Power distribution network fault monitoring method and system

Abstract

The invention discloses a power distribution network fault monitoring method and system, and relates to the technical field of power systems, and the method comprises the steps: collecting environment parameters in an island micro-grid ring main unit, and obtaining the charge discharge current data of an insulation support member grounding path after a power frequency voltage crosses a zero point; constructing an environment-electrical response phase space trajectory locally, and judging whether the insulation support member has a critical instability trend or not to obtain an insulation state abnormal signal; the current dew point temperature is calculated, whether the temperature in the ring main unit falls into the condensation high-risk interval or not is judged, and a condensation risk signal is obtained; if the insulation state abnormal signal and the condensation risk signal exist at the same time, high-frequency transient signal collection is started, partial discharge feature extraction is carried out on the collected high-frequency transient signal, judgment is carried out in combination with the degradation trend of the charge discharge time constant, and a power distribution network fault monitoring result is generated. According to the invention, a completely autonomous, low-power-consumption and high-robustness in-situ fault monitoring capability is provided for the island micro-grid.

Description

Technical Field

[0001] This application relates to the field of power system technology, and in particular to a method and system for monitoring faults in a distribution network. Background Technology

[0002] As a critical node device in the power distribution network, the health of the internal insulation support components directly affects the reliability of system operation. To assess insulation performance, some monitoring schemes collect the charge discharge current waveform of the grounding path after the power frequency voltage crosses zero, and simultaneously acquire temperature and humidity data inside the cabinet to analyze whether the insulation material is showing signs of moisture absorption or deterioration.

[0003] This type of solution calculates the time constant by fitting the exponential decay curve of the charge discharge current, and compares the current time constant with historical data to determine whether the insulation condition is abnormal. Simultaneously, it calculates the dew point temperature based on temperature and humidity data, and generates an environmental risk signal when the cabinet temperature falls into a preset high-risk condensation range. The fault warning logic typically combines the abnormal insulation condition signal and the environmental risk signal for judgment, and relies on the communication link to upload local measurement data to the master station to complete the final decision.

[0004] In island microgrids far from the mainland, the power supply system operates in isolation for extended periods. Ring main units are deployed in dispersed locations and lack stable communication infrastructure. Simultaneously, the system frequency frequently deviates from the power frequency reference due to fluctuations in renewable energy output. Under these conditions, the aforementioned monitoring schemes relying on centralized processing at the main station cannot provide early warning of intermittent high-resistance grounding faults induced by insulation degradation at the ring main unit. Summary of the Invention

[0005] In view of the aforementioned problems, this application is hereby filed.

[0006] Therefore, this application provides a method and system for monitoring faults in a power distribution network, which can solve the problems mentioned in the background art.

[0007] To solve the above-mentioned technical problems, this application provides the following technical solution: In the first aspect, this application provides a method for monitoring faults in a power distribution network, including: collecting environmental parameters within the ring network cabinet of an island microgrid and obtaining charge discharge current data of the grounding path of the insulating support after the power frequency voltage crosses zero. Based on the characteristics of charge discharge current and environmental parameters, an environmental-electrical response phase space trajectory is constructed locally, and the geometric shape of the trajectory is used to determine whether the insulation support has a critical instability trend, thus obtaining an abnormal insulation state signal. Calculate the current dew point temperature based on environmental data, and determine whether the temperature inside the ring main unit falls into the high-risk range for condensation, thus obtaining a condensation risk signal; If an insulation condition abnormality signal and a condensation risk signal coexist, high-frequency transient signal acquisition is initiated. Partial discharge characteristics are extracted from the acquired high-frequency transient signals, and the degradation trend of the charge discharge time constant is used for judgment to generate distribution network fault monitoring results.

[0008] Preferably, the charge discharge current data of the grounding path of the insulating support after the power frequency voltage crosses zero is obtained, including: The temperature and relative humidity of the air inside the cabinet are collected to obtain environmental parameters; The three-phase power frequency voltage waveform is detected, and the zero-crossing detection circuit identifies the moment when each phase voltage crosses zero from negative to positive, thus obtaining the local real-time power frequency voltage zero-crossing moment. Within a preset time window after the zero-crossing moment of any phase power frequency voltage is detected, the charge discharge current waveform of the grounding path of the insulating support is collected, and the time axis of the charge discharge current waveform is aligned with the zero-crossing moment of the corresponding phase as the origin to obtain the charge discharge current data.

[0009] Preferably, the environmental-electrical response phase space trajectory is constructed locally, including: Fit an exponential decay curve to the charge discharge current data and calculate the charge discharge time constant; The relative humidity value obtained after each zero crossing of the power frequency voltage is combined with the initial rising slope of the charge discharge current to form a phase space data point, which is then stored in the local non-volatile memory. Read all phase space data points with relative humidity in the pre-condensation range from the local non-volatile memory, and sort them from low to high relative humidity values ​​to form a phase space trajectory sequence; Calculate the rate of change of slope between adjacent data points in the phase space trajectory sequence, and identify whether there are two or more consecutive positive abrupt change segments where the rate of change of slope exceeds a preset threshold. When a positive abrupt change segment is identified, the degradation trend of the charge discharge time constant is determined to reflect the true critical instability of the insulation, generating an abnormal insulation state signal.

[0010] Preferably, the identification process includes identifying whether there are two or more consecutive positive abrupt change segments where the rate of change of slope exceeds a preset threshold, including: In the phase space trajectory sequence, identify all consecutive data point pairs that satisfy the condition that the initial upward slope of the subsequent data point is greater than that of the previous data point and the relative humidity difference between the two points is less than the width of the preset humidity window, and use them as candidate jump intervals. For each candidate jump interval, the ratio of the change in the initial upward slope between the termination point and the starting point to the change in relative humidity is calculated as the jump steepness. When the steepness of at least one candidate leap interval exceeds a preset steepness threshold, it is determined that there is a positive abrupt change segment in the phase space trajectory that represents critical instability.

[0011] As a preferred option, condensation risk signals are obtained, including: Based on the real-time temperature and relative humidity values ​​collected inside the ring network cabinet, the current dew point temperature is calculated through a pre-stored dew point temperature mapping relationship. Compare the current temperature inside the ring main unit with the dew point temperature, and calculate the absolute value of the temperature difference between the two. When the absolute value of the temperature difference is less than the preset condensation threshold, the ring main unit is determined to be in a high-risk condensation zone, and a condensation risk signal is generated.

[0012] Preferably, the generated distribution network fault monitoring results include: When the insulation condition abnormality signal and the condensation risk signal are both valid, the high-frequency transient signal acquisition is triggered to record the high-frequency current pulse waveform on the grounding path. The temporal and polarity features of partial discharge events are extracted from high-frequency current pulse waveforms to form a partial discharge feature set; Cross-validate the partial discharge feature set with the historical degradation trend of the charge discharge time constant. If the two are consistent in time and physical logic, then generate the distribution network fault monitoring result.

[0013] Preferably, a partial discharge feature set is formed, including: Pulse detection is performed on high-frequency current pulse waveforms to identify all transient pulse events whose amplitude exceeds the noise floor and meets the pulse width and monotonicity conditions; For each pulse event, calculate the rise time and peak polarity as time-domain and polarity features; The rise time and peak polarity of all valid pulse events are combined to form a partial discharge feature set.

[0014] Preferably, the partial discharge feature set is cross-validated with the historical degradation trend of the charge discharge time constant, including: Determine whether there are pulse events in the partial discharge feature set whose rise time is less than the threshold for the discharge front and whose peak polarity is consistent with the polarity of the current power frequency voltage half-cycle; Determine whether there are multiple consecutive periods of monotonically increasing time constants in the historical sequence of charge discharge time constants, and confirm the validity of the monotonically increasing trend through positive abrupt change segments of the phase space trajectory; When both of the above judgment results are true, the distribution network fault monitoring result is generated.

[0015] Preferably, the zero-crossing detection circuit identifies the moment when each phase voltage crosses zero from negative to positive, including: The analog voltage signal output by the voltage transformer is sequentially subjected to bandpass filtering and hysteresis comparison to obtain a preliminary zero-crossing pulse signal; The initial zero-crossing pulse signal is logically ANDed with the differential signal of the original voltage signal. Only when the differential signal is positive is the initial zero-crossing pulse signal confirmed as a valid crossover event from negative to positive. The time interval between the effective crossing event and the previous in-phase effective crossing event is measured, and the zero-crossing time corresponding to the effective crossing event whose time interval falls within the preset power frequency cycle tolerance range is recorded as the local real-time power frequency voltage zero-crossing time.

[0016] Secondly, this application also provides a power distribution network fault monitoring system, including: a data acquisition module, which collects environmental parameters in the ring network cabinet of the island microgrid and obtains the charge discharge current data of the grounding path of the insulation support after the power frequency voltage crosses zero; The anomaly detection module constructs an environmental-electrical response phase space trajectory locally based on the characteristics of charge discharge current and environmental parameters, and determines whether the insulation support has a critical instability trend based on the geometric shape of the trajectory, thereby obtaining an insulation state anomaly signal. The condensation risk module calculates the current dew point temperature based on environmental data and determines whether the temperature inside the ring main unit falls into the high-risk range of condensation, thus obtaining a condensation risk signal. If an insulation condition abnormality signal and a condensation risk signal coexist in the fault monitoring module, high-frequency transient signal acquisition will be initiated. The acquired high-frequency transient signals will be partially discharged, and the deterioration trend of the charge discharge time constant will be used to make a judgment, generating a distribution network fault monitoring result.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: collecting environmental parameters inside the ring network cabinet of the island microgrid and obtaining charge discharge current data of the grounding path of the insulating support after the power frequency voltage crosses zero. Based on the characteristics of charge discharge current and environmental parameters, an environmental-electrical response phase space trajectory is constructed locally, and the geometric shape of the trajectory is used to determine whether the insulation support has a critical instability trend, thus obtaining an abnormal insulation state signal. Calculate the current dew point temperature based on environmental data, and determine whether the temperature inside the ring main unit falls into the high-risk range for condensation, thus obtaining a condensation risk signal; If an insulation condition abnormality signal and a condensation risk signal coexist, high-frequency transient signal acquisition is initiated. Partial discharge characteristics are extracted from the acquired high-frequency transient signals, and the degradation trend of the charge discharge time constant is used for judgment to generate distribution network fault monitoring results.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps: Collect environmental parameters inside the ring network cabinet of the island microgrid, and obtain charge discharge current data of the grounding path of the insulating support after the power frequency voltage crosses zero; Based on the characteristics of charge discharge current and environmental parameters, an environmental-electrical response phase space trajectory is constructed locally, and the geometric shape of the trajectory is used to determine whether the insulation support has a critical instability trend, thus obtaining an abnormal insulation state signal. Calculate the current dew point temperature based on environmental data, and determine whether the temperature inside the ring main unit falls into the high-risk range for condensation, thus obtaining a condensation risk signal; If an insulation condition abnormality signal and a condensation risk signal coexist, high-frequency transient signal acquisition is initiated. Partial discharge characteristics are extracted from the acquired high-frequency transient signals, and the degradation trend of the charge discharge time constant is used for judgment to generate distribution network fault monitoring results.

[0019] Implementing this application will have the following beneficial effects: This application provides a method and system for monitoring faults in a power distribution network. 1. This invention constructs an environmental-electrical response phase space trajectory on-site in the ring main unit of an island microgrid and identifies the critical instability trend of insulation based on the steepness of the trajectory. This effectively distinguishes between true insulation degradation caused by salt spray deposition and parameter drift caused by ordinary moisture or sensor noise. At the same time, high-frequency transient acquisition is only initiated when condensation risk signals and insulation abnormality signals coexist, and cross-validation is performed by combining the degradation trend of charge discharge time constant. This avoids false alarms and high power consumption caused by traditional methods due to single threshold criteria or continuous monitoring, and improves the accuracy and energy efficiency of early warning of intermittent high-resistance grounding faults.

[0020] 2. This invention introduces a triple anti-interference mechanism of bandpass filtering, hysteresis comparison, and differential polarity verification in power frequency voltage zero-crossing detection. It can still accurately obtain the local phase reference under islanded operation conditions with frequent system frequency fluctuations. Based on this, the charge discharge current waveform is strictly aligned, ensuring the reliability of phase space trajectory and time constant calculation. It solves the monitoring blind zone problem caused by synchronization failure in the existing technology in environments with no communication and high humidity and high salt, and provides island microgrids with a completely autonomous, low-power, and highly robust local fault monitoring capability. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1This is an overall flowchart of a power distribution network fault monitoring method involved in this application; Figure 2 This is a flowchart of a power distribution network fault monitoring method involved in this application; Figure 3 This is a schematic diagram of the overall structure of an overhead line insulation modification coating method involved in this application; Figure 4 This is a computer equipment diagram of a coating method for insulation modification of overhead lines, which is the subject of this application. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0024] In one exemplary embodiment, such as Figure 1 As shown, a method for monitoring faults in a distribution network is provided, including: S100: Collect environmental parameters inside the ring network cabinet of the island microgrid and obtain charge discharge current data of the grounding path of the insulation support after the power frequency voltage crosses zero; It should be noted that in the isolated operation scenario of island microgrids, the key ring main unit refers to the local switchgear deployed at the diesel generator outlet, photovoltaic grid connection point, or energy storage access point. The insulating support refers to the epoxy resin or ceramic insulators inside the ring main unit used to fix the busbars or cable terminals. The charge discharge current data refers to the microampere-level transient current waveform formed by the residual charge flowing to the grounding point through the deteriorated path on the surface or inside the insulating support after the power frequency voltage crosses zero. Traditional ring main unit insulation monitoring schemes typically use a fixed sampling rate to periodically collect the charge discharge current and assume the system frequency is stable at 50Hz to determine the analysis window after the voltage crosses zero. However, in isolated island microgrids, due to the lag in diesel generator speed regulation response, harmonic injection from photovoltaic inverters, and sudden load changes, the system frequency often fluctuates between 49.0Hz and 51.5Hz. This causes the sampling window based on a fixed time interval to shift from the actual voltage phase, resulting in inaccurate truncation of the charge discharge waveform and consequently affecting the accuracy of the time constant calculation. This application uses the locally detected zero-crossing moment of the power frequency voltage as the synchronous trigger reference. After each zero-crossing event, it dynamically initiates charge discharge current acquisition and aligns the time axis of the acquired data with the zero-crossing moment as the origin. This mechanism frees the entire monitoring process from dependence on system frequency stability, ensuring accurate capture of charge discharge characteristics strictly correlated with voltage phase even under islanded operation conditions with continuous frequency fluctuations, providing a reliable data foundation for subsequent insulation status determination.

[0025] In some embodiments, step S100 includes steps S110, S120, and S130, as follows: S110: Environmental parameters are obtained by real-time collection of temperature and relative humidity values ​​of the air inside the ring network cabinet through temperature and humidity sensors deployed inside the cabinet.

[0026] Understandably, the temperature and humidity sensor is a digital output sensor module installed in the ring main unit near the insulating support to continuously monitor the local microenvironment. The collected temperature values ​​are in degrees Celsius, and the relative humidity values ​​are in percentage. Together, they constitute the input data required for subsequent dew point calculations.

[0027] S120: Utilizes a voltage transformer to detect the three-phase power frequency voltage waveform, and identifies the moment when each phase voltage crosses zero from negative to positive based on the zero-crossing detection circuit, thus obtaining the local real-time power frequency voltage zero-crossing moment.

[0028] Understandably, the zero-crossing detection circuit uses a hardware circuit with a hysteresis comparator or a digital algorithm based on software threshold discrimination to suppress false triggering caused by high-frequency noise in the voltage waveform; the zero-crossing time of each of the three phases A, B, and C is recorded separately to ensure that the phase reference of each phase voltage can still be accurately captured under asymmetrical system operation conditions.

[0029] In some embodiments, step S120 includes steps S121-S123, as follows: S121: The analog voltage signal output by the voltage transformer is sequentially processed by bandpass filtering and hysteresis comparison to obtain a preliminary zero-crossing pulse signal. The passband range of the bandpass filter is 45Hz to 55Hz, and the upper threshold of the hysteresis comparison is +0.5V and the lower threshold is -0.5V.

[0030] The voltage transformer is installed on the three-phase busbar inside the ring main unit, and the secondary side output signal is connected to the analog front-end circuit of the local monitoring device. The bandpass filter is implemented by a second-order active analog filter to suppress the switching harmonics of the photovoltaic inverter and the excitation noise of the diesel generator. The hysteresis comparison is implemented by a Schmitt trigger composed of an operational amplifier. The difference between the positive and negative flip thresholds is 1.0V to prevent the voltage from repeatedly oscillating near zero due to noise, resulting in multiple output pulses.

[0031] Understandably, in island microgrids operating in isolation, the system voltage is often superimposed with high-frequency interference ranging from hundreds of hertz to several kHz. Directly using a zero-threshold comparator would generate a large number of false zero-crossing pulses. This application uses a 45–55Hz bandpass filter to retain the fundamental frequency component, followed by a ±0.5V hysteresis comparator. Only when the voltage signal stably crosses this window is a clean initial zero-crossing pulse signal output, effectively suppressing false triggering caused by high-frequency oscillations and plateau regions.

[0032] For example, in a field test of a microgrid on a certain island, a sudden change in photovoltaic output caused the A-phase voltage waveform to exhibit an oscillation plateau lasting approximately 800 μs near the zero-crossing point, with an amplitude of ±0.8V. If a traditional 0V threshold comparator is used, more than five false flips will occur; however, after using the 45–55Hz bandpass filter of this application, the high-frequency oscillation is attenuated to below 0.1V, and after ±0.5V hysteresis comparison, only one valid preliminary zero-crossing pulse signal is output, significantly improving the detection reliability.

[0033] S122: Perform a logical AND operation between the preliminary zero-crossing pulse signal and the differential signal of the original voltage signal. Only when the differential signal is positive, confirm that the preliminary zero-crossing pulse signal is a valid crossover event from negative to positive.

[0034] The differential signal of the original voltage signal is obtained through an analog differentiating circuit or a digital differential algorithm, with the differential time constant set to 20μs. The logical AND operation is implemented by digital logic gate circuits or bit operation instructions of the embedded processor to ensure that the initial zero-crossing pulse is marked as a valid crossing event only when the instantaneous rate of change of voltage is positive.

[0035] Understandably, voltage waveforms may experience localized oscillations in the zero-crossing region due to insulation flashover or load switching, causing the signal to cross zero multiple times in inconsistent directions. Only a "negative-to-positive" crossing corresponds to the start of the positive half-cycle of the power frequency, serving as the correct synchronization benchmark for subsequent charge discharge analysis. By introducing polarity verification using a differential signal, false zero-crossings caused by falling edge bounce or negative disturbances can be eliminated, ensuring that the captured events strictly correspond to the voltage rising edge.

[0036] For example, during a diesel generator parallel switching process, the B-phase voltage experiences a brief negative spike after crossing zero, causing the initial zero-crossing pulse to be triggered prematurely. At this time, the differential signal is negative, and the AND logic output is low, so the event is judged as an invalid crossover. Only when the voltage truly recovers and crosses zero does the differential signal turn positive, the AND logic output goes high, and the event is confirmed as a valid crossover, thus avoiding synchronization reference offset.

[0037] S123: After confirming a valid crossover event, start the high-resolution time-to-digital converter, use the system master clock as a reference, measure the time interval between the event and the previous in-phase zero-crossing event, and record the zero-crossing edge time of the event whose time interval falls within the preset power frequency cycle tolerance range as the local real-time power frequency voltage zero-crossing time.

[0038] The high-resolution time-to-digital converter has a resolution of 10 nanoseconds and a system master clock frequency of 100MHz. The preset power frequency cycle tolerance range is 19 milliseconds to 21 milliseconds, corresponding to a system frequency of 47.6Hz to 52.6Hz, covering the typical island operation fluctuation range of island microgrids.

[0039] It is understandable that islanded systems may experience abnormal lengthening or shortening of single-cycle periods during sudden load changes or fluctuations in renewable energy output. For example, if a voltage dip occurs due to inverter protection activation, without a verification mechanism, such abnormal zero-crossing events will serve as the synchronization reference, leading to misalignment of subsequent charge discharge windows. This application achieves inherent consistency verification of the synchronization reference by measuring the interval between consecutive in-phase zero-crossings and only accepting events falling within a physically reasonable range, without relying on external clocks or communication calibration.

[0040] For example, in an event where cloud cover caused a sudden 30% drop in photovoltaic output, the C-phase voltage cycle momentarily extended to 23ms. Since 23ms exceeded the 19–21ms tolerance range, this zero-crossing event was discarded. The system continued to use the previous valid cycle to calculate the next expected zero-crossing window and re-locked the synchronization reference when the next normal cycle arrived, ensuring the continuity and accuracy of the monitoring process.

[0041] Preferably, step S121 effectively suppresses the interference of high-frequency harmonics and voltage oscillations on zero-crossing detection in island environments through hardware-level anti-interference design using bandpass filtering and hysteresis comparison; step S122 ensures that only true zero-crossing events with the rising edge of the voltage are captured through differential polarity verification, avoiding misjudgments caused by waveform sway; step S123 filters abnormal zero-crossing events through power frequency cycle tolerance verification, maintaining the long-term stability of the local synchronization reference. The synergistic effect of these three aspects enables the entire zero-crossing moment identification mechanism to continuously output a high-precision and high-reliability local phase reference in island microgrids with no communication, no external synchronization source, and severe voltage distortion, providing a solid foundation for the phase alignment of subsequent charge discharge current and partial discharge signals.

[0042] S130: Within a preset time window after detecting the zero-crossing moment of any phase power frequency voltage, the charge discharge current waveform of the grounding path of the insulating support is acquired by a high-precision micro-current sensor, and the time axis of the waveform is aligned with the zero-crossing moment of the corresponding phase as the origin to obtain the charge discharge current data.

[0043] Understandably, the preset time window is 0 to 1 millisecond, covering the main transient process of charge discharge; the range of the high-precision microcurrent sensor is 0 to 100 microamps, and the resolution is not less than 0.1 microamps; the timestamps of the acquired waveforms are all normalized relative to the zero-crossing moment of the phase voltage that triggered the acquisition, thereby eliminating the influence of system frequency fluctuations on the waveform truncation position.

[0044] Preferably, steps S110 to S130, by synchronously acquiring environmental parameters and phase-aligned charge discharge current waveforms on-site, solve the problem of the failure of traditional monitoring methods caused by frequency drift and communication loss in island microgrids, and provide a time-consistent and phase-accurate original data foundation for subsequent insulation status trend analysis.

[0045] Preferably, step S100 achieves a self-synchronization sampling mechanism without external clock dependence by using the actual voltage zero-crossing time as the local event benchmark, enabling the entire fault monitoring process to be stably executed under extreme conditions of island operation and no master station support, significantly improving the autonomous early warning capability of key nodes in the island distribution network.

[0046] In some embodiments, after obtaining the charge discharge current data, this application further includes: The ratio of the peak-to-peak value of the charge discharge current waveform to the standard deviation of the baseline noise within the analysis interval is calculated as the signal-to-noise ratio indicator. The presence of a monotonically rising segment before the peak and a monotonically falling segment after the peak in the waveform is used as an indicator of morphological integrity. When the signal-to-noise ratio (SNR) is greater than the preset SNR threshold and the morphological integrity meets the monotonicity requirement, the waveform is determined to be a valid waveform; otherwise, it is marked as an invalid waveform and discarded.

[0047] It should be noted that in the isolated operation scenario of island microgrids, the discharge current waveform within the ring main unit is susceptible to external electromagnetic disturbances. Sudden changes in the diesel generator excitation system, high-frequency switching of the photovoltaic inverter, or switching of nearby high-power loads can all introduce non-insulation degradation-related current pulses or baseline drift after the power frequency voltage crosses zero. Such interference signals superimposed on the actual discharge waveform can cause non-monotonic oscillations, local plateau jumps, or a significant decrease in signal-to-noise ratio. If the contaminated waveform is directly used for subsequent insulation status analysis, it will lead to distortion in the calculation of time constant or half-life characteristics, resulting in misjudgment. After completing the discharge current data acquisition for S100, this application performs validity screening on the discharge current waveform for each cycle. First, the ratio of the peak-to-peak value of the waveform within the analysis interval to the baseline noise standard deviation is calculated as the signal-to-noise ratio indicator. The baseline noise standard deviation is obtained by calculating the standard deviation of the sampling points in the stable region at the end of the waveform, and the peak-to-peak value is the difference between the maximum and minimum values ​​within the analysis interval. When the signal-to-noise ratio (SNR) is greater than the preset SNR threshold, the waveform is considered to have sufficient signal strength to support feature extraction.

[0048] Furthermore, the waveform is tested to determine whether there is a strictly monotonically increasing trend from the starting point to the peak value, and whether there is a strictly monotonically decreasing trend from the peak value to the end of the analysis interval, serving as a morphological integrity indicator. Monotonicity is determined by comparing the magnitudes of adjacent sampled values ​​point by point. If any subsequent sampled value in the rising segment is less than the previous sampled value, or any subsequent sampled value in the falling segment is greater than the previous sampled value, then the monotonicity requirement is not met. When the signal-to-noise ratio (SNR) is greater than a preset SNR threshold and the morphological integrity indicator meets the monotonicity requirement, the charge discharge current waveform is determined to be a valid waveform and retained for subsequent insulation state analysis in S200. Otherwise, it is marked as an invalid waveform and discarded, and does not participate in the historical trend sequence update.

[0049] The preset signal-to-noise ratio (SNR) threshold is an empirical value calibrated based on the sensor accuracy of the ring main unit's local device and the noise level of a typical island environment. In this application, it is primarily determined based on the ratio of the background noise of the high-precision micro-current sensor to the amplitude of the typical discharge current in the early stages of insulation degradation. The sensor background noise is characterized by the standard deviation of multiple samples under no-voltage excitation conditions, and the typical discharge current amplitude is obtained through laboratory simulations of salt spray deposition on insulators after the power frequency zero-crossing. The preset SNR threshold can be set as a certain proportion of the ratio of the typical discharge current amplitude to the background noise, adjusted according to the actual deployment environment to ensure that only physically credible discharge events are retained.

[0050] Understandably, by using both signal-to-noise ratio (SNR) and morphological integrity as criteria, this application effectively filters out false discharge waveforms caused by external disturbances. For example, due to synchronous switching of the photovoltaic inverter, a high-frequency oscillation lasting 300 microseconds occurs after the A-phase voltage crosses zero, resulting in a multi-peak structure in the charge discharge waveform with an SNR of only 2.3. Since the waveform does not meet either the SNR threshold or the monotonicity requirement, it is judged as an invalid waveform and discarded. Meanwhile, the unaffected waveforms in adjacent cycles have an SNR of 6.8 and good monotonicity, and are retained for trend analysis. This avoids the long-term impact of a single interference event on insulation condition assessment and significantly improves the reliability of fault precursor identification.

[0051] S200: Based on the characteristics of charge discharge current and environmental parameters, an environmental-electrical response phase space trajectory is constructed locally, and the geometric shape of the trajectory is used to determine whether the insulation support has a critical instability trend, thus obtaining an abnormal insulation state signal.

[0052] It should be noted that in the isolated operation scenario of island microgrids, the fault evolution of insulation support components within the ring main unit is not a uniform degradation process, but rather a nonlinear abrupt change under specific environmental conditions. When surface salt spray deposition reaches a critical humidity level, the insulation performance rapidly transitions from a high-resistivity state to a locally conductive state. This transition manifests in the electrical response as a step-like increase in the sensitivity of charge discharge characteristics to environmental humidity. Traditional insulation monitoring methods only focus on the current amplitude or time constant at a single moment, failing to capture the changes in system behavior before critical instability.

[0053] This application, based on the synchronous acquisition of environmental parameters and charge discharge current data in S100, no longer assesses the strength of a single discharge event in isolation. Instead, it treats the relative humidity and the initial rise slope of the charge discharge current synchronously acquired during multiple power frequency zero-crossing events as a set of correlated observations, constructing an environmental-electrical response phase space trajectory locally. By analyzing the geometric changes of the phase space trajectory, it identifies whether there are nonlinear transition characteristics caused by the activation of ion migration channels on the surface of the insulating material. Specifically, the local device continuously records the relative humidity value measured after each voltage zero-crossing and the corresponding initial rise slope of the charge discharge current, forming a series of phase space data points. When the ambient humidity naturally rises, if the insulation is healthy, the phase space trajectory shows a gentle upward trend; if a humidity-sensitive defect exists, the trajectory will show a steep rise near a certain humidity threshold. This application determines that the system has entered the critical instability region by detecting whether the rate of change of the slope between adjacent data points exceeds a preset threshold, and combining this with the condition that the rise direction is positive, thereby generating an abnormal insulation state signal before condensation actually occurs. The entire solution does not rely on external communication, does not use predictive models, and does not preset fixed thresholds. Instead, it uses the daily natural temperature and humidity cycles in the island environment as probes. By constructing and analyzing the topology of phase space trajectories on-site, it achieves the prospective identification of precursors to intermittent high-resistance grounding faults, significantly improving the autonomy and reliability of the monitoring system.

[0054] In some embodiments, step S200 includes steps S210-S240, as follows: S210: Combine the relative humidity value synchronously acquired after each zero crossing of the power frequency voltage with the initial rising slope of the charge discharge current into a phase space data point and store it in the local non-volatile memory.

[0055] Specifically, the zero-crossing moment of each power frequency voltage is captured in real time by the zero-crossing detection circuit. The system completes the synchronous acquisition of the relative humidity value and the charge discharge current waveform within 800 microseconds after the zero crossing, and calculates the initial rise slope based on the 10% interval from the start of the waveform to the peak value. The relative humidity value and the initial rise slope are combined into a two-dimensional phase space data point and written into the ferroelectric memory of the local device.

[0056] Furthermore, each phase space data point is stored in the form of a structure, which includes the relative humidity value, initial rise slope, collection timestamp, and corresponding phase identifier, ensuring data traceability and supporting subsequent filtering.

[0057] Furthermore, the write operation adopts a circular buffer management strategy. When the memory is full, it automatically overwrites the earliest written valid data point. The buffer capacity is set to 500 data points, which is sufficient to cover all valid events within the typical day-night temperature and humidity cycle of the island.

[0058] It is easy to understand that the data point writing operation is only performed when the charge discharge current waveform meets the conditions of a signal-to-noise ratio greater than four and good morphological monotonicity, so as to avoid low-quality events from contaminating the phase space trajectory.

[0059] Ideally, the relative humidity value is in percentage, and the initial rise rate is in microamps per microsecond, both stored in a 16-bit fixed-point format to balance accuracy and memory efficiency.

[0060] Preferably, the phase identifier is represented by an enumeration type, with phase A being zero, phase B being one, and phase C being two, which facilitates fast program indexing and phase consistency verification.

[0061] S220: Read all phase space data points with relative humidity in the pre-condensation range from the local non-volatile memory, and sort them from low to high relative humidity values ​​to form a phase space trajectory sequence.

[0062] Specifically, the pre-condensation interval is defined as a relative humidity of 60% to 80%. The system traverses all phase space data points in the local non-volatile memory, selects the set of points whose relative humidity values ​​fall within this interval, and arranges them in ascending order according to the relative humidity values, forming a monotonically increasing phase space trajectory sequence.

[0063] Furthermore, the sorting is implemented using the insertion sort algorithm, which is suitable for small datasets and requires no additional memory. The sorting results are stored in the form of an array, and the array index order corresponds to the environmental evolution order from low to high humidity.

[0064] Furthermore, the complete structural information of each data point is preserved during the sorting process, including the initial rising slope, collection timestamp, and phase identifier, ensuring that the trajectory sequence not only reflects the humidity-slope relationship but also supports time dimension backtracking and phase consistency verification.

[0065] It is easy to understand that for multiple data points with the same relative humidity value, they are sorted twice in ascending order according to the collection timestamp to ensure that the trajectory sequence still has a unique and definite arrangement order during the humidity plateau period.

[0066] Preferably, the upper and lower limits of the pre-condensation range can be adjusted on-site to adapt to the condensation characteristics under different island climate conditions.

[0067] Ideally, the sorted phase space trajectory sequence retains only data points from the most recent 24 hours, and points outside the time range are filtered out during the reading phase to reduce invalid calculations.

[0068] S230: Calculate the rate of change of slope between adjacent data points in the phase space trajectory sequence, and identify whether there are two or more consecutive positive abrupt change segments where the rate of change of slope exceeds a preset threshold. In some embodiments, step S230 includes steps S231-S233, as follows: S231: In the phase space trajectory sequence, identify all consecutive data point pairs that satisfy the condition that "the initial upward slope of the next data point is greater than that of the previous data point and the relative humidity difference between the two points is less than the preset humidity window width" as candidate jump intervals.

[0069] Specifically, the system traverses all adjacent data point pairs in the phase space trajectory sequence, compares the initial upward slope of the next data point with the initial upward slope of the previous data point, and calculates the relative humidity difference between the two points.

[0070] Furthermore, the data point pair is marked as a candidate jump interval only when the initial upward slope shows an upward trend and the relative humidity difference is less than the preset humidity window width.

[0071] Furthermore, the preset humidity window width is two percentage points of relative humidity, which reflects the typical humidity variation range required for the formation of conductive channels during salt spray deliquescence.

[0072] It is easy to understand that if the relative humidity difference between two adjacent points is zero or negative, then that point pair is skipped to ensure that only the response characteristics during the monotonic increase of humidity are analyzed.

[0073] Ideally, the preset humidity window width can be configured on-site according to the annual average salt spray deposition rate of the island where the ring main unit is located. The higher the deposition rate, the narrower the window width.

[0074] Ideally, the identification process for candidate leap intervals is performed incrementally after each new data point is added to the trajectory sequence, avoiding full recalculation and improving the response speed of the local device.

[0075] S232: For each candidate jump interval, calculate the ratio of the change in the initial upward slope between the termination point and the starting point to the change in relative humidity, and use this ratio as the jump steepness.

[0076] Specifically, for each candidate jump interval, the system reads the initial rise slope and relative humidity value at its starting and ending points, and calculates the change in the initial rise slope and the change in relative humidity respectively.

[0077] Furthermore, the initial change in the upward slope is divided by the change in relative humidity, and the resulting quotient is the steepness of the jump in the candidate jump range, expressed in microamps per microsecond per percentage humidity.

[0078] Furthermore, the division operation is implemented using fixed-point division, with both the numerator and denominator extended to 32-bit integers to avoid precision loss, and the result retained to 16 decimal places.

[0079] It is easy to understand that if the change in relative humidity is less than 0.5 percentage points, it is considered as environmental fluctuation noise, and this candidate rise interval is not included in the rise steepness calculation.

[0080] Preferably, the steepness calculation result is cached in a temporary register in floating-point format for subsequent threshold comparison, avoiding repeated memory reads.

[0081] Ideally, when the initial change in the upward slope is negative, even if the S231 condition is met, the steepness of the jump should be forced to zero to prevent misjudgment of the falling edge interference.

[0082] S233: When the steepness of at least one candidate leap interval exceeds a preset steepness threshold, it is determined that there is a positive abrupt change segment in the phase space trajectory that represents critical instability.

[0083] Specifically, the system compares the steepness of each candidate leap interval with a preset steepness threshold. If any leap steepness is greater than the threshold, a positive mutation segment determination is triggered.

[0084] Furthermore, the preset steepness threshold is set to 0.5 microamps per microsecond per percentage humidity, a value derived from laboratory statistics on the minimum steepness of salt spray-contaminated insulators before critical flashover.

[0085] Furthermore, the judgment result is stored in the form of a Boolean flag. When the flag is true, it indicates that the phase space trajectory has entered the critical instability region, and when the flag is false, it indicates that the insulation state is stable.

[0086] It is easy to understand that once a positive mutation segment is identified, the result will remain valid until the next complete phase space trajectory reconstruction is completed, thus avoiding frequent flips.

[0087] Ideally, the preset steepness threshold supports remote configuration updates, allowing maintenance personnel to dynamically adjust the sensitivity based on historical early warning accuracy.

[0088] Ideally, if multiple candidate jump intervals exceed the threshold simultaneously, only the time and parameters of the first triggering event should be recorded for log tracing to avoid redundant alarms.

[0089] S240: When a positive abrupt change segment is detected, it is determined that the environmental-electrical response relationship of the insulating support has entered the critical instability region, and an abnormal insulation state signal is generated.

[0090] Specifically, once a positive abrupt change segment indicating critical instability is detected, the local device immediately sets the insulation state abnormality flag and outputs a high-level digital signal as the insulation state abnormality signal.

[0091] Furthermore, an abnormal insulation status signal triggers a local early warning process, including illuminating the yellow LED indicator on the ring main unit panel, recording the event log, and initiating the standby state for high-frequency transient signal acquisition.

[0092] Furthermore, the event logs are written to a dedicated area of ​​non-volatile memory in read-only mode to prevent them from being overwritten by subsequent data. Maintenance personnel can read historical warning records during inspections via infrared interface or Bluetooth module.

[0093] It is easy to understand that the insulation condition abnormality signal is only generated after the positive change segment is confirmed. If the slope drops due to a subsequent decrease in humidity, the generated abnormality signal is not canceled to ensure that no fault precursor events are missed.

[0094] Ideally, after an insulation condition abnormality signal is generated, the system automatically increases the frequency of charge discharge current acquisition from once per power frequency cycle to three times every five power frequency cycles, thereby enhancing the subsequent monitoring density.

[0095] Ideally, if a partial discharge pulse is detected within two hours of the abnormal signal being generated, it will automatically be upgraded to a distribution network fault monitoring result and a trip warning command will be output.

[0096] S300: Calculates the current dew point temperature based on environmental data and determines whether the temperature inside the ring main unit falls into the high-risk range for condensation, thus obtaining a condensation risk signal.

[0097] It should be noted that in the isolated operation scenario of island microgrids, the temperature and humidity of the air inside the ring main unit are directly affected by the maritime climate. Large diurnal temperature differences and relative humidity consistently exceeding 80% make condensation highly likely to form on the metal and insulating surfaces inside the unit. Condensation itself is not a fault, but when salt spray deposits or dirt are present on the surface of the insulating support components, the condensation dissolves the electrolyte, forming a conductive water film. This significantly reduces the local insulation resistance, potentially triggering intermittent high-resistance grounding discharges under power frequency voltage. Therefore, accurately identifying high-risk condensation conditions is a crucial prerequisite for early fault warning.

[0098] Existing technologies typically treat dew point calculation as an independent environmental monitoring function, used only for recording or remote alarms, without deeply integrating condensation risk assessment with insulation status analysis locally. Some solutions set fixed temperature difference thresholds to assess condensation risk, but fail to consider factors such as frequent temperature fluctuations and sensor response lags in island environments, leading to false alarms or missed alarms.

[0099] This application calculates the current dew point temperature based on real-time collected temperature and relative humidity data, using a pre-stored dew point temperature mapping relationship, and compares the current temperature inside the cabinet with the dew point temperature. When the temperature difference between the two is less than a preset condensation threshold, a condensation risk signal is immediately generated. This signal is not used for standalone alarms, but rather as one of the necessary enabling conditions for activating high-frequency transient signal acquisition, ensuring that the high-power monitoring module is activated only during truly high-risk periods, balancing early warning sensitivity and energy efficiency. The entire process requires no communication support and is completed locally within the ring main unit, adapting to the operational constraints of island microgrids with no master station and low power consumption.

[0100] In some embodiments, step S310 includes steps S310-S330, as follows: S310: Based on the real-time temperature and relative humidity values ​​collected inside the ring network cabinet, the current dew point temperature is calculated through a pre-stored dew point temperature mapping relationship; Specifically, the pre-stored dew point temperature mapping relationship is stored in the non-volatile memory of the local device in the form of a two-dimensional lookup table. The horizontal axis represents the temperature value, the vertical axis represents the relative humidity value, and each cell in the table stores the dew point temperature under the corresponding conditions.

[0101] Furthermore, the lookup table has a temperature resolution of one degree Celsius and a relative humidity resolution of five percent, covering a temperature range of 0 to 50 degrees Celsius and a relative humidity range of 10 percent to 100 percent.

[0102] Furthermore, when calculating the current dew point temperature, bilinear interpolation is used to interpolate between four neighboring points in the lookup table, thereby improving the calculation accuracy.

[0103] It is easy to understand that when the collected temperature or relative humidity exceeds the range of the lookup table, the boundary value extrapolation method is used to determine the dew point temperature to ensure the continuity of the calculation.

[0104] Ideally, the lookup table is obtained through laboratory calibration, by measuring the actual dew point under different temperature and humidity combinations in a standard temperature and humidity chamber, and then discretizing and storing the resulting measured dataset.

[0105] Ideally, bilinear interpolation is performed in the fixed-point domain, avoiding the overhead of floating-point operations and making it suitable for low-power embedded platforms.

[0106] S320: Compare the current temperature value inside the ring main unit with the dew point temperature, and calculate the absolute value of the temperature difference between the two. Specifically, the system reads the current temperature sensor output value and the dew point temperature calculated by S310, performs a subtraction operation to obtain the temperature difference, and then takes the absolute value to obtain the absolute value of the temperature difference.

[0107] Furthermore, the absolute value of the temperature difference is represented by a sixteen-digit unsigned integer, in units of one-tenth of a degree Celsius, with one decimal place of precision.

[0108] Furthermore, the two temperature values ​​are aligned and calibrated before the subtraction operation to compensate for system errors caused by sensor zero-point drift.

[0109] It is easy to understand that the absolute value of the temperature difference is used to determine the risk of condensation. The smaller the value, the closer the temperature inside the cabinet is to the dew point, and the higher the probability of condensation.

[0110] Ideally, the temperature sensor and humidity sensor share the same sampling clock to ensure that their data are synchronized and to avoid distortion in temperature difference calculations due to asynchronous sampling.

[0111] Ideally, intermediate variables should be cleared immediately after the absolute value of the temperature difference is calculated to release register resources and reduce memory usage.

[0112] S330: When the absolute value of the temperature difference is less than the preset condensation threshold, the ring main unit is determined to be in a high-risk condensation zone, and a condensation risk signal is generated.

[0113] Specifically, the preset condensation threshold is stored in the configuration register of the local device, with a default value of five, in units of one-tenth of a degree Celsius, or 0.5 degrees Celsius.

[0114] Furthermore, the system compares the absolute value of the temperature difference with the preset condensation threshold. If the absolute value of the temperature difference is less than the preset condensation threshold, the condensation risk flag is set.

[0115] Furthermore, after the condensation risk flag is set, a high-level digital signal is output as a condensation risk signal for use by subsequent logic.

[0116] It is easy to understand that condensation risk signals only reflect the current environmental risk status and do not include historical trend information.

[0117] Ideally, the preset condensation threshold can be updated on-site, allowing maintenance personnel to adjust the sensitivity according to seasonal changes on the island, lowering it during the rainy season and raising it during the dry season.

[0118] Ideally, after the condensation risk signal is generated, the high-frequency environmental data recording mode is automatically activated, shortening the temperature and humidity sampling cycle from one minute to ten seconds, and continuing for ten minutes, which facilitates post-event analysis.

[0119] S400: If an insulation condition abnormality signal and a condensation risk signal coexist, high-frequency transient signal acquisition is initiated. The acquired high-frequency transient signal is subjected to partial discharge feature extraction, and the judgment is made in combination with the deterioration trend of the charge discharge time constant to generate the distribution network fault monitoring result.

[0120] It should be noted that in the isolated operation scenario of island microgrids, salt spray deposited on the surface of the internal insulation support components of the ring main unit can easily form local conductive channels in high humidity environments. However, these channels only conduct momentarily under condensation conditions, resulting in intermittent high-resistance grounding discharges. These discharges are weak in energy and short in duration, and partially overlap in the frequency spectrum with external electromagnetic interference such as diesel generator start-up and shutdown, and photovoltaic inverter switching. Relying solely on the amplitude or frequency of high-frequency transient signals for judgment can easily lead to false alarms. Existing technologies typically treat partial discharge monitoring as an independent functional module, continuously activating high-frequency acquisition, resulting in excessive power consumption of local devices; or triggering alarms after detecting a single abnormal signal, lacking cross-verification of the fault's authenticity. Especially under conditions without communication, it is impossible to rely on multi-source data fusion from the master station, requiring the local judgment logic to have high confidence.

[0121] This application initiates high-frequency transient signal acquisition only when both insulation condition abnormality signals and condensation risk signals are present, avoiding energy waste caused by continuous monitoring. The acquired high-frequency transient signals are not only used to extract partial discharge pulse characteristics, but also further cross-referenced with the historical degradation trend of the charge discharge time constant: only when a local discharge event occurs against the backdrop of continuously deteriorating insulation performance is it determined to be a valid fault precursor. This dual verification mechanism effectively distinguishes between discharges caused by actual insulation defects and external electromagnetic interference, improving the accuracy and reliability of local fault monitoring in island microgrids.

[0122] S410: When the insulation condition abnormality signal and the condensation risk signal are both valid, the high-frequency transient signal acquisition is triggered, and the high-frequency current pulse waveform on the grounding path is recorded. Specifically, the start of high-frequency transient signal acquisition is controlled by a dual-signal synchronous detection circuit. The dual-signal synchronous detection circuit adopts a hardware AND gate structure, and only outputs an enable pulse when the insulation state abnormal signal and the condensation risk signal overlap in time and the duration exceeds the minimum window required for the physical formation of condensation.

[0123] Furthermore, the enable pulse activates the power supply link between the low-noise analog front-end and the high-speed ADC, while simultaneously disabling the clock gating of non-critical digital modules, thereby achieving dynamic power consumption management and adapting to the energy-constrained environment of island ring network cabinets that rely on photovoltaic-battery power supply.

[0124] Furthermore, the acquisition window is strictly aligned with the zero-crossing moment of the power frequency voltage, and the local phase reference extracted from S120 is used as the sampling start trigger to eliminate the waveform truncation offset caused by system frequency drift.

[0125] It is easy to understand that if any signal fails during the enable setup period, the power supply of the high-frequency acquisition link will be immediately cut off and the system will fall back to sleep mode to prevent unnecessary power consumption caused by signal jitter.

[0126] Preferably, the length of the preset synchronization window is set according to the typical condensation formation dynamics of the island, ensuring that insulation degradation and moisture activation are truly coupled on a physical time scale, and avoiding false triggering caused by instantaneous signal overlap.

[0127] Preferably, the power supply for high-frequency transient signal acquisition is provided by an independent low-noise LDO, whose enable signal is controlled by a dual-signal AND gate for insulation abnormality and condensation risk, achieving hardware-level energy saving and significantly extending the operating life of the ring main unit's local device in photovoltaic-battery power supply mode.

[0128] S420: Extract the time-domain and polarity features of partial discharge events from high-frequency current pulse waveforms to form a partial discharge feature set; Specifically, step S420 includes steps S421 to S423: S421: Performs pulse detection on high-frequency current pulse waveforms to identify all transient pulse events with amplitudes exceeding three times the noise floor. Specifically, the pulse detection uses an adaptive threshold comparator, whose threshold is dynamically set by the real-time estimated baseline noise standard deviation, and an upper limit clamping mechanism is introduced to deal with sudden strong interference.

[0129] Furthermore, the baseline noise estimation is based on sliding window statistics of the stable region at the end of the waveform, which is far from the power frequency zero-crossing point, to avoid noise assessment contamination from the charge discharge main pulse.

[0130] Furthermore, candidate pulse events must simultaneously meet the triple conditions of amplitude, width, and rise time monotonicity to exclude spurious pulses caused by switching transients or sensor ringing.

[0131] It is easy to understand that if the interval between adjacent pulses is less than the typical discharge repetition period, they are merged into a composite event to reflect the spatial coupling characteristics of multi-point micro-discharge in the salt spray environment of the island.

[0132] Preferably, the upper limit of the dynamic threshold is set according to the baseline level of the electromagnetic environment in the ring main unit. This baseline is determined by long-term monitoring of the maximum interference amplitude under diesel generator start-up and shutdown and inverter switching events to prevent loss of sensitivity under strong interference.

[0133] Preferably, the merging logic for candidate pulse events takes into account the physical characteristics of multi-point discharge in an island environment. When multiple micro-discharge sources are spatially adjacent, their electromagnetic coupling will form a composite pulse. Merging processing can avoid duplicate counting.

[0134] S422: Calculate the rise time and peak polarity for each pulse event as time-domain and polarity characteristics; Specifically, the rise time is calculated using a high-precision time interpolation algorithm. This algorithm reconstructs the local waveform by utilizing the slope information between sampling points, breaking through the limitations of the original sampling rate and accurately capturing nanosecond-level leading-edge changes.

[0135] Furthermore, the peak polarity determination is tied to the half-cycle state of the power frequency voltage, which is derived from the zero-crossing time sequence in S120, ensuring that the discharge phase assignment is consistent with the direction of the actual applied electric field.

[0136] Furthermore, a partial discharge is considered valid only when the pulse occurs after the rising edge of the voltage half-cycle and the polarity is matched, thus eliminating negative disturbances or external electromagnetic crosstalk.

[0137] It is easy to understand that rise time and polarity together constitute the "fingerprint characteristics" of the discharge source, which are used to distinguish between internal air gap discharge, surface discharge, and external interference.

[0138] Ideally, the rise time interpolation calculation is combined with sampling clock phase correction to compensate for the sampling point offset caused by system frequency fluctuations, ensuring that the physical scale of the discharge power source can still be accurately reflected under islanded operation conditions.

[0139] Ideally, the determination of peak polarity is strictly aligned with the zero-crossing detection result of power frequency voltage, using the same hardware comparator output as a phase reference to eliminate polarity misjudgment caused by digital processing delay.

[0140] S423: Combine the rise time and peak polarity of all valid pulse events into a partial discharge feature set; Specifically, the partial discharge feature set is organized in the form of structured data blocks, including the occurrence time, phase, rise time, polarity, and original waveform index of each pulse, supporting fast retrieval and backtracking.

[0141] Furthermore, the feature set is stored in a dedicated non-volatile memory partition, which employs a wear leveling strategy to extend the lifespan of the local device in scenarios with frequent alarms.

[0142] Furthermore, after the feature set is generated, the feature validity is automatically verified. The verification includes the number of pulses, the uniformity of phase distribution, and the energy concentration, and filters out isolated noise events.

[0143] It is easy to understand that if the feature set passes the verification, it is passed to S430 for cross-validation; otherwise, the current acquisition result is discarded and the system returns to the low-power monitoring state.

[0144] Preferably, the memory region of the partial discharge feature set adopts a double buffer structure. When the current acquisition is written to buffer A, the CPU can read buffer B for analysis, realizing parallel acquisition and processing and improving the real-time performance of the local device.

[0145] Preferably, the remote reading interface of the feature set adopts infrared modulation encoding to avoid introducing an additional radio frequency transmission module and meet the electromagnetic compatibility requirements of the island ring main unit.

[0146] S430: Cross-validate the partial discharge feature set with the historical degradation trend of the charge discharge time constant. If the two are consistent in time and physical logic, generate the distribution network fault monitoring result.

[0147] Specifically, step S430 includes steps S431 to S433: S431: Determine whether there are pulse events in the partial discharge feature set whose rise time is less than the threshold for the discharge front and whose peak polarity is consistent with the polarity of the current power frequency voltage half-cycle; Specifically, polarity consistency verification is achieved through a hardware phase latch, which captures the voltage polarity each time the power frequency crosses zero and maintains a stable output during high-frequency acquisition for pulse polarity comparison.

[0148] Furthermore, the rise time threshold is set according to the type and structure of the insulation material. Epoxy resin supports use a shorter threshold to capture internal defects, while ceramic insulators use a slightly wider threshold to contain surface discharges.

[0149] Furthermore, the discharge event is only confirmed as valid when multiple pulses with the same phase exist within the same power frequency cycle, thus enhancing the confidence level of the judgment through repeatability.

[0150] It is easy to understand that pulses with inconsistent polarity or excessively long rise times are marked as interference and are not included in subsequent cross-validation.

[0151] The discharge front discrimination threshold is set based on the type of insulation material and typical discharge modes. For epoxy resin insulating supports, the rise time of internal air gap discharge is typically less than 50 nanoseconds, while surface discharge is slightly longer; the rise time of external electromagnetic interference is generally greater than 100 nanoseconds. Therefore, the discharge front discrimination threshold is set between that of typical internal discharge and external interference to effectively distinguish between actual insulation defect discharge and environmental noise.

[0152] In the high-salt-fog environment of islands, this threshold can be further adaptively adjusted by combining historical discharge data to ensure that the discrimination sensitivity is maintained at different aging stages.

[0153] Preferably, the derivation of the half-cycle polarity of the power frequency voltage is based on the accurate recording of the zero-crossing time in S120, without relying on the 50Hz assumption, ensuring that the discharge phase assignment can still be correctly determined when the frequency drifts to 49.2Hz or 51.3Hz.

[0154] Ideally, the rise time threshold setting distinguishes between two fault modes: internal insulation discharge and surface discharge. The former corresponds to a shorter rise time, while the latter is slightly longer. This dual-threshold mechanism enhances the fault type identification capability.

[0155] S432: Determine whether the historical degradation trend of the charge discharge time constant is in a state of continuous deterioration; Specifically, the degradation trend is determined using a sliding window monotonicity test, with the window length dynamically matched to the validity period of the condensation risk signal to ensure that the assessment is based on data within the same environmental excitation cycle.

[0156] Furthermore, the monotonicity test requires that all adjacent time constants within the window satisfy an increasing relationship, and that the overall slope is greater than the noise fluctuation level, thus excluding the influence of random drift.

[0157] Furthermore, the historical data comes from the phase space trajectory constructed in S200, ensuring that it is of the same physical and spatial origin as the current partial discharge event.

[0158] It is easy to understand that if the time constant shows a plateau or a downward trend, the insulation condition is considered stable, and no alarm will be triggered even if partial discharge is detected.

[0159] Ideally, the determination of the monotonically increasing trend of the time constant introduces a sliding window consistency test. The deterioration state is only confirmed when all adjacent point pairs within the window satisfy the increasing relationship, thus eliminating misjudgments caused by single-point disturbances.

[0160] Ideally, the assessment period for historical degradation trends should be dynamically aligned with the validity period of condensation risk signals to ensure that cross-validation is based on data within the same environmental stimulus window, thereby enhancing physical and logical consistency.

[0161] S433: When the judgment results of S431 and S432 are both true, the distribution network fault monitoring result is generated.

[0162] Specifically, the results of power distribution network fault monitoring are output as high-level digital signals through a hardware state machine. This state machine is driven by three conditions: insulation abnormality, condensation risk, and local discharge polarity consistency, ensuring that the alarm path is activated only when the physical mechanism is self-consistent.

[0163] Furthermore, the fault state triggers a local alarm process, which adopts a hierarchical response mechanism: first, a red LED indicator is lit to provide a visual warning; at the same time, the current environmental parameters, charge discharge characteristics, and high-frequency transient waveform snapshots are written to the read-only log area to prevent subsequent data overwriting; and the last effective buffer of the high-frequency acquisition module is locked for on-site diagnosis by maintenance personnel.

[0164] Furthermore, once the fault monitoring result is set, it enters a locked state, which is unaffected by subsequent signal fluctuations and can only be released by a physical reset button or a remotely authenticated command, ensuring that fault evidence can still be preserved in the event of communication interruption or program crash.

[0165] It is easy to understand that if either criterion S431 or S432 is not met, the state machine remains in the monitoring ready state, the alarm path is not activated, and the system continues to perform the low-power background monitoring task to avoid malfunctions caused by a single interference.

[0166] Preferably, the fault flag register's locking mechanism employs hardware write protection, allowing reset only via a dedicated physical button or encrypted command, preventing alarm loss due to software malfunctions.

[0167] Ideally, optocoupler isolation and RC filtering should be added before the relay contact output to suppress electromagnetic interference from the high-frequency transient signal acquisition circuit to the protection output, ensuring reliable operation.

[0168] It should be understood that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0169] Based on the same inventive concept, this application also provides a distribution network fault monitoring system. The solution provided by this system is similar to the solution described in the above method. Therefore, the specific limitations of one or more distribution network fault monitoring system embodiments provided below can be found in the limitations of the distribution network fault monitoring method above, and will not be repeated here.

[0170] In one exemplary embodiment, such as Figure 3 As shown, a power distribution network fault monitoring system is provided, including: The data acquisition module collects environmental parameters within the ring network cabinet of the island microgrid and obtains charge discharge current data of the grounding path of the insulating support after the power frequency voltage crosses zero. The anomaly detection module constructs an environmental-electrical response phase space trajectory locally based on the characteristics of charge discharge current and environmental parameters, and determines whether the insulation support has a critical instability trend based on the geometric shape of the trajectory, thereby obtaining an insulation state anomaly signal. The condensation risk module calculates the current dew point temperature based on environmental data and determines whether the temperature inside the ring main unit falls into the high-risk range of condensation, thus obtaining a condensation risk signal. If an insulation condition abnormality signal and a condensation risk signal coexist in the fault monitoring module, high-frequency transient signal acquisition will be initiated. The acquired high-frequency transient signals will be partially discharged, and the deterioration trend of the charge discharge time constant will be used to make a judgment, generating a distribution network fault monitoring result.

[0171] The modules in the aforementioned power distribution network fault monitoring system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0172] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for detecting faults in power distribution lines. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0173] Those skilled in the art will understand that Figure 4 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.

[0174] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0175] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0176] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0177] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0178] 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. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0179] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0180] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of monitoring a power distribution network for faults, characterised by, The application relates to a power distribution network fault monitoring method based on a local environment-electricity response phase space trajectory. The method comprises the following steps: environmental parameters in an island micro-grid ring network cabinet are collected, and charge discharge current data of an insulation support grounding path after a power frequency voltage zero-crossing point are obtained; based on charge discharge current characteristics and environmental parameters, a local environment-electricity response phase space trajectory is constructed, and whether the insulation support has a critical instability trend is judged according to the geometric shape of the trajectory, so that an insulation state abnormal signal is obtained; current dew point temperature is calculated according to environmental data, and whether the temperature in the ring network cabinet falls into a high risk interval of condensation is judged, so that a condensation risk signal is obtained; 2. A method of supervising a power distribution network for faults as claimed in claim 1, characterized in that: if the insulation state abnormal signal and the condensation risk signal exist at the same time, high-frequency transient signal collection is started, partial discharge characteristics of the collected high-frequency transient signal are extracted, and a power distribution network fault monitoring result is generated by combining a deterioration trend of a charge discharge time constant. The charge discharge current data of the insulation support grounding path after the power frequency voltage zero-crossing point are obtained, and the method comprises the following steps: temperature and relative humidity values of air in the cabinet are collected to obtain environmental parameters; three-phase power frequency voltage waveforms are detected, and the time when each phase voltage crosses zero from negative to positive is identified based on a zero-crossing detection circuit to obtain a local real-time power frequency voltage zero-crossing time; 3. A method of supervising a power distribution network for faults as claimed in claim 1, characterized in that: within a preset time window after the zero-crossing time of any phase power frequency voltage is detected, charge discharge current waveforms of the insulation support grounding path are collected, and the time axis of the charge discharge current waveforms is aligned with the zero-crossing time of the corresponding phase as the origin, so that the charge discharge current data are obtained. The local environment-electricity response phase space trajectory is constructed, and the method comprises the following steps: an exponential decay curve is fitted to the charge discharge current data, and a charge discharge time constant is calculated; a relative humidity value obtained after the power frequency voltage zero-crossing and an initial rising slope of the charge discharge current are combined into a phase space data point, and the phase space data point is stored in a local nonvolatile memory; all phase space data points with relative humidity in a pre-condensation interval are read from the local nonvolatile memory, and are sorted from low to high according to the relative humidity values to form a phase space trajectory sequence; a slope change rate between adjacent data points in the phase space trajectory sequence is calculated, and whether there is a positive mutation segment with more than two continuous slope change rates exceeding a preset jump threshold value is identified; 4. A power distribution network fault monitoring method as claimed in claim 3, characterized in that: when the positive mutation segment is identified, a deterioration trend of the charge discharge time constant is determined, and an insulation state abnormal signal is generated. The method for identifying whether there is a positive mutation segment with more than two continuous slope change rates exceeding a preset jump threshold value comprises the following steps: in the phase space trajectory sequence, all continuous data point pairs satisfying that the initial rising slope of the next data point is greater than that of the previous data point and the relative humidity difference between the two points is less than a preset humidity window width are identified as candidate jump intervals; for each candidate jump interval, a ratio of an initial rising slope change amount between a terminal point and a starting point to a relative humidity change amount is calculated as a jump steepness; 5. A method of supervising a power distribution network for faults as claimed in claim 1, characterized in that: when the jump steepness of at least one candidate jump interval exceeds a preset steepness threshold value, it is determined that there is a positive mutation segment representing critical instability in the phase space trajectory. The condensation risk signal is obtained, and the method comprises the following steps: based on the temperature and relative humidity values collected in the ring network cabinet, a current dew point temperature is calculated through a pre-stored dew point temperature mapping relationship; Compare the current temperature value in the ring main unit with the dew point temperature, and calculate the absolute value of the temperature difference between the current temperature value and the dew point temperature; When the absolute value of the temperature difference is less than the preset condensation threshold, it is determined that the ring main unit is in a high risk interval of condensation, and a condensation risk signal is generated.

6. A method of supervising a power distribution network for faults as claimed in claim 1, characterized in that: Generate a power distribution network fault monitoring result, including: When the insulation state abnormal signal and the condensation risk signal are valid at the same time, trigger high-frequency transient signal acquisition, and record the high-frequency current pulse waveform on the grounding path; Extract the time domain and polarity characteristics of partial discharge events from the high-frequency current pulse waveform to form a partial discharge feature set; Cross-verify the partial discharge feature set with the historical degradation trend of the charge leakage time constant. If they are consistent in time and physical logic, generate a power distribution network fault monitoring result.

7. A method of supervising a power distribution network for faults as claimed in claim 6, characterized in that: Forming a partial discharge feature set includes: Detect the high-frequency current pulse waveform to identify all transient pulse events with an amplitude exceeding the noise floor and meeting the pulse width and rising edge monotonicity conditions; Calculate the rise time and peak polarity for each transient pulse event, and remove transient pulse events with peak polarity inconsistent with the current power frequency voltage half cycle polarity to obtain valid transient pulse events; Combine the rise time and peak polarity of all valid transient pulse events into a partial discharge feature set.

8. A method of supervising a power distribution network for faults as claimed in claim 6, characterized in that: Cross-verify the partial discharge feature set with the historical degradation trend of the charge leakage time constant, including: Determine whether there are pulse events in the partial discharge feature set with a rise time less than the discharge front threshold and a peak polarity consistent with the current power frequency voltage half cycle polarity; Determine whether there are consecutive multiple periods of time constants that are monotonically increasing in the historical sequence of the charge leakage time constant, and the validity of the monotonic increasing trend has been confirmed by the positive mutation segment of the phase space trajectory; When both of the above two determinations are true, generate a power distribution network fault monitoring result.

9. A method of supervising a power distribution network for faults as claimed in claim 2, wherein: Identify the time when each phase voltage crosses zero from negative to positive based on a zero-crossing detection circuit, including: Band-pass filter and hysteresis comparison processing are sequentially performed on the analog voltage signal output by the voltage transformer to obtain a preliminary zero-crossing pulse signal; Perform a logical AND operation on the preliminary zero-crossing pulse signal and the differential signal of the original voltage signal. Only when the differential signal is positive, the preliminary zero-crossing pulse signal is confirmed as a valid crossing event from negative to positive; Measure the time interval between the valid crossing event and the previous same-phase valid crossing event, and record the zero-crossing edge time corresponding to the valid crossing event whose time interval falls within the preset power frequency period tolerance range as the local real-time power frequency voltage zero-crossing time.

10. A power distribution network failure monitoring system employing the power distribution network failure monitoring method according to any one of claims 1 to 9, characterized by It includes: A data acquisition module acquires environmental parameters in the ring main unit of the island microgrid and obtains charge leakage current data of the grounding path of the insulating support after the power frequency voltage zero-crossing point; An abnormality judgment module constructs an environmental-electric response phase space trajectory based on the charge leakage current characteristics and the environmental parameters, and determines whether the insulating support has a critical instability trend according to the geometric shape of the trajectory to obtain an insulation state abnormal signal; A condensation risk module calculates the current dew point temperature according to the environmental data, and determines whether the temperature in the ring main unit falls into a high risk interval of condensation to obtain a condensation risk signal; The fault monitoring module starts high-frequency transient signal collection if the insulation state abnormal signal and the condensation risk signal exist simultaneously, extracts partial discharge characteristics from the collected high-frequency transient signal, and judges in combination with a deterioration trend of a charge discharge time constant to generate a distribution network fault monitoring result.

Images