A micro-grid fault diagnosis method and diagnosis system

Abstract

The application discloses a micro-grid fault diagnosis method and a diagnosis system, relates to the technical field of power grid fault diagnosis, and is used for improving the accuracy of diagnosing high-resistance faults of a micro-grid. In the method, a gate driver of an existing micro-grid inverter is used as a signal collection source, and no additional sensor needs to be additionally arranged, so that interference of high-frequency background noise of the micro-grid on fault feature extraction is avoided from a signal collection link. A miller platform time is used as a heat-sensitive quantity of a power switch device junction temperature, a three-phase low-frequency power loss fluctuation sequence is obtained through envelope extraction and transient thermal impedance network model inversion, and a negative sequence loss amplitude is extracted by using a symmetrical component decomposition, so that asymmetric detection of three-phase electrical quantities is converted into detection of internal device heat state differences of the inverter. In a fault determination link, a double-condition joint determination of the negative sequence loss amplitude and an intermittent fluctuation characteristic value is introduced, so that a high-resistance fault and an ordinary load imbalance can be distinguished, and a phase of the high-resistance fault occurrence can be located.

Description

Technical Field

[0001] This application relates to the field of power grid fault diagnosis technology, and in particular to a microgrid fault diagnosis method and system. Background Technology

[0002] With the large-scale grid connection of distributed renewable energy into modern power systems, local microgrids have become an important unit for supporting the safe operation of the main grid and improving the grid fault defense architecture on the distribution side. In the actual operating environment of microgrids, distribution lines often face complex conditions such as tree branches touching the ground or conductors breaking and falling into sand or mud, which can lead to high-resistance faults and pose risks to the overall power grid's power supply stability and fault defense system.

[0003] To detect high-resistivity faults in microgrids, relevant technologies primarily rely on installing high-precision voltage and current transformers or miniature phasor measurement units at the grid-side line nodes to directly acquire macroscopic voltage and current signals from the grid side. After acquiring these macroscopic electrical quantities, these technologies typically employ signal processing algorithms (such as wavelet transform) to extract high-frequency transient harmonic components, zero-sequence current components, or waveform distortion features from the voltage and current signals. These macroscopic electrical features are then input into a pre-trained pattern recognition classifier, and the presence of a high-resistivity fault is determined by comparing the feature data.

[0004] However, with the increasing proportion of power electronic inverters connected to the grid in microgrids, the background harmonics and high-frequency electromagnetic noise in microgrids have increased significantly. Since the weak transient characteristics of high-resistance faults and the broadband noise generated by the high-frequency switching action of inverters are highly overlapping in the frequency domain, the high-precision current transformers of related technologies will introduce strong background noise when acquiring signals, resulting in low accuracy in diagnosing high-resistance faults in microgrids. Summary of the Invention

[0005] This application provides a microgrid fault diagnosis method and system to improve the accuracy of diagnosing high-resistance faults in microgrids.

[0006] Firstly, a microgrid fault diagnosis method is provided, applied to a diagnostic system. This method includes: real-time acquisition of the gate-source voltage signal of the three-phase bridge arm power switching devices during the conduction transient process via the gate driver of the microgrid inverter; calculation of the duration of the gate-source voltage signal within a preset Miller voltage range, generating a three-phase Miller plateau time series synchronized with the inverter switching frequency, where the preset Miller voltage range is the range of gate-source voltage when the power switching devices are in the Miller effect stage; envelope extraction processing of the three-phase Miller plateau time series, filtering out high-frequency components corresponding to the switching frequency to obtain a low-frequency pulsating envelope with a frequency twice the fundamental frequency of the microgrid, where the low-frequency pulsating envelope characterizes the alternating fluctuations in the junction temperature of the power switching devices caused by asymmetrical operating conditions of the microgrid; and based on a preset device thermal... Using a thermosensitive coefficient and a preset transient thermal impedance network model, the low-frequency pulsating envelope of the three phases is inverted and analyzed into a three-phase low-frequency power loss fluctuation sequence. The preset device thermosensitive coefficient defines the linear mapping relationship between the junction temperature change of the power switching device and the time change of the Miller plateau. Symmetric component decomposition is performed on the three-phase low-frequency power loss fluctuation sequence to extract the negative sequence loss amplitude, and the time derivative of the negative sequence loss amplitude within a preset time window is calculated to obtain the intermittent fluctuation characteristic value characterizing the intermittentity of the fault arc. When the negative sequence loss amplitude is greater than the preset unbalance reference threshold and the intermittent fluctuation characteristic value is greater than the preset arc distortion threshold, a high-resistance fault is determined to have occurred in the microgrid. The phase in which the high-resistance fault occurs is determined based on the target phase with the largest amplitude change in the three-phase low-frequency power loss fluctuation sequence.

[0007] By adopting the above technical solution, the existing gate driver of the microgrid inverter is used as the signal acquisition source, eliminating the need for additional sensors on the grid side and avoiding interference from high-frequency background noise in the microgrid for fault feature extraction. Using Miller plateau time as the thermally sensitive quantity of the junction temperature of the power switching devices, the three-phase low-frequency power loss fluctuation sequence is obtained through envelope extraction and transient thermal impedance network model inversion. Then, symmetrical component decomposition is used to extract the negative sequence loss amplitude, transforming the detection of three-phase electrical quantity asymmetry into the detection of differences in the thermal state of internal inverter components. In the fault determination stage, a dual-condition joint determination based on the negative sequence loss amplitude and intermittent fluctuation characteristic value is introduced. This not only distinguishes between high-resistance faults and ordinary load imbalances but also locates the phase where the high-resistance fault occurs, improving the accuracy of high-resistance fault diagnosis in microgrids.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, the step of calculating the duration of the gate-source voltage signal within a preset Miller voltage range and generating a three-phase Miller plateau time series synchronized with the inverter switching frequency specifically includes: calculating the voltage change rate of the gate-source voltage signal between adjacent sampling points to obtain a voltage change rate sequence; identifying continuous time periods in the voltage change rate sequence where the absolute value of the voltage change rate is lower than a preset smoothing threshold; using the start and end times of the continuous time periods as the boundary points for the power switching devices to enter and exit the Miller effect stage, respectively; calculating the time difference between the start and end times, and splicing the time differences corresponding to each switching cycle in chronological order to generate a three-phase Miller plateau time series.

[0009] By adopting the above technical solution, the boundary of the Miller effect stage is identified based on the rate of change sequence between adjacent sampling points of the gate-source voltage. Compared with the direct voltage threshold judgment method, it has a certain suppression effect on the instantaneous voltage spikes introduced by electromagnetic interference or switching oscillation in the acquired waveform. This helps to more accurately determine the start and end times of the Miller platform in complex noise environments and improve the accuracy of the Miller platform time calculation results.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, the step of extracting the envelope of the three-phase Miller platform time series and filtering out the high-frequency components corresponding to the switching frequency to obtain a low-frequency pulsating envelope with a frequency twice that of the microgrid fundamental frequency specifically includes: synchronously resampling the three-phase Miller platform time series based on the periodic trigger signal of the inverter switching frequency to obtain discrete platform data with equal time intervals; using a preset sliding polynomial fitting algorithm to perform local feature smoothing on the discrete platform data to obtain a fitting baseline that eliminates the switching transient extreme points; and extracting the low-frequency fluctuation profile of the discrete platform data relative to the fitting baseline as the low-frequency pulsating envelope.

[0011] By adopting the above technical solution, synchronous resampling reduces the time jitter introduced by asynchronous acquisition, enabling the sequence data to be distributed at equal intervals along the time axis. A pre-defined sliding polynomial fitting algorithm is used to remove transient extreme points of the switching states while preserving the low-frequency fluctuation trend, allowing the extracted low-frequency pulsation envelope to more accurately reflect the low-frequency alternating fluctuation characteristics of the device junction temperature.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, before the step of inverting and resolving the three-phase low-frequency pulsating envelope into a three-phase low-frequency power loss fluctuation sequence based on a preset device thermistor coefficient and a preset transient thermal impedance network model, the method further includes: obtaining the heat sink substrate temperature and the zero-crossing time of the three-phase output current corresponding to each phase arm under steady-state symmetrical operation of the microgrid; extracting the steady-state platform time reference value of the three-phase Miller platform time series near the zero-crossing time; and performing online calibration and updating of the thermal resistance parameters in the preset transient thermal impedance network model based on the heat sink substrate temperature and the steady-state platform time reference value.

[0013] By adopting the above technical solution, the thermal resistance parameters are calibrated and updated online based on the measured heat sink substrate temperature and the steady-state platform time reference value under the steady-state symmetrical operation condition of the microgrid. This enables the preset transient thermal impedance network model to compensate for parameter deviations caused by device aging or changes in the operating environment, which helps to maintain the accuracy of power loss inversion during long-term system operation.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the step of inverting and resolving the three-phase low-frequency pulsating envelope into a three-phase low-frequency power loss fluctuation sequence based on preset device thermistor coefficients and preset transient thermal impedance network models specifically includes: synchronously acquiring the three-phase output current of the inverter and extracting the fundamental effective value of the three-phase output current; inputting the fundamental effective value into a preset current-thermistor coefficient mapping table, and interpolating to match the dynamic thermistor coefficient corresponding to the current load level; using the dynamic thermistor coefficient as the preset device thermistor coefficient, and combining it with the preset transient thermal impedance network model, inversely deriving the three-phase low-frequency pulsating envelope into a three-phase low-frequency power loss fluctuation sequence.

[0015] By adopting the above technical solution, the thermistor coefficient is interpolated and matched by querying the preset current and thermistor coefficient mapping table to obtain the dynamic thermistor coefficient corresponding to the current load level. This corrects the nonlinear error caused by the thermistor coefficient changing with the load current, and improves the accuracy of power loss inversion results under load change conditions.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the step of calculating the time derivative of the negative-sequence loss amplitude within a preset time window to obtain an intermittent fluctuation characteristic value characterizing the intermittency of the fault arc specifically includes: calculating the instantaneous rate of change sequence of the negative-sequence loss amplitude within the preset time window; performing statistical variance calculation on the instantaneous rate of change sequence to obtain a divergence value characterizing the severity of loss fluctuation; calculating the arithmetic mean of the negative-sequence loss amplitude within the preset time window; and using the ratio of the divergence value to the arithmetic mean as the intermittent fluctuation characteristic value characterizing the intermittency of the fault arc.

[0017] By adopting the above technical solution, the ratio of the variance of the instantaneous rate of change sequence to the arithmetic mean of the negative sequence loss amplitude is used as the intermittent fluctuation characteristic value. The normalization process reduces the influence of the basic magnitude of the negative sequence loss amplitude on the characteristic value calculation, so that the intermittent fluctuation characteristic value has a more stable discrimination ability for the quantification of the intermittent characteristics of the fault arc under different operating conditions.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, when it is determined that the negative sequence loss amplitude is greater than a preset unbalance reference threshold and the intermittent fluctuation characteristic value is greater than a preset arc distortion threshold, the step of determining that a high-resistance fault has occurred in the microgrid and determining the phase in which the high-resistance fault occurs based on the target phase with the largest amplitude in the three-phase low-frequency power loss fluctuation sequence specifically includes: when it is determined that the negative sequence loss amplitude is greater than the preset unbalance reference threshold, activating the fault monitoring state; in the fault monitoring state, performing time integration on the intermittent fluctuation characteristic value that is greater than the preset arc distortion threshold to obtain an intermittent integral quantity characterizing the accumulation of arc energy; when the intermittent integral quantity reaches a preset energy confirmation threshold, determining that a high-resistance fault has occurred in the microgrid and determining the phase in which the high-resistance fault occurs based on the target phase with the largest amplitude in the three-phase low-frequency power loss fluctuation sequence.

[0019] By adopting the above technical solution, the intermittent fluctuation characteristic value that meets the threshold condition is accumulated over time. The intermittent arc characteristic is required to accumulate continuously in the time dimension until the preset energy confirmation threshold is reached before the fault judgment conclusion can be output. This reduces the probability of false alarms caused by the intermittent fluctuation characteristic value briefly exceeding the threshold due to a single or brief occasional disturbance, and improves the stability of high resistance fault judgment.

[0020] In a second aspect, embodiments of this application provide a diagnostic system comprising: one or more processors and a memory; the memory is coupled to the one or more processors and is used to store computer program code, the computer program code including computer instructions, wherein the one or more processors invoke the computer instructions to cause the diagnostic system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a diagnostic system, cause the diagnostic system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a diagnostic system, cause the diagnostic system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the diagnostic system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. By using the existing gate driver of the microgrid inverter as the signal acquisition source, no additional sensors are needed on the grid side, thus avoiding interference from high-frequency background noise in the microgrid for fault feature extraction. Miller plateau time is used as the thermally sensitive quantity of the junction temperature of the power switching devices. The three-phase low-frequency power loss fluctuation sequence is obtained through envelope extraction and transient thermal impedance network model inversion. Then, symmetrical component decomposition is used to extract the negative sequence loss amplitude, transforming the detection of three-phase electrical quantity asymmetry into the detection of differences in the thermal state of internal inverter components. A dual-condition joint judgment of negative sequence loss amplitude and intermittent fluctuation characteristic value is introduced in the fault determination stage. This not only distinguishes between high-resistance faults and ordinary load imbalances but also locates the phase where the high-resistance fault occurs, improving the accuracy of high-resistance fault diagnosis in microgrids.

[0026] 2. The Miller effect stage boundary is identified based on the rate of change sequence between adjacent sampling points of the gate-source voltage. Compared with the direct voltage threshold judgment method, it has a certain suppression effect on the instantaneous voltage spikes introduced by electromagnetic interference or switching oscillation in the acquired waveform. This helps to more accurately determine the start and end times of the Miller platform in complex noise environments and improve the accuracy of the Miller platform time calculation results.

[0027] 3. By performing time integration and accumulation on the intermittent fluctuation characteristic values ​​that meet the threshold conditions, the intermittent arc characteristics are required to accumulate continuously in the time dimension until the preset energy confirmation threshold is reached before the fault judgment conclusion can be output. This reduces the probability of false alarms caused by the intermittent fluctuation characteristic values ​​briefly exceeding the threshold due to a single or brief occasional disturbance, and improves the stability of high resistance fault judgment. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a microgrid fault diagnosis method in an embodiment of this application.

[0029] Figure 2 This is another flowchart illustrating a microgrid fault diagnosis method in an embodiment of this application.

[0030] Figure 3 This is a schematic diagram of the physical device structure of a diagnostic system in an embodiment of this application. Detailed Implementation

[0031] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification and appended claims of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0032] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0033] This application provides a microgrid fault diagnosis method and system to improve the accuracy of diagnosing high-resistance faults in microgrids.

[0034] Please see Figure 1 This is a flowchart illustrating a microgrid fault diagnosis method in an embodiment of this application.

[0035] S101. The gate-source voltage signal of the three-phase bridge arm power switching device during the conduction transient process is acquired in real time through the gate driver of the microgrid inverter.

[0036] In a microgrid, an inverter is a power electronic device that performs DC / AC power conversion in a microgrid, converting DC power from distributed power sources into AC power that meets the grid's frequency and voltage requirements. A gate driver is a drive circuit unit connected between the control signal source and the gate of the power switching device, amplifying the control signal and providing drive current to the gate of the power switching device to achieve precise control over its on / off state. A three-phase bridge arm is a basic switching unit in the inverter's three-phase output circuit, where each phase consists of two power switching devices connected in series. The three-phase bridge arms work together to synthesize a three-phase AC output. A power switching device is a semiconductor device in the inverter bridge arm that performs power switching conversion, typically an IGBT or MOSFET. The conduction transient process refers to the transient time period during which a power switching device transitions from the off state to the fully on state. This process includes a delay phase, a current rise phase, a Miller effect phase, and a fully on phase. The gate-source voltage signal refers to the waveform of the voltage between the gate and source (or emitter, or gate-emitter voltage for IGBTs) of a power switching device as a function of time. This signal exhibits a phased change characteristic during the turn-on transient process, with the Miller effect phase corresponding to a plateau region where the voltage is approximately constant.

[0037] Specifically, the diagnostic system acquires the gate-source voltage waveforms of each power switching device in the three-phase bridge arm during each turn-on operation using the existing gate driver in the inverter. Since the gate driver is directly connected to the gate and source of the device, it can acquire transient gate-source voltage signals in the nanosecond to microsecond range without introducing additional sensors, thus avoiding background interference introduced by installing transformers on the grid side in the high-noise environment of a microgrid. During the turn-on transient process, the gate-source voltage sequentially undergoes a voltage rise phase (charging the gate-source capacitance to the threshold voltage), a Miller effect phase (the gate current is drawn by the Miller capacitance, and the gate-source voltage maintains a nearly constant plateau), and a subsequent charging phase to the drive voltage. The duration of the gate-source voltage plateau during the Miller effect phase corresponds to the drain voltage drop process of the device. The duration of this phase is closely related to the device's power loss and junction temperature. The acquisition of data from the three-phase bridge arm is performed synchronously to ensure consistency of the three-phase data in terms of time reference.

[0038] In some embodiments, the gate-source voltage signal during the conduction transient can be acquired in real time through various methods. Optionally, a high-speed differential voltage sampling circuit can be connected in parallel at the output of the gate driver. A high-speed analog-to-digital converter (ADC) is used to continuously sample the gate-source voltage at a sampling frequency that meets nanosecond-level time resolution. The acquisition window is activated when the trigger signal arrives in each switching cycle to record the conduction transient waveform. The acquired digital quantity is then transmitted to the diagnostic system through an isolation interface, and the processing unit performs time alignment on the three-phase data before storing it. Optionally, an analog signal conditioning circuit can also be integrated inside the gate driver. Using a circuit structure combining peak hold and comparator, hardware-level level judgment is performed on the gate-source voltage. When the gate-source voltage enters a preset Miller voltage range, a timer is triggered to start counting. When the gate-source voltage leaves the range, the counting stops. The Miller plateau time within a single switching cycle is measured directly in hardware. The timing result is transmitted to the processing unit in the form of a digital signal, thereby reducing the dependence on the high-speed ADC and reducing the consumption of digital computing resources.

[0039] S102. Calculate the duration of the gate-source voltage signal within the preset Miller voltage range, and generate a three-phase Miller plateau time series synchronized with the inverter switching frequency.

[0040] The preset Miller voltage range refers to the range of gate-source voltage values ​​when the power switching device is in the Miller effect stage during the turn-on transient. This range is predetermined based on the device model's datasheet parameters and actual calibration results, and typically corresponds to the voltage range of an approximate horizontal plateau segment in the gate-source voltage waveform. The Miller effect stage refers to the phenomenon during the turn-on transient where, due to charge transfer on the parasitic capacitance between the gate and drain (Miller capacitance), the drive current is almost entirely occupied by the Miller capacitance, causing the gate-source voltage to remain approximately constant for a period of time. During this stage, the drain-source voltage continuously decreases, and the device's current approaches the load current; this is the period with the most concentrated switching losses. The Miller plateau time refers to the duration of the gate-source voltage in the Miller effect stage during a single turn-on transient of the power switching device. Physically, it corresponds to the time it takes for the drain voltage to drop from the off state voltage to the on state voltage. This duration is directly related to the device junction temperature, load current, and operating conditions. A three-phase Miller plateau time series is a discrete data sequence formed by arranging the Miller plateau time values ​​calculated by the power switching devices of each of the three phase arms over multiple consecutive switching cycles in chronological order. Its sampling frequency is synchronized with the inverter switching frequency. The switching frequency, which refers to the number of times the inverter's power switching devices complete turn-on and turn-off actions per unit time, determines the time resolution of the three-phase Miller plateau time series.

[0041] Specifically, for the gate-source voltage waveform acquired in each switching cycle, the waveform is first divided into segments according to a preset Miller voltage range, identifying the continuous time period within which the gate-source voltage falls. The preset Miller voltage range is determined in advance based on the device model's datasheet parameters and actual calibration results, typically corresponding to the voltage range of an approximate horizontal plateau segment in the gate-source voltage waveform. After identifying this continuous time period, the time difference between its start and end times is calculated, which is the Miller plateau time corresponding to this switching cycle. The Miller plateau time values ​​corresponding to each phase of the three-phase bridge arm in each switching cycle are arranged sequentially according to the time interval defined by the inverter switching frequency, forming Miller plateau time sequences for phases A, B, and C, respectively. Since the sequence generation frequency is synchronized with the switching frequency, the time interval between adjacent points in the sequence is equal to the switching cycle, thereby improving the accuracy of frequency analysis in subsequent signal processing steps.

[0042] In some embodiments, Miller plateau time can be calculated and a three-phase Miller plateau time series can be generated in various ways. Optionally, a voltage threshold dual-boundary judgment method can be used. The upper and lower bound voltage thresholds of the Miller voltage range are set in advance according to the device calibration data. In the acquisition waveform of each switching cycle, the voltage values ​​of each sampling point are scanned sequentially. The moment when the gate-source voltage first enters the range is recorded as the start time, and the moment when the gate-source voltage first leaves the range is recorded as the end time. The difference between the two moments is calculated to obtain the Miller plateau time of this cycle, and the three-phase results are appended to their respective sequence arrays in chronological order. Optionally, the voltage change rate between adjacent sampling points can be calculated point by point on the acquired gate-source voltage waveform to construct a voltage change rate sequence. Then, a continuous time period in which the absolute value of the voltage change rate is continuously lower than a preset smoothing threshold can be identified. The preset smoothing threshold is a pre-set upper limit of the absolute value of the voltage change rate used to determine whether the gate-source voltage change rate is sufficiently smooth and meets the characteristics of the Miller effect stage when using the voltage change rate criterion to identify the Miller effect stage. When the absolute value of the voltage change rate is continuously lower than this threshold, the gate-source voltage is considered to be in the approximately constant Miller plateau region. This threshold is pre-calibrated and determined according to the power switching device model and drive circuit parameters. The start and end times of this continuous segment are used as the boundary points for the device to enter and exit the Miller effect stage. The time difference between the boundary points is the Miller plateau time of this cycle. The results of each switching cycle are then sequentially spliced ​​to form a three-phase Miller plateau time sequence. This method can further improve noise immunity.

[0043] S103. The envelope of the three-phase Miller platform time series is extracted and the high-frequency component corresponding to the switching frequency is filtered out to obtain a low-frequency pulsating envelope with a frequency twice that of the microgrid fundamental frequency.

[0044] The high-frequency component corresponding to the switching frequency refers to the high-frequency periodic fluctuations in the three-phase Miller plateau time series, introduced by the inverter switching operation itself, with the switching frequency as the fundamental frequency. This component reflects the natural fluctuations in the Miller plateau time caused by PWM modulation within each switching cycle and is unrelated to fault characteristics. The microgrid fundamental frequency refers to the basic frequency of the microgrid AC power, typically 50 Hz or 60 Hz. The low-frequency pulsating envelope refers to the low-frequency slowly varying waveform retained after envelope extraction. Its frequency is twice the microgrid fundamental frequency (i.e., 100 Hz or 120 Hz). This characteristic frequency originates from the fact that the amplitude of the three-phase inverter output current undergoes two complete changes in each fundamental cycle, causing the thermal state of the power switching devices to fluctuate periodically at twice the fundamental frequency. Asymmetrical operating conditions refer to the operating state in which the three-phase load or operating conditions of a microgrid are different, resulting in unequal amplitudes of the three-phase output current of the inverter. When a high-resistance fault occurs, the equivalent impedance of the faulty phase increases, thereby causing asymmetry in the three-phase current, which in turn causes the junction temperature of the power switching devices in each phase to exhibit different low-frequency pulsation characteristics.

[0045] Specifically, the three-phase Miller platform time series contains both high-frequency fluctuation components caused by switching frequency and low-frequency components caused by microgrid operating condition changes. Since the amplitude of the high-frequency components is much larger than that of the low-frequency components, direct analysis of the original series makes it difficult to identify low-frequency thermal characteristics. Therefore, envelope extraction processing is performed on the three-phase Miller platform time series to filter out the high-frequency components corresponding to the switching frequency. First, the original series is synchronously resampled based on the periodic trigger signal of the inverter switching frequency, ensuring that the data points in the series are evenly distributed on the time axis, thus eliminating the time jitter error introduced by asynchronous acquisition. Subsequently, a pre-defined sliding polynomial fitting algorithm is used to locally smooth the resampled discrete data. This algorithm uses a predetermined sliding window length and polynomial order as configuration parameters to perform least-squares polynomial fitting on the discrete data within a progressively moving time window. The window length and polynomial order are predetermined based on the ratio of the inverter switching frequency to the microgrid fundamental frequency. The window length needs to cover a sufficient number of switching cycles to effectively smooth high-frequency fluctuations, and the polynomial order needs to be low enough so that the fitted curve only tracks low-frequency trends and does not fit high-frequency fluctuations. While eliminating extreme points caused by switching transients, the low-frequency trend of the data is preserved, resulting in a fitting baseline. Finally, the low-frequency fluctuation profile of the discrete data relative to the fitting baseline is extracted, which is the required low-frequency pulsation envelope. Under normal symmetrical operating conditions, the amplitudes of the three-phase low-frequency pulsation envelopes should be approximately equal. However, when a high-resistance fault occurs in one phase, causing three-phase asymmetry, the envelope amplitude of the faulty phase will be significantly different from the other two phases, thus forming an identifiable asymmetry feature. It should be noted that since the three-phase bridge arms of an inverter typically share a heat sink, there is an interphase thermal coupling effect. However, the heat sink has a relatively large thermal time constant, acting as a thermal low-pass filter. Interphase heat conduction mainly affects the steady-state reference temperature of the device, while having minimal impact on alternating fluctuations at frequencies of 100Hz or 120Hz. This application extracts a low-frequency pulsation envelope at twice the fundamental frequency, naturally isolating thermal crosstalk between adjacent phases in the frequency domain.

[0046] In some embodiments, the envelope of the three-phase Miller platform time series can be extracted in various ways. Optionally, a weighted moving average of the data within the window can be calculated point-by-point for the three-phase Miller platform time series, using a sliding window length that is several times the period corresponding to the inverter switching frequency. The low-pass characteristic of the moving average is used to filter out high-frequency fluctuations caused by the switching frequency, and the sliding average result is used as the low-frequency pulsation envelope. The selection of the window length must ensure that the cutoff frequency is lower than the switching frequency and higher than twice the fundamental frequency to retain the low-frequency characteristics around 100 Hz. Optionally, a method of synchronous resampling combined with polynomial fitting can also be used. First, the Miller platform time series is synchronously resampled at equal time intervals according to the trigger signal of the inverter switching frequency to eliminate time jitter. Then, low-order polynomial least squares fitting is performed on the discrete platform data within a preset sliding data window. The polynomial fitting curve is used as a local baseline, and the difference waveform between the original data and the baseline is extracted. The low-frequency pulsation component in the difference waveform is further filtered to form the low-frequency pulsation envelope.

[0047] S104. Based on the preset device thermistor coefficient and preset transient thermal impedance network model, the low-frequency pulsation envelope of the three phases is inverted and analyzed into a three-phase low-frequency power loss fluctuation sequence.

[0048] The preset device thermistor coefficient is a proportional coefficient reflecting the linear mapping relationship between the junction temperature change and the Miller plateau time change of a power switching device. Its physical meaning is that the gate-drain Miller capacitance and transconductance of the power switching device are affected by the junction temperature, causing the charging and discharging time corresponding to the Miller effect stage to change as the junction temperature rises under the same load current conditions. Therefore, the Miller plateau time becomes a thermally sensitive parameter of the junction temperature, and the thermistor coefficient quantifies the slope of this linear relationship. The preset transient thermal impedance network model is an equivalent circuit model describing the thermal conduction characteristics of each layer of material from the junction to the heat sink substrate in a power switching device. It is usually represented by a multi-order Foster or Cauer RC network. This model uses power loss as the excitation input and the temperature difference between each layer as the response output, reflecting the transient thermal response process of the device under dynamically changing power input. The three-phase low-frequency power loss fluctuation sequence refers to the time series that characterizes the periodic power loss fluctuation of three-phase power switching devices, obtained by inversion calculation through the low-frequency pulsation envelope. Its essence is to inversely deduce the power loss change corresponding to the change in thermal state by using the Miller plateau time change in the thermal-time domain through the thermal impedance network model and thermistor coefficient.

[0049] Specifically, the low-frequency pulsation envelope reflects the low-frequency alternating fluctuation amplitude of the device junction temperature, and the relationship between the change in junction temperature and the change in power loss is described by a preset transient thermal impedance network model. The diagnostic system first converts the Miller plateau time variation in the three-phase low-frequency pulsation envelope into the corresponding junction temperature fluctuation based on the preset device thermistor coefficient. Then, based on the preset transient thermal impedance network model, it reverse-engineers the junction temperature fluctuation into the corresponding power loss fluctuation, thus obtaining the three-phase low-frequency power loss fluctuation sequence. It should be noted that the thermistor coefficient is related to the load current level of the device. Therefore, the diagnostic system can simultaneously acquire the fundamental effective value of the three-phase output current of the inverter, and obtain the dynamic thermistor coefficient corresponding to the current load level by interpolating through a preset current-thermistor coefficient mapping table. The preset current-thermistor coefficient mapping table is a lookup table that is established in advance through experimental calibration and describes the variation of the thermistor coefficient of the power switching device with the load current. This mapping table stores the calibrated value of the thermistor coefficient at the corresponding operating point for different load current levels (usually indexed by the fundamental effective value). During system operation, the dynamic thermistor coefficient corresponding to the current load level is obtained by interpolation calculation. The purpose of establishing this mapping table is to correct the nonlinear error caused by the change of the thermistor coefficient with the load and improve the accuracy of power loss inversion.

[0050] In some embodiments, before the step of inverting and resolving the three-phase low-frequency pulsating envelope into a three-phase low-frequency power loss fluctuation sequence based on the preset device thermistor coefficient and the preset transient thermal impedance network model, during the steady-state symmetrical operation of the microgrid, the heat sink substrate temperature corresponding to each phase arm and the zero-crossing time of the three-phase output current can be obtained, the steady-state platform time reference value of the three-phase Miller platform time series near the zero-crossing time can be extracted, and the thermal resistance parameters in the preset transient thermal impedance network model can be calibrated and updated online based on the heat sink substrate temperature and the steady-state platform time reference value to correct the model parameter deviation caused by device aging or environmental changes.

[0051] In some embodiments, within each calculation cycle, a preset current-thermal coefficient mapping table can be queried based on the measured effective value of the fundamental wave of the three-phase output current. The dynamic thermistor coefficient under the current load level can be obtained by linear interpolation. The thermistor coefficient is multiplied by the amplitude of the low-frequency pulsating envelope to obtain the junction temperature fluctuation. Then, the power loss is inverted by combining the transient thermal impedance network model. The inversion result of the previous calculation cycle is used as the initial value and substituted into the calculation of the current cycle to improve the inversion accuracy in an iterative manner.

[0052] S105. Perform symmetrical component decomposition on the three-phase low-frequency power loss fluctuation sequence, extract the negative sequence loss amplitude, and calculate the time derivative of the negative sequence loss amplitude within a preset time window to obtain the intermittent fluctuation characteristic value that characterizes the intermittency of the fault arc.

[0053] Symmetrical component decomposition refers to a mathematical method of decomposing three-phase asymmetrical physical quantities into three sets of three-phase symmetrical components. Specifically, any set of asymmetrical three-phase quantities can be decomposed into a superposition of positive-sequence components (three-phase amplitudes equal, phases differing by 120°, and phase sequence consistent with the normal operating phase sequence of the system), negative-sequence components (three-phase amplitudes equal, phases differing by 120°, and phase sequence opposite to the positive-sequence), and zero-sequence components (three-phase amplitudes equal and in phase). The physical meaning of the negative-sequence loss amplitude is that when the three-phase power losses are equal during three-phase symmetrical operation, the negative-sequence component amplitude is close to zero. However, when a high-resistance fault occurs in a phase, causing asymmetrical three-phase losses, the negative-sequence component amplitude will increase significantly. The preset time window length comprehensively considers the fundamental frequency period of the microgrid and the typical intermittent period of the fault arc. It is usually set to a time length that includes several fundamental frequency periods to ensure that the complete fluctuation characteristics of the intermittent arc can be captured within the window, while having sufficient time resolution for online real-time detection.

[0054] The typical electrical characteristic of high-resistivity faults is that a grounding path is formed between the faulty phase and the ground through high-resistivity media such as branches, sand, and soil, causing a decrease in the load current of that phase and resulting in three-phase asymmetry. Simultaneously, high-resistivity faults are usually accompanied by intermittent arcing. The random extinction and reignition of the arc cause random changes in the fault impedance within a short period, leading to irregular fluctuations in power loss. Specifically, the diagnostic system applies a symmetric component transformation matrix to the data within the current time window of the three-phase low-frequency power loss fluctuation sequence to extract the negative-sequence power loss component, the amplitude of which is the negative-sequence loss amplitude. Subsequently, within a preset time window, the time derivative between adjacent sampling times is calculated point-by-point for the negative-sequence loss amplitude sequence; that is, the difference in negative-sequence loss amplitude between adjacent sampling points is divided by the sampling time interval, yielding a derivative sequence reflecting the rate of change of the negative-sequence loss amplitude. Statistical analysis of this derivative sequence extracts statistical features that quantify the severity of amplitude changes within the window, which are the intermittent fluctuation characteristic values. When the fault arc experiences intermittent extinction and reignition, the time derivative of the negative sequence loss amplitude will exhibit a large degree of dispersion and frequent fluctuations within the window, making the intermittent fluctuation characteristic value significantly different from the numerical level under normal operation or steady-state load imbalance conditions.

[0055] In some embodiments, intermittent fluctuation characteristic values ​​can be extracted based on time derivatives in various ways. Optionally, the maximum derivative peak value normalization method can be used. Within a preset time window, the difference between adjacent sampling points in the negative sequence loss amplitude sequence is calculated and divided by the sampling interval to obtain the instantaneous rate of change sequence. The maximum absolute value in this sequence is extracted, and the ratio of this maximum value to the arithmetic mean of the negative sequence loss amplitude within the window is used as the intermittent fluctuation characteristic value. This normalization eliminates the influence of the basic loss magnitude on the characteristic value under different operating conditions. When a high-resistance fault triggers an intermittent arc, the random extinction and reignition process of the arc will cause a local abrupt change in the negative sequence loss amplitude, significantly increasing this ratio. Optionally, the derivative zero-crossing frequency method can also be used. The number of zero-crossings in the time derivative sequence of the negative sequence loss amplitude within a preset time window is calculated, i.e., the number of times the derivative sign changes from positive to negative or from negative to positive. The number of zero-crossings per unit time is used as the intermittent fluctuation characteristic value. The intermittent arcing associated with high-resistance faults will cause frequent reverse changes in the amplitude of negative sequence losses, making the zero-crossing frequency significantly higher than that under normal operation and steady-state unbalanced conditions.

[0056] S106. When it is determined that the negative sequence loss amplitude is greater than the preset unbalance reference threshold and the intermittent fluctuation characteristic value is greater than the preset arc distortion threshold, the microgrid is judged to have a high resistance fault, and the phase in which the high resistance fault occurs is determined based on the target phase with the largest amplitude in the three-phase low-frequency power loss fluctuation sequence.

[0057] The preset unbalance reference threshold refers to a threshold for judging the negative sequence loss amplitude based on the maximum permissible unbalance of the three-phase load under normal microgrid operating conditions, combined with the measurement error of the system's three-phase losses. Exceeding this threshold indicates that the asymmetry of the three-phase power loss has exceeded the normal operating range, requiring further assessment to determine if a fault exists. The preset arc distortion threshold is a threshold for judging intermittent fluctuation characteristics, used to distinguish between intermittent arc characteristics and ordinary load fluctuation characteristics. Exceeding this threshold indicates that the power loss fluctuation has the random intermittent nature characteristic of a fault arc.

[0058] Specifically, the diagnostic system continuously monitors the negative sequence loss amplitude and intermittent fluctuation characteristic value in each calculation cycle. When the negative sequence loss amplitude exceeds the preset unbalance reference threshold and the intermittent fluctuation characteristic value simultaneously exceeds the preset arc distortion threshold, it determines that a high-resistance fault has occurred in the microgrid. Based on the target phase with the largest amplitude change in the three-phase low-frequency power loss fluctuation sequence, the phase in which the high-resistance fault occurs is determined. This is because, under the condition of a high-resistance fault, the equivalent impedance of the faulty phase changes significantly, resulting in a decrease in current. The change in the power loss fluctuation amplitude of the faulty phase (or the degree of deviation from the normal symmetrical reference value) is the largest relative to the other two phases. Thus, the location of the fault can be directly located, providing phase information for subsequent operation and maintenance.

[0059] In the above embodiments, the existing gate driver of the microgrid inverter is used as the signal acquisition source, eliminating the need for additional sensors on the grid side and avoiding interference from high-frequency background noise in the microgrid for fault feature extraction. Miller plateau time is used as the thermally sensitive quantity for the junction temperature of the power switching devices. The three-phase low-frequency power loss fluctuation sequence is obtained through envelope extraction and transient thermal impedance network model inversion. Then, symmetrical component decomposition is used to extract the negative sequence loss amplitude, transforming the detection of three-phase electrical quantity asymmetry into the detection of differences in the thermal state of internal inverter components. A dual-condition joint judgment based on the negative sequence loss amplitude and intermittent fluctuation characteristic value is introduced in the fault determination stage. This distinguishes between high-resistance faults and ordinary load imbalances while locating the phase where the high-resistance fault occurs, improving the accuracy of high-resistance fault diagnosis in microgrids.

[0060] However, when non-fault factors (such as sudden load changes or inverter control actions) cause the negative-sequence loss amplitude and intermittent fluctuation characteristic value to briefly exceed their respective thresholds within the same calculation cycle, the above method may generate false alarms. Furthermore, the calculation of the intermittent fluctuation characteristic value in the above embodiments relies on the statistical analysis of the time derivative. When the basic magnitude of the negative-sequence loss amplitude changes with operating conditions, the margin between the characteristic value and the threshold may change accordingly, affecting the stability assessment. Therefore, the following further optimized implementation scheme is proposed.

[0061] Please see Figure 2 This is another flowchart illustrating a microgrid fault diagnosis method in an embodiment of this application.

[0062] S201. The gate-source voltage signal of the three-phase bridge arm power switching device during the conduction transient process is acquired in real time through the gate driver of the microgrid inverter.

[0063] S202. Calculate the duration of the gate-source voltage signal within the preset Miller voltage range, and generate a three-phase Miller plateau time series synchronized with the inverter switching frequency.

[0064] S203. The envelope of the three-phase Miller platform time series is extracted and the high-frequency component corresponding to the switching frequency is filtered out to obtain a low-frequency pulsating envelope with a frequency twice that of the microgrid fundamental frequency.

[0065] S204. Based on the preset device thermistor coefficient and preset transient thermal impedance network model, the low-frequency pulsation envelope of the three phases is inverted and analyzed into a three-phase low-frequency power loss fluctuation sequence.

[0066] S205. Perform symmetrical component decomposition on the three-phase low-frequency power loss fluctuation sequence to extract the negative sequence loss amplitude.

[0067] Step S201 is similar to step S101, step S202 is similar to step S102, step S203 is similar to step S103, step S204 is similar to step S104, and step S205 is similar to step S105, so they will not be repeated here.

[0068] S206. Calculate the instantaneous rate of change sequence of negative sequence loss amplitude within a preset time window.

[0069] S207. Perform statistical variance calculation on the instantaneous rate of change sequence to obtain the divergence value that characterizes the severity of loss fluctuation.

[0070] S208. Calculate the arithmetic mean of the negative sequence loss amplitude within a preset time window.

[0071] S209. The ratio of the divergence value to the arithmetic mean is used as the intermittent fluctuation characteristic value that characterizes the intermittency of the fault arc.

[0072] Steps S206 to S209 aim to reduce the impact of the basic magnitude of the negative-sequence loss amplitude on the characteristic value calculation results by introducing a normalized statistical fluctuation index. Specifically, in step S206, the diagnostic system calculates the difference between adjacent sampling times for each point in the negative-sequence loss amplitude sequence within a preset time window and divides it by the sampling time interval to obtain the instantaneous rate of change sequence. In step S207, statistical variance is calculated for all elements in the instantaneous rate of change sequence to quantify the dispersion of the rate of change of the negative-sequence loss amplitude within the window, obtaining a divergence value. In step S208, the arithmetic mean of all data points of the negative-sequence loss amplitude within the preset time window is calculated to characterize the average intensity of the three-phase asymmetry within that time period. In step S209, the divergence value is divided by the arithmetic mean to obtain the intermittent fluctuation characteristic value; the physical meaning of this ratio is the normalized degree of change severity. When a high-resistance fault is accompanied by intermittent arcing, the random extinction and reignition of the arc cause frequent abrupt changes in the amplitude of the negative-sequence loss, resulting in a larger variance in the instantaneous rate of change sequence and a relatively stable mean amplitude, thus significantly increasing the intermittent fluctuation characteristic value. In contrast, under steady-state unbalanced load conditions, the amplitude of the negative-sequence loss changes slowly, the variance of the instantaneous rate of change sequence is smaller, and the intermittent fluctuation characteristic value is lower, thus allowing it to be distinguished from a high-resistance fault.

[0073] S210. When it is determined that the negative sequence loss amplitude is greater than the preset unbalanced reference threshold, the fault monitoring state is activated.

[0074] S211. Under fault monitoring status, the intermittent fluctuation characteristic value that is greater than the preset arc distortion threshold is integrated over time to obtain the intermittent integral quantity characterizing the accumulation of arc energy.

[0075] S212. When the intermittent integral reaches the preset energy confirmation threshold, it is determined that a high-resistance fault has occurred in the microgrid, and the phase in which the high-resistance fault occurs is determined based on the target phase with the largest amplitude in the three-phase low-frequency power loss fluctuation sequence.

[0076] Among them, the preset energy confirmation threshold refers to the intermittent integral threshold value required to confirm the occurrence of a high-resistance fault. By setting this threshold in advance, false alarms can be avoided before a single or brief intermittent fluctuation triggers a false alarm before the accumulated amount reaches a level sufficient to characterize a real fault. Its specific value is determined comprehensively based on the energy accumulation characteristics of intermittent arcs under typical high-resistance fault scenarios and the maximum allowable detection delay.

[0077] Specifically, in step S210, when the calculated result of the negative sequence loss amplitude exceeds the preset unbalance reference threshold, the diagnostic system switches the system state to fault monitoring state. In fault monitoring state, the diagnostic system performs real-time judgment on the intermittent fluctuation characteristic values ​​output in each calculation cycle. Only characteristic values ​​greater than the preset arc distortion threshold are accumulated over time; characteristic values ​​that do not meet the threshold condition are not accumulated into the integral. This achieves selective accumulation of intermittent arc energy, eliminating interference from normal operation fluctuations below the threshold on the integral. In step S212, the diagnostic system continuously monitors the intermittent integral. When it first reaches the preset energy confirmation threshold, it determines that a high-resistance fault has occurred in the microgrid. Simultaneously, it extracts the amplitude of the three-phase low-frequency power loss fluctuation sequence at that moment, and outputs the high-resistance fault occurrence phase information as the target phase, using the phase with the largest amplitude change. In addition, if the negative sequence loss amplitude drops below the preset unbalanced reference threshold before the intermittent integral amount reaches the confirmation threshold, the fault monitoring state is exited and the intermittent integral amount is reset to zero. The integration is restarted when the negative sequence loss amplitude exceeds the threshold again, so as to improve the stability of the diagnostic conclusion.

[0078] In the above embodiments, by introducing a normalized intermittent fluctuation characteristic value calculation method based on the ratio of the instantaneous rate of change sequence variance to the arithmetic mean, the impact of the basic magnitude change of the negative sequence loss amplitude on the stability of characteristic value discrimination is reduced. Simultaneously, an integral accumulation confirmation mechanism is introduced, requiring that the intermittent arc characteristics accumulate continuously over time to a preset energy confirmation threshold before a fault judgment conclusion can be output. This reduces the probability of false alarms caused by brief, occasional disturbances, further improving the stability and reliability of high-resistance fault diagnosis in microgrids.

[0079] The above describes a microgrid fault diagnosis method in the embodiments of this application. The following describes an exemplary diagnostic system 300 provided in the embodiments of this application.

[0080] Figure 3This is an exemplary hardware structure diagram of the diagnostic system 300 provided in an embodiment of this application. In some embodiments, the diagnostic system 300 is a computer device. The computer device includes a processor, a memory, and a network interface connected via a system bus. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with other external terminals or servers via a network connection. In some embodiments, the network interface can be a wired network interface; in some embodiments, the network interface can also be a wireless network interface. When the computer program is executed by the processor, it implements a microgrid fault diagnosis method according to an embodiment of this application.

[0081] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0082] In some embodiments of this application, a computer-readable storage medium is also provided, including instructions that, when executed on the diagnostic system 300, cause the diagnostic system 300 to perform a microgrid fault diagnosis method according to an embodiment of this application.

[0083] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0084] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0085] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive), etc.

[0086] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A microgrid fault diagnosis method, characterized in that, Applied to a diagnostic system, the method includes: The gate-source voltage signal of the three-phase bridge arm power switching device during the turn-on transient process is acquired in real time by the gate driver of the microgrid inverter. The duration of the gate-source voltage signal within a preset Miller voltage range is calculated, and a three-phase Miller plateau time series synchronized with the inverter switching frequency is generated. The preset Miller voltage range is the range of gate-source voltage when the power switching device is in the Miller effect stage. Envelope extraction processing is performed on the three-phase Miller platform time series to filter out high-frequency components corresponding to the switching frequency, resulting in a low-frequency pulsating envelope with a frequency twice that of the microgrid fundamental frequency. The low-frequency pulsating envelope characterizes the alternating fluctuation of the junction temperature of the power switching device caused by the asymmetrical operating conditions of the microgrid. Based on the preset device thermistor coefficient and the preset transient thermal impedance network model, the low-frequency pulsation envelope of the three phases is inverted and analyzed into a three-phase low-frequency power loss fluctuation sequence. The preset device thermistor coefficient defines the linear mapping relationship between the junction temperature change of the power switching device and the Miller plateau time change. Symmetric component decomposition is performed on the three-phase low-frequency power loss fluctuation sequence to extract the negative sequence loss amplitude, and the time derivative of the negative sequence loss amplitude within a preset time window is calculated to obtain the intermittent fluctuation characteristic value characterizing the intermittency of the fault arc. If the negative sequence loss amplitude is greater than the preset unbalance reference threshold and the intermittent fluctuation characteristic value is greater than the preset arc distortion threshold, a high-resistance fault is determined to have occurred in the microgrid, and the phase in which the high-resistance fault occurs is determined based on the target phase with the largest amplitude change in the three-phase low-frequency power loss fluctuation sequence.

2. The method according to claim 1, characterized in that, The step of calculating the duration of the gate-source voltage signal within a preset Miller voltage range and generating a three-phase Miller plateau time series synchronized with the inverter switching frequency specifically includes: Calculate the rate of change of the gate-source voltage signal between adjacent sampling points to obtain a voltage rate of change sequence; Identify consecutive time periods in the voltage change rate sequence where the absolute value of the voltage change rate is lower than a preset smoothing threshold; The start and end times of the continuous time period are respectively used as the boundary points for the power switching device to enter and exit the Miller effect stage. Calculate the time difference between the start time and the end time, and concatenate the time differences corresponding to each switching cycle in chronological order to generate the three-phase Miller platform time series.

3. The method according to claim 1, characterized in that, The step of extracting the envelope of the three-phase Miller platform time series, filtering out the high-frequency components corresponding to the switching frequency, and obtaining a low-frequency pulsating envelope with a frequency twice that of the microgrid fundamental frequency, specifically includes: Based on the periodic trigger signal of the inverter switching frequency, the three-phase Miller platform time series is synchronously resampled to obtain discrete platform data with equal time intervals. The discrete platform data is smoothed locally using a preset sliding polynomial fitting algorithm to obtain a fitting baseline that excludes switch transient extreme points. The low-frequency fluctuation profile of the discrete platform data relative to the fitted baseline is extracted as the low-frequency pulsation envelope.

4. The method according to claim 1, characterized in that, Before the step of inverting and resolving the low-frequency pulsation envelope of the three phases into a three-phase low-frequency power loss fluctuation sequence based on the preset device thermistor coefficient and the preset transient thermal impedance network model, the method further includes: Under steady-state symmetrical operation of the microgrid, the temperature of the heat sink substrate corresponding to each phase arm and the zero-crossing time of the three-phase output current are obtained. Extract the steady-state platform time reference value of the three-phase Miller platform time series near the zero-crossing time; Based on the heat sink substrate temperature and the steady-state platform time reference value, the thermal resistance parameters in the preset transient thermal impedance network model are calibrated and updated online.

5. The method according to claim 1, characterized in that, The step of inverting and analyzing the low-frequency pulsating envelope of the three phases into a three-phase low-frequency power loss fluctuation sequence based on the preset device thermistor coefficient and preset transient thermal impedance network model specifically includes: The three-phase output current of the inverter is collected synchronously, and the fundamental effective value of the three-phase output current is extracted. Input the fundamental effective value into a preset current and thermistor coefficient mapping table, and interpolate to match the dynamic thermistor coefficient corresponding to the current load level. Using the dynamic thermistor coefficient as the preset device thermistor coefficient, and combining it with the preset transient thermal impedance network model, the low-frequency pulsation envelope of the three phases is reverse-derived into the low-frequency power loss fluctuation sequence of the three phases.

6. The method according to claim 1, characterized in that, The step of calculating the time derivative of the negative sequence loss amplitude within a preset time window to obtain the intermittent fluctuation characteristic value characterizing the intermittency of the fault arc specifically includes: Calculate the instantaneous rate of change sequence of the negative sequence loss amplitude within the preset time window; Statistical variance was calculated on the instantaneous rate of change sequence to obtain the divergence value characterizing the severity of loss fluctuations; Calculate the arithmetic mean of the negative sequence loss amplitude within the preset time window; The ratio of the divergence value to the arithmetic mean is used as the intermittent fluctuation characteristic value that characterizes the intermittency of the fault arc.

7. The method according to claim 6, characterized in that, The step of determining a high-resistance fault in the microgrid when the negative sequence loss amplitude is greater than a preset unbalance reference threshold and the intermittent fluctuation characteristic value is greater than a preset arc distortion threshold, and determining the phase in which the high-resistance fault occurs based on the target phase with the largest amplitude in the three-phase low-frequency power loss fluctuation sequence, specifically includes: If the magnitude of the negative sequence loss is determined to be greater than the preset unbalance reference threshold, the fault monitoring state is activated. Under the fault monitoring state, the intermittent fluctuation characteristic value that is greater than the preset arc distortion threshold is integrated over time to obtain the intermittent integral quantity characterizing the accumulation of arc energy. When the intermittent integral reaches a preset energy confirmation threshold, a high-resistance fault is determined to have occurred in the microgrid, and the phase in which the high-resistance fault occurs is determined based on the target phase with the largest amplitude change in the three-phase low-frequency power loss fluctuation sequence.

8. A diagnostic system, characterized in that, The diagnostic system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the diagnostic system to perform the method as described in any one of claims 1-7.

9. A computer program product containing instructions, characterized in that, When the computer program product is run on a diagnostic system, the diagnostic system performs the method as described in any one of claims 1-7.

10. A computer-readable storage medium comprising instructions, characterized in that, When the instructions are executed on the diagnostic system, the diagnostic system performs the method as described in any one of claims 1-7.

Images