A transformer area electric leakage intelligent detection and fault positioning system

Abstract

This invention belongs to the field of power system detection technology, specifically providing an intelligent detection and fault location system for transformer substation leakage current. The system includes a distributed synchronous monitoring network, an edge intelligent analysis unit, a central diagnostic and location platform, a data visualization and human-machine interface, and a system power supply and communication module. The distributed synchronous monitoring network is deployed at key nodes in the transformer substation to collect synchronous current data; the edge intelligent analysis unit performs vector analysis and manages dynamic baselines; the central diagnostic and location platform diagnoses leakage current using an adaptive threshold judgment algorithm and utilizes a collaborative location engine that combines topology and vector differences to locate fault branches step by step. This invention achieves real-time online detection and accurate fault location of low-voltage transformer substation leakage current, and can adapt to changes in leakage current characteristics under different operating conditions.

Description

Technical Field

[0001] This invention belongs to the field of power system detection technology, specifically referring to an intelligent detection and fault location system for transformer substation leakage. Background Technology

[0002] Low-voltage distribution transformer substations are the final link in the power system supplying electricity to end users, and their power supply security directly affects the quality and safety of electricity use for users. Due to factors such as aging lines, damaged insulation, poor grounding, corrosion from humid environments, and leakage from user-side electrical equipment, leakage faults frequently occur in these substations. Leakage not only leads to wasted energy and abnormal line heating, but can also cause serious safety accidents such as electric shock and electrical fires.

[0003] Currently, the detection and location of leakage current in transformer substations mainly rely on two methods: regular manual inspections and power outage detection. Manual inspections typically use clamp meters to measure the line point by point, relying on the experience of maintenance personnel to determine the leakage point. This method suffers from low efficiency, high missed detection rate, and inability to monitor in real time. While power outage detection can accurately measure insulation resistance, it requires interrupting power supply, affecting users' normal electricity use, and cannot reflect the true leakage current situation under energized conditions. It is also difficult to detect intermittent leakage current or leakage current caused by dynamic load changes.

[0004] With the development of smart grids and distribution network automation, some online monitoring devices have emerged that can monitor and alarm on the total residual current in a distribution area. However, most of these devices are single-function, only capable of alarming when the total leakage current exceeds the limit. They cannot distinguish between leakage at the user side and leakage within the line itself, nor can they accurately locate the leakage branch or specific fault point. When a leakage alarm occurs, maintenance personnel still need to conduct manual on-site inspections, and the efficiency of fault location has not been fundamentally improved. Existing technologies often use fixed thresholds for judgment, making it difficult to adapt to changes in leakage characteristics under different seasons and load conditions, easily leading to false alarms or missed alarms.

[0005] Therefore, there is a lack of intelligent detection and positioning systems in the current technology that can detect leakage current in low-voltage distribution areas in real time, online and accurately without power interruption, and can automatically analyze and locate faulty branches or even specific sections, so as to improve operation and maintenance efficiency, ensure power supply safety and reduce personal and property risks. Summary of the Invention

[0006] This invention overcomes the shortcomings of existing technologies and provides a smart detection and fault location system for transformer substation leakage current. By integrating high-precision current sensors and intelligent detection terminals at multiple monitoring points (such as transformer outlets, branch lines, and user access points), a distributed synchronous monitoring network is constructed. Employing a collaborative workflow of "multi-dimensional synchronous current acquisition—vector synthesis and dynamic baseline analysis—adaptive threshold determination under energized mode—multi-source data fusion positioning," the system achieves real-time perception of transformer substation leakage current status, intelligent anomaly identification, and precise step-by-step fault location. The core innovation of the system lies in the deep coupling of energized detection mode, vector analysis algorithm, dynamic baseline correction mechanism, and multi-node collaborative positioning logic. This ensures that the functional modules no longer operate in isolation but form an interconnected and mutually reinforcing organic whole through closed-loop feedback of data flow and decision flow. This significantly improves the sensitivity of leakage current detection, the accuracy of location, and the system's adaptability to different transformer substation operating conditions without requiring power outages.

[0007] The technical solution adopted by this invention is as follows: This solution provides a transformer substation leakage intelligent detection and fault location system, which is applied to real-time leakage monitoring and fault location in low-voltage transformer substations. The system includes a distributed synchronous monitoring network, an edge intelligent analysis unit, a central diagnostic and location platform, a data visualization and human-machine interface, and a system power supply and communication module.

[0008] The distributed synchronous monitoring network is used to synchronously collect synchronous current datasets from various monitoring points. The network consists of multiple intelligent detection terminals deployed at key nodes within the target transformer area. Each intelligent detection terminal includes a current acquisition module, which comprises a phase current acquisition unit and a residual current acquisition unit. The phase current acquisition unit acquires the instantaneous values ​​of the three-phase phase currents, while the residual current acquisition unit acquires the instantaneous values ​​of the residual currents. The distributed synchronous monitoring network incorporates a high-precision clock synchronization module, ensuring that the current acquisition actions of all intelligent detection terminals are performed under a unified time scale, thereby obtaining time-aligned synchronous current datasets.

[0009] The edge intelligent analysis unit is used to receive and process synchronous current datasets. The edge intelligent analysis unit is locally integrated with each intelligent detection terminal or connected nearby via wired / wireless means. The edge intelligent analysis unit includes a vector analysis engine and a dynamic baseline management module. The vector analysis engine calculates the instantaneous vector sum of the three-phase currents in real time based on the instantaneous values ​​of the three-phase currents in the synchronous current dataset, and compares the instantaneous vector sum of the three-phase currents with the synchronously acquired instantaneous values ​​of the residual current to generate vector analysis results. The dynamic baseline management module learns the current characteristics of the transformer area under normal operating conditions based on historical data, establishes and continuously updates a dynamic current baseline that represents the state without leakage current, and provides an adaptive reference benchmark for anomaly judgment.

[0010] The central diagnostic and location platform is deployed in the cloud or a regional master station and communicates with all edge intelligent analysis units. The central diagnostic and location platform includes a fault diagnosis core and a collaborative location engine. The fault diagnosis core receives vector analysis results and real-time current data reported by each node, combines them with dynamic current baselines, and applies an adaptive threshold judgment algorithm to determine whether leakage has occurred and the severity level of leakage, generating a leakage diagnosis conclusion. After receiving the leakage diagnosis conclusion, the collaborative location engine starts the location process: first, based on the amplitude and phase relationship of the instantaneous residual current values ​​of each node, it initially determines the main area where leakage occurs; then, it retrieves the synchronous current dataset of relevant branches, analyzes the differences and correlations of current vectors between different nodes, and obtains the transformer area topology information, gradually narrowing down the scope, and finally outputs fault location information. The fault location information includes at least the suspected leakage branch identifier and location confidence.

[0011] The data visualization and human-computer interaction interface is used to display system status, real-time current data, vector analysis results, leakage diagnosis conclusions, fault location information and historical records to operation and maintenance personnel; it supports the generation of current vector diagrams, leakage trend curves and location topology diagrams, and provides parameter setting, data export and alarm management functions.

[0012] The system power supply and communication module provides working power and data communication channels for the distributed synchronous monitoring network and edge intelligent analysis unit; it adopts a voltage-mode power supply method to ensure continuous operation of the system when the transformer area is energized; the communication methods support power line carrier, low-power wireless, LoRa, NB-IoT or 4G / 5G networks to meet the transmission needs of different field environments.

[0013] Furthermore, the dynamic baseline management module's operation includes the following steps:

[0014] Step B1: After the system starts running or is reset, start the baseline learning phase, continuously collect the synchronous current dataset within the preset learning period, and record the corresponding total load level and environmental parameters of the transformer area;

[0015] Step B2: Clean and extract features from normal data (without alarm triggering) within the learning cycle, calculate the instantaneous vector sum of three-phase current and the statistical distribution model of the instantaneous value of residual current at each monitoring point under various typical load sections, and establish an initial static baseline library.

[0016] Step B3: After the system enters normal operation, the dynamic baseline management module continuously monitors the load changes in the transformer area and seasonal environmental factors; when a significant change in load pattern or environmental conditions is detected, the baseline dynamic update mechanism is triggered.

[0017] Step B4: The baseline dynamic update mechanism uses the latest normal operation data to smoothly correct the baseline parameters under the corresponding conditions in the static baseline library, so that the dynamic current baseline can adaptively track the natural drift of the normal operation status of the transformer area and reduce misjudgments caused by load changes.

[0018] Furthermore, the collaborative localization engine's operation includes the following steps:

[0019] Step L1: When the fault diagnosis core generates a leakage diagnosis conclusion confirming the occurrence of leakage, the collaborative positioning engine is activated;

[0020] Step L2: Obtain complete topology information of the distribution area, including the deployment location relationships of transformers, main lines, branch lines, and intelligent detection terminals;

[0021] Step L3: Read the synchronous current dataset of all intelligent detection terminals during the current leakage period, especially the amplitude and phase information of the instantaneous residual current of each node;

[0022] Step L4: Based on Kirchhoff's current law and the transformer substation topology, locate the primary area: compare the phase difference between the total residual current at the transformer outlet and the residual current of each main branch, and identify the leakage current as the "source" of the injected fault current. By analyzing the flow direction of the fault current, the main area where the leakage occurs can be preliminarily located.

[0023] Step L5: Within the locked main branch area, perform secondary branch location: Retrieve the synchronous current dataset of all downstream intelligent detection terminals within the locked main branch area and apply the differential vector comparison algorithm. The differential vector comparison algorithm calculates the change in current vectors (especially the instantaneous value of residual current and the instantaneous vector sum of the three-phase currents) between adjacent upstream and downstream nodes. If the instantaneous value of residual current at a downstream node of a branch has a significant abrupt change relative to the upstream node, and this abrupt change cannot be explained by the normal load change of the branch, then the branch is marked as a high-suspected leakage branch;

[0024] Step L6: Combine the historical leakage records of the high-suspected leakage branch, the historical insulation data of the line, and the environmental information to calculate the final location confidence, integrate it into the fault location information and output it. The fault location information includes the high-suspected leakage branch identifier, the suggested investigation section, and the location confidence.

[0025] Furthermore, the voltage-mode power supply method is as follows: the current acquisition module of the intelligent detection terminal adopts an open-type clamp-on current transformer. The open-type clamp-on current transformer is clamped onto the conductor under test without disconnecting the line to achieve live detection; the system power supply and communication module obtains working power from the line under test through induction or small and micro current transformers to ensure that the entire detection process does not require power outages.

[0026] Furthermore, the data visualization and human-computer interaction interface also provides a dynamic vector display function, which can dynamically draw the position and trajectory of the instantaneous vector sum of the three-phase current in polar coordinates or rectangular coordinates based on real-time or historical synchronous current datasets, and superimpose the vector of the instantaneous value of the residual current in the same vector diagram, making it easier for maintenance personnel to intuitively analyze the current balance state and the abnormal changes of the vector when leakage occurs.

[0027] Compared with the prior art, the beneficial effects of the present invention are:

[0028] (1) By adopting open-type clamp-on current transformers with voltage mode power supply and a distributed synchronous monitoring network, the system can continue to work under normal power supply conditions in the transformer area, completely avoiding the impact on users' power consumption caused by the traditional method of power outage detection. Combining vector analysis engine, collaborative positioning engine and dynamic baseline management, the entire process from leakage current detection to branch positioning is intelligent, and the positioning accuracy is much higher than that of existing devices that only provide total leakage current alarm, which greatly improves the operation and maintenance response efficiency.

[0029] (2) By designing a dynamic baseline management module, a real-time updated reference benchmark is provided for the adaptive threshold judgment algorithm, enabling the threshold to be adaptively adjusted according to the working conditions, thereby improving the accuracy of the judgment. The results of the vector analysis engine directly serve the fault diagnosis core and the collaborative positioning engine. When performing positioning analysis, the collaborative positioning engine will call back and compare the synchronous current datasets of multiple nodes, forming a closed-loop intelligent chain of "data acquisition → feature analysis → intelligent diagnosis → precise positioning → feedback verification".

[0030] (3) By introducing a dynamic baseline management module and an adaptive threshold determination algorithm, the system can learn and track the natural changes in the normal operating status of the transformer area, automatically correct the determination benchmark, and effectively overcome the defects of the fixed threshold method that is prone to false alarms or missed alarms during load fluctuations and seasonal changes. The high-precision clock synchronization module ensures the time consistency of the entire network data, laying a reliable data foundation for subsequent phase analysis, differential comparison and other precise positioning algorithms, and enhancing the credibility of the positioning results.

[0031] (4) Through data visualization and human-computer interaction interface, maintenance personnel can remotely and intuitively grasp the overall current balance status of the transformer area, the development process of leakage events, the results of automatic system positioning, and historical trends. The dynamic display function of vector graphics transforms abstract current data into intuitive graphics, greatly reducing the analysis threshold, assisting manual in-depth diagnosis and decision-making, and improving the level of intelligent operation and maintenance management. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the overall architecture of a transformer substation leakage current intelligent detection and fault location system proposed in this invention;

[0033] Figure 2 This is a flowchart illustrating the logic of multi-level fault location using the collaborative positioning engine in this invention.

[0034] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. Detailed Implementation

[0035] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0036] Example 1:

[0037] Please see Figures 1-2 This embodiment presents a smart detection and fault location system for transformer substation leakage, applied to a typical low-voltage distribution transformer substation. The system aims to achieve 24 / 7 uninterrupted online monitoring and intelligent location of transformer substation leakage, and includes a distributed synchronous monitoring network, an edge intelligent analysis unit, a central diagnostic and location platform, a data visualization and human-machine interface, and a system power supply and communication module.

[0038] A distributed synchronous monitoring network is used to synchronously collect synchronous current datasets from various monitoring points. This network consists of intelligent detection terminals deployed at key monitoring points within the target distribution area. Each intelligent detection terminal includes a current acquisition module containing four high-precision open-type clamp-on current transformers: three of these open-type clamp-on current transformers serve as phase current acquisition units, clamped onto the three-phase conductors to measure the instantaneous values ​​of the three-phase currents; the other open-type clamp-on current transformer serves as a residual current acquisition unit, clamped onto the neutral line to measure the instantaneous value of the residual current. The four high-precision open-type clamp-on current transformers are configured with a range of "clamp meter 50A" to accommodate the current range of the low-voltage distribution area. Each intelligent detection terminal integrates a high-precision clock synchronization module. In this embodiment, the high-precision clock synchronization module uses a BeiDou-based timing chip to ensure that the current acquisition actions of all intelligent detection terminals are performed under a unified time scale, thereby obtaining accurate synchronous current datasets.

[0039] The edge intelligent analysis unit receives and processes the synchronous current dataset; it is integrated within each intelligent detection terminal. The edge intelligent analysis unit includes a vector analysis engine, which processes the synchronous current dataset acquired by the intelligent detection terminal. For each frame of data, the vector analysis engine calculates the instantaneous vector sum of the three-phase currents in real time based on the instantaneous values ​​of the three-phase currents in the synchronous current dataset, and compares this instantaneous vector sum with the synchronously acquired instantaneous values ​​of the residual current to generate a vector analysis result; this result includes statistical characteristics of the deviation.

[0040] The edge intelligent analysis unit also includes a dynamic baseline management module, whose operation process includes the following steps:

[0041] Step B1: After the system starts running, start the baseline learning phase, continuously collect the synchronous current dataset within the preset learning period, and record the corresponding total load level and environmental parameters of the transformer area;

[0042] Step B2: Record the normal current characteristics of the monitoring points under different time periods (such as valley, flat and peak load periods), calculate the instantaneous vector sum of the three-phase current and the statistical distribution model of the instantaneous value of the residual current at each monitoring point under various typical load sections, and establish an initial static baseline library.

[0043] Step B3: After the system enters normal operation, the dynamic baseline management module continuously monitors the load changes in the transformer area and seasonal environmental factors; when a significant change in load pattern or environmental conditions is detected, the baseline dynamic update mechanism is triggered.

[0044] Step B4: The baseline dynamic update mechanism uses the latest normal operation data to smoothly correct the baseline parameters under the corresponding conditions in the static baseline library, so that the dynamic current baseline can adaptively track the natural drift of the normal operation status of the transformer area and reduce misjudgments caused by load changes.

[0045] The central diagnostic and location platform is deployed on the main cloud server and communicates with all edge intelligent analysis units. The central diagnostic and location platform includes a fault diagnosis core and a collaborative location engine.

[0046] The fault diagnosis core receives vector analysis results and real-time current data, combined with dynamic current baseline and ambient humidity data (environmental correction factor) obtained from the current weather system. ) and total load data of transformer areas (load correction factor) The adaptive threshold determination algorithm is run. If a node is determined to have leakage, a leakage diagnosis conclusion is generated, which includes the leakage node ID, the occurrence time, and the leakage severity level (e.g., general or severe).

[0047] After receiving the leakage diagnosis conclusion, the collaborative positioning engine starts the positioning process: First, based on the amplitude and phase relationship of the instantaneous residual current values ​​of each node, it initially determines the main area where the leakage occurred; then, it retrieves the synchronous current dataset of the relevant branches, analyzes the differences and correlations of current vectors between different nodes, and obtains the transformer area topology information, gradually narrowing down the scope, and finally outputs the fault location information. The fault location information includes at least the suspected leakage branch identifier and the location confidence.

[0048] The data visualization and human-machine interface communicates with the central diagnostic and positioning platform to display system status, real-time current data, vector analysis results, leakage diagnosis conclusions, fault location information, and historical records to maintenance personnel. It supports the generation of current vector diagrams, leakage trend curves, and positioning topology maps, and provides parameter settings, data export, and alarm management functions. The homepage of the data visualization and human-machine interface displays a map of the distribution area, marking the location and status of each intelligent detection terminal. Clicking on any intelligent detection terminal allows viewing real-time current data and historical curves. When a leakage alarm occurs, the data visualization and human-machine interface automatically pops up an alarm window, displaying the leakage diagnosis conclusion and fault location information, and highlighting the suspected fault section on the distribution area single-line diagram. The data visualization and human-machine interface also provides a dynamic vector diagram display function, which can dynamically draw the position and trajectory of the instantaneous vector sum of the three-phase current in polar or rectangular coordinates based on real-time or historical synchronous current datasets, and overlay the instantaneous value vector of the residual current in the same vector diagram, facilitating maintenance personnel's intuitive analysis of the current balance state and abnormal vector changes when leakage occurs.

[0049] The system power supply and communication module provides operating power and data communication channels for the distributed synchronous monitoring network and edge intelligent analysis unit. The system power supply and communication module adopts a voltage-mode power supply method, which is as follows: the current acquisition module of the intelligent detection terminal uses an open-type clamp-on current transformer. The open-type clamp-on current transformer is clamped onto the conductor under test without disconnecting the line to realize live detection. The system power supply and communication module obtains operating power from the line under test through a small micro current transformer, ensuring that the entire detection process does not require power outages.

[0050] Furthermore, the working process of the vector analysis engine includes the following steps:

[0051] Step V1: At each sampling time t, obtain the instantaneous values ​​of the three-phase currents measured by the current intelligent detection terminal from the synchronous current dataset. , , ;

[0052] Step V2: Calculate the instantaneous vector sum of the three-phase currents:

[0053] ;

[0054] Step V3: Obtain the instantaneous value of the residual current collected at the same time. ;

[0055] Step V4: Calculate the instantaneous vector sum of the three-phase currents and the instantaneous deviation from the instantaneous value of the residual current:

[0056] ;

[0057] Step V5: For Perform statistical analysis within a continuous time window to calculate its mean, standard deviation, and deviation from the dynamic current baseline, forming vector analysis results. The vector analysis results include vector sum characteristics, residual current characteristics, deviation statistical characteristics, and baseline deviation index.

[0058] Furthermore, the specific operation process of the adaptive threshold determination algorithm includes:

[0059] Step T1: Receive vector analysis results from the vector analysis engine, focusing on the statistical characteristics of the deviation and the baseline deviation index;

[0060] Step T2: Obtain the dynamic current baseline under the current operating condition from the dynamic baseline management module, and extract the corresponding normal deviation range threshold from it. ;

[0061] Step T3: Introduce environmental correction factors (e.g., the effects of humidity and temperature on insulation) and load correction factors (Reflecting the current load factor), the baseline threshold is dynamically adjusted to generate an adaptive judgment threshold:

[0062] ;

[0063] Step T4: Compare the deviation statistical characteristics with the adaptive judgment threshold; if the deviation statistical characteristics exceed the adaptive judgment threshold for a preset number of consecutive times, it is judged as leakage current abnormality, and the severity level of leakage current is determined according to the proportion of deviation statistical characteristics exceeding the adaptive judgment threshold, and a leakage current diagnosis conclusion is generated.

[0064] Furthermore, the collaborative localization engine's operation includes the following steps:

[0065] Step L1: When the fault diagnosis core generates a leakage diagnosis conclusion confirming the occurrence of leakage, the collaborative positioning engine is activated;

[0066] Step L2: Obtain complete topology information of the distribution area, including the deployment location relationships of transformers, main lines, branch lines, and intelligent detection terminals;

[0067] Step L3: Read the synchronous current dataset of all intelligent detection terminals during the current leakage period, especially the amplitude and phase information of the instantaneous residual current of each node;

[0068] Step L4: Based on Kirchhoff's current law and the transformer substation topology, locate the primary area: compare the phase difference between the total residual current at the transformer outlet and the residual current of each main branch, and identify the leakage current as the "source" of the injected fault current. By analyzing the flow direction of the fault current, the main area where the leakage occurs can be preliminarily located.

[0069] Step L5: Within the locked main line area, perform secondary branch location: Retrieve the synchronization current datasets of all downstream intelligent detection terminals (numbered DTU-02, DTU-04, DTU-05) within the locked main line area, and apply the differential vector comparison algorithm. The differential vector comparison algorithm calculates the amplitude and phase difference of the residual current between DTU-02 and DTU-04. If an abnormal increase in residual current is found from DTU-02 to DTU-04 and the phase matches the characteristics of fault current, while the current of DTU-05 downstream of DTU-04 returns to normal, it is preliminarily determined that the leakage occurs on the line between DTU-02 and DTU-04, or at the user entrance monitored by DTU-04. This branch is then marked as a high-probability leakage branch.

[0070] Step L6: Combine the historical leakage records of the high-suspected leakage branches, the historical insulation data of the line, and the environmental information to calculate the final location confidence, integrate it into the fault location information and output it. The fault location information includes the high-suspected leakage branches, the suggested investigation sections, and the location confidence.

[0071] Through the deployment and operation of the system in this embodiment, maintenance personnel can remotely monitor the leakage status of the distribution area from their office without having to go to the site for power outage testing. They can also quickly obtain accurate location guidance when leakage occurs, which greatly improves the efficiency of handling leakage faults in low-voltage distribution areas and the level of power supply safety.

[0072] Example 2:

[0073] This embodiment is based on Embodiment 1 and focuses on illustrating a specific implementation of the baseline dynamic update mechanism in the dynamic baseline management module, as well as the method for obtaining the correction factor in the adaptive threshold determination algorithm.

[0074] Specific implementation of the dynamic baseline update mechanism:

[0075] The dynamic baseline management module divides a day into 24 time periods (one time period per hour) and establishes a baseline model for each time period. The baseline model includes the expected value of the statistical characteristics of the deviation within that time period under normal, leakage-free conditions. and standard deviation .

[0076] Update trigger conditions include:

[0077] (1) Periodic triggering: The baseline data for all time periods is updated every 7 days, and the normal data for that time period in the past 7 days is recalculated. and An exponentially weighted moving average method is used to smooth the transition and avoid abrupt changes.

[0078] (2) Event Trigger: When a large-scale load adjustment is carried out in the distribution area (such as adding a new user access point) or after experiencing special weather (such as continuous heavy rain), the system records the event. After the event ends, a targeted baseline fast relearning is initiated for the data of the affected period.

[0079] Specific implementation of the correction factor acquisition method:

[0080] Environmental Correction Factors The system accesses real-time current data from a local weather station. An empirical correlation table between humidity and insulation resistance is established. In this embodiment, when humidity remains above 80%, the insulation level is considered to potentially decrease, and a threshold is set... (That is, lower the judgment threshold and increase sensitivity); when the humidity is below 50%, set... (That is, appropriately relax the threshold to reduce false alarms).

[0081] Load correction factor Calculate the current load factor based on the total transformer outlet current, and establish a model relating the load factor to the normal current fluctuation range. When the load factor is below 30%, the normal current fluctuation is small, and a setting is made accordingly. When the load factor is between 70% and 100%, the normal current fluctuates greatly, and the setting... .

[0082] The adaptive decision threshold is then calculated as follows:

[0083] ;

[0084] in The initial threshold is obtained based on the dynamic baseline of the current time period.

[0085] Example 3:

[0086] This embodiment, based on Embodiment 1, details a specific application of the differential vector comparison algorithm in secondary branch localization within the collaborative localization engine.

[0087] Assuming the main area is locked, there are 3 monitoring terminals downstream: DTU-A (branch head), DTU-B (sub-branch 1 head), and DTU-C (sub-branch 2 head). The topology is DTU-A→(line L1)→DTU-B and DTU-A→(line L2)→DTU-C.

[0088] When a leakage occurs, the collaborative positioning engine acquires the synchronous current dataset of the three monitoring terminals for a period of time before and after the fault.

[0089] The specific steps of the differential vector alignment algorithm are as follows:

[0090] Calculate the fault characteristic current: For each terminal, take the instantaneous vector sum of its three-phase currents and the instantaneous deviation from the instantaneous value of the residual current. (The phase information, which is a complex number, is retained here), denoted as the fault characteristic current. Under normal circumstances without leakage, Theoretically, it should be zero or close to zero (containing only measurement errors and minor imbalances). When leakage occurs, the monitoring point upstream of the leakage point will detect it. .

[0091] Perform differential comparison:

[0092] Calculate the difference between paths AB: Theoretically, if the leakage occurs upstream of DTU-A or at DTU-A itself, and The values ​​should be similar, with a small difference. If the leakage occurs on line L1 or downstream of DTU-B, then... It should be significantly greater than The difference is large and the direction is the same as the difference. similar.

[0093] Similarly, calculate the difference of path AC: .

[0094] Analysis and Judgment:

[0095] like and If all values ​​are very small, it can be determined that the leakage current may be upstream of DTU-A (i.e., the part further upstream of the main trunk), and further analysis is needed in conjunction with the location results of the primary area.

[0096] like It is very big, and If the leakage is very small, it strongly indicates that the leakage is downstream of line L1 or DTU-B. Furthermore, if there are no other monitoring points downstream of DTU-B, the location result is "the user under the jurisdiction of line L1 or DTU-B".

[0097] like and If both values ​​are very large, it may indicate that there is leakage near DTU-A (e.g., insulation problems with the DTU-A mounting bracket), or that there are multiple leakage points, requiring a more complex correlation analysis combining amplitude ratio and topology.

[0098] The collaborative localization engine integrates the differential comparison results of all paths and combines them with topological logic to ultimately provide the most likely high-probability leakage branch.

[0099] Example 4:

[0100] This embodiment demonstrates the application process and output results of the system in a real leakage current event.

[0101] In a coastal area with high humidity, the system has been running stably for a month, and baseline learning is complete.

[0102] One afternoon, the fault diagnosis core of the central diagnostic and positioning platform issued an alarm based on the adaptive threshold judgment algorithm: monitoring point DTU-08 (located at the power supply branch entrance of a certain farm) continuously detected leakage current, level "severe".

[0103] The collaborative positioning engine is started.

[0104] Primary area location: Analysis shows that the leakage current phase is highly correlated with the main branch monitored by DTU-03, and the area is initially identified as branch F3.

[0105] Secondary branch location: Branch F3 mainly has two branches: one leading to the residential area (terminal DTU-07), and the other leading to the farm (terminal DTU-08). Data from DTU-03, DTU-07, and DTU-08 were retrieved. Differential comparison revealed:

[0106] The fault characteristic current of DTU-07 is very similar to that of DTU-03.

[0107] The fault characteristic current of DTU-08 is much larger than that of DTU-03, and the phase of the fault characteristic current of DTU-08 is basically consistent with the phase of the alarm residual current.

[0108] Location information generated: The collaborative positioning engine determined that the leakage occurred on the line downstream of DTU-03 to DTU-08, or inside the farm under the jurisdiction of DTU-08. Based on the topology, this line is an overhead line of approximately 500 meters. Fault location information output: "Highly suspected leakage branch: 'Farm dedicated line' under branch F3 (between DTU-03 and DTU-08); Location confidence: 90%; It is recommended to immediately inspect this section of the line and check the distribution box and water pumps and other equipment in humid environments inside the farm."

[0109] Upon receiving the information, maintenance personnel went directly to the designated line section and the fish farm for inspection. They quickly discovered that the insulation tape at a line joint located near the fishpond was damaged, causing a short circuit due to moisture. They immediately applied insulation and the system monitoring showed that the leakage had disappeared.

[0110] Throughout the entire process, the power supply to the transformer substation remained uninterrupted. From system alarms to fault location, and then to manual on-site confirmation and handling, the efficiency was greatly improved compared to traditional blind inspections.

[0111] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A smart system for detecting and locating leakage current in transformer substations, characterized in that, It includes a distributed synchronous monitoring network, an edge intelligent analysis unit, and a central diagnostic and positioning platform; the distributed synchronous monitoring network consists of intelligent detection terminals, which are deployed at key nodes in the target area to synchronously collect synchronous current datasets from each monitoring point, including the instantaneous value of the residual current; The edge intelligent analysis unit is connected to the intelligent detection terminal to receive and process synchronous current datasets; The edge intelligent analysis unit includes a vector analysis engine and a dynamic baseline management module; The vector analysis engine generates vector analysis results; the dynamic baseline management module collects synchronous current datasets within a preset learning period and establishes a static baseline library. The system monitors load changes and seasonal environmental factors in the transformer area. When a change is detected, a dynamic baseline update mechanism is triggered. Based on the latest normal operation data, the baseline parameters in the static baseline library under the corresponding conditions are smoothly corrected to generate a dynamic current baseline. The central diagnostic and positioning platform is connected to the edge intelligent analysis unit, which includes a fault diagnosis core and a collaborative positioning engine. The fault diagnosis core includes an adaptive threshold determination algorithm and receives the deviation statistical characteristics from the vector analysis results. The system acquires a dynamic current baseline and extracts the corresponding normal deviation range threshold from it. It introduces environmental correction factors and load correction factors to adjust the normal deviation range threshold and generates an adaptive judgment threshold. It compares the statistical characteristics of the deviation with the adaptive judgment threshold to generate a leakage current diagnosis conclusion. The collaborative positioning engine receives the leakage current diagnosis conclusion and acquires the transformer area topology information. Based on the amplitude and phase relationship of the instantaneous residual current values ​​of each node, the main area where leakage occurs is located; the synchronous current dataset of the downstream intelligent detection terminal in the located main area is retrieved, and the change in current vector between adjacent nodes is calculated using the differential vector comparison algorithm to mark the branches with high suspicion of leakage. Calculate the location reliability by combining historical data and output the fault location information.

2. The intelligent detection and fault location system for transformer substation leakage current according to claim 1, characterized in that: The distributed synchronous monitoring network has a built-in high-precision clock synchronization module, which ensures that the current acquisition actions of all intelligent detection terminals are carried out under a unified time scale, so as to obtain a time-aligned synchronous current dataset.

3. The intelligent detection and fault location system for transformer substation leakage current according to claim 2, characterized in that: Each intelligent detection terminal includes a current acquisition module, which includes a phase current acquisition unit and a residual current acquisition unit; the phase current acquisition unit acquires the instantaneous values ​​of the three-phase phase currents, and the residual current acquisition unit acquires the instantaneous values ​​of the residual current; The vector analysis engine calculates the instantaneous vector sum of the three-phase currents in real time based on their instantaneous values, and compares the instantaneous vector sum with the synchronously acquired instantaneous value of the residual current to generate the vector analysis result.

4. The intelligent detection and fault location system for transformer substation leakage current according to claim 1, characterized in that: The system also includes a system power supply and communication module. The system power supply and communication module adopts a voltage-mode power supply method and uses an open-type clamp-on current transformer to clamp onto the conductor under test without disconnecting the line to realize live detection. It also obtains working power through a small micro current transformer to ensure that the entire detection process does not require power outages.

5. The intelligent detection and fault location system for transformer substation leakage current according to claim 3, characterized in that: The system also includes a data visualization and human-computer interaction interface, which provides a dynamic vector display function. It draws the position and trajectory of the instantaneous vector sum of the three-phase currents based on the synchronous current dataset, and displays the instantaneous value vector of the remaining current superimposed on the same vector.

Images