Carrier modulation method and modulation-demodulation system of MMC (Modular Multilevel Converter) circulation suppressor

Abstract

The invention provides a carrier modulation method and modulation-demodulation system for an MMC circulation suppressor, and the method comprises the following steps: obtaining MMC operation log data, carrying out the spectrum analysis, and extracting the spectrum amplitude distribution characteristics; identifying a bridge arm circulation condition based on the amplitude distribution characteristics, and further detecting bridge arm load abnormity and unit energy abnormity conditions; identifying a circulating current frequency component according to an energy abnormal condition, performing electric energy quality attenuation detection and initial circulating current position tracing, and determining a circulating current generation mechanism; and finally, based on the risk assessment result and the generation mechanism data, constructing a modulation parameter, and generating an adaptive circulation suppressor carrier modulation strategy. Compared with the prior art, the method has the following beneficial effects: through modulation of the circulating current suppressor, modulation of the circulating current suppressor is more accurate, so that the problems of inaccurate analysis of a bridge arm circulating current generation mechanism and inaccurate analysis of an energy abnormal condition of the MMC unit existing in carrier modulation of a traditional circulating current suppressor are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of circulating current suppressors, and in particular to a carrier modulation method and a modulation and demodulation system of an MMC circulating current suppressor. Background Art

[0002] Due to the coupling between the upper and lower bridge arms in the MMC structure, improper current balance control can cause circulating current. Circulating current refers to internal reactive current flowing through the bridge arm but not through the load. It is primarily caused by factors such as DC bias, unbalanced modulation, harmonic injection, and module voltage differences. Circulating current increases losses and heat loads in bridge arm components, reducing component lifespan. It also degrades system power quality, manifesting as increased harmonic distortion, voltage instability, and output power fluctuations. In severe cases, it can even lead to grid instability or system protection activation. Existing technologies primarily employ current-controlled strategies, including proportional resonant control, virtual impedance injection, and frequency-locked modulation. However, these methods generally rely on model prediction or feedback control of circulating currents, resulting in slow response, sensitivity to operational disturbances, and reliance on manual setting of control parameters. These methods struggle to meet the requirements for stable operation under high-speed dynamic conditions. However, the carrier modulation of traditional circulating current suppressors suffers from inaccurate analysis of the generation mechanism of circulating current in the bridge arm and inaccurate analysis of MMC unit energy anomalies. Summary of the Invention

[0003] In view of the shortcomings of the prior art described above, the object of the present invention is to provide a carrier modulation method and modulation and demodulation system for an MMC circulating current suppressor, which is used to solve the problems of inaccurate analysis of the bridge arm circulating current generation mechanism and inaccurate analysis of MMC unit energy abnormalities in the carrier modulation of the traditional circulating current suppressor.

[0004] To achieve the above-mentioned and other related purposes, the present invention provides the following technical solutions:

[0005] A carrier modulation method for an MMC circulating current suppressor comprises the following steps:

[0006] Step S1: Acquire MMC operation log data; Perform spectrum analysis based on MMC operation log data to obtain MMC operation bridge arm current spectrum data; Extract MMC operation spectrum amplitude distribution characteristics based on MMC operation bridge arm current spectrum data; Step S2: Determine MMC bridge arm circulation status based on MMC operation spectrum amplitude distribution characteristics; Detect MMC bridge arm load abnormality based on MMC bridge arm circulation status; Determine MMC unit energy abnormality based on MMC bridge arm load abnormality; Step S3: Determine circulation frequency component based on MMC unit energy abnormality; Detect MMC output according to circulation frequency component Output power quality attenuation situation; perform MMC initial circulation position tracing processing according to the MMC output power quality attenuation situation to obtain MMC initial circulation position tracing data; determine the bridge arm circulation generation mechanism according to the MMC initial circulation position tracing data to obtain the bridge arm circulation generation mechanism data; step S4: evaluate the abnormal risk situation of the bridge arm circulation according to the bridge arm circulation generation mechanism data; construct MMC carrier modulation parameters based on the abnormal risk situation of the bridge arm circulation and the bridge arm circulation generation mechanism data; generate an adaptive circulation suppressor carrier modulation strategy according to the MMC carrier modulation parameters to obtain the adaptive circulation suppressor carrier modulation strategy data.

[0007] Preferably, step S1 includes the following steps: step S11: obtaining MMC operation log data; step S12: collecting the MMC operation bridge arm current operating status according to the MMC operation log data; step S13: performing spectrum analysis on the MMC operation bridge arm current operating status to obtain MMC operation bridge arm current spectrum data; step S14: extracting the MMC operation spectrum amplitude distribution characteristics according to the MMC operation bridge arm current spectrum data.

[0008] Preferably, step S13 includes the following steps: step S131: performing current timing analysis on the operating status of the MMC operating bridge arm current to obtain the MMC operating bridge arm current timing data; step S132: performing current timing fluctuation analysis based on the MMC operating bridge arm current timing data to obtain the bridge arm current timing fluctuation data; step S133: performing statistics on the periodic change of the bridge arm current based on the bridge arm current timing fluctuation data to obtain the bridge arm current periodic change data; step S134: performing bridge arm current fluctuation amplitude measurement based on the bridge arm current timing fluctuation data to obtain the bridge arm current fluctuation amplitude data; step S135: performing spectrum analysis based on the bridge arm current periodic change data and the bridge arm current fluctuation amplitude data to obtain the MMC operating bridge arm current spectrum data.

[0009] Preferably, step S2 includes the following steps: step S21: detecting the MMC circulating current characteristic harmonic condition according to the MMC operation spectrum amplitude distribution characteristics; step S22: determining the MMC bridge arm circulating current condition according to the MMC operation spectrum amplitude distribution characteristics and the MMC circulating current characteristic harmonic condition; step S23: detecting the MMC bridge arm load abnormality condition based on the MMC bridge arm circulating current condition to obtain the MMC bridge arm load abnormality condition data; step S24: identifying the MMC unit energy abnormality according to the MMC bridge arm load abnormality condition to obtain the MMC unit energy abnormality data.

[0010] Preferably, step S23 includes the following steps: step S231: performing bridge arm current abnormality superposition detection according to the MMC bridge arm circulation condition to obtain bridge arm current abnormality superposition data; step S232: determining the bridge arm current transient fluctuation according to the bridge arm current abnormality superposition data; step S233: identifying the instantaneous growth of the current peak based on the bridge arm current transient fluctuation and the bridge arm current abnormality superposition data; step S234: predicting the saturation condition of the filter inductor core according to the instantaneous growth of the current peak; step S235: determining the bridge arm structure loss growth according to the instantaneous growth of the current peak and the saturation condition of the filter inductor core; step S236: performing MMC bridge arm load abnormality detection based on the bridge arm structure loss growth to obtain MMC bridge arm load abnormality data.

[0011] Preferably, step S24 includes the following steps: step S241: testing the current imbalance condition of adjacent bridge arms according to the abnormal condition of the MMC bridge arm load; step S242: detecting the strengthening trend of the bridge arm circulation component according to the adjacent bridge arm current imbalance condition and the abnormal condition of the MMC bridge arm load; step S243: predicting the positive feedback effect of the bridge arm circulation according to the strengthening trend of the bridge arm circulation component and the current imbalance condition of the adjacent bridge arm; step S244: identifying the invalid energy circulation condition inside the bridge arm based on the positive feedback effect of the bridge arm circulation; step S245: predicting the enhanced electromagnetic interference of the bridge arm operation according to the invalid energy circulation condition inside the bridge arm; step S246: determining the abnormal condition of the bridge arm power exchange efficiency based on the enhanced electromagnetic interference of the bridge arm operation; step S247: identifying the MMC unit energy abnormality according to the abnormal condition of the bridge arm power exchange efficiency and the enhanced condition of the bridge arm operation electromagnetic interference, and obtaining the MMC unit energy abnormality data.

[0012] Preferably, step S3 includes the following steps: step S31: determining the circulating current frequency component based on the MMC unit energy abnormality to obtain the circulating current frequency component situation; step S32: performing MMC output power quality attenuation detection according to the circulating current frequency component situation to obtain the MMC output power quality attenuation situation; step S33: performing MMC initial circulating current position tracing processing according to the MMC output power quality attenuation situation to obtain MMC initial circulating current position tracing data; step S34: determining the bridge arm circulating current generation mechanism according to the MMC initial circulating current position tracing data and the circulating current frequency component situation to obtain the bridge arm circulating current generation mechanism data.

[0013] Preferably, step S32 includes the following steps: step S321: detecting the instability of the internal circulating current component of the bridge arm according to the circulating current frequency component; step S322: determining the bridge arm output spectrum broadening condition based on the instability of the internal circulating current component of the bridge arm; step S323: using the bridge arm output spectrum broadening condition to test the waveform distortion condition of the bridge arm output end; step S324: predicting the bridge arm output heat loss cumulative data based on the bridge arm output end waveform distortion condition; step S325: evaluating the bridge arm grid-connected stability attenuation trend according to the bridge arm output heat loss cumulative data; step S326: performing MMC output power quality attenuation detection based on the bridge arm grid-connected stability attenuation trend and the instability of the internal circulating current component of the bridge arm to obtain the MMC output power quality attenuation condition.

[0014] Preferably, step S4 includes the following steps: step S41: calculating the circulation generation probability according to the bridge arm circulation generation mechanism data to obtain the bridge arm circulation generation probability data; step S42: evaluating the bridge arm circulation abnormality risk situation based on the bridge arm circulation generation probability data; step S43: constructing MMC carrier modulation parameters based on the bridge arm circulation abnormality risk situation and the bridge arm circulation generation mechanism data; step S44: constructing an MMC suppressor modulation matrix according to the MMC carrier modulation parameters; step S45: generating an adaptive circulation suppressor carrier modulation strategy according to the MMC suppressor modulation matrix to obtain adaptive circulation suppressor carrier modulation strategy data.

[0015] A carrier modulation system for an MMC circulating current suppressor is used to execute the carrier modulation method for the MMC circulating current suppressor described above. The carrier modulation system for the MMC circulating current suppressor includes:

[0016] The spectrum feature extraction module is used to obtain the MMC operation log data; perform spectrum analysis based on the MMC operation log data to obtain the MMC operation bridge arm current spectrum data; extract the MMC operation spectrum amplitude distribution characteristics based on the MMC operation bridge arm current spectrum data; the energy anomaly determination module is used to determine the MMC bridge arm circulation status based on the MMC operation spectrum amplitude distribution characteristics; detect the MMC bridge arm load anomaly based on the MMC bridge arm circulation status; determine the MMC unit energy anomaly based on the MMC bridge arm load anomaly; the circulation mechanism identification module is used to determine the circulation frequency component based on the MMC unit energy anomaly; and determine the MMC unit energy anomaly based on the circulation frequency component. The invention can detect the attenuation of MMC output power quality according to the condition; perform MMC initial circulating current position tracing processing according to the attenuation of MMC output power quality to obtain MMC initial circulating current position tracing data; determine the bridge arm circulating current generation mechanism according to the MMC initial circulating current position tracing data to obtain bridge arm circulating current generation mechanism data; a modulation strategy generation module is used to evaluate the abnormal risk of bridge arm circulating current according to the bridge arm circulating current generation mechanism data; construct MMC carrier modulation parameters based on the abnormal risk of bridge arm circulating current and the bridge arm circulating current generation mechanism data; generate an adaptive circulating current suppressor carrier modulation strategy according to the MMC carrier modulation parameters to obtain adaptive circulating current suppressor carrier modulation strategy data.

[0017] As described above, the carrier modulation method and modulation and demodulation system of an MMC circulating current suppressor of the present invention have the following beneficial effects: the present invention forms a complete set of circulating current analysis and modulation control mechanisms by constructing a multi-level spectrum analysis, anomaly detection, mechanism identification and modulation strategy generation process. Compared with the existing scheme based only on model derivation or linear feedback control, this method has a higher degree of systematicity and automation in structural design and parameter reasoning, and can achieve refined tracing and adaptive modulation response of the circulating current generation mechanism during the operation of the MMC, thereby improving the pertinence and control accuracy of the overall circulating current suppression. The present invention introduces a multi-dimensional feature extraction process of the bridge arm current spectrum, and obtains statistically significant frequency domain feature parameters by performing time series fluctuation measurement, periodic change analysis and spectrum amplitude extraction on the bridge arm current information in the MMC operation log data, providing original support for subsequent anomaly analysis. By clarifying the logical relationship between the current fluctuation amplitude and the periodic parameters, it ensures that the extracted spectrum features are systematic and comparable, providing a solid data foundation for bridge arm circulating current identification. The present invention establishes a technical path extending from circulating current feature identification to energy anomaly measurement. This step uses harmonic detection and overlay analysis of spectral amplitude distribution characteristics, combined with transient arm current fluctuations and filter device responses, to gradually identify arm current anomalies, structural loss trends, and decreased switching efficiency, ultimately identifying the energy status of the MMC unit. This locates localized load anomalies caused by circulating currents and simultaneously assesses the energy circulation imbalance and electromagnetic interference accumulation characteristics within the arm, enabling dynamic determination and structural identification of energy anomalies. Based on the energy anomaly identification results, a mechanism for determining circulating current frequency components and a power quality analysis process were further established. Frequency component analysis and spectrum broadening feature detection, combined with heat loss prediction at the arm output and grid stability trend assessment, enable a comprehensive assessment of power quality degradation. This result is then used to identify the initial circulating current source. In particular, the parallel processing of output waveform distortion and heat loss accumulation enables more targeted source identification, effectively supporting the subsequent derivation of the arm circulating current generation mechanism and ensuring the physical interpretability and modulation feasibility of the mechanism data. By establishing a modulation parameter construction and risk prediction method centered on data on the generation mechanism of circulating currents in the bridge arms, the adaptability of the modulation strategy has been significantly improved. Specifically, by calculating the probability of circulating current generation and assessing abnormal risks, a carrier modulation parameter set is constructed in conjunction with the system's operating status. This is then converted into a suppressor modulation matrix, enabling fully automatic generation of an adaptive modulation strategy. This modulation strategy not only considers the frequency response and harmonic component changes within the bridge arm but also introduces logical control of energy circulation and loss paths, ensuring a close match between the suppressor output and the operating status, enabling active regulation of the circulating current frequency components and energy distribution.Therefore, the present invention optimizes the carrier modulation of the traditional circulating current suppressor, solves the problem of inaccurate analysis of the bridge arm circulating current generation mechanism and the problem of inaccurate analysis of MMC unit energy abnormality in the carrier modulation of the traditional circulating current suppressor, and improves the accuracy of the analysis of the bridge arm circulating current generation mechanism and the accuracy of the analysis of MMC unit energy abnormality. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 A schematic flow chart of the steps of a carrier modulation method for an MMC circulating current suppressor;

[0019] Figure 2 for Figure 1 Detailed implementation steps of step S2 in FIG.

[0020] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. DETAILED DESCRIPTION

[0021] The following describes the embodiments of the present invention through specific examples. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that the following embodiments and features in the embodiments can be combined with each other unless there is a conflict.

[0022] See also Figures 1 to 3 The present invention provides a carrier modulation method for an MMC circulating current suppressor, comprising the following steps:

[0023] Step S1: obtaining MMC operation log data; performing spectrum analysis based on the MMC operation log data to obtain MMC operation bridge arm current spectrum data; extracting MMC operation spectrum amplitude distribution characteristics based on the MMC operation bridge arm current spectrum data;

[0024] In an embodiment of the present invention, in a high-voltage direct current transmission system, historical operation log data of a modular multilevel converter (MMC) during actual operation is obtained as the basis for data analysis. The source of the operation log data includes the data acquisition interface of the MMC control system, where the acquired signal covers the operating parameters of all bridge arm currents. The log data is accurately recorded according to the timestamp, the sampling frequency is not less than 10kHz, and the sampling accuracy is not less than 0.1%. The complete export of the operation log database is completed through an industrial communication interface (such as Modbus, IEC61850), and time series alignment and outlier removal are performed. The removal standard is the triple standard deviation principle or sliding window denoising. After the data preparation is completed, a fast Fourier transform (FFT) analysis is performed on the current time series data of each bridge arm to extract the spectrum information of the bridge arm current in the frequency range of 0Hz to 5kHz. The frequency resolution is set to 1Hz to capture the fine fluctuations of the frequency components. The obtained spectrum data includes the amplitude information of the corresponding frequency points, and a spectrum amplitude distribution vector is constructed for each bridge arm. The spectral amplitude distribution characteristics of each bridge arm are then calculated using methods including spectrum peak detection (to extract the amplitude of the main frequency component), bandwidth energy concentration analysis (e.g., the width of the frequency interval where 90% of the energy is concentrated), and total harmonic distortion (THD) statistics. The resulting spectral amplitude distribution feature vectors for multiple bridge arms contain statistics such as the main frequency amplitude, characteristic harmonic amplitudes, and frequency band energy distribution. These serve as the input data for subsequent circulating current identification and status diagnosis.

[0025] Step S2: determining the MMC arm circulation condition according to the MMC operation spectrum amplitude distribution characteristics; detecting the MMC arm load abnormality based on the MMC arm circulation condition; and determining the MMC unit energy abnormality according to the MMC arm load abnormality.

[0026] In this embodiment of the present invention, the bridge arm circulating current condition is identified based on the spectrum amplitude distribution characteristics extracted in Example 1. Characteristic harmonics in the spectrum characteristics of each bridge arm are classified and statistically analyzed, and their amplitudes and relative energy proportions are analyzed for typical circulating current frequency points (such as 150Hz, 300Hz, and 450Hz). A threshold judgment method is used. If the characteristic harmonic amplitude exceeds 5% of the fundamental amplitude, a significant circulating current frequency component is determined to be present. Furthermore, combined with the phase difference between the bridge arms, it is confirmed to be common-mode or differential-mode circulating current. Based on the identified bridge arm circulating current conditions, the bridge arm load is then detected for abnormality. The output currents of each bridge arm are superimposed and analyzed, and their transient fluctuation rates and peak drift trends are calculated. Sliding window variance and peak transient increment are used to identify abnormal current superposition. If the current peak value is detected to be continuously above the upper limit of the normal range (predetermined by the equipment operating specifications) and accompanied by an increase in transient oscillation frequency, it can be further determined to be an abnormal load condition. After obtaining the abnormal bridge arm load condition data, the impact of this condition on the energy distribution of the MMC unit is analyzed. Using the recorded values of each submodule's capacitor voltage, the energy variation and fluctuation trend per unit time are calculated. By coupling the capacitor voltage variation with abnormal load conditions, abnormal energy units within the bridge arm are identified. The output includes the unit number with the most severe energy fluctuation and the corresponding abnormal energy amplitude, which serves as input for the next step of frequency component analysis.

[0027] Step S3: determining the circulating current frequency component based on the MMC unit energy anomaly; detecting the MMC output power quality attenuation based on the circulating current frequency component; performing MMC initial circulating current position tracing processing based on the MMC output power quality attenuation to obtain MMC initial circulating current position tracing data; determining the bridge arm circulating current generation mechanism based on the MMC initial circulating current position tracing data to obtain bridge arm circulating current generation mechanism data;

[0028] In an embodiment of the present invention, based on the energy anomaly data obtained in step S2, the impact of the abnormal unit on the overall bridge arm circulating frequency component is analyzed. A frequency domain filter is used to extract the current in the circulating path, and the circulating current before and after the abnormal energy unit is band-pass filtered (the center frequency is set to the abnormal harmonic frequency point) to analyze the distribution of the circulating frequency component. The circulating frequency component evolution spectrum is further constructed based on the energy changes of different frequency components at different times. After obtaining the circulating frequency component, the attenuation of the MMC output power quality is detected. The output voltage and current waveforms are measured. THD (total harmonic distortion), power factor, harmonic energy spectrum distribution and other indicators are measured, and the output end heat loss, voltage swing and other performance are quantified; if the THD exceeds 5% and the power factor is lower than 0.9, it is judged that there is obvious power quality attenuation. After the power quality attenuation is confirmed, the initial circulating position is traced and analyzed. The initial excitation source of the harmonic components of the bridge arm's circulating current is tracked, and correlation matching is performed based on the timing of the occurrence of different frequency points and the corresponding bridge arm energy anomaly moments. Event time series matching technology is used to determine the bridge arm and module locations where the anomaly occurred, forming the initial circulating current location traceability data. The internal current path of the bridge arm, the module energy change mechanism, and the control command response are further analyzed based on the traceability data. The control command log and current response data before and after the fault occur are combined to comprehensively determine the circulating current generation mechanism. The relevant causal relationship is established as a bridge arm circulating current generation mechanism data structure, which includes the generation time point, the affected frequency component, the energy flow direction, the bridge arm location, and the module number.

[0029] Step S4: Evaluate the abnormal risk of arm circulation according to the data of the bridge arm circulation generation mechanism; construct MMC carrier modulation parameters based on the abnormal risk of arm circulation and the data of the bridge arm circulation generation mechanism; generate an adaptive circulation suppressor carrier modulation strategy according to the MMC carrier modulation parameters to obtain adaptive circulation suppressor carrier modulation strategy data.

[0030] In an embodiment of the present invention, after obtaining data on the generation mechanism of arm circulation current, an abnormal arm circulation risk assessment is performed. By comparing similar energy fluctuation characteristics and harmonic structure characteristics from historical similar events, a deterministic statistical matching method is used to calculate the probability of the current arm circulation recurrence, generating arm circulation probability data. This data, combined with the variation patterns of the arm spectrum, is used to determine whether there is a persistent or periodic risk, thereby generating abnormal arm circulation risk data. Subsequently, carrier modulation parameters are constructed based on the abnormal arm circulation risk and the generation mechanism data. These parameters involve the following modulation factors: modulation frequency fine-tuning (in 1Hz steps), pulse width variation ratio (pulse width offset within a unit cycle), and control delay compensation value (set based on the sampling period and calculation delay). These modulation parameters are input into the carrier modulation parameter generation function to complete the adaptive modulation coefficient matching process. Based on these carrier modulation parameters, an MMC carrier modulation matrix is constructed, consisting of multiple weights and modulation coefficients. This matrix has a two-dimensional structure, with rows corresponding to each arm unit and columns representing modulation coefficients for different frequency bands, arranged according to modulation frequency and harmonic influence weights. Finally, a complete carrier modulation strategy is designed based on the modulation matrix. The strategy content includes: bridge arm control priority, modulation frequency dynamic adjustment path, modulation amplitude update mechanism, etc. All of them are compiled into a fixed-format data structure and written into the controller instruction cache to form the adaptive circulating current suppressor carrier modulation strategy data for real-time scheduling and updating during system operation.

[0031] Preferably, step S1 includes the following steps:

[0032] Step S11: Obtain MMC operation log data;

[0033] In an embodiment of the present invention, during the operation of a modular multilevel converter (MMC) in a high-voltage direct current transmission system, historical data during the operation process is collected and stored in real time to form operation log data containing complete electrical variable records. The log data is collected by a data acquisition module embedded in the converter control system, and the acquisition frequency is set to 20kHz to ensure sufficient analysis capabilities for high-frequency disturbances and harmonic components. The collected content includes but is not limited to each bridge arm current, bus voltage, unit capacitor voltage, load current, phase voltage, control command, electromagnetic interference waveform and system event identifier. The data is stored in a structured time series database, such as InfluxDB, so that it can be indexed by timestamp later. During the acquisition process, a data integrity check mechanism is used to perform frame check on the fields in each sampling period based on the CRC-16 algorithm, and data frames with packet loss and synchronization anomalies are eliminated to ensure the continuity and accuracy of the log data, providing high-confidence raw data support for subsequent steps.

[0034] Step S12: collecting the MMC bridge arm current operating status according to the MMC operation log data;

[0035] In this embodiment of the present invention, bridge arm current data within a target time period is extracted from a log database by timestamp. The bridge arm current data includes upper and lower arm currents, covering all three phases (A, B, and C), totaling six bridge arms. To prevent transient fluctuations from interfering with data stability, a moving average filtering algorithm is used to preprocess the current data. The filter window width is set to 50ms. After filtering, the current curve that truly reflects the steady-state operating characteristics of the system is retained. The sampling accuracy is required to be within ±0.1A, and the data unit is uniformly amperes (A). After processing, the sampled data of each bridge arm forms a structured current operation data set containing a timestamp, current value, and phase identifier. The current operation status of each bridge arm is further reconstructed into a uniformly distributed current time series signal at 1000 sampling points within one minute, ensuring that the time domain resolution and frequency domain bandwidth of the subsequent spectrum analysis meet the requirements for resolving the frequency components of the internal circulating current of the bridge arm. The processed result is output as a bridge arm current operation status dataset. The dataset structure includes: bridge arm identifier, sampling timestamp sequence, current amplitude sequence, and phase synchronization identifier sequence.

[0036] Step S13: performing spectrum analysis on the MMC operating arm current operating status to obtain MMC operating arm current spectrum data;

[0037] In an embodiment of the present invention, based on the bridge arm current operating status data set obtained in S12, a fast Fourier transform (FFT) algorithm is used to perform spectrum decomposition processing on each bridge arm current sequence. The FFT analysis module is implemented in the MATLAB environment, the sampling window is set to 2048 points, and the window function type is selected as the Hanning window to reduce the interference of sidelobe leakage on the frequency domain amplitude. After the FFT operation, the spectrum results are amplitude normalized, and the spectrum data with a frequency distribution range of 0Hz to 5kHz is retained. The analysis frequency interval is 2.5Hz, and the amplitude response curve of each bridge arm within this frequency bandwidth is obtained. The spectrum data is represented in matrix form, where each column corresponds to a frequency point and each row corresponds to the amplitude response of a bridge arm current, in amperes (A). Low-order harmonics (e.g., 100Hz, 150Hz, 300Hz) above the fundamental frequency (50Hz or 60Hz) in the spectrum data will be used as key indicators for subsequent identification of circulating current energy distribution, and the spectrum data will be retained intact until the next step. The output spectrum data structure is: bridge arm number, frequency sequence, amplitude sequence and time label corresponding to the frequency distribution, forming the MMC operation bridge arm current spectrum dataset.

[0038] Step S14: extracting the MMC operation spectrum amplitude distribution characteristics according to the MMC operation bridge arm current spectrum data.

[0039] In an embodiment of the present invention, a spectrum amplitude distribution feature extraction operation is performed on the spectrum data set obtained in S13. Five characteristic frequency band intervals are set in the frequency domain: 0-50Hz, 50-500Hz, 500-1000Hz, 1000-2000Hz, and 2000-5000Hz, representing the fundamental wave, low-order harmonics, medium and high frequency disturbances, and high frequency noise areas, respectively. The corresponding amplitude integral value is calculated in each frequency band, and the amplitude envelope area is estimated using the trapezoidal integration method to quantify the energy distribution characteristics in each frequency band. At the same time, the maximum amplitude frequency point in each frequency band is extracted, and its frequency position and peak value are recorded as characteristic point indicators. In addition, the frequency offset of each bridge arm current spectrum is calculated, that is, the maximum offset amplitude between the reference frequency (50Hz or 60Hz), which is used to reflect whether there is low-frequency oscillation or AC component offset in the system. The resulting spectrum amplitude distribution data includes the following: the integrated amplitude values for the five frequency bands, the frequency and amplitude of the maximum amplitude point in each band, the overall frequency offset index, and the RMS amplitude of each bridge arm spectrum. This distribution characteristic result is stored in matrix form and serves as input data for bridge arm circulation identification and analysis. It is then passed to the subsequent implementation process corresponding to step S2 to ensure data chain consistency.

[0040] Preferably, step S13 includes the following steps:

[0041] Step S131: performing a current time series analysis on the current operating status of the MMC operating bridge arm to obtain the current time series data of the MMC operating bridge arm;

[0042] In an embodiment of the present invention, during the operation of the MMC system, a high-speed sampling current sensor (such as a Hall sensor or a fluxgate sensor) is deployed at each bridge arm bus to continuously collect the AC current signal of each bridge arm at a sampling frequency of more than 10kHz. The acquisition time window is set to 10 seconds, covering the complete low-frequency, medium-frequency and harmonic signal segments. The collected current signal is converted into a digital signal via a high-precision analog-to-digital conversion module (ADC module with a resolution of 16bit) and then transmitted to the data processing module. To ensure the integrity and anti-interference performance of the current signal, the sampled data is first subjected to noise preprocessing, a bandpass filter is used (to filter out DC bias and high-frequency interference, and the bandwidth is set to 1Hz to 10kHz), and then a linear detrending process is performed to eliminate long-period background changes and retain stable periodic signals. After the above operations, structured bridge arm current time series data is generated. The data structure includes: timestamp, bridge arm number, phase identifier, current value, which is stored in matrix form as the data basis for subsequent time series fluctuation analysis.

[0043] Step S132: performing current timing fluctuation analysis based on the MMC operation bridge arm current timing data to obtain bridge arm current timing fluctuation data;

[0044] In an embodiment of the present invention, based on the bridge arm current time series data generated in step S131, a sliding window variance analysis method is used to evaluate the current fluctuation. The sliding window width is set to 100ms and the step length is 10ms. Local statistics are performed on each bridge arm current time series segment, and the current mean and standard deviation in the window are calculated. By calculating the mean difference and variance change rate between each time window and the previous window, and constructing a first-order difference sequence and a second-order difference sequence, the time series fluctuation points are extracted. A volatility index matrix is further constructed, which includes: time index, maximum fluctuation amplitude in the window, fluctuation duration, peak-to-valley frequency, and local jitter density. This index matrix is the bridge arm current time series fluctuation data, which directly reflects the non-stationarity and modulation trend of the current signal, and provides a reference for periodic change analysis.

[0045] Step S133: performing statistics on periodic changes of the bridge arm current according to the bridge arm current time series fluctuation data to obtain bridge arm current periodic change data;

[0046] In an embodiment of the present invention, the fluctuation index matrix extracted in step S132 is used to perform period detection processing. A method combining autocorrelation function analysis and fast Fourier transform (FFT) is used to identify the main frequency and secondary frequency components of the current signal. Autocorrelation analysis is used to determine the existence of periodic changes and their repetitive patterns. By judging the stability of the fluctuation peak and interval, the preliminary characteristics of the signal period are determined; then FFT processing is performed on the fluctuation signal, and the spectrum resolution is set to 0.1Hz. The main frequency value and the second highest frequency point are extracted, and the main frequency amplitude and the period corresponding time interval data are combined to form a periodic change feature set. The bridge arm current periodic change data consists of the following fields: main period value, frequency domain peak frequency, amplitude weight, period stability, and frequency drift range, which are used for subsequent fluctuation amplitude measurement and spectrum extraction.

[0047] Step S134: measuring the bridge arm current fluctuation amplitude based on the bridge arm current time series fluctuation data to obtain the bridge arm current fluctuation amplitude data;

[0048] In an embodiment of the present invention, the time series fluctuation data obtained in step S132 is used as input to extract the maximum amplitude, minimum amplitude and zero crossing amplitude of the current signal in each cycle. A peak tracking algorithm is used to identify the peak and trough positions, and combined with the periodic boundary data, the current amplitude difference is calculated by period segmentation. The sliding extreme value envelope method (EnvelopeDetection) is introduced to construct the upper envelope and the lower envelope, and their average difference is calculated as the periodic fluctuation amplitude. At the same time, the Fourier amplitude accumulation method is used to integrate the amplitude distribution within the 1st to 3rd main harmonic frequency band to form a frequency-weighted fluctuation value. The bridge arm current fluctuation amplitude data is output, and its structure includes the maximum current value, minimum current value, amplitude fluctuation range, average harmonic amplitude ratio, and fluctuation amplitude normalization factor corresponding to each cycle, providing a data basis for spectrum calculation.

[0049] Step S135: performing spectrum analysis based on the bridge arm current periodic variation data and the bridge arm current fluctuation amplitude data to obtain MMC operation bridge arm current spectrum data.

[0050] In this embodiment of the present invention, the periodic variation data and fluctuation amplitude data extracted in steps S133 and S134 are combined to enhance the accuracy and feature expression capabilities of spectrum analysis. A high-resolution short-time Fourier transform (STFT) is used, with a Hamming window selected as the window function, a window length of 256ms, and a 50% overlap ratio to construct a time-frequency distribution diagram. A frequency band weighting mechanism based on periodicity guidance is introduced during spectrum processing. This weights frequency bands with prominent periodicity to increase the spectrum energy threshold and reduce background noise interference. The target frequency range for spectrum analysis is set to 0Hz to 5kHz, and the amplitude-frequency distribution curve and energy density spectrum of the bridge arm current signal within this frequency range are output to form the MMC operating bridge arm current spectrum data. This data is presented as a two-dimensional array, with the frequency points on the X-axis and the corresponding amplitudes on the Y-axis. Additional fields include the main peak frequency of the spectrum, the amplitude peak width, and the frequency offset index.

[0051] Preferably, step S2 includes the following steps:

[0052] Step S21: detecting the MMC circulating current characteristic harmonic condition according to the MMC operation spectrum amplitude distribution characteristics;

[0053] In an embodiment of the present invention, when the MMC operating spectrum amplitude distribution characteristics are used to detect the circulating current harmonic components, the frequency components and corresponding amplitude information of the bridge arm current are extracted from the spectrum amplitude distribution characteristic data obtained in step S14. A bandpass filter group is used to construct fixed bandwidth windows at specific low-order harmonic frequency positions such as the 1st, 2nd, 3rd, 6th, and 12th orders, and the spectrum amplitude data is extracted by frequency band. Each bandpass window is determined by a constant center frequency and a fixed sideband frequency, and is constructed using a Butterworth bandpass filter. The filter order is set to 4th order, and the bandwidth range is set to ±5% of the center frequency. After completing the filtering operation, the amplitude peak, mean, and effective value corresponding to each harmonic frequency band are counted to construct a harmonic distribution vector. The amplitude mean is compared with the theoretical fundamental wave proportional amplitude of the corresponding frequency band, and the THD (Total Harmonic Distortion) calculation formula is used to perform normalized harmonic distortion calculation. Based on the typical performance of low-order characteristic harmonics (such as 6f and 12f components) in the circulating current, a harmonic characteristic index list is constructed. Output MMC circulating current characteristic harmonic status data. The data format is a structure table of {frequency component: amplitude, normalized amplitude, distortion rate}.

[0054] Step S22: determining the MMC bridge arm circulating current status according to the MMC operation spectrum amplitude distribution characteristics and the MMC circulating current characteristic harmonic status;

[0055] In an embodiment of the present invention, the bridge arm circulation condition is determined using the spectrum amplitude distribution characteristics and the circulation harmonic characteristics as dual input conditions. The frequency distribution curve in the spectrum amplitude distribution characteristics is aligned with the characteristic harmonic frequency response data obtained in step S21, and the amplitude trend of the current harmonic surge frequency is obtained through frequency consistency mapping. Using the time domain-frequency domain fusion algorithm, an amplitude response weight curve is constructed in the frequency domain space, and a joint evaluation index is constructed by superimposing the fluctuation trend curve of the circulating harmonics. In the joint evaluation process, a circulation severity factor (Circulating Current Severity Index, CCSI) is introduced. This index is composed of the average growth rate and instantaneous change rate of the characteristic harmonic amplitude. When the CCSI is greater than the set threshold of 0.35, it is determined that the circulation is significantly abnormal; when it is greater than 0.2 and less than 0.35, it is a slight abnormality; when it is less than 0.2, it is considered that the bridge arm circulation is within the normal controllable range. The output result is the bridge arm circulation condition data, and the data format is {CCSI value, harmonic amplitude list, circulation status level}.

[0056] Step S23: performing MMC bridge arm load abnormality detection based on the MMC bridge arm circulation status to obtain MMC bridge arm load abnormality data;

[0057] In an embodiment of the present invention, after obtaining the bridge arm circulating current status data, the bridge arm circulating current fluctuation characteristics need to be mapped to the actual load change trend of the corresponding bridge arm. By comparing the equivalent load impedance changes of the submodules where the bridge arm is located and combining the instantaneous active power change curve within the sampling period, the load stability is evaluated. The instantaneous current change rate di / dt and voltage change rate du / dt measured in the bridge arm are used to construct a dynamic load response curve, and the instantaneous derivative curve dZ / dt of the load impedance Z = U / I is further solved. During the period when there is obvious circulating current fluctuation in the bridge arm, if the dZ / dt change amplitude is greater than the threshold value of 0.1Ω / ms, it is determined that there is a load anomaly in this period. The load anomaly period is further classified and marked using a time series sliding window to obtain the bridge arm load anomaly status data, which includes: {abnormal bridge arm number, abnormal duration period, maximum instantaneous dZ / dt value, average load amplitude fluctuation rate, and abnormality judgment level}.

[0058] Step S24: performing MMC unit energy abnormality identification according to the MMC bridge arm load abnormality to obtain MMC unit energy abnormality data.

[0059] In an embodiment of the present invention, the load abnormality data output in step S23 is subjected to energy sampling processing of the submodule corresponding to each bridge arm, and the sampling period is set to 20ms. The energy change rate per unit time of all submodules is counted during the abnormal load period to obtain a time series distribution diagram of energy increase and decrease. Comparing the energy change curves of the submodules in adjacent non-abnormal periods, if the energy change rate of a submodule is greater than the upper limit of the normal deviation (such as ±10%), the submodule is marked as having energy abnormal behavior. Further, the energy abnormality data is output in combination with its position number and the corresponding bridge arm number, and the data structure is: {submodule number, bridge arm number, energy change rate, abnormal time interval, abnormal level}. This data serves as a direct input for subsequent circulating current frequency component analysis and power quality analysis to ensure that the data chain is logically rigorous.

[0060] Preferably, step S23 includes the following steps:

[0061] Step S231: performing bridge arm current abnormality superposition detection according to the MMC bridge arm circulation status to obtain bridge arm current abnormality superposition data;

[0062] In an embodiment of the present invention, an abnormal superposition detection of the bridge arm current is performed on the MMC bridge arm circulation condition, and the actual measurement data of the bridge arm output current is selected, which comes from the continuous time series signal recorded by the current sensor installed in the bridge arm. The signal is compared with the standard bridge arm harmonic reference waveform identified in the spectrum amplitude characteristics, and the transient deviation trend is analyzed by time domain superposition processing. In order to enhance the detection accuracy, the comparison process adopts the moving window superposition averaging technology to extract the degree of deviation of the actual current amplitude in the abnormal frequency band. All time periods in which abnormal superposition phenomena occur with the reference harmonic waveform in amplitude, phase or period are marked as abnormal points. The superposition analysis results form bridge arm current abnormal superposition data, which is used to reflect the difference and interference characteristics between the current bridge arm current and the ideal state, and serve as the basic data for subsequent transient fluctuation identification.

[0063] Step S232: determining the transient fluctuation of the bridge arm current according to the bridge arm current abnormality superposition data;

[0064] In an embodiment of the present invention, based on the abnormal superposition data of the bridge arm current, the transient fluctuation changes of the bridge arm current within a specific period range are further extracted. In the processing method, a time series segment of several milliseconds before and after each abnormal superposition event is first selected, and then the short-time Fourier transform technology is applied to convert it into a time-frequency joint view. By detecting the sudden change points of amplitude changes in different frequency bands, a transient fluctuation profile is constructed. In particular, at the position of the frequency component where a sharp increase or decrease occurs, the corresponding time point and peak range are marked, and the transient fluctuation of the bridge arm current is obtained after integration. This data reflects the area where instability or intense energy transfer occurs during the operation of the bridge arm, and can be used to evaluate the risk level of instantaneous current growth later.

[0065] Step S233: identifying the instantaneous increase of the current peak based on the transient fluctuation of the bridge arm current and the superimposed data of the abnormal bridge arm current;

[0066] In an embodiment of the present invention, based on the transient fluctuation of the bridge arm current and the superposition data of the current anomaly, the instantaneous growth of the current peak is identified. The identification method is to select the current amplitude sequence corresponding to the fluctuation time point, calculate the maximum rising rate per unit time, and locate the rapid growth phenomenon by extracting the rate mutation point. In addition, the different bridge arm current peak growth periods are compared and analyzed to identify the distribution range of the corresponding bridge arm inductance structure to determine whether there is a local overload trend. The instantaneous growth of the current peak is obtained by integrated analysis. This data reflects the dynamic severity of the current deviation from the steady-state working area, and is subsequently used to determine the operating status of the filter inductor.

[0067] Step S234: predicting the saturation state of the filter inductor core according to the instantaneous growth of the current peak value;

[0068] In an embodiment of the present invention, the instantaneous growth of the current peak is combined with the material parameters and magnetic permeability characteristic curve of the filter inductor to predict the saturation condition of the filter inductor core. In the specific operation, the current value corresponding to the known inductor current saturation point is selected as the saturation critical value, and the current growth current peak is compared with the critical value for judgment. With the support of the core characteristic test data, the degree of convergence in different amplitude regions is used to determine whether the saturation working area has been entered, and the duration and impact range are further calculated based on the saturation trend. The output filter inductor core saturation condition is a data structure that combines qualitative and quantitative data, marking whether the current operation exceeds the effective linear working range of the magnetic element.

[0069] Step S235: determining the increase in loss of the bridge arm structure according to the instantaneous increase in the current peak value and the saturation condition of the filter inductor core;

[0070] In an embodiment of the present invention, the loss growth of the bridge arm structure is measured based on the instantaneous growth of the current peak and the saturation condition of the filter inductor core. By comparing the copper loss of the bridge arm during normal operation with the difference in the heating rate caused by the increase in current in the current state, combined with the eddy current loss model after the inductor core enters the saturation region, the current loss change trend is inferred based on the current-temperature curve in the historical operation cycle. Based on the heat integration data recorded by the temperature measuring device inside the bridge arm module, the thermal load of key components (such as power components and heat dissipation structures) under the action of current impact is calculated to generate data on the loss growth of the bridge arm structure as a basis for determining whether the load abnormality forms substantial structural stress.

[0071] Step S236: performing MMC bridge arm load abnormality detection based on the bridge arm structure loss growth, and obtaining MMC bridge arm load abnormality data.

[0072] In an embodiment of the present invention, based on the data on the growth of the loss of the bridge arm structure, combined with the frequency and duration of the instantaneous increase of the current peak and the operating threshold data of the bridge arm components, the MMC bridge arm load abnormality condition detection is carried out. During the detection process, the loss growth threshold and the structural tolerance data comparison method are used to determine whether the current structural stress has reached the critical standard that can be regarded as a load abnormality. All time periods that exceed the standard and the corresponding bridge arm position data are sorted out to form the MMC bridge arm load abnormality condition data. The data content includes the abnormal bridge arm number, the start and end of the detection period, and the abnormality type classification (such as overcurrent, magnetic saturation, loss surge, etc.), which serves as an important input for subsequent energy abnormality analysis.

[0073] Preferably, step S24 includes the following steps:

[0074] Step S241: testing the current imbalance condition of adjacent bridge arms according to the abnormal load condition of the MMC bridge arm;

[0075] In an embodiment of the present invention, the current imbalance condition of adjacent bridge arms is tested based on the abnormal load condition of the MMC bridge arm. This operation is based on the abnormal load data of the bridge arm obtained in the previous step, and by comparing the measured current data with the reference value, it is identified whether there is a current imbalance phenomenon. In specific implementation, the current time series is extracted from the current sensor of each bridge arm, the current value of each bridge arm at the same time point is calculated, and the current difference of adjacent bridge arms is analyzed by the standard deviation method. If the current difference between the two bridge arms exceeds the set threshold, it is considered that there is a current imbalance. In order to improve the detection accuracy, low-pass filtering technology is used to denoise the signal during analysis to ensure that the identification of the imbalance condition has a high accuracy. The current imbalance condition data will be used for subsequent analysis of the trend of enhanced circulating current components.

[0076] Step S242: detecting an increasing trend of a bridge arm circulating current component according to an adjacent bridge arm current imbalance condition and an MMC bridge arm load abnormality condition;

[0077] In an embodiment of the present invention, the strengthening trend of the circulating current component of the bridge arm is detected based on the current imbalance status of the adjacent bridge arms and the abnormal load status of the MMC bridge arm. During operation, the strengthening of the circulating current component is analyzed by comparing the current imbalance data of the adjacent bridge arms and the abnormal load data. The unbalanced current is analyzed by Fourier transform to determine the current increment in the characteristic frequency band in its spectrum. If the current amplitude increases significantly in a specific frequency band, it indicates that the circulating current component is strengthening. By analyzing the peak position in the current waveform and combining the abnormal load status, the aggravation trend of the circulating current is evaluated. This trend reflects the phenomenon of increased circulating current in the system. Based on this data, the subsequent positive feedback effect of the circulating current is further predicted.

[0078] Step S243: predicting the positive feedback effect of the bridge arm circulating current according to the strengthening trend of the bridge arm circulating current component and the current imbalance status of adjacent bridge arms;

[0079] In an embodiment of the present invention, the positive feedback effect of the bridge arm circulating current is predicted based on the enhancement trend of the bridge arm circulating current component and the current imbalance condition of the adjacent bridge arm. This step uses the known enhancement trend of the circulating current component and the current imbalance condition to evaluate their potential impact on the stability of the system, and inputs the dynamic data of the current imbalance and the circulation enhancement into the prediction algorithm, which establishes a circulation positive feedback model based on historical operation data. The model determines whether the circulation enters a positive feedback state by identifying the nonlinear change trend of the current waveform when the circulation increases. The positive feedback effect of the circulation is usually manifested as a continuous growth or loss of control of the circulation, leading to problems such as overheating of the equipment and increased losses. In this step, the feedback effect of the circulation is further simulated using the time series analysis method to form a specific prediction result.

[0080] Step S244: identifying the invalid energy circulation condition inside the bridge arm based on the positive feedback effect of the bridge arm circulation;

[0081] In an embodiment of the present invention, based on the positive feedback effect of the bridge arm circulation, the invalid energy circulation status inside the bridge arm is identified. During the identification process, the positive feedback effect data of the circulation obtained in step S243 is used to analyze the invalid energy area in the current waveform to identify the time period in which the invalid energy cycle exists. By detecting the power factor of the current waveform, it is further determined whether the electric energy is effectively converted into output power. If the power factor is low and the phase lag of the current is obvious, it indicates that an invalid energy cycle has occurred. The internal energy invalid cycle is accurately located by using the time domain analysis method, combining the real-time data of the current waveform with the preset invalid cycle standard. This data is used to evaluate the system efficiency loss and the occurrence of faults.

[0082] Step S245: predicting the electromagnetic interference enhancement of the bridge arm operation according to the invalid energy circulation status inside the bridge arm;

[0083] In an embodiment of the present invention, the enhanced electromagnetic interference during the operation of the bridge arm is predicted based on the invalid energy circulation status within the bridge arm. In this step, the current waveform is Fourier transformed to analyze the amplitude changes of different frequency components to determine whether there is a trend of increasing electromagnetic interference. In particular, the amplitude changes in the low-frequency and high-frequency regions are related to the invalid energy circulation status. If the amplitude of the interference signal increases significantly within a specific frequency band, it indicates that the electromagnetic interference has increased. Based on the interference frequency characteristics, its impact on the components is further evaluated. This predicted data helps determine whether the electromagnetic interference will interfere with the normal operation of the equipment and provides a basis for subsequent abnormality identification.

[0084] Step S246: determining abnormality of the bridge arm power exchange efficiency based on the enhanced electromagnetic interference during bridge arm operation;

[0085] In an embodiment of the present invention, based on the enhanced electromagnetic interference during the operation of the bridge arm, the abnormality of the bridge arm's battery exchange efficiency is determined. Changes in battery exchange efficiency are usually closely related to electromagnetic interference, because electromagnetic interference can cause instability or reduction in power transmission. Therefore, in this step, the electromagnetic interference signal is analyzed to identify its frequency components and amplitude changes, and then the electromagnetic interference data is compared with the actual power conversion efficiency during the bridge arm's battery exchange process. If a significant deviation is found in the relationship between electromagnetic interference and battery exchange efficiency, and the interference amplitude increases, it means that the battery exchange efficiency has become abnormal. This data reflects potential problems in the energy conversion process within the system and is used for subsequent energy anomaly identification.

[0086] Step S247: Identify MMC unit energy anomaly based on the abnormality of the bridge arm battery replacement efficiency and the enhanced electromagnetic interference of the bridge arm operation to obtain MMC unit energy anomaly data.

[0087] In an embodiment of the present invention, based on the abnormality of the bridge arm battery exchange efficiency and the enhanced electromagnetic interference during the operation of the bridge arm, the MMC unit energy anomaly identification is performed to obtain the MMC unit energy anomaly data, and the battery exchange efficiency abnormality data in step S246 is combined with the electromagnetic interference enhanced data to calculate the total energy output and input of the MMC unit through the energy balance algorithm. If there is a significant mismatch between the energy output and input, and there is an obvious trend of electromagnetic interference and battery exchange efficiency abnormality, it is determined that the unit has energy anomaly. By calculating the deviation between the energy loss rate and the output power, the energy anomaly data of the MMC unit is identified, and this data provides a basis for further maintenance and optimization.

[0088] Preferably, step S3 includes the following steps:

[0089] Step S31: determining the circulating current frequency component based on the abnormal energy condition of the MMC unit to obtain the circulating current frequency component condition;

[0090] In an embodiment of the present invention, based on the obtained MMC unit energy anomaly data, the energy fluctuation sequence of each MMC submodule is extracted, and a synchronous waveform curve of the submodule output voltage and the submodule energy storage capacitor voltage is constructed based on time. The energy fluctuation sequence is subjected to fast Fourier transform processing using a time series analysis method to extract the main frequency component and subharmonic component contained therein. A frequency domain analysis tool, such as a spectrum analyzer equipped with a multi-resolution analysis module, is used to perform segmented spectrum analysis on the extracted data, and the energy distribution in each frequency band is mapped to the frequency axis to obtain the circulating current frequency component contained in the bridge arm current and coupled with the energy fluctuation. With 3kHz as the lower limit and 50kHz as the upper limit, a frequency band window is constructed to confirm the dominant frequency and power density of the circulating current in each frequency band, forming circulating current frequency component situation data as the data basis for subsequent power quality analysis and traceability analysis.

[0091] Step S32: performing MMC output power quality attenuation detection according to the circulating current frequency component to obtain the MMC output power quality attenuation;

[0092] In an embodiment of the present invention, the circulating current frequency component data obtained in step S31 is compared with the real-time voltage and current waveforms at the output end of the MMC system. The output power indicators under normal working conditions are compared by using the harmonic distortion analysis method, and the harmonic voltage content (THDv) and harmonic current content (THDi) are used as the power quality evaluation benchmark. Combined with the three-phase output voltage waveform collected in the power quality monitoring platform, the voltage distortion rate at different frequencies is compared item by item. The voltage distortion rate threshold is set to 5%, and the current distortion rate threshold is set to 8%. The frequency component exceeding the threshold is determined to be the power quality attenuation source caused by the circulating current. The wavelet packet decomposition method is used to further confirm the concentrated frequency band of the harmonic interference, and the standard deviation analysis is performed on the power output power fluctuation characteristics to comprehensively form the MMC output power quality attenuation data, and a dynamic attenuation trend chart is constructed with the time axis as the horizontal axis for subsequent source positioning.

[0093] Step S33: performing MMC initial circulation position tracing processing according to the MMC output power quality attenuation to obtain MMC initial circulation position tracing data;

[0094] In an embodiment of the present invention, after obtaining the power quality attenuation data, the position tracking processing of the three-phase asymmetry rate of the voltage is performed according to the three-phase imbalance analysis standard of the output end. By comparing the time series of the bridge arm current and the bus voltage, the time and position of the power reversal in each bridge arm unit are identified. Using the data tracing algorithm, the current corresponding to the distortion peak is transmitted back to the corresponding sub-module unit number, and combined with the bridge arm topology diagram, the physical position of the initial circulation path is locked. Through the synchronous voltage and current waveforms obtained by the unit status monitoring sensor arranged in the MMC bridge arm, the direction of the circulating energy feedback at the current moment is judged, and the error term is calibrated in combination with the energy abnormality data to achieve accurate positioning of the initial circulating current excitation unit. The MMC initial circulating current position traceability data including the initial position number, timestamp, and bridge arm number is formed.

[0095] Step S34: determining the generation mechanism of the bridge arm circulation current according to the MMC initial circulation position tracing data and the circulation frequency component, and obtaining the bridge arm circulation current generation mechanism data.

[0096] In an embodiment of the present invention, the initial circulation position tracing data obtained in step S33 is correlated with the circulation frequency component formed in step S31, and the generation path and formation mechanism of each circulation component in the bridge arm are clarified by constructing a ternary causal mapping matrix of "starting position-frequency characteristics-energy propagation direction". The frequency-space mapping technology is used to convert the frequency component into a power density trajectory on the propagation path in the corresponding bridge arm, and the topological channel in which the circulation is reversely conducted along the sub-module or circulates and feeds back between multiple bridge arms is identified. A multi-channel time-domain convolution analysis method is used to track the timing offset of the circulation energy between different bridge arms, and the energy accumulation point and release point in a short time are identified by combining the capacitor voltage drop slope and the inductor magnetic saturation trend, forming the bridge arm circulation generation mechanism data, including multiple dimensional information such as the trigger source position, propagation path, frequency range, load coupling mode, feedback channel structure, and instantaneous energy conversion characteristics, providing complete theoretical support for the circulation elimination path planning of the subsequent modulation strategy.

[0097] Preferably, step S32 includes the following steps:

[0098] Step S321: detecting the instability of the circulating current component inside the bridge arm according to the circulating current frequency component;

[0099] In this embodiment of the present invention, the circulating current frequency component data is obtained in step S31. The circulating current frequency component data includes the low-frequency common-mode components, higher-order harmonic components, and frequency stability parameters of each of the three-phase bridge arms. This data is subjected to differential frequency tracking analysis, and a short-time Fourier transform method is used to perform sliding window amplitude statistics on the low-frequency components and specific high-order frequency components in the spectrum. A spectrum stability evaluation matrix is constructed by combining the mean square error of the amplitude changes at each frequency point. Frequency component amplitudes that fluctuate beyond a set threshold within a specific period are marked in the matrix to form a sequence of unstable circulating current frequency component labels within the bridge arm, which are then output as the internal circulating current instability status of the bridge arm.

[0100] Step S322: determining the broadening of the output spectrum of the bridge arm based on the instability of the circulating current component inside the bridge arm;

[0101] In this embodiment of the present invention, the arm internal circulating current instability flag sequence obtained in step S321 is combined with the arm output current waveform data during the current operating cycle to perform high-resolution spectrum reconstruction of the arm output signal. Wavelet packet decomposition is then used to perform partitioned statistics on the output current spectral energy. Using the center of the stable frequency band as a reference, a quantitative assessment is performed on the energy center offset, the rate of change of the spectrum bandwidth, and the noise floor variation trend. By constructing a spectrum broadening coefficient index matrix, spectrum broadening status data corresponding to each arm is output, which serves as the basis for subsequent bridge arm waveform distortion detection.

[0102] Step S323: testing the waveform distortion condition of the bridge arm output end using the bridge arm output spectrum broadening condition;

[0103] In this embodiment of the present invention, based on the spectrum broadening coefficient output in step S322, the time-domain waveforms of the bridge arm output voltage and output current are subjected to corresponding Fourier transform and harmonic distortion (THD) analysis. The voltage and current signals of each bridge arm in the current cycle are expanded into the fundamental wave and harmonic components of each order. The energy contribution of each order of distortion is determined by combining the spectrum broadening bandwidth and energy weight ratio, and expressed as a waveform distortion index. The severity of the distortion at the bridge arm output is determined based on the absolute value and change trend of the waveform distortion index, and the bridge arm output waveform distortion status data is obtained.

[0104] Step S324: predicting the cumulative data of heat loss output by the bridge arm based on the waveform distortion condition of the bridge arm output end;

[0105] In this embodiment of the present invention, the amplitude and duration of each harmonic order of the bridge arm output current are obtained based on the waveform distortion status data output in step S323. Heat loss is estimated according to the thermal power equivalence principle, with reference to the thermal power consumption characteristic curves of the actual bridge arm main switching devices and filters. Specifically, the additional losses of each order of current harmonics in the conductive devices, filter inductors, and busbars are superimposed, and the thermal power per cycle is integrated and accumulated to form a bridge arm output heat loss accumulation curve. This accumulated bridge arm output heat loss data is output to provide a heat loss basis for subsequent grid stability assessment.

[0106] Step S325: Evaluate the attenuation trend of the bridge arm grid-connected stability based on the accumulated data of the bridge arm output heat loss;

[0107] In this embodiment of the present invention, the accumulated heat loss data output in step S324 is compared with the heat loss benchmark for the initial phase of the bridge arm grid-connected operation. Combined with the real-time recorded data on voltage and current phase stability and the impact of harmonic currents on bus voltage fluctuations, a bridge arm grid-connected stability trend assessment model is constructed. The slope of the heat loss curve is extracted using a time series fitting method, and the rates of change of the phase and frequency perturbation amplitudes are superimposed. The cumulative offsets over multiple sampling periods are compared and analyzed, and the output is the bridge arm grid-connected stability attenuation trend indicator data, including the power factor attenuation rate, bus voltage jitter frequency, and phase lock stability change curve.

[0108] Step S326: Based on the attenuation trend of the bridge arm grid-connected stability and the instability of the internal circulating current component of the bridge arm, the MMC output power quality attenuation detection is performed to obtain the MMC output power quality attenuation status.

[0109] In this embodiment of the present invention, the bridge arm circulating current frequency component instability flag output from step S321 and the bridge arm grid-connected stability attenuation trend data output from step S325 are combined to construct a power quality status assessment system based on multi-source data fusion. The frequency instability flag is logically superimposed with the grid-connected power factor fluctuation rate, and a weighted decision-making method is used to determine whether there is a persistent output quality degradation phenomenon. The output result is represented by the MMC output power quality attenuation level identification, which includes four status categories: stable, mild attenuation, moderate attenuation, and severe attenuation. It also includes the influencing factor data for each bridge arm and the attenuation trigger cause identification. The output is the MMC output power quality attenuation status.

[0110] Preferably, step S4 includes the following steps:

[0111] Step S41: Calculating the probability of circulation generation based on the bridge arm circulation generation mechanism data to obtain bridge arm circulation generation probability data;

[0112] In an embodiment of the present invention, quantitative discretization processing is performed based on the data of the bridge arm circulation generation mechanism. The frequency component combination data, initial circulation excitation position data and instantaneous phase information of the grid-connected node in the bridge arm circulation generation mechanism data are called to dynamically reconstruct the bridge arm circulation excitation logic for the entire cycle, and the time domain sliding window method is used to attribute and mark the circulation excitation factors within multiple control cycles. Subsequently, the number of occurrences of each type of circulation excitation factor is normalized and counted within a unit time, and a multivariate matching analysis is performed in combination with the bridge arm operation spectrum within a specific time. The probability density histogram estimation technology is used to accumulate the number of bridge arm circulations caused by each excitation factor to generate circulation generation distribution data. In this process, the FFT spectrum extraction module is used to obtain the bridge arm operation frequency spectrum, and the excitation conditions are dynamically tracked based on the time domain sampling rate of 5kHz. The matching threshold filtering method is used to eliminate irrelevant disturbance data to ensure that the circulation probability assessment data is only linked to factors significantly correlated with physical characteristics. The triggering times of each circulation excitation factor are statistically normalized to generate the probability data of bridge arm circulation generation. The unit is the average number of triggering times per minute. The data structure is encapsulated in JSON structure for subsequent calls.

[0113] Step S42: assessing the abnormal risk of the bridge arm circulation based on the bridge arm circulation generation probability data;

[0114] In an embodiment of the present invention, the assessment of the abnormal risk of the bridge arm circulation is based on the bridge arm circulation generation probability data output in step S41. This process extracts the excitation path data that is higher than the set abnormal threshold (for example, the trigger frequency is greater than 10 times / minute) in the aforementioned probability data, and constructs a risk weight matrix in combination with the bridge arm structure type, grid-connected capacity level and the number of circulation coupling channels. The weight factors include operating constraint factors such as the bridge arm inductance type, the commutation unit topology and the commutation instruction cycle. A multi-factor weighted accumulation algorithm is used to perform point multiplication on the trigger probability of each circulation excitation path and its corresponding weight factor, and then sum them up to output the risk intensity index of the bridge arm within the unit cycle. The risk intensity data is quantitatively expressed in percentage, and each bridge arm generates a set of independent risk data. The data structure is indexed by the bridge arm number and includes the trigger probability, risk weighting index and cumulative risk coefficient. The output bridge arm circulation abnormal risk situation data is provided as input to the next step for modulation parameter construction.

[0115] Step S43: constructing MMC carrier modulation parameters based on the abnormal risk of the bridge arm circulating current and the data on the generation mechanism of the bridge arm circulating current;

[0116] In an embodiment of the present invention, the construction of MMC carrier modulation parameters is based on the abnormal risk situation of the bridge arm circulation and the bridge arm circulation generation mechanism data. The main control frequency component, excitation phase information and coupling submodule number information in the circulation excitation mechanism are called, and the frequency range of the carrier modulation band is reconstructed in combination with the high-risk trigger path identified in the risk situation. This process uses a bandpass filter design module to shield all identified abnormal frequency components in the excitation frequency band in the frequency domain, and sets the frequency shift factor to adjust the carrier modulation base frequency to avoid overlap with the abnormal circulation frequency. Subsequently, a multi-bridge arm modulation coordination algorithm is used to set independent modulation parameters for each bridge arm, including modulation depth factor, modulation offset angle and phase shift between bridge arms. All parameters are uniformly encapsulated into a modulation parameter table, and the table structure includes bridge arm number, base frequency adjustment value, modulation depth proportional coefficient and suppression offset angle. The generated MMC carrier modulation parameters are output in matrix form.

[0117] Step S44: constructing an MMC suppressor modulation matrix according to the MMC carrier modulation parameters;

[0118] In this embodiment of the present invention, an MMC suppressor modulation matrix is constructed based on the MMC carrier modulation parameters from step S43. During the construction process, the bridge arm number is used as the row index and the modulation parameter is used as the column index. The modulation matrix is initialized as an empty matrix. Subsequently, the modulation depth factor is mapped to a normalized value interval between 0 and 1 to control the carrier amplitude dynamic coefficient within each modulation cycle; the modulation offset angle is mapped to a range of 0 to 360 degrees to serve as the phase offset of the sinusoidal modulation signal. Independent modulation function expression instructions are then generated for each bridge arm. An embedded modulation module is used to align the carrier signals of each bridge arm with the system clock reference, and an inter-bridge arm coupling modulation control strategy is introduced. This strategy achieves adaptive phase compensation between multiple bridge arms by reading the modulation parameters of adjacent bridge arms, avoiding resonant modulation caused by phase overlap. The resulting MMC suppressor modulation matrix structure is a three-dimensional matrix, with dimensions corresponding to the bridge arm number, modulation time, and modulation factor. The matrix data is output at the hardware level as a multi-channel PWM modulation sequence.

[0119] Step S45: Generate an adaptive circulating current suppressor carrier modulation strategy according to the MMC suppressor modulation matrix to obtain adaptive circulating current suppressor carrier modulation strategy data.

[0120] In an embodiment of the present invention, an adaptive circulating current suppressor carrier modulation strategy is generated based on the MMC suppressor modulation matrix. This process extracts the modulation matrix moment by moment and dynamically adjusts the modulation weights based on the current bridge arm operating state (including temperature, voltage, current, etc.). A dynamic modulation parameter loading mechanism is used to load the control variables in the suppression matrix into the carrier modulation control module in real time, enabling periodic modulation curve reconstruction. Real-time monitoring data of the bridge arm output spectrum is read and compared with historical circulating current frequency data to identify the effectiveness of frequency domain modulation. Based on this, a feedback control channel is constructed to dynamically modify the modulation matrix based on the real-time bridge arm operating state, ensuring that the modulation strategy maintains minimum coupling with abnormal frequency components during the operating cycle. The constructed adaptive circulating current suppressor carrier modulation strategy data is output in the form of a control instruction table. Each instruction includes the bridge arm number, modulation mode number, current carrier base frequency, modulation phase angle, and modulation amplitude control signal. This strategy can be directly loaded into the FPGA modulation control module of the MMC control system to drive the generation of modulation signals for each bridge arm.

[0121] The present invention also provides a carrier modulation system for an MMC circulating current suppressor, which is used to execute the carrier modulation method for the MMC circulating current suppressor described above. The carrier modulation system for the MMC circulating current suppressor includes:

[0122] The spectrum feature extraction module is used to obtain MMC operation log data; perform spectrum analysis based on the MMC operation log data to obtain MMC operation bridge arm current spectrum data; and extract MMC operation spectrum amplitude distribution characteristics based on the MMC operation bridge arm current spectrum data;

[0123] The energy anomaly determination module is used to determine the MMC bridge arm circulation condition based on the MMC operation spectrum amplitude distribution characteristics; detect the MMC bridge arm load anomaly based on the MMC bridge arm circulation condition; and determine the MMC unit energy anomaly based on the MMC bridge arm load anomaly;

[0124] The circulating current mechanism identification module is used to determine the circulating current frequency component based on the MMC unit energy anomaly; detect the MMC output power quality degradation based on the circulating current frequency component; perform MMC initial circulating current position tracing based on the MMC output power quality degradation to obtain MMC initial circulating current position tracing data; determine the bridge arm circulating current generation mechanism based on the MMC initial circulating current position tracing data to obtain bridge arm circulating current generation mechanism data;

[0125] The modulation strategy generation module is used to evaluate the abnormal risk of bridge arm circulation current based on the data of the bridge arm circulation current generation mechanism; construct the MMC carrier modulation parameters based on the abnormal risk of bridge arm circulation current and the data of the bridge arm circulation current generation mechanism; generate the adaptive circulation suppressor carrier modulation strategy based on the MMC carrier modulation parameters to obtain the adaptive circulation suppressor carrier modulation strategy data.

[0126] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any equivalent modifications or variations made by persons skilled in the art without departing from the spirit and technical concepts disclosed herein shall be encompassed by the claims of the present invention.

Claims

1. A carrier modulation method for an MMC circulating current suppressor, characterized in that: The following steps are included Step S1: obtaining MMC operation log data; performing spectrum analysis based on the MMC operation log data to obtain MMC operation bridge arm current spectrum data; extracting MMC operation spectrum amplitude distribution characteristics based on the MMC operation bridge arm current spectrum data; Step S2: determining the MMC arm circulation condition according to the MMC operation spectrum amplitude distribution characteristics; detecting the MMC arm load abnormality based on the MMC arm circulation condition; and determining the MMC unit energy abnormality according to the MMC arm load abnormality. Step S3: determining the circulating current frequency component based on the MMC unit energy anomaly; detecting the MMC output power quality attenuation based on the circulating current frequency component; performing MMC initial circulating current position tracing processing based on the MMC output power quality attenuation to obtain MMC initial circulating current position tracing data; determining the bridge arm circulating current generation mechanism based on the MMC initial circulating current position tracing data to obtain bridge arm circulating current generation mechanism data; Step S4: Assessing the abnormal risk of the bridge arm circulation based on the bridge arm circulation generation mechanism data; The MMC carrier modulation parameters are constructed based on the abnormal risk situation of the bridge arm circulation current and the bridge arm circulation current generation mechanism data; the adaptive circulation suppressor carrier modulation strategy is generated according to the MMC carrier modulation parameters to obtain the adaptive circulation suppressor carrier modulation strategy data.

2. The carrier modulation method of the MMC circulating current suppressor according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: Obtain MMC operation log data; Step S12: collecting the MMC bridge arm current operating status according to the MMC operation log data; Step S13: performing spectrum analysis on the MMC operating arm current operating status to obtain MMC operating arm current spectrum data; Step S14: extracting the MMC operation spectrum amplitude distribution characteristics according to the MMC operation bridge arm current spectrum data.

3. The carrier modulation method of the MMC circulating current suppressor according to claim 2, characterized in that: Step S13 includes the following steps: Step S131: performing a current time series analysis on the operating status of the MMC operating bridge arm current to obtain the MMC operating bridge arm current time series data, specifically: using a high-speed sampling current sensor deployed at each bridge arm busbar to continuously collect the AC current signal of each bridge arm at a sampling frequency of more than 10 kHz, and setting the collection time window to 10 seconds, covering the complete low-frequency, medium-frequency and harmonic signal segments; the collected current signal is converted into a digital signal by a high-precision analog-to-digital conversion module, and transmitted to the data processing module; the sampled data is first subjected to noise preprocessing, and a bandpass filter is used to perform linear detrending processing to eliminate long-period background changes and retain stable periodic signals, thereby generating the MMC operating bridge arm current time series data; Step S132: performing a current time series fluctuation analysis based on the MMC operation bridge arm current time series data to obtain the bridge arm current time series fluctuation data; specifically, using a sliding window variance analysis method to evaluate the current fluctuation based on the MMC operation bridge arm current time series data, setting the sliding window width to 100ms and the step length to 10ms, performing local statistics on each bridge arm current time series segment, calculating the current mean and standard deviation within the window, calculating the mean difference and variance change rate between each time window and the previous window, and constructing a first-order difference sequence and a second-order difference sequence, extracting the time series fluctuation points, and constructing a volatility index matrix to obtain the bridge arm current time series fluctuation data; Step S133: performing statistics on periodic changes of the bridge arm current according to the bridge arm current time series fluctuation data to obtain bridge arm current periodic change data; Step S134: measuring the bridge arm current fluctuation amplitude based on the bridge arm current time series fluctuation data to obtain the bridge arm current fluctuation amplitude data; Step S135: Perform spectrum analysis based on the periodic change data of the bridge arm current and the bridge arm current fluctuation amplitude data to obtain the MMC operation bridge arm current spectrum data; specifically: use high-resolution short-time Fourier transform, select the Hamming window as the window function, set the window length to 256ms, and the overlap rate to 50% to construct a time-frequency distribution diagram. Based on the periodicity-guided frequency band weighting mechanism, weight the periodically prominent frequency band to increase the spectrum energy threshold and reduce background noise interference. The target frequency range of the spectrum analysis is set to 0Hz to 5kHz, and the amplitude-frequency distribution curve and energy density spectrum of the bridge arm current signal within this frequency range are output to constitute the MMC operation bridge arm current spectrum data.

4. The carrier modulation method of the MMC circulating current suppressor according to claim 1, characterized in that: Step S2 includes the following steps: Step S21: detecting the MMC circulating current characteristic harmonic condition according to the MMC operation spectrum amplitude distribution characteristics; Step S22: determining the MMC bridge arm circulating current status according to the MMC operation spectrum amplitude distribution characteristics and the MMC circulating current characteristic harmonic status; Step S23: performing MMC bridge arm load abnormality detection based on the MMC bridge arm circulation status to obtain MMC bridge arm load abnormality data; Step S24: performing MMC unit energy abnormality identification according to the MMC bridge arm load abnormality to obtain MMC unit energy abnormality data.

5. The carrier modulation method of the MMC circulating current suppressor according to claim 4, characterized in that: Step S23 includes the following steps: Step S231: performing bridge arm current abnormality superposition detection according to the MMC bridge arm circulation status to obtain bridge arm current abnormality superposition data; Step S232: determining the transient fluctuation of the bridge arm current according to the bridge arm current abnormality superposition data; Step S233: identifying the instantaneous increase of the current peak based on the transient fluctuation of the bridge arm current and the superimposed data of the abnormal bridge arm current; Step S234: predicting the saturation state of the filter inductor core according to the instantaneous growth of the current peak value; Step S235: determining the increase in loss of the bridge arm structure according to the instantaneous increase in the current peak value and the saturation condition of the filter inductor core; Step S236: performing MMC bridge arm load abnormality detection based on the bridge arm structure loss growth, and obtaining MMC bridge arm load abnormality data.

6. The carrier modulation method of the MMC circulating current suppressor according to claim 4, characterized in that: Step S24 includes the following steps: Step S241: testing the current imbalance condition of adjacent bridge arms according to the abnormal load condition of the MMC bridge arm; Step S242: detecting an increasing trend of a bridge arm circulating current component according to an adjacent bridge arm current imbalance condition and an abnormal MMC bridge arm load condition; Step S243: predicting the positive feedback effect of the bridge arm circulating current according to the strengthening trend of the bridge arm circulating current component and the current imbalance status of adjacent bridge arms; Step S244: identifying the invalid energy circulation condition inside the bridge arm based on the positive feedback effect of the bridge arm circulation; Step S245: predicting the electromagnetic interference enhancement of the bridge arm operation according to the invalid energy circulation status inside the bridge arm; Step S246: determining abnormality of the bridge arm power exchange efficiency based on the enhanced electromagnetic interference during bridge arm operation; Step S247: Identify MMC unit energy anomaly based on the abnormality of the bridge arm battery replacement efficiency and the enhanced electromagnetic interference of the bridge arm operation to obtain MMC unit energy anomaly data.

7. The carrier modulation method of the MMC circulating current suppressor according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: determining the circulating current frequency component based on the abnormal energy condition of the MMC unit to obtain the circulating current frequency component condition; Step S32: performing MMC output power quality attenuation detection according to the circulating current frequency component to obtain the MMC output power quality attenuation; Step S33: performing MMC initial circulation position tracing processing according to the MMC output power quality attenuation to obtain MMC initial circulation position tracing data; Step S34: determining the generation mechanism of the bridge arm circulation current according to the MMC initial circulation position tracing data and the circulation frequency component, and obtaining the bridge arm circulation current generation mechanism data.

8. The carrier modulation method of the MMC circulating current suppressor according to claim 7, characterized in that: Step S32 includes the following steps: Step S321: detecting the instability of the circulating current component inside the bridge arm according to the circulating current frequency component; Step S322: determining the broadening of the output spectrum of the bridge arm based on the instability of the circulating current component inside the bridge arm; Step S323: testing the waveform distortion condition of the bridge arm output end using the bridge arm output spectrum broadening condition; Step S324: predicting the cumulative data of heat loss output by the bridge arm based on the waveform distortion condition of the bridge arm output end; Step S325: Evaluate the attenuation trend of the bridge arm grid-connected stability based on the accumulated data of the bridge arm output heat loss; Step S326: Based on the attenuation trend of the bridge arm grid-connected stability and the instability of the internal circulating current component of the bridge arm, the MMC output power quality attenuation detection is performed to obtain the MMC output power quality attenuation status.

9. The carrier modulation method of the MMC circulating current suppressor according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: Calculating the probability of circulation generation based on the bridge arm circulation generation mechanism data to obtain bridge arm circulation generation probability data; Step S42: assessing the abnormal risk of the bridge arm circulation based on the bridge arm circulation generation probability data; Step S43: constructing MMC carrier modulation parameters based on the abnormal risk of the bridge arm circulating current and the data on the generation mechanism of the bridge arm circulating current; Step S44: constructing an MMC suppressor modulation matrix according to the MMC carrier modulation parameters; Step S45: Generate an adaptive circulating current suppressor carrier modulation strategy according to the MMC suppressor modulation matrix to obtain adaptive circulating current suppressor carrier modulation strategy data.

10. A carrier modulation system for an MMC circulating current suppressor, characterized in that: For executing the carrier modulation method of the MMC circulating current suppressor according to claim 1, the carrier modulation system of the MMC circulating current suppressor comprises: The spectrum feature extraction module is used to obtain MMC operation log data; perform spectrum analysis based on the MMC operation log data to obtain MMC operation bridge arm current spectrum data; and extract MMC operation spectrum amplitude distribution characteristics based on the MMC operation bridge arm current spectrum data; The energy anomaly determination module is used to determine the MMC bridge arm circulation condition based on the MMC operation spectrum amplitude distribution characteristics; detect the MMC bridge arm load anomaly based on the MMC bridge arm circulation condition; and determine the MMC unit energy anomaly based on the MMC bridge arm load anomaly; The circulating current mechanism identification module is used to determine the circulating current frequency component based on the MMC unit energy anomaly; detect the MMC output power quality degradation based on the circulating current frequency component; perform MMC initial circulating current position tracing based on the MMC output power quality degradation to obtain MMC initial circulating current position tracing data; determine the bridge arm circulating current generation mechanism based on the MMC initial circulating current position tracing data to obtain bridge arm circulating current generation mechanism data; The modulation strategy generation module is used to evaluate the abnormal risk of bridge arm circulation current based on the data of the bridge arm circulation current generation mechanism; construct the MMC carrier modulation parameters based on the abnormal risk of bridge arm circulation current and the data of the bridge arm circulation current generation mechanism; generate the adaptive circulation suppressor carrier modulation strategy based on the MMC carrier modulation parameters to obtain the adaptive circulation suppressor carrier modulation strategy data.

Images