Method and system for monitoring residual current of alternating current power supply for station, equipment and storage medium

Abstract

The invention provides a station AC power supply residual current monitoring method and system, equipment and a storage medium, and belongs to the technical field of power monitoring, and the method comprises the steps: obtaining first residual current monitoring data of a zero-sequence current sensor and second residual current monitoring data of a leakage current sensor; performing feature extraction on the first residual current monitoring data to obtain a first feature vector; performing feature extraction on the second residual current monitoring data to obtain a second feature vector; fusing the first feature vector and the second feature vector to obtain a target feature vector of the residual current of the station AC power supply; and analyzing the residual current of the station AC power supply based on the target feature vector, and determining the cause of the residual current of the station AC power supply. According to the station AC power supply residual current monitoring method and system, the equipment and the storage medium provided by the invention, the accuracy of fault judgment can be improved.

Description

Technical Field

[0001] This application belongs to the field of power monitoring technology, and more specifically, relates to a method, system, equipment, and storage medium for monitoring the residual current of station AC power supplies. Background Technology

[0002] The AC power supply system for substations provides power to control, protection, and communication equipment within the substation, and its safe and reliable operation is crucial to the entire substation. Residual current (also known as leakage current) monitoring is a key technology for ensuring the safety of low-voltage AC systems, enabling timely detection of insulation faults and grounding anomalies. Traditional residual current monitoring often relies on a single sensor (such as a zero-sequence current transformer or residual current operated protective device) to monitor three-phase unbalanced current or leakage current, which has certain limitations: for example, measurement accuracy may be affected by interference in complex electromagnetic environments, a single sensor failure can lead to monitoring failure, and it is difficult to distinguish between different types of leakage current (AC and DC components). With the development of smart grids, the integration of distributed power sources and power electronic equipment has made the residual current composition more complex, including multiple components such as AC and DC. A single sensor often cannot fully reflect this complex residual current characteristic, thus affecting the accuracy of fault diagnosis. Summary of the Invention

[0003] The purpose of this application is to provide a method, system, device, and storage medium for monitoring residual current of station AC power supplies, so as to improve the accuracy of fault diagnosis.

[0004] A first aspect of this application provides a method for monitoring the residual current of a station AC power supply, comprising: Acquire the first residual current monitoring data of the zero-sequence current sensor and the second residual current monitoring data of the leakage current sensor; the zero-sequence current sensor is installed on the low-voltage side of the transformer of the station AC power system, and the leakage current sensor is installed in the feeder circuit of the station AC power system, and the feeder circuit contains multiple sensors. Feature extraction is performed on the first residual current monitoring data to obtain first feature data, and the first feature data is transformed to obtain a first feature vector corresponding to the first residual current monitoring data; feature extraction is performed on the second residual current monitoring data to obtain second feature data, and the second residual current monitoring data is transformed to obtain a second feature vector corresponding to the second residual current monitoring data. The first and second eigenvectors are fused to obtain the target eigenvector of the residual current of the station AC power supply. The residual current of the station AC power supply is analyzed based on the target feature vector to determine the cause of the residual current in the station AC power supply.

[0005] A second aspect of this application provides a residual current monitoring system for station AC power supplies, comprising: The data acquisition module is used to acquire the first residual current monitoring data of the zero-sequence current sensor and the second residual current monitoring data of the leakage current sensor. The zero-sequence current sensor is installed on the low-voltage side of the transformer of the station AC power supply system, and the leakage current sensor is installed in the feeder circuit of the station AC power supply system, and the feeder circuit contains multiple sensors. The feature extraction module is used to extract features from the first residual current monitoring data to obtain first feature data, and to transform the first feature data to obtain a first feature vector corresponding to the first residual current monitoring data; and to extract features from the second residual current monitoring data to obtain second feature data, and to transform the second residual current monitoring data to obtain a second feature vector corresponding to the second residual current monitoring data. The feature fusion module is used to fuse the first feature vector and the second feature vector to obtain the target feature vector of the residual current of the station AC power supply. The analysis module is used to analyze the residual current of the station AC power supply based on the target feature vector and determine the cause of the residual current in the station AC power supply.

[0006] A third aspect of this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the above-described method for monitoring the residual current of a station AC power supply.

[0007] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described method for monitoring the residual current of a station AC power supply.

[0008] The beneficial effects of the residual current monitoring method, system, equipment, and storage medium for station AC power supplies provided in this application are as follows: This application's embodiments acquire the total residual current of the system by setting a zero-sequence current sensor on the low-voltage side of the transformer, while simultaneously deploying leakage current sensors in multiple feeder circuits to collect local data, forming a three-dimensional monitoring network of overall and local components. This avoids monitoring failure caused by a single sensor malfunction and can offset the interference of complex electromagnetic environments through cross-verification of multi-source data. Secondly, by extracting features from the two types of data and converting them into vectors, the complex residual current characteristics containing AC and DC components can be accurately captured, overcoming the shortcomings of traditional methods in distinguishing current types. Furthermore, the target feature vector obtained through feature fusion integrates the overall leakage trend and the local contributions of each circuit. Combined with electrical principles and intelligent analysis, it can accurately identify whether the leakage originates from a specific circuit equipment fault, the superposition of normal leakage currents from multiple circuits, or a hidden circuit problem. This process, from multi-source acquisition to feature fusion to intelligent diagnosis, comprehensively reflects the complex residual current characteristics, ultimately significantly improving the accuracy of fault diagnosis. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A flowchart illustrating a method for monitoring the residual current of a station AC power supply according to an embodiment of this application; Figure 2 A structural block diagram of a station AC power supply residual current monitoring system provided in an embodiment of this application; Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0011] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0013] Please refer to Figure 1 , Figure 1 This is a flowchart illustrating a method for monitoring the residual current of a station AC power supply according to an embodiment of this application. The method can be executed by an electronic device and may include: S101: Acquire the first residual current monitoring data of the zero-sequence current sensor and the second residual current monitoring data of the leakage current sensor; the zero-sequence current sensor is installed on the low-voltage side of the transformer of the station AC power system, and the leakage current sensor is installed in the feeder circuit of the station AC power system, and the feeder circuit contains multiple sensors.

[0014] In this embodiment, the zero-sequence current sensor is installed on the low-voltage side of the transformer. Its function is to monitor the total residual current (first residual current monitoring data) of the entire station AC power supply system. Under normal circumstances, the vector sum of the three-phase currents is zero, and the zero-sequence current sensor has no output. When there is a ground leakage current in the station AC power supply system, the three-phase currents are unbalanced, and the zero-sequence current sensor can sense the signal related to the leakage current, reflecting the overall leakage status of the station AC power supply system.

[0015] In this embodiment, the raw zero-sequence current data collected by the zero-sequence current sensor can be processed and transformed into first residual current monitoring data that reflects the actual leakage current state of the station's AC power supply.

[0016] For example, the raw zero-sequence current data collected by the zero-sequence current sensor is preprocessed to eliminate interference from normal loop current: the raw zero-sequence current data collected by the zero-sequence current sensor is subjected to 50Hz bandpass filtering to retain the power frequency leakage current signal and filter out high-frequency switching noise (such as high-frequency harmonics generated by inverters and rectifiers) and DC bias interference.

[0017] Feature separation is performed on the preprocessed signal to extract the effective residual current component: Wavelet transform is used to perform time-frequency decomposition on the preprocessed signal, decomposing it into a fundamental frequency component (50Hz), harmonic components, and transient components. The fundamental frequency component mainly reflects continuous leakage, while the transient components reflect sudden leakage (such as the instantaneous insulation breakdown). A zero-sequence current threshold is set for normal operation; when the amplitude of a component exceeds this threshold, it is determined to be an effective residual current component. After the above processing, the first residual current monitoring data is finally output.

[0018] In this embodiment, multiple leakage current sensors are included, with one sensor installed on each feeder circuit. The leakage current sensors are used to collect the local residual current (second residual current monitoring data) of each branch circuit. Each feeder circuit is a sub-unit of the station AC power supply system. When equipment in a feeder circuit experiences insulation damage, line aging, or other faults, local leakage may occur. The leakage current sensor can be used to locate the leakage situation of a specific branch.

[0019] In this embodiment, the first residual current monitoring data can reflect whether there is leakage in the AC power supply system of the station, while the second residual current monitoring data of each feeder circuit can further narrow down the fault range and avoid the generality caused by judging only by the total current.

[0020] S102: Extract features from the first residual current monitoring data to obtain first feature data, and transform the first feature data to obtain a first feature vector corresponding to the first residual current monitoring data; extract features from the second residual current monitoring data to obtain second feature data, and transform the second residual current monitoring data to obtain a second feature vector corresponding to the second residual current monitoring data.

[0021] In this embodiment, both the first and second residual current monitoring data are raw monitoring data. Raw monitoring data often contains noise or redundant information, and directly using it for analysis can affect accuracy. Therefore, feature extraction can be performed on these raw monitoring data, that is, screening out key features related to residual current faults from a large amount of raw monitoring data.

[0022] In this embodiment, due to the presence of noise sources such as electromagnetic interference and harmonics within the substation, the first residual current monitoring data from the zero-sequence current sensor and the second residual current monitoring data from the leakage current sensor may contain noise and interference components. Therefore, this embodiment further includes preprocessing the first and second residual current monitoring data before feature extraction.

[0023] In this embodiment, the preprocessing mainly includes steps such as filtering, noise reduction, and time synchronization.

[0024] First, digital filtering is applied to the first and second residual current monitoring data to remove high-frequency noise and harmonic components other than the power frequency. For example, a 50Hz bandpass filter can be used to retain the power frequency leakage current signal and filter out high-frequency switching noise; for leakage current containing a DC component, a high-pass or DC bias correction circuit can be added to accurately extract the DC leakage current component.

[0025] Secondly, denoising and outlier removal are performed to further suppress random noise and impulse interference, improving signal smoothness and reliability. Simultaneously, the sampled data undergoes a rationality check, removing obviously abnormal outliers (such as data exceeding the sensor's range or significantly distorted data) to avoid the impact of abnormal data on subsequent analysis.

[0026] Secondly, ensure that the data from all sensors are synchronized in time. Since different sensors may acquire data through different acquisition channels, even with synchronized sampling in hardware, there will still be slight time deviations. By embedding timestamps in the data or using a synchronization trigger signal, the first residual current monitoring data and the second residual current monitoring data are aligned to a unified time axis.

[0027] After preprocessing, the initial data corresponding to the zero-sequence current sensor and the initial data corresponding to the leakage current sensor are obtained.

[0028] In this embodiment, the raw data collected by the zero-sequence current sensor is a continuous current signal (such as a current waveform that changes over time). It is necessary to extract features (first feature data) reflecting the overall leakage state of the station AC power supply system from the initial data corresponding to the zero-sequence current sensor. For example: time-domain features (such as the effective value reflecting the average intensity of the residual current and the peak value reflecting the intensity of the instantaneous leakage current) and frequency-domain features (such as the frequency distribution bandwidth reflecting the spectral complexity of the signal).

[0029] The data collected by the leakage current sensor for each feeder loop needs to have features (secondary feature data) that reflect the characteristics of the local loop extracted from the initial data corresponding to the leakage current sensor. For example, the transient characteristics of the leakage current.

[0030] Finally, the first feature data and the second feature data are transformed to obtain the corresponding first feature vector and second feature vector.

[0031] Because the feature data comes from different sensors, it has different forms (such as spectral distribution, transient characteristics, statistics, etc.). For example: The frequency components extracted from the zero-sequence current are a set of discrete values ​​(such as the amplitude of each harmonic); the transient features extracted from the leakage current are the rate of change or peak value of the time series. After converting these different types of feature data into feature vectors, all features are unified into an ordered array of values ​​(vectors), enabling multi-dimensional information to be integrated into the same mathematical framework, facilitating subsequent fusion.

[0032] S103: The first and second eigenvectors are fused to obtain the target eigenvector of the residual current of the station AC power supply.

[0033] In this embodiment, both the total residual current feature (first feature vector) and the local loop feature (second feature vector) have limitations. For example, the total feature cannot locate a specific loop, and the local feature may ignore the overall system correlation. The purpose of fusion is to integrate the information from both through an algorithm to form a more comprehensive and robust integrated feature (target feature vector).

[0034] In this embodiment, the weight of the first feature vector can be determined based on the deviation between the total residual current of the station AC power supply system and the safety threshold, and the weight of the second feature vector can be determined based on the equipment importance level of the feeder circuit. The first feature vector and the second feature vector are then weighted and fused to obtain the target feature vector.

[0035] By using weighted summation, the target feature vector simultaneously encompasses both the overall trend of the station's AC power supply system and the details of key local circuits, avoiding the one-sidedness of a single feature. For example, when the total residual current of the station's AC power supply system exceeds the standard but the transient characteristics of a certain important circuit are abnormal, the fused vector can highlight both dimensions simultaneously, reflecting the fault state more comprehensively.

[0036] S104: Analyze the residual current of the station AC power supply based on the target feature vector to determine the cause of the residual current in the station AC power supply.

[0037] In this embodiment, different causes of leakage current correspond to specific combinations of features. This can be achieved through rule-based matching, which pre-defines the correspondence rules between common causes of leakage current and feature vectors (such as expert experience), and directly matches the target feature vector.

[0038] For example: if the peak value of the residual current of a certain feeder circuit in the target feature vector is high and persistent, and the total residual current is approximately equal to this value, then the cause is insulation breakdown of the equipment in that circuit; if the residual current fluctuates with time, and the residual current of multiple feeder circuits simultaneously exhibits intermittent peak values ​​with high harmonic components, it is due to the superposition of normal leakage currents caused by the simultaneous operation of electronic equipment (such as frequency converters) in multiple circuits; if the total residual current is much greater than the sum of the residual currents of each feeder circuit, then there is leakage in a circuit without a sensor installed.

[0039] As can be seen from the above, this embodiment obtains the total residual current of the system by setting a zero-sequence current sensor on the low-voltage side of the transformer, and simultaneously deploys leakage current sensors in multiple feeder circuits to collect local data, forming a three-dimensional monitoring network of overall and local components. This not only avoids monitoring failure caused by the failure of a single sensor, but also offsets the interference of complex electromagnetic environments through cross-verification of multi-source data. Secondly, by extracting features from the two types of data and converting them into vectors, the complex residual current characteristics containing AC and DC components can be accurately captured, overcoming the deficiency of traditional methods in distinguishing current types. Furthermore, the target feature vector obtained by feature fusion integrates the overall leakage current trend and the local contribution of each circuit. Combining electrical principles and intelligent analysis, it can accurately identify whether the leakage current originates from a fault in a specific circuit device, the superposition of normal leakage currents in multiple circuits, or a problem in a hidden circuit. This process from multi-source acquisition to feature fusion and then to intelligent diagnosis comprehensively reflects the complex residual current characteristics, ultimately greatly improving the accuracy of fault diagnosis.

[0040] In one embodiment of this application, feature extraction is performed on the first residual current monitoring data to obtain first feature data, including: Perform wavelet transform on the first monitoring data to obtain the spectral distribution corresponding to the first residual current monitoring data, obtain the frequency component corresponding to the first residual current monitoring data based on the spectral distribution, and use the frequency component as the first feature data. The second feature data is obtained by feature extraction from the second residual current monitoring data, including: The first-order difference calculation is performed on the second residual current monitoring data to obtain the first transient characteristic component corresponding to the second residual current monitoring data, and the first transient characteristic component is used as the second characteristic data.

[0041] In this embodiment, feature extraction for the first residual current monitoring data: Since the zero-sequence current sensor monitors the total residual current of the station's AC power supply system, its data not only includes the power frequency leakage current signal but also contains rich frequency components due to factors such as harmonic interference and changes in equipment operating status within the system. These frequency components can reflect the overall leakage characteristics of the system; for example, different fault types may correspond to specific harmonic distributions. Therefore, this embodiment uses wavelet transform to process the first residual current monitoring data. Wavelet transform has good time-frequency localization characteristics, which can decompose the current signal in the time domain into different frequency scales, thereby obtaining the spectral distribution corresponding to the first residual current monitoring data. From this spectral distribution, information such as the amplitude corresponding to each frequency point can be extracted. This information constitutes the frequency components corresponding to the first residual current monitoring data, i.e., the first feature data. Through the frequency components, the power frequency component and various harmonic components contained in the total residual current can be effectively characterized, providing a key basis for subsequent judgment of the overall leakage status of the system.

[0042] For example, suppose that in a 220kV substation station AC power supply system, a zero-sequence current sensor is installed on the low-voltage side of a 10kV transformer to continuously collect the total residual current signal. After preprocessing (50Hz bandpass filtering, outlier removal, and time synchronization), a smooth current waveform is obtained from the raw data.

[0043] When performing wavelet transform on the waveform, the db4 wavelet basis function was selected to decompose the signal into 5 scales (corresponding to different frequency ranges). By analyzing the wavelet coefficients at each scale, the spectral distribution was obtained: a main frequency component with an amplitude of 30mA was detected at scale 3 (corresponding to the vicinity of 50Hz power frequency), a 250Hz harmonic component with an amplitude of 2mA was present at scale 1 (high frequency band), and a DC offset component with an amplitude of 0.5mA was present at scale 5 (low frequency band).

[0044] These frequency components (50Hz / 30mA, 250Hz / 2mA, DC 0.5mA) are integrated into the first feature data, i.e. the first feature vector is [30,2,0.5] (unit: mA). This vector clearly reflects the frequency composition of the total residual current of the system, in which the power frequency component is dominant, and the high-frequency harmonics and DC components are weak, indicating that the overall leakage current is mainly normal power frequency leakage.

[0045] In this embodiment, feature extraction for the second residual current monitoring data: Leakage current sensors are used to monitor the local residual current in each feeder circuit. Faults such as insulation damage and poor line contact in feeder circuits often cause sudden changes in leakage current within a short period. This transient change is a crucial characteristic for locating local faults. First-order differential calculation reflects the rate of change of the data sequence at adjacent moments, precisely capturing this instantaneous change in leakage current. Therefore, in this embodiment, first-order differential calculation is performed on the second residual current monitoring data. The result obtained is the first transient characteristic component corresponding to the second residual current monitoring data, which is also the second characteristic data. The first transient characteristic component clearly reflects the amplitude and trend of the leakage current's sudden change over time. For example, when equipment suddenly experiences insulation breakdown, the leakage current increases sharply, and its first-order differential result will show a significant peak. This peak value can be used to quickly locate the faulty feeder circuit.

[0046] For example, the 10kV feeder circuit of this substation contains 8 branches, each of which is equipped with a leakage current sensor. Taking feeder circuit #3 (connected to the UPS power system) as an example, the raw data collected by its leakage current sensor is preprocessed to obtain a continuous 10-second leakage current time series (sampling frequency 1kHz).

[0047] First-order difference calculation was performed on the time series: at 5.2 seconds, the leakage current suddenly increased from 1.2mA to 8.7mA, and the difference between adjacent sampling points was 7.5mA / ms; in the following 3 seconds, the leakage current fluctuated in the range of 8-9mA, and the first-order difference result stabilized in the range of ±0.3mA / ms.

[0048] These differential results are used as the first transient feature component (second feature data). The transformed second feature vector is [7.5, 0.3] (unit: mA / ms). The peak value of 7.5 mA / ms reflects the leakage current change that occurred in the #3 feeder circuit at 5.2 seconds. The fluctuation value of 0.3 mA / ms reflects the stable state after the change, indicating that the circuit may have entered a state of continuous leakage after a transient insulation fault.

[0049] As can be seen from the above, this embodiment obtained key feature data from two dimensions: the frequency characteristics of the total residual current and the transient change characteristics of the local residual current. This lays a solid foundation for converting them into feature vectors and performing fusion analysis, which helps to more comprehensively and accurately reflect the residual current status of the station AC power supply system.

[0050] In one embodiment of this application, the first feature vector and the second feature vector are fused to obtain a target feature vector of the residual current of the station AC power supply, including: The weights corresponding to the first feature vector are determined based on the deviation between the first residual current monitoring data and the preset safety threshold. The weights of the second feature vector are determined based on the importance level of the electrical equipment under different feeder circuits; The target feature vector of the residual current of the station AC power supply is obtained by weighted fusion based on the first feature vector and the second feature vector, and the weights corresponding to the first feature vector and the second feature vector.

[0051] In this embodiment, a safety threshold for the total residual current of the station AC power supply system can be preset based on experience. The first feature vector reflects the key characteristics of the total residual current of the system, and its weight is positively correlated with the deviation between the first residual current monitoring data (total residual current) and the preset safety threshold. The larger the deviation, the higher the overall leakage risk of the station AC power supply system, the stronger the importance of the first feature vector, and the greater the weight allocation; conversely, the smaller the deviation, the lower the weight accordingly.

[0052] For example, setting a security threshold as The actual total residual current is Deviation value Then the weight of the first eigenvector It can be calculated using the formula: Where K is a constant (used to avoid...) When the value is 0, the weight is 0), and this is ensured through normalization. .

[0053] In this embodiment, the second feature vector reflects the local leakage characteristics of each feeder circuit. Its weight is divided according to the importance level of the electrical equipment in the feeder circuit. The higher the importance level (such as the circuit involving the main transformer cooling system and relay protection device), the greater the weight of the corresponding second feature vector; the weight of secondary circuits (such as lighting and auxiliary ventilation) is smaller.

[0054] For example, the importance of equipment can be divided into three levels: Level 1 consists of core equipment such as protection devices and monitoring systems, with a weight of 0.6; Level 2 consists of auxiliary equipment such as lighting and ventilation, with a weight of 0.3; and Level 3 consists of temporary electrical equipment, with a weight of 0.1. The sum of the weights of all second feature vectors is... (make sure ).

[0055] The target feature vector is obtained by weighted summation of the first and second feature vectors, using the following formula: Target feature vector = ×first eigenvector+ × Second eigenvector. This process preserves the influence of the overall leakage current trend of the system while highlighting the local characteristics of the critical circuit, avoiding the one-sidedness of information from a single dimension.

[0056] As can be seen from the above, this embodiment determines the weight of the first feature vector based on the deviation between the total residual current and the safety threshold, determines the weight of the second feature vector based on the importance of the feeder circuit equipment, and then obtains the target vector through weighted fusion. This approach highlights both the overall leakage risk of the system and focuses on the status of key circuits, improving the accuracy of fault location, reducing false positives and false negatives, and enhancing the reliability and practicality of residual current monitoring for station AC power supplies.

[0057] In one embodiment of this application, the residual current of the station AC power supply is analyzed based on the target feature vector to determine the cause of the residual current in the station AC power supply, including: Perform a Fourier transform on the target feature vector to obtain the spectral feature components; The residual current in the station AC power supply is determined to be caused by nonlinear load interference, since the proportions of the 3rd and 5th harmonics in the spectral characteristic components are greater than their respective preset proportions. Since the proportions of the 3rd and 5th harmonics in the spectral characteristic components are less than or equal to their respective preset proportions, it is determined that the cause of the residual current in the station AC power supply is not nonlinear load interference.

[0058] In this embodiment, the core logic of analyzing the cause of residual current based on the target feature vector is to identify nonlinear load interference through spectral features. The target feature vector integrates the frequency characteristics of the total residual current of the station AC power supply system with the transient characteristics of each feeder circuit. By performing a Fourier transform on it, the time-domain information can be converted into spectral feature components in the frequency domain, clearly showing the proportion of each harmonic.

[0059] Nonlinear loads (such as frequency converters and rectifiers) generate a large number of odd harmonics during operation, with the 3rd and 5th harmonics being typical and accounting for a relatively high proportion. Therefore, preset thresholds for the proportion of the 3rd and 5th harmonics (e.g., 15% and 10% respectively) are set. If both proportions exceed the corresponding thresholds, it indicates that the residual current is mainly caused by harmonic interference from the nonlinear load. If neither exceeds the threshold, nonlinear load interference is ruled out, and further analysis of the fault cause (such as equipment insulation damage, line aging, etc.) is required in conjunction with other characteristics.

[0060] For example, suppose that in a 110kV substation station AC power supply system, the target feature vector obtained through the aforementioned steps is [28, 5.2, 3.8] (unit: mA), corresponding to the 50Hz power frequency component, the 3rd harmonic component (150Hz), and the 5th harmonic component (250Hz), respectively. The preset threshold for the proportion of the 3rd harmonic is 15%, and the threshold for the proportion of the 5th harmonic is 10%.

[0061] Spectral feature extraction: Perform Fourier transform on the target feature vector and calculate the proportion of each harmonic. Total residual current RMS value = ≈28.7mA; The proportion of the third harmonic is approximately 18.1% (5.2 / 28.7) × 100%. The proportion of the 5th harmonic is approximately 13.2% (3.8 / 28.7) × 100%.

[0062] Threshold judgment: The proportion of 3rd harmonic (18.1%) is greater than the preset threshold of 15%, and the proportion of 5th harmonic (13.2%) is greater than the preset threshold of 10%.

[0063] Cause identified: The condition that the proportions of the 3rd and 5th harmonics both exceed the preset values ​​is met, therefore the cause of the residual current is determined to be nonlinear load interference. On-site verification revealed that the harmonic superposition generated during the operation of feeder circuit #2 of the substation (connected to the SVG reactive power compensation device, containing numerous rectifier modules) caused the proportion of higher harmonics in the system's residual current to exceed the standard, consistent with the analysis results.

[0064] As can be seen from the above, by performing a Fourier transform on the target feature vector, extracting the proportions of the 3rd and 5th harmonics and comparing them with preset values, the residual current caused by nonlinear load interference can be quickly identified. This not only accurately locates specific harmonic sources and reduces misjudgments of other fault causes, but also improves the pertinence and efficiency of residual current cause analysis, providing a reliable basis for judgment on the stable operation of substation power systems.

[0065] In one embodiment of this application, after determining that the cause of the residual current in the station AC power supply is not nonlinear load interference, the method further includes: Extract the second transient feature component from the target feature vector; If the ratio of the absolute value of the temperature change rate of the feeder circuit to the absolute value of the transient characteristic component change rate within a preset time is greater than a preset ratio, it is determined that the cause of residual current in the station AC power supply is line insulation damage; the temperature change rate is the temperature fluctuation amplitude of the feeder circuit within a preset time. If the ratio of the absolute value of the rate of change of temperature in the feeder circuit to the absolute value of the rate of change of transient characteristic component within a preset time is less than or equal to a preset ratio, it is determined that the cause of residual current in the station AC power supply is not line insulation damage.

[0066] In this embodiment, the temperature change rate refers to the temperature fluctuation range of the feeder circuit per unit time, reflecting the dynamic trend of ambient or line temperature changes. For example, if the temperature rises from 30°C to 35°C within 5 minutes, the temperature change rate is 1°C / minute. This indicator can be collected and calculated in real time by a temperature sensor to evaluate the impact of temperature on the insulation performance of the line.

[0067] After eliminating nonlinear load interference, this embodiment extracts the second transient feature component (reflecting the dynamic rate of change of the feeder circuit leakage current) from the target feature vector and performs cross-analysis by combining it with the temperature change rate of the feeder circuit within a preset time (e.g., 5 minutes).

[0068] The second transient characteristic component differs from the first transient characteristic component in that the first transient characteristic component focuses on the amplitude of the sudden change in leakage current, reflecting the intensity of the fault when it occurs; while the second transient characteristic component focuses on the rate of change of leakage current over time, reflecting the dynamic trend of fault development.

[0069] By calculating the ratio of the absolute value of the temperature change rate to the absolute value of the change rate of the second transient characteristic component within a preset time period, if the ratio is greater than the preset value (e.g., 1.5), it indicates that the temperature change has a significant impact on the leakage current change within the duration period, which is consistent with the gradual characteristic that the line insulation is accelerated to deteriorate as the temperature rises, so it is determined to be line insulation damage; otherwise, this cause is ruled out, and other factors (such as poor contact) need to be further investigated.

[0070] For example, suppose that after eliminating nonlinear load interference, further fault analysis is carried out on the #2 feeder circuit (connecting to the terminal box heating circuit) of a 110kV substation: The second transient characteristic component in the target feature vector is the leakage current change rate: 0.6 mA / s (reflecting the dynamic change trend of leakage current over time). The fiber optic temperature sensor configured in this feeder loop collects a temperature change rate of 1.2℃ / s (the ambient temperature rises due to the operation of the heating device).

[0071] The ratio of the absolute value of the temperature change rate (1.2℃ / s) to the absolute value of the change rate of the second transient characteristic component (0.6mA / s) is 2.0. The preset threshold for this ratio is 1.2 (based on thermal aging test data of insulating materials, this ratio is usually >1.2 when insulation is damaged).

[0072] Since 1.2 / 0.6 = 2.0 > 1.2, which matches the characteristic that increased temperature accelerates insulation damage and leads to increased leakage current, the cause of the residual current is determined to be insulation damage in the #2 feeder circuit.

[0073] As can be seen from the above, this embodiment can accurately identify line insulation damage by extracting the second transient characteristic component and combining it with the ratio of the feeder circuit temperature change rate to the transient characteristic change rate. It utilizes the characteristics of temperature's influence on insulation and reduces subjective judgment by quantifying the ratio, improving the accuracy of fault location and providing a reliable basis for targeted maintenance, thus ensuring the safety of the power supply system.

[0074] In one embodiment of this application, after determining that the cause of the residual current in the station AC power supply is not line insulation damage, the method further includes: Input the target feature vector into the pre-trained fault classification model to determine whether the residual current in the station AC power supply is caused by a ground fault or equipment leakage. The pre-trained fault classification model is trained based on historical fault data that includes equipment leakage and grounding faults.

[0075] In this embodiment, after ruling out line insulation damage as a cause, a pre-trained fault classification model is used to further and more accurately determine whether the cause of the residual current is a grounding fault or equipment leakage.

[0076] The pre-trained fault classification model is built upon historical fault data from substations. This data covers a large number of cases of equipment leakage and grounding faults, and each case contains a corresponding feature vector and a clear fault cause label. During the model training phase, the feature vectors from the historical fault data are used as input, and the corresponding fault causes are used as output. Through repeated iterative training, the model learns the mapping relationship between different fault types and feature vectors, thereby gaining the ability to classify and judge new target feature vectors.

[0077] Specifically, grounding faults and equipment leakage currents exhibit different characteristics in their feature vectors. For example, grounding faults may show a large residual current value accompanied by specific frequency characteristics, while equipment leakage currents have a relatively small residual current value and different transient characteristics. The model learns these differences to form a discrimination criterion for the two fault types.

[0078] When the target feature vector is input into the model, the model will analyze and match the target feature vector based on the discrimination criteria learned during the training process, and then output the judgment result of whether the cause of the residual current is a grounding fault or equipment leakage.

[0079] For example, suppose that after excluding nonlinear load interference and line insulation damage, a pre-trained fault classification model is used to further diagnose the cause of residual current in the #5 feeder circuit (connected to the GIS equipment control circuit) of a 220kV substation.

[0080] The pre-trained model uses the random forest algorithm, and the training data includes 1200 historical fault cases from the past 5 years: 650 ground fault cases (feature label: residual current > 50mA, including a specific 100Hz grounding arc characteristic frequency); and 550 equipment leakage cases (feature label: residual current 2-20mA, transient change rate < 0.5mA / s). The model has undergone 10-fold cross-validation and achieved an accuracy of 92.3%, and has been deployed to the substation monitoring system.

[0081] After preliminary fusion processing, the target feature vector of the #5 feeder circuit is [18, 0.3, 0.2] (unit: mA, mA / s, Hz), which includes: 18mA of residual current at power frequency; 0.3mA / s of the second transient feature component (leakage current change rate); and 50Hz of dominant frequency (no special harmonics).

[0082] After the model extracts vector features, it is compared with the feature distribution learned during training: the residual current of 18mA is in the typical range of equipment leakage (2-20mA); the transient rate of change is 0.3mA / s < the ground fault threshold (0.5mA / s); there is no 100Hz ground arc characteristic frequency.

[0083] The model outputs a judgment result of equipment leakage with a confidence level of 94.7%. On-site monitoring by maintenance personnel revealed that the relay insulation pads in the GIS equipment control circuit were aging, resulting in continuous micro-leakage (17.8mA), consistent with the model diagnosis.

[0084] As can be seen from the above, this embodiment, through a pre-trained fault classification model, can accurately distinguish between grounding faults and equipment leakage after ruling out the causes of residual current damage to the line insulation. The model is trained based on historical data, which can capture subtle feature differences, compensate for the limitations of rule-based judgments, improve the accuracy and efficiency of fault diagnosis in complex scenarios, and provide reliable guidance for rapid repair.

[0085] In one embodiment of this application, the method further includes: Based on the determined cause of the residual current in the station's AC power supply and the degree of anomaly in the target characteristic vector, the fault risk level is classified. Based on the fault risk level and the determination of the cause of residual current in the station's AC power supply, a corresponding control strategy is generated.

[0086] In this embodiment, the fault risk level is divided into two dimensions: cause severity and feature anomaly degree. Cause severity is based on fault type (e.g., ground fault has a base weight of 0.6, equipment leakage has a base weight of 0.3). Feature anomaly degree is calculated by the deviation of the target feature vector from the normal threshold (e.g., exceeding the standard by 20% is counted as 0.2, exceeding the standard by 50% is counted as 0.5). The two are weighted and summed to form three levels (Level I ≥ 0.8, Level II 0.4-0.8, Level III < 0.4).

[0087] Control strategies are matched according to risk level: Level I (high risk, such as grounding fault and current exceeding the limit by 50%) triggers immediate tripping and issues emergency maintenance instructions; Level II (medium risk, such as equipment leakage exceeding the limit by 30%) activates circuit current limiting protection and pushes planned maintenance reminders; Level III (low risk, such as slight nonlinear load interference) only records anomalies and continuously monitors them.

[0088] For example, if a circuit is identified as having a ground fault (basic weight 0.6), and the target feature vector shows that the current exceeds the standard by 60% (abnormality 0.6), with a total score of 1.2, it belongs to the level I risk. The station's AC power system should immediately cut off the power supply to the circuit and notify the maintenance personnel for emergency handling.

[0089] As can be seen from the above, this embodiment combines the cause of the fault with the degree of anomaly in the target feature vector to classify the risk level, and then generates the corresponding control strategy. This not only achieves accurate quantification of fault risk, but also enables the implementation of immediate tripping, current limiting protection, or monitoring measures according to the level, avoiding over- or under-handling, improving the scientific nature and timeliness of power system fault response, and ensuring operational safety.

[0090] In one embodiment of this application, the method further includes: Based on the output of the fault risk level, residual current cause and fault impact range prediction model, the execution priority and operation sequence of the control strategy are dynamically adjusted. During the execution of the control strategy, execution feedback data is collected in real time and compared with the preset effect threshold. If the deviation exceeds the threshold, the strategy is iteratively optimized until the residual current is restored to the safe range.

[0091] In this embodiment, the fault impact range prediction model can be a model built based on the topology of the station AC power supply system (such as circuit connection relationship and equipment dependency relationship) and the fault propagation law (such as the possible propagation path of leakage current) to predict the range of equipment or circuits that may be affected if the fault is not dealt with in time.

[0092] Execution priority refers to the order in which control strategies are executed. High-priority strategies (such as cutting off faulty circuits that endanger the main equipment) take precedence over low-priority strategies (such as recording minor anomalies) to avoid wasting resources or delaying handling.

[0093] Operation timing is the execution sequence of multiple control strategies (such as isolating the fault before starting the backup power supply) to prevent secondary problems caused by strategy conflicts (such as current surges during power switching).

[0094] Execution feedback data is real-time status data (such as equipment action feedback signals) collected during the execution of the control strategy, used to evaluate the effectiveness of the strategy.

[0095] The preset effect threshold is a benchmark value for judging whether the control strategy is effective (such as the residual current dropping below the safety threshold within 10 minutes), which is determined by historical operation and maintenance data and equipment safety parameters.

[0096] Strategy iteration optimization is a dynamic improvement process that adjusts strategy parameters (such as modifying tripping thresholds or optimizing protection action logic) based on feedback data when the execution effect of the control strategy does not meet expectations, until the fault is eliminated.

[0097] In this embodiment, the execution priority and operation sequence of the control strategy are adjusted in real time by combining the fault risk level, residual current causes (such as grounding faults and equipment leakage) and the output of the fault impact range prediction model (such as the number of circuits that may be affected and key equipment).

[0098] Priority: Strategies with high risk levels and impacting core equipment (such as the main transformer cooling system) should be implemented first; Operation sequence: Avoid system fluctuations caused by multiple strategies running in parallel (e.g., disconnect the faulty circuit first, and then start the backup power supply).

[0099] During the execution of the control strategy, feedback data such as changes in residual current and equipment status are collected in real time and compared with preset effect thresholds (such as a residual current reduction of ≥50% and no new alarms generated). If the deviation is ≤ the threshold, the strategy continues to be executed; if the deviation is > the threshold (such as a residual current reduction of less than 30% after the strategy is executed), the strategy iteration is immediately triggered (such as adjusting the tripping sequence and adding current limiting parameters) until the residual current returns to the safe range.

[0100] As can be seen from the above, this embodiment, by dynamically adjusting the priority and timing of the control strategy and combining it with closed-loop feedback optimization, can adapt to real-time changes in the power grid. It ensures that high-risk faults are handled first, avoiding strategy conflicts, and improves control performance through iterative optimization, solving the problem of poor adaptability of traditional fixed strategies, efficiently eliminating residual current hazards, and ensuring the stable operation of the power system.

[0101] Corresponding to the residual current monitoring method for station AC power supply in the above embodiment, Figure 2 This is a structural block diagram of a station AC power supply residual current monitoring system according to an embodiment of this application. For ease of explanation, only the parts relevant to the embodiment of this application are shown. References Figure 2 The AC power residual current monitoring system 20 used at this station includes: a data acquisition module 21, a feature extraction module 22, a feature fusion module 23, and an analysis module 24.

[0102] Among them, the data acquisition module 21 is used to acquire the first residual current monitoring data of the zero-sequence current sensor and the second residual current monitoring data of the leakage current sensor; the zero-sequence current sensor is installed on the low-voltage side of the transformer of the station AC power supply system, and the leakage current sensor is installed in the feeder circuit of the station AC power supply system, and the feeder circuit includes multiple sensors. The feature extraction module 22 is used to extract features from the first residual current monitoring data to obtain first feature data, and to transform the first feature data to obtain a first feature vector corresponding to the first residual current monitoring data; to extract features from the second residual current monitoring data to obtain second feature data, and to transform the second residual current monitoring data to obtain a second feature vector corresponding to the second residual current monitoring data. The feature fusion module 23 is used to fuse the first feature vector and the second feature vector to obtain the target feature vector of the residual current of the station AC power supply. Analysis module 24 is used to analyze the residual current of the station AC power supply based on the target feature vector and determine the cause of the residual current in the station AC power supply.

[0103] In one embodiment of this application, the feature extraction module 22 is specifically used for: Perform wavelet transform on the first monitoring data to obtain the spectral distribution corresponding to the first residual current monitoring data, obtain the frequency component corresponding to the first residual current monitoring data based on the spectral distribution, and use the frequency component as the first feature data. The first-order difference calculation is performed on the second residual current monitoring data to obtain the first transient characteristic component corresponding to the second residual current monitoring data, and the first transient characteristic component is used as the second characteristic data.

[0104] In one embodiment of this application, the feature fusion module 23 is specifically used for: The weights corresponding to the first feature vector are determined based on the deviation between the first residual current monitoring data and the preset safety threshold. The weights of the second feature vector are determined based on the importance level of the electrical equipment under different feeder circuits; The target feature vector of the residual current of the station AC power supply is obtained by weighted fusion based on the first feature vector and the second feature vector, and the weights corresponding to the first feature vector and the second feature vector.

[0105] In one embodiment of this application, the analysis module 24 is specifically used for: Perform a Fourier transform on the target feature vector to obtain the spectral feature components; The residual current in the station AC power supply is determined to be caused by nonlinear load interference, since the proportions of the 3rd and 5th harmonics in the spectral characteristic components are greater than their respective preset proportions. Since the proportions of the 3rd and 5th harmonics in the spectral characteristic components are less than or equal to their respective preset proportions, it is determined that the cause of the residual current in the station AC power supply is not nonlinear load interference.

[0106] In one embodiment of this application, the analysis module 24 is further configured to: Extract the second transient feature component from the target feature vector; If the ratio of the absolute value of the rate of change of temperature in the feeder circuit to the absolute value of the rate of change of transient characteristic component within a preset time is greater than a preset ratio, it is determined that the cause of residual current in the station AC power supply is line insulation damage. If the ratio of the absolute value of the rate of change of temperature in the feeder circuit to the absolute value of the rate of change of transient characteristic component within a preset time is less than or equal to a preset ratio, it is determined that the cause of residual current in the station AC power supply is not line insulation damage.

[0107] In one embodiment of this application, the analysis module 24 is further configured to: Input the target feature vector into the pre-trained fault classification model to determine whether the residual current in the station AC power supply is caused by a ground fault or equipment leakage. The pre-trained fault classification model is trained based on historical fault data that includes equipment leakage and grounding faults.

[0108] In one embodiment of this application, the station AC power supply residual current monitoring system 20 further includes: a control module; specifically used for: Based on the determined cause of the residual current in the station's AC power supply and the degree of anomaly in the target characteristic vector, the fault risk level is classified. Based on the fault risk level and the determination of the cause of residual current in the station's AC power supply, a corresponding control strategy is generated.

[0109] See Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided according to an embodiment of this application. Figure 3 The electronic device 300 in this embodiment may include one or more processors 301, one or more input devices 302, one or more output devices 303, and one or more memories 304. The processors 301, input devices 302, output devices 303, and memories 304 communicate with each other via a communication bus 305. The memories 304 store computer programs, including program instructions. The processors 301 execute the program instructions stored in the memories 304. Specifically, the processors 301 are configured to invoke the program instructions to perform the functions of the modules in the aforementioned system embodiments, for example... Figure 2 The functions of the data acquisition module 21, feature extraction module 22, feature fusion module 23, and analysis module 24 are shown.

[0110] It should be understood that, in the embodiments of this application, the processor 301 may be a central processing unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.

[0111] Input device 302 may include a touchpad, a fingerprint sensor (for collecting the user's fingerprint information and fingerprint orientation information), a microphone, etc., and output device 303 may include a display (LCD, etc.), a speaker, etc.

[0112] The memory 304 may include read-only memory and random access memory, and provides instructions and data to the processor 301. A portion of the memory 304 may also include non-volatile random access memory. For example, the memory 304 may also store information such as preset percentages and preset ratios.

[0113] In specific implementations, the processor 301, input device 302, and output device 303 described in the embodiments of this application can execute the implementation method described in the station AC power supply residual current monitoring method provided in the embodiments of this application, or they can execute the implementation method of the electronic equipment described in the embodiments of this application, which will not be repeated here.

[0114] In another embodiment of this application, a computer-readable storage medium is provided. This computer-readable storage medium stores a computer program, which includes program instructions. When executed by a processor, the program instructions implement all or part of the processes in the methods described above. Alternatively, the computer program can instruct related hardware to implement these processes. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include any entity or system capable of carrying computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0115] The computer-readable storage medium can be an internal storage unit of the electronic device in any of the foregoing embodiments, such as a hard disk or memory of the electronic device. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device. Furthermore, the computer-readable storage medium can include both internal and external storage units of the electronic device. The computer-readable storage medium is used to store computer programs and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0116] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0117] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the electronic devices and units described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0118] In the several embodiments provided in this application, it should be understood that the disclosed electronic devices and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connections shown or discussed may be indirect coupling or communication connections through some interfaces or units, or they may be electrical, mechanical, or other forms of connection.

[0119] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of this application, depending on actual needs.

[0120] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0121] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A residual current monitoring method for an alternating current power supply for a station, characterized by, The method comprises: obtaining first residual current monitoring data of a zero sequence current sensor and second residual current monitoring data of a leakage current sensor; the zero sequence current sensor is arranged at the low-voltage side of a transformer of an AC power supply system for a station, and the leakage current sensor is arranged at a feeder loop of the AC power supply system for the station, the feeder loop comprising a plurality of feature extraction is performed on the first residual current monitoring data to obtain first feature data, and the first feature data is converted to obtain a first feature vector corresponding to the first residual current monitoring data; feature extraction is performed on the second residual current monitoring data to obtain second feature data, and the second residual current monitoring data is converted to obtain a second feature vector corresponding to the second residual current monitoring data; the first feature vector and the second feature vector are fused to obtain a target feature vector of residual current of the AC power supply for the station; based on the target feature vector, the residual current of the AC power supply for the station is analyzed to determine the cause of the residual current of the AC power supply for the station.

2. The residual current monitoring method of the AC power supply for the station according to claim 1, wherein the feature extraction performed on the first residual current monitoring data to obtain the first feature data comprises: wavelet transform is performed on the first monitoring data to obtain a frequency spectrum distribution corresponding to the first residual current monitoring data, and a frequency component corresponding to the first residual current monitoring data is obtained based on the frequency spectrum distribution, and the frequency component is taken as the first feature data; the feature extraction performed on the second residual current monitoring data to obtain the second feature data comprises: first-order difference calculation is performed on the second residual current monitoring data to obtain a first transient feature component corresponding to the second residual current monitoring data, and the first transient feature component is taken as the second feature data.

3. The method of claim 1, wherein the step of monitoring the residual current of the AC power source for the station is performed by a residual current monitor. the fusion of the first feature vector and the second feature vector to obtain the target feature vector of the residual current of the AC power supply for the station comprises: a weight corresponding to the first feature vector is determined based on a deviation of the first residual current monitoring data from a preset safety threshold; weights of the second feature vector are determined based on importance levels of electrical equipment under different feeder loops; the first feature vector and the second feature vector, and the weights corresponding to the first feature vector and the second feature vector are weighted and fused to obtain the target feature vector of the residual current of the AC power supply for the station.

4. The residual current monitoring method of the AC power supply for the station according to claim 1, wherein the analysis of the residual current of the AC power supply for the station based on the target feature vector to determine the cause of the residual current of the AC power supply for the station comprises: Fourier transform is performed on the target feature vector to obtain a frequency spectrum feature component; in response to proportions of the 3rd harmonic and the 5th harmonic in the frequency spectrum feature component being greater than respective preset proportions, it is determined that the cause of the residual current of the AC power supply for the station is nonlinear load interference. In response to the proportion of the third harmonic and the fifth harmonic in the spectral feature component being less than or equal to the respective preset proportion, it is determined that the cause of the occurrence of the residual current of the station AC power supply is not nonlinear load interference.

5. The method of claim 4, wherein the step of monitoring the residual current of the AC power source for the station is performed by the monitoring unit. After determining that the cause of the occurrence of the residual current of the station AC power supply is not nonlinear load interference, the method further comprises: extracting a second transient feature component in the target feature vector; In response to the ratio of the absolute value of the temperature change rate of the feeder circuit to the absolute value of the transient feature component change rate within the preset time being greater than a preset ratio, it is determined that the cause of the occurrence of the residual current of the station AC power supply is line insulation damage; the temperature change rate is the temperature fluctuation amplitude of the feeder circuit within the preset time; In response to the ratio of the absolute value of the temperature change rate of the feeder circuit to the absolute value of the transient feature component change rate within the preset time being less than or equal to a preset ratio, it is determined that the cause of the occurrence of the residual current of the station AC power supply is not line insulation damage.

6. The method of monitoring residual current of an AC power source for a station according to claim 5, wherein After determining that the cause of the occurrence of the residual current of the station AC power supply is not line insulation damage, the method further comprises: inputting the target feature vector into a pre-trained fault classification model to determine whether the cause of the occurrence of the residual current of the station AC power supply is a grounding fault or equipment leakage; The pre-trained fault classification model is trained based on historical fault data containing equipment leakage and grounding fault.

7. The method of claim 1, wherein the step of monitoring the residual current of the AC power source for the station comprises the steps of: monitoring the residual current of the AC power source for the station; and determining whether the monitored residual current is within a predetermined range. Further comprising: dividing a fault risk level according to the cause of the occurrence of the residual current of the station AC power supply and the abnormality degree of the target feature vector; generating a corresponding control strategy based on the fault risk level and the cause of the occurrence of the residual current of the station AC power supply.

8. An alternating current power supply residual current monitoring system for a station, characterized by, Comprise: a data acquisition module configured to acquire first residual current monitoring data of a zero sequence current sensor and second residual current monitoring data of a leakage current sensor; the zero sequence current sensor is arranged at the low-voltage side of a transformer of a station AC power supply system, and the leakage current sensor is arranged at a feeder circuit of the station AC power supply system, the feeder circuit comprising a plurality of a feature extraction module configured to extract first feature data from the first residual current monitoring data, and convert the first feature data to obtain a first feature vector corresponding to the first residual current monitoring data; extract first feature data from the second residual current monitoring data, and convert the second residual current monitoring data to obtain a second feature vector corresponding to the second residual current monitoring data; a feature fusion module configured to fuse the first feature vector and the second feature vector to obtain a target feature vector of the residual current of the station AC power supply; an analysis module configured to analyze the residual current of the station AC power supply based on the target feature vector to determine the cause of the occurrence of the residual current of the station AC power supply.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor executes the computer program to realize the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 7.

Images