitance value and resonant frequency of the LC circuit. Extraction module 022 is used to extract the resonant current component of the current signal based on the resonant frequency; Calculation module 023 is used to calculate the damping power based on the resonant current component and convert the damping power into a voltage compensation command in the αβ coordinate system; Generation module 024 is used to generate PWM control signals based on voltage compensation instructions and voltage reference instructions; The switching control module 025 is used to control the switching of the inverter based on the PWM control signal and inject power in the LC circuit that is opposite to the resonant current component to produce a virtual damping effect.

[0041] In this embodiment, the specific implementation of each unit in the above device embodiment is described in the above method embodiment, and will not be repeated here.

[0042] This invention employs a three-stage cascaded observer structure. Through the synergistic effect of the front-stage linear ESO and the rear-stage nonlinear ESO, it can accurately track bus voltage ripple and resonant disturbance components, exhibiting higher detection accuracy and stronger anti-interference capability compared to traditional single-stage observers. Based on an online parameter identification mechanism using the recursive least squares method, it can track changes in LC circuit parameters in real time. Through variable forgetting factor adjustment and boundary constraint protection, it ensures the accuracy and numerical stability of the identification results, effectively adapting to changes in system operating conditions. A two-stage design using adaptive notch filters and virtual impedance achieves precise suppression of specific resonant frequencies while maintaining low impedance characteristics at the fundamental frequency, avoiding adverse effects on normal power transmission. A power allocation strategy in the αβ coordinate system is adopted. By accurately calculating damping power and rationally allocating it to each axis component, it ensures accurate matching between injected power and resonant current, achieving efficient energy dissipation. Through voltage command superposition and seven-segment switching sequence generation, the virtual impedance characteristics are seamlessly integrated into the inverter control, ensuring rapid response throughout the entire process from resonance detection to power injection, effectively suppressing resonant peaks and improving current quality.

[0043] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or method that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, apparatus, article, or method. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or method that includes that element.

[0044] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for suppressing LC resonance with active damping control, characterized in that, include: Based on the bus voltage, parameter identification processing is performed to obtain the inductance, capacitance and resonant frequency of the LC circuit; The resonant current component of the current signal is extracted based on the resonant frequency; The damping power is calculated based on the resonant current components, and the damping power is converted into a voltage compensation command in the αβ coordinate system; A PWM control signal is generated based on the voltage compensation command and the voltage reference command; The inverter is switched on and off based on the PWM control signal, and power opposite to the resonant current component is injected into the LC circuit to produce a virtual damping effect.

2. The active damping controlled LC resonance suppression method according to claim 1, characterized in that, The parameter identification process based on the bus voltage to obtain the inductance, capacitance, and resonant frequency of the LC circuit includes: The state equation of the first-order linear extended state observer is established based on the state matrix containing the LC natural frequency and quality factor. The bus voltage is input into the first-order linear extended state observer to obtain the first-order observation result. The bus voltage ripple component, voltage rate of change component and resonant disturbance component in the first-order observation results are used as state variables for linear tracking to obtain the second-order observation results. The second-order observation results are input into the third-order nonlinear extended state observer for nonlinear compensation in the resonant frequency band to obtain the third-order observation results; The third-order observation results are subjected to state separation to obtain the bus ripple signal and the resonant disturbance component; Based on the bus ripple signal and the resonant disturbance component, parameter identification processing is performed to obtain the inductance value, capacitance value and resonant frequency of the LC circuit.

3. The active damping controlled LC resonance suppression method according to claim 2, characterized in that, The second-order observation results are obtained by linearly tracking the bus voltage ripple component, voltage rate of change component, and resonant disturbance component from the first-order observation results as state variables, including: Extract the bus voltage ripple component, voltage rate of change component, and resonant disturbance component from the first-order observation results; The second-order extended state vector is constructed by combining the bus voltage ripple component, the voltage change rate component, and the resonant disturbance component. Based on the second-order extended state vector, the linear relationship between each state variable is described by the state transition matrix and recursively calculated to obtain the state tracking equation. Solve the state tracking equation, update the second-order extended state vector, and output the second-order observation result.

4. The active damping controlled LC resonance suppression method according to claim 3, characterized in that, The parameter identification processing based on the bus ripple signal and the resonant disturbance component to obtain the inductance, capacitance, and resonant frequency of the LC circuit includes: The bus ripple signal is used as the observation output vector, the resonant disturbance component is constructed as a regression vector, and a parameter vector containing inductance, capacitance and impedance values ​​is established. The parameter vector is corrected using the covariance matrix and forgetting factor to obtain the updated parameter identification results. Based on the updated parameter identification results, the forgetting factor is dynamically adjusted and the covariance matrix is ​​updated synchronously using a variable forgetting factor adjustment mechanism to obtain the target parameters. The inductance and capacitance values ​​of the LC circuit are extracted from the target parameters, and the resonant frequency is obtained by taking the square root of the inductance and capacitance components.

5. The active damping controlled LC resonance suppression method according to claim 1, characterized in that, The extraction of the resonant current component of the current signal based on the resonant frequency includes: The center frequency of the bandpass filter is set according to the resonant frequency, and the damping coefficient is set at the same time. A resonant frequency selection filter is constructed based on the center frequency and the damping coefficient. The current signal is input into the resonant frequency selection filter for frequency domain separation to obtain the filtered current in the resonant frequency band; The amplitude and phase of the filtered current in the resonant frequency band are detected to obtain the amplitude and phase angle of the resonant current. The phase delay of the filter is calculated based on the transfer function of the resonant frequency selective filter. The phase angle of the resonant current and the phase delay of the filter are phase compensated. The resonant current component is generated by reconstructing the phase angle and the amplitude of the resonant current using a cosine function.

6. The active damping controlled LC resonance suppression method according to claim 1, characterized in that, The step of calculating the damping power based on the resonant current component and converting the damping power into a voltage compensation command in the αβ coordinate system includes: The amplitude of the resonant current component is extracted and multiplied by the resonant current component to obtain the square value of the resonant current. Then, the square value of the resonant current is multiplied by the virtual impedance to obtain the damping power. The three-phase current signal is transformed from the abc three-phase coordinate system to the αβ stationary coordinate system by the Clarke transformation matrix, and the α-axis current component and the β-axis current component are obtained. Based on the resonant frequency, the α-axis current component and the β-axis current component are extracted by bandpass filtering to obtain the α-axis resonant current component and the β-axis resonant current component. The sum of the squares of the α-axis resonant current component and the β-axis resonant current component is calculated as the denominator value. When the denominator value is less than a preset minimum threshold, the preset minimum threshold is used as the denominator protection value. The damping power is multiplied by the α-axis resonant current component and the β-axis resonant current component respectively, and then divided by the denominator protection value to obtain the voltage compensation command in the αβ coordinate system.

7. The active damping controlled LC resonance suppression method according to claim 1, characterized in that, The step of generating a PWM control signal based on the voltage compensation command and the voltage reference command includes: The α-axis voltage compensation command and the α-axis voltage reference command are added together to obtain the α-axis total voltage command. At the same time, the β-axis voltage compensation command and the β-axis voltage reference command are added together to obtain the β-axis total voltage command. The total voltage command of the α-axis and the total voltage command of the β-axis are superimposed to obtain the first voltage command; Obtain the voltage amplitude of the first voltage command; when the voltage amplitude exceeds the maximum output voltage of the inverter, scale the α-axis total voltage command and the β-axis total voltage command proportionally to obtain the second voltage command. Based on the second voltage command, determine the sector where the voltage vector is located and calculate the duration of action of adjacent basic vectors; Based on the stated duration, the zero vector is evenly distributed across both ends of the switching cycle, and the switching sequence of the basic vector is arranged to obtain the PWM control signal.

8. The active damping controlled LC resonance suppression method according to claim 7, characterized in that, The step of uniformly distributing the zero vector to both ends of the switching cycle and arranging the switching sequence of the basic vector according to the action time to obtain the PWM control signal includes: Calculate the total action time of the zero vector based on the aforementioned action time; The total action time of the zero vector is symmetrically allocated to obtain the initial zero vector time and the final zero vector time; Based on the α-axis total voltage command and β-axis total voltage command in the second voltage command, the arctangent function is performed to obtain the voltage vector angle. According to the correspondence between the voltage vector angle and the six 60-degree intervals, the sector where the current voltage vector is located is determined. The timing arrangement of the first basic vector and the second basic vector is arranged according to the principle of minimum switching of adjacent vectors to obtain the vector switching sequence. A PWM control signal is generated based on the start zero vector time, the vector switching sequence, and the end zero vector time.

9. The active damping controlled LC resonance suppression method according to claim 1, characterized in that, The switching control of the inverter based on the PWM control signal, and the injection of power opposite to the resonant current component into the LC circuit to generate a virtual damping effect, include: The PWM control signal is input into the inverter's drive circuit to control the power transistor switching, thereby obtaining the switching state. Based on the switching state, the DC voltage is converted into an inverter output voltage containing virtual impedance characteristics; The inverter output voltage is applied to the LC circuit, causing the compensation component in the inverter output voltage to exchange power with the resonant current component in opposite phase, thus obtaining reverse injected power. The reverse injected power generates a virtual damping effect in the LC circuit.

10. An LC resonance suppression device with active damping control, characterized in that, The steps for implementing the LC resonance suppression method with active damping control according to any one of claims 1 to 9 include: The parameter identification and processing module is used to perform parameter identification and processing based on the bus voltage to obtain the inductance value, capacitance value and resonant frequency of the LC circuit. An extraction module is used to extract the resonant current component of the current signal based on the resonant frequency; The calculation module is used to calculate the damping power based on the resonant current component and convert the damping power into a voltage compensation command in the αβ coordinate system; The generation module is used to generate a PWM control signal based on the voltage compensation command and the voltage reference command; The switching control module is used to control the switching of the inverter based on the PWM control signal and inject power in the LC circuit that is opposite to the resonant current component to generate a virtual damping effect.

Images