10kV voltage transformer grounding fault real-time monitoring method and system

Abstract

The invention relates to the technical field of power system fault monitoring, and discloses a 10kV voltage transformer ground fault real-time monitoring method and system, and the system comprises a ground fault data collection module, a waveform distortion analysis module, a transient signal detection module and the like. Three-phase current, zero-sequence current and voltage waveform data are collected, a current harmonic component proportion, a zero-sequence current high-frequency transient energy value and the like are analyzed, a fault type is determined by combining a phase difference, and a fault position is positioned according to topological parameters and line impedance. And the temperature of the grounding down lead can be monitored, and the fault probability is evaluated and the threshold value is dynamically adjusted by machine learning. According to the invention, real-time accurate monitoring can be realized, fault types and positions can be accurately diagnosed, potential hidden dangers can be comprehensively monitored, the intelligence of a monitoring system is improved, the operation and maintenance cost and risks are effectively reduced, and safe and stable operation of the 10kV power distribution network is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of power system fault monitoring, and in particular to a 10kV voltage transformer ground fault real-time monitoring method and system. Background Art

[0002] In modern power systems, 10kV distribution networks are widely used in urban and rural power supply networks. The stability and reliability of their operation are directly related to the quality of electricity consumption of users and the economic benefits of power companies. As a key device in the 10kV distribution network, the 10kV voltage transformer is mainly used to measure voltage, provide protection and metering functions. However, due to the complex operating environment of the 10kV distribution network, it is affected by lightning strikes, dirt, external force damage and other factors, and voltage transformer grounding faults occur from time to time.

[0003] If the grounding fault cannot be discovered and handled in time, it will pose a serious threat to the safe and stable operation of the power system. On the one hand, the fault may cause the voltage transformer to burn out, affecting the measurement, protection and metering functions of the power system, thereby causing damage to other electrical equipment and expanding the scope of power outages; on the other hand, the grounding fault may also generate zero-sequence current and zero-sequence voltage, interfere with nearby communication lines, and even endanger personal safety. Therefore, real-time monitoring of 10kV voltage transformer grounding faults, timely and accurate fault detection and determination of fault type and location are of great significance to ensure the safe and stable operation of the power system.

[0004] The traditional 10kV voltage transformer grounding fault detection method has many limitations. In the early days, faults were mainly discovered by manual inspection. Operation and maintenance personnel regularly inspected substations and lines to observe the appearance of the voltage transformer and listen to the sound, but this method was inefficient and subjective, making it difficult to find some hidden faults, and it was not possible to monitor the occurrence of faults in real time. With the development of technology, detection methods based on electrical tests such as insulation resistance testing and dielectric loss measurement have emerged. Although these methods can detect insulation defects of voltage transformers to a certain extent, they require power outages for testing, which affects the normal power supply of the power system, and the detection cycle is long, so sudden grounding faults cannot be captured in time.

[0005] In recent years, some online monitoring technologies based on electrical quantity analysis have been gradually applied to the detection of 10kV voltage transformer grounding faults. For example, the amplitude and phase relationship of the three-phase voltage and current of the voltage transformer are monitored to determine whether there is a grounding fault. However, these methods are greatly affected by various interference factors in the power system in practical applications, and are prone to misjudgment and missed judgment. At the same time, these methods can only determine whether there is a fault, but it is difficult to accurately determine the type and location of the fault, which brings difficulties to the handling of the fault.

[0006] In addition, existing monitoring systems often lack consideration of environmental factors. The line impedance in the 10kV distribution network will change with changes in temperature and humidity, and the traditional fault location method does not take into account the influence of these factors, resulting in low accuracy of fault location. Moreover, there is a lack of effective monitoring methods for overheating problems that may be caused by ground faults, and potential safety hazards cannot be discovered in a timely manner.

[0007] With the development of artificial intelligence technology, although some fault diagnosis methods based on machine learning have begun to be applied to power systems, in the field of 10kV voltage transformer grounding fault monitoring, these methods are still in the exploratory stage. There are problems such as insufficient model training data and poor generalization ability, which makes it difficult to meet the needs of actual engineering applications. Summary of the invention

[0008] The object of the present invention is to provide a 10kV voltage transformer ground fault real-time monitoring method and system thereof to solve the problems raised in the above background technology.

[0009] To achieve the above object, the present invention provides the following technical solution: a 10kV voltage transformer ground fault real-time monitoring method and system, the system comprising: The ground fault data acquisition module is used to obtain the three-phase current, zero-sequence current and voltage waveform data of the target 10kV voltage transformer in real time, and mark the target voltage transformer as a monitoring object; A waveform distortion analysis module is used to extract the three-phase current waveform distortion rate of the monitored object, calculate the proportion of harmonic components of each phase current, and screen out abnormal current waveform phases; A transient signal detection module, used to perform time-frequency decomposition on the zero-sequence current signal of the monitored object based on wavelet transform, extract the energy value of the high-frequency transient component, and determine the ground fault triggering condition according to a preset energy threshold; The specific determination method of the transient signal detection module is: obtain the zero-sequence current baseline energy spectrum density of the target voltage transformer under normal working conditions from the local database, calculate the deviation between the real-time high-frequency transient component energy value and the baseline energy spectrum density ;when When the preset deviation threshold is exceeded, a ground fault trigger signal is generated; A phase difference calculation module is used to calculate the phase difference between the zero-sequence current and the voltage of the monitored object. , combining with a preset fault type mapping table, determining the ground fault type; The fault location unit is used to calculate the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution, and determine the coordinates of the ground fault position.

[0010] Preferably, the specific method of calculating the proportion of harmonic components of each phase current includes: Perform fast Fourier transform on the three-phase current of the monitored object to extract the fundamental amplitude and the kth harmonic amplitude ,in, N, N is the preset maximum harmonic order; Calculate the proportion of each harmonic ,statistics A set S of harmonic orders exceeding a preset threshold; When the cardinality of the set S is greater than or equal to 2, the phase is determined to be a phase with abnormal current waveform.

[0011] Preferably, the specific method for determining the ground fault type includes: Obtain the phase difference range corresponding to different ground fault types from the local database, including metallic grounding, arc grounding and high-resistance grounding; The real-time phase difference Match the phase difference range of each fault type. , it is determined to be metallic grounding; if , it is judged as arc grounding; if , it is judged as high resistance grounding, where is the lowest value of the metallic ground fault phase difference range, It is the highest value of the phase difference range of metallic ground fault and also the lowest value of the phase difference range of arc ground fault. It is the lowest value of the phase difference range of high-resistance grounding fault and the highest value of the phase difference range of arc grounding fault, which satisfies .

[0012] Preferably, the fault locating unit further includes: Dynamic impedance correction module, used to collect the line temperature and ambient humidity of the monitored object in real time, and calculate the line impedance temperature correction coefficient and humidity correction factor ; Based on the corrected line impedance value , update the equivalent impedance calculation model, where is the line impedance value before correction, is the temperature change, Humidity change.

[0013] Preferably, the system further comprises: The temperature rise monitoring module is used to obtain the infrared thermal image of the grounding down conductor of the monitored object and extract the highest temperature point and temperature rise rate ; when Exceeds a preset temperature threshold or When the preset rate threshold is exceeded, an overheat alarm signal is generated and associated with the fault location result.

[0014] Preferably, the extraction temperature rise rate The specific method is: In continuous time series Get the temperature data series ; Calculate the temperature change at adjacent time points , and fitted the temperature rise trend curve, whose slope is ,in .

[0015] Preferably, the system further comprises: The machine learning classification module is used to classify the harmonic component proportion and high-frequency transient energy deviation of the monitored object. and phase difference Input the pre-trained support vector machine model and output the ground fault probability score

[0016] when When the preset probability threshold is exceeded, a multi-dimensional fault verification process is triggered.

[0017] Preferably, the training method of the support vector machine model includes: Obtain a data set marked with fault types from the historical database, and extract the harmonic component proportion, high-frequency transient energy deviation and phase difference as feature vectors; Kernel function mapping is used to transform feature vectors into high-dimensional space, and the classification hyperplane parameters are optimized through cross-validation.

[0018] Preferably, the system further comprises: The adaptive threshold adjustment module is used to collect statistics on the historical operation data of the monitored object within a preset time window and calculate the dynamic mean of harmonic components, phase difference and high-frequency energy. and standard deviation ; Update the real-time monitoring threshold to , where m is the preset sensitivity coefficient.

[0019] Preferably, the method for using the system includes: The ground fault data acquisition module is used to obtain the three-phase current, zero-sequence current and voltage waveform data of the target 10kV voltage transformer in real time, and the target voltage transformer is marked as a monitoring object; The waveform distortion analysis module is used to extract the three-phase current waveform distortion rate of the monitored object, calculate the proportion of harmonic components of each phase current, and screen out abnormal current waveform phases; The transient signal detection module is used to perform time-frequency decomposition of the zero-sequence current signal of the monitored object based on wavelet transform, extract the energy value of the high-frequency transient component, obtain the zero-sequence current baseline energy spectrum density of the target voltage transformer under normal working conditions from the local database, and calculate the deviation between the real-time high-frequency transient component energy value and the baseline energy spectrum density. ,when When the preset deviation threshold is exceeded, a ground fault trigger signal is generated; The phase difference calculation module is used to calculate the phase difference between the zero-sequence current and the voltage of the monitored object. , combining with a preset fault type mapping table, determining the ground fault type; The fault location unit is used to calculate the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution, and determine the coordinates of the ground fault position.

[0020] Compared with the prior art, the present invention has the following beneficial effects: With the help of the ground fault data acquisition module, the system can obtain the three-phase current, zero-sequence current and voltage waveform data of the 10kV voltage transformer in real time to ensure the timeliness of monitoring. The waveform distortion analysis module uses the fast Fourier transform algorithm to accurately calculate the proportion of harmonic components in each phase current, which can keenly capture tiny current waveform anomalies and effectively identify early fault signs. The transient signal detection module performs time-frequency decomposition of the zero-sequence current signal based on wavelet transform, accurately extracts the energy value of the high-frequency transient component, and compares it with the baseline energy spectrum density under normal working conditions. Once the deviation exceeds the preset threshold, a ground fault trigger signal is immediately generated, which greatly shortens the fault detection time and realizes real-time monitoring and rapid warning of faults.

[0021] The phase difference calculation module can accurately determine the type of grounding fault and distinguish between metallic grounding, arc grounding and high-resistance grounding based on the phase difference between zero-sequence current and voltage, combined with a preset fault type mapping table. This provides a key basis for subsequent targeted fault handling, avoids blind troubleshooting, and improves fault handling efficiency. Compared with the traditional method of judging the fault type based on manual experience or a single electrical quantity, the fault diagnosis accuracy and reliability of the present invention are greatly improved.

[0022] The fault location unit calculates the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution, and then determines the coordinates of the ground fault location. At the same time, the dynamic impedance correction module collects the line temperature and ambient humidity in real time, and dynamically corrects the line impedance, fully considering the impact of environmental factors on the line impedance, and significantly improving the accuracy of fault location. This enables operation and maintenance personnel to quickly arrive at the fault site for repair, effectively shortening the power outage time and reducing the economic losses caused by the power outage.

[0023] The temperature rise monitoring module obtains the infrared thermal image of the grounding down conductor to extract the highest temperature point and the temperature rise rate. When the temperature or temperature rise rate exceeds the preset threshold, an overheating alarm signal is generated and associated with the fault location result, which can timely detect overheating problems that may be caused by grounding faults, prevent serious accidents such as equipment burning in advance, and ensure the safe and stable operation of the power system. This comprehensive monitoring function for potential hidden dangers makes up for the defect that traditional monitoring technology only focuses on the fault itself and ignores the problems derived from the fault.

[0024] The machine learning classification module inputs the harmonic component proportion, high-frequency transient energy deviation and phase difference into the pre-trained support vector machine model, and outputs the ground fault probability score. When the score exceeds the preset probability threshold, the multi-dimensional fault verification process is triggered, further improving the accuracy of fault judgment. The adaptive threshold adjustment module dynamically adjusts the monitoring threshold according to the historical operating data of the monitored object, so that the system can better adapt to changes in the operating status of the power system and improve the overall performance and adaptability of the monitoring system.

[0025] The monitoring system of the present invention can detect faults in a timely and accurate manner and provide detailed fault information, helping operation and maintenance personnel to quickly formulate maintenance plans, reducing unnecessary inspection and maintenance workload, and reducing operation and maintenance costs. At the same time, it effectively avoids large-scale power outages caused by the expansion of faults, reduces the risk of power system operation, ensures the quality of electricity used by users, and improves the economic and social benefits of power companies. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a working principle diagram of the 10kV voltage transformer ground fault real-time monitoring method of the present invention; Figure 2 A diagram showing the steps for calculating the proportion of harmonic components in each phase current; Figure 3 Diagram of the implementation steps for determining the type of ground fault. DETAILED DESCRIPTION

[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0028] See also Figure 1-3 The present invention provides a technical solution: a 10kV voltage transformer grounding fault real-time monitoring system, the system comprising: Ground fault data acquisition module: Use high-precision current sensors and voltage sensors to collect three-phase current, zero-sequence current and voltage waveform data of the target 10kV voltage transformer in real time. These sensors are precisely installed on the relevant line nodes of the voltage transformer to ensure the accuracy and real-time nature of the collected data. After collecting the data, the system marks the target voltage transformer as a monitoring object, laying the foundation for the subsequent monitoring process.

[0029] Waveform distortion analysis module: For the marked monitoring object, this module extracts the current waveform distortion rate from the collected three-phase current data. Specifically, the fast Fourier transform algorithm is used to process the three-phase current. Through this algorithm, the fundamental amplitude can be accurately extracted. Pass Subharmonic amplitude (in, , The maximum harmonic order is preset according to the harmonic characteristics of the power system. For example, in a common 10kV power system, Then, according to the formula , calculate the proportion of each harmonic. Then, calculate the proportion of harmonics. The harmonic order set that exceeds the preset threshold (the threshold is set according to the harmonic level during normal operation and industry standards, such as 5% of the fundamental amplitude can generally be used as the threshold) When the collection When the cardinality is greater than or equal to 2, the phase is determined to be a phase with abnormal current waveform, thereby screening out the current phase that may have hidden dangers of fault.

[0030] Transient signal detection module: Based on wavelet transform technology, the zero-sequence current signal of the monitored object is decomposed in time and frequency. Wavelet transform can effectively capture the characteristics of the signal at different times and frequencies, and extract the energy value of the high-frequency transient component from it. The system obtains the zero-sequence current baseline energy spectrum density of the target voltage transformer under normal working conditions from the local database. This baseline data is obtained through long-term monitoring and statistical analysis of voltage transformer data under normal operating conditions. Then, the deviation between the real-time high-frequency transient component energy value and the baseline energy spectrum density is calculated. .when When the deviation exceeds a preset threshold (the threshold is verified and determined through a large number of experiments and actual operation data, such as 10%), a ground fault trigger signal is generated to promptly prompt the system that a ground fault may have occurred.

[0031] Phase difference calculation module: by measuring the phase difference between the zero-sequence current and voltage of the monitored object , and combined with the preset fault type mapping table to determine the ground fault type. The fault type mapping table is established based on theoretical analysis, experimental testing and actual operation experience, covering the phase difference range corresponding to different ground fault types.

[0032] Fault location unit: Calculate the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution. Use the line impedance distribution model and topological structure information to obtain the equivalent impedance value through relevant algorithms (such as impedance calculation algorithm). Then, based on the corresponding relationship between the equivalent impedance and the fault location, determine the coordinates of the ground fault location to achieve accurate fault location, providing an accurate basis for subsequent fault repair.

[0033] The present invention will be further described below in conjunction with Examples 1 to 5: Embodiment 1:

[0034] This embodiment elaborates on the specific method of calculating the proportion of harmonic components of each phase current and screening abnormal phases of current waveform, which provides technical support for more accurate analysis of current waveform distortion and helps to discover potential ground fault hazards at an early stage.

[0035] In the waveform distortion analysis module, the fast Fourier transform (FFT) algorithm is used to process the three-phase current of the monitored object. For example, its discrete time series is , , , , ( is the number of sampling points). The principle of fast Fourier transform is to convert the time domain signal to the frequency domain, and obtain the spectrum information in the frequency domain by operating on discrete data. The transformation formula is:

[0036] in is an imaginary unit. After FFT operation, the fundamental amplitude can be extracted from the obtained spectrum. and Subharmonic amplitude Assume that the current data of a certain phase collected at a certain moment is processed by FFT to obtain the fundamental amplitude , the third harmonic amplitude , the 5th harmonic amplitude .

[0037] According to the formula , calculate the proportion of each harmonic: The third harmonic ratio

[0038] The proportion of the fifth harmonic

[0039] Assuming the preset harmonic ratio threshold is 5%, then and If both exceed the preset threshold, the two harmonic orders are recorded in the collection middle, . Because the collection The cardinality of is 2, which satisfies the condition of being greater than or equal to 2, so the phase is determined to be the phase with abnormal current waveform. In this way, the harmonic situation of the current can be systematically analyzed, and the current phase with possible problems can be effectively screened out, providing a strong basis for subsequent fault judgment. Embodiment 2:

[0040] This embodiment clarifies the specific process of determining the type of ground fault according to the phase difference between zero-sequence current and voltage, so that after the ground fault is detected, the specific type of the fault can be further accurately determined, providing key information for formulating targeted fault handling measures.

[0041] In the phase difference calculation module, when the transient signal detection module generates a ground fault trigger signal, the system immediately obtains the phase difference between the zero-sequence current and voltage of the monitored object. The local database stores the phase difference ranges corresponding to different ground fault types, which are obtained through a large amount of theoretical research, laboratory tests and actual power system operation data.

[0042] For example, the phase difference range corresponding to metallic grounding is , assuming that in actual operation, , ; The phase difference range corresponding to arc grounding is ,set up ; The phase difference range corresponding to high-resistance grounding is .

[0043] When the phase difference is measured in real time When, due to , so the system determines that the type of ground fault occurring at this time is metallic grounding. ,because , it is judged as arc grounding. ,because , and it is determined to be high-resistance grounding. Through this precise phase difference matching method, the type of grounding fault can be quickly and accurately identified, which improves the accuracy and reliability of fault diagnosis and facilitates operation and maintenance personnel to adopt appropriate response strategies. Embodiment 3:

[0044] This embodiment adds a dynamic impedance correction module to the fault locating unit. Taking into account the influence of line temperature and ambient humidity on line impedance, the equivalent impedance can be calculated more accurately by collecting relevant data in real time and performing corrections, thereby improving the accuracy of ground fault locating.

[0045] In the dynamic impedance correction module of the fault location unit, high-precision temperature sensors and humidity sensors are used to collect the line temperature and environmental humidity data of the monitored object in real time. The temperature sensor uses a thermistor sensor, whose resistance value changes with temperature. By measuring the resistance value and calibrating and converting it, accurate line temperature data can be obtained; the humidity sensor can use a capacitive humidity sensor to determine the environmental humidity based on the change in capacitance value.

[0046] Assume that the line temperature collected at a certain moment is After a period of time, the temperature becomes , then the temperature change ; The ambient humidity collected at the same time is , which then becomes , humidity change .

[0047] According to the characteristics of the line material and related research data, the line impedance temperature correction coefficient is determined in advance and humidity correction factor For example, for a line of a certain material, , (Dimensionless, this is an example value, the actual value is determined based on the line material).

[0048] Known line impedance value before correction , according to the formula , calculate the corrected line impedance value: .

[0049] The corrected line impedance value Substitute the equivalent impedance calculation model to recalculate the equivalent impedance from the fault point to the monitoring point, and then more accurately determine the coordinates of the ground fault location. This method of dynamically correcting line impedance effectively overcomes the influence of environmental factors on fault location and improves the practicality and reliability of the entire monitoring system. Embodiment 4:

[0050] This embodiment adds a temperature rise monitoring module, which obtains the highest temperature point and temperature rise rate information by analyzing the infrared thermal image of the grounding lead, so as to promptly detect overheating problems that may be caused by grounding faults, and associate the overheating alarm with the fault location results, so as to provide more comprehensive fault information for operation and maintenance personnel so that more effective measures can be taken.

[0051] In the temperature rise monitoring module, an infrared thermal imager is used to monitor the grounding down conductor in real time. The infrared thermal imager receives infrared radiation emitted by an object and converts it into a thermal image, thereby intuitively displaying the temperature distribution on the surface of the object.

[0052] In continuous time series For example, the temperature data is collected every 1 minute to obtain a series of temperature data. . Assume that in a certain monitoring, hour, ; hour, ; hour, .

[0053] Calculate the temperature change at adjacent time points:

[0054]

[0055] By analyzing the temperature changes at multiple time points, the temperature rise trend curve is fitted. The least squares method is used for curve fitting. Assume that the equation of the temperature rise trend curve obtained by fitting is (in, is the temperature, For time, is the slope, The slope is obtained by calculating The value of is the temperature rise rate In this example, assume that the fit yields , that is, the temperature rise rate .

[0056] At the same time, the highest temperature point is extracted from the infrared thermal image , assuming that in a certain monitoring, The preset temperature threshold is , the preset rate threshold is .because and At this time, the system generates an overheating alarm signal and associates the alarm signal with the fault location result. After receiving the alarm information, the operation and maintenance personnel can quickly understand the fault location and overheating situation, and take corresponding cooling and maintenance measures in time to ensure the safe and stable operation of the power system. Embodiment 5:

[0057] This embodiment includes a machine learning classification module and an adaptive threshold adjustment module. The machine learning classification module uses a support vector machine model to analyze the monitoring data and outputs a ground fault probability score to improve the accuracy of fault judgment; the adaptive threshold adjustment module dynamically adjusts the monitoring threshold according to historical operating data, so that the system can better adapt to changes in the operating status of the power system and further improve the performance of the monitoring system.

[0058] In the machine learning classification module, the support vector machine (SVM) model is used. First, a large number of data sets with fault types marked are obtained from the historical database. These data cover the voltage transformer monitoring data under different operating conditions. The proportion of harmonic components and high-frequency transient energy deviation are extracted from the data set. and phase difference as feature vectors. For example, the proportion of harmonic components extracted from a set of historical data is , ; High frequency transient energy deviation ; Phase difference , forming the feature vector .

[0059] Kernel function mapping (such as radial basis kernel function) is used to transform feature vectors into high-dimensional space, so that data that is linearly inseparable in low-dimensional space becomes linearly separable in high-dimensional space. Classification hyperplane parameters are optimized through cross-validation (such as 10-fold cross-validation) to improve the generalization ability and classification accuracy of the model. After a large amount of data training, a trained support vector machine model is obtained.

[0060] In actual operation, the harmonic component proportion and high-frequency transient energy deviation of the monitored object are monitored in real time. and phase difference Input the trained SVM model, and the model outputs the ground fault probability score . Suppose after a certain input data, the model output , the preset probability threshold is 0.7, because , triggering a multi-dimensional fault verification process to further verify the fault situation.

[0061] In the adaptive threshold adjustment module, the system collects historical operating data of the monitored object within a preset time window (such as the past 24 hours). For example, the system collects harmonic components, phase difference, and high-frequency energy data of a monitored object every 10 minutes in the past 24 hours. The dynamic mean of these data is calculated. and standard deviation Assume that for a certain harmonic component (such as the proportion of the third harmonic), the dynamic mean is calculated , standard deviation .

[0062] Preset sensitivity coefficients , then the real-time monitoring threshold is updated to ,Right now , and the new monitoring threshold range is The system uses the updated threshold to analyze and judge subsequent monitoring data, so that the monitoring system can dynamically adjust the monitoring standards according to the actual operation of the power system, better adapt to different operating conditions, and improve the accuracy and reliability of monitoring.

[0063] The specific steps of the multi-dimensional fault verification process include: Step 1: Call the waveform distortion analysis module, recalculate the proportion of the three-phase current harmonic components of the monitored object, and filter the current abnormal phase set. If the cardinality of the current abnormal phase set is still greater than or equal to 2, and the matching degree of its harmonic components with the historical fault data pattern exceeds 80%, the harmonic anomaly is determined to be valid; otherwise, it is marked as a suspected false trigger.

[0064] Step 2: Through the transient signal detection module, the zero-sequence current signal is decomposed by wavelet packets to extract the time domain energy distribution curve of the high-frequency transient component. If the energy peak value continues to exceed the range of 3 standard deviations of the baseline energy spectrum density and the duration is greater than 10ms, the transient abnormal signal is confirmed to be a real fault feature.

[0065] Step 3: The phase difference calculation module updates the phase difference between zero-sequence current and voltage in real time , and combined with the current phase difference range threshold provided by the adaptive threshold adjustment module. If it exceeds the dynamic threshold range and coincides with the metallic grounding, arc grounding or high-resistance grounding characteristic interval in the fault type mapping table, the phase difference is determined to be valid.

[0066] Step 4: Obtain the infrared thermal image of the grounding down conductor collected by the temperature rise monitoring module and extract the maximum temperature near the predicted position of the fault location unit and temperature rise rate .like Exceeding the ambient temperature by 20°C and >2℃ / s, or the deviation from the temperature rise curve under normal historical conditions exceeds 50%, it is confirmed that the fault point is overheated, which enhances the credibility of the positioning result.

[0067] Step 5: The dynamic impedance correction module updates the equivalent impedance model based on the real-time line temperature and humidity data, and recalculates the equivalent impedance of the fault point. If the impedance difference before and after correction is less than 5% of the total line impedance, and the matching error between the fault location coordinates and the topology parameters is within 10 meters, the fault location result is determined to be consistent and valid.

[0068] Step 6: Input the above verification results into the decision logic unit. If at least three verification conditions are met (for example, three of the steps 1 to 5 are "valid"), the ground fault is confirmed to be real, and the final fault alarm signal is generated, and the fault type, location coordinates and temperature rise alarm information are output; otherwise, it is marked as a false trigger and fed back to the adaptive threshold adjustment module to optimize the model parameters. Through the above multi-dimensional cross-validation, the system effectively reduces the misjudgment rate of a single detection module and improves the accuracy and reliability of ground fault diagnosis.

[0069] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0070] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A 10kV voltage transformer ground fault real-time monitoring system, characterized by: The ground fault data acquisition module is used to obtain the three-phase current, zero-sequence current and voltage waveform data of the target 10kV voltage transformer in real time, and mark the target voltage transformer as a monitoring object; A waveform distortion analysis module is used to extract the three-phase current waveform distortion rate of the monitored object, calculate the proportion of harmonic components of each phase current, and screen out abnormal current waveform phases; A transient signal detection module, used to perform time-frequency decomposition on the zero-sequence current signal of the monitored object based on wavelet transform, extract the energy value of the high-frequency transient component, and determine the ground fault triggering condition according to a preset energy threshold; The specific determination method of the transient signal detection module is: obtain the zero-sequence current baseline energy spectrum density of the target voltage transformer under normal working conditions from the local database, calculate the deviation between the real-time high-frequency transient component energy value and the baseline energy spectrum density ;when When the preset deviation threshold is exceeded, a ground fault trigger signal is generated; A phase difference calculation module is used to calculate the phase difference between the zero-sequence current and the voltage of the monitored object. , combining with a preset fault type mapping table, determining the ground fault type; The fault location unit is used to calculate the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution, and determine the coordinates of the ground fault position.

2. The monitoring system according to claim 1, characterized in that: The specific method for calculating the proportion of harmonic components of each phase current includes: Perform fast Fourier transform on the three-phase current of the monitored object to extract the fundamental amplitude and the kth harmonic amplitude ,in, N, N is the preset maximum harmonic order; Calculate the proportion of each harmonic ,statistics A set S of harmonic orders exceeding a preset threshold; When the cardinality of the set S is greater than or equal to 2, the phase is determined to be a phase with abnormal current waveform.

3. The monitoring system according to claim 2, characterized in that: The specific method for determining the ground fault type includes: Obtain the phase difference range corresponding to different ground fault types from the local database, including metallic grounding, arc grounding and high-resistance grounding; The real-time phase difference Match the phase difference range of each fault type. , it is determined to be metallic grounding; if , it is judged as arc grounding; if , it is judged as high resistance grounding, where is the lowest value of the metallic ground fault phase difference range, It is the highest value of the phase difference range of metallic ground fault and also the lowest value of the phase difference range of arc ground fault. It is the lowest value of the phase difference range of high-resistance grounding fault and the highest value of the phase difference range of arc grounding fault, which satisfies .

4. The monitoring system according to claim 3, characterized in that: The fault location unit also includes: Dynamic impedance correction module, used to collect the line temperature and ambient humidity of the monitored object in real time, and calculate the line impedance temperature correction coefficient and humidity correction factor ; Based on the corrected line impedance value , update the equivalent impedance calculation model, where is the line impedance value before correction, is the temperature change, Humidity change.

5. The monitoring system according to claim 4, characterized in that: Also includes: The temperature rise monitoring module is used to obtain the infrared thermal image of the grounding down conductor of the monitored object and extract the highest temperature point and temperature rise rate ; when Exceeds a preset temperature threshold or When the preset rate threshold is exceeded, an overheat alarm signal is generated and associated with the fault location result.

6. The monitoring system according to claim 5, characterized in that: The extraction temperature rise rate The specific method is: In continuous time series Get the temperature data series ; Calculate the temperature change at adjacent time points , and fitted the temperature rise trend curve, whose slope is ,in .

7. The monitoring system according to claim 6, characterized in that: Also includes: The machine learning classification module is used to classify the harmonic component proportion and high-frequency transient energy deviation of the monitored object. and phase difference Input the pre-trained support vector machine model and output the ground fault probability score ; when When the preset probability threshold is exceeded, a multi-dimensional fault verification process is triggered.

8. The monitoring system according to claim 7, characterized in that: The training method of the support vector machine model includes: Obtain a data set marked with fault types from the historical database, and extract the harmonic component proportion, high-frequency transient energy deviation and phase difference as feature vectors; Kernel function mapping is used to transform feature vectors into high-dimensional space, and the classification hyperplane parameters are optimized through cross-validation.

9. The monitoring system according to claim 8, characterized in that: Also includes: The adaptive threshold adjustment module is used to collect statistics on the historical operation data of the monitored object within a preset time window and calculate the dynamic mean of harmonic components, phase difference and high-frequency energy. and standard deviation ; Update the real-time monitoring threshold to , where m is the preset sensitivity coefficient.

10. A method for using the monitoring system according to claim 1, characterized in that: The method of use includes: The ground fault data acquisition module is used to obtain the three-phase current, zero-sequence current and voltage waveform data of the target 10kV voltage transformer in real time, and the target voltage transformer is marked as a monitoring object; The waveform distortion analysis module is used to extract the three-phase current waveform distortion rate of the monitored object, calculate the proportion of harmonic components of each phase current, and screen out abnormal current waveform phases; The transient signal detection module is used to perform time-frequency decomposition of the zero-sequence current signal of the monitored object based on wavelet transform, extract the energy value of the high-frequency transient component, obtain the zero-sequence current baseline energy spectrum density of the target voltage transformer under normal working conditions from the local database, and calculate the deviation between the real-time high-frequency transient component energy value and the baseline energy spectrum density. ,when When the preset deviation threshold is exceeded, a ground fault trigger signal is generated; The phase difference calculation module is used to calculate the phase difference between the zero-sequence current and the voltage of the monitored object. , combining with a preset fault type mapping table, determining the ground fault type; The fault location unit is used to calculate the equivalent impedance from the fault point to the monitoring point based on the topological parameters of the monitored object and the line impedance distribution, and determine the coordinates of the ground fault position.

Images