Low-voltage distribution area electric leakage fault positioning and early warning method and product based on edge computing

Abstract

The application discloses a low-voltage distribution area electric leakage fault positioning and early warning method and product based on edge calculation, and relates to the technical field of area operation and maintenance.The application deploys a data collection unit integrated with an edge calculation module on the low-voltage side of a transformer, supports data analysis and processing at the edge, effectively reduces delay to improve response, deploys corresponding branch detection units on each branch line, deploys corresponding intelligent circuit breakers on each user meter, and through multi-point synchronous sampling and analysis, can position electric leakage faults to specific user meter boxes.The application constructs a tree topology of the entire area and real-time monitors the total residual current of the area, through the cooperation of setting early warning conditions and early warning pre-judgment, effectively realizes early warning while reducing false signal false positives.

Description

Technical Field

[0001] This invention relates to the field of transformer substation operation and maintenance technology, specifically to: 1. a method for locating and warning of leakage faults in low-voltage distribution transformer substations based on edge computing; 2. a computer program product. Background Technology

[0002] Currently, low-voltage distribution substations (such as public transformers and rural power grid substations) generally use a main circuit breaker (i.e., a main residual current operated protective device) as the primary means of protection against electric shock and equipment leakage. This approach has the following significant drawbacks: 1. Poor fault location accuracy: When the main circuit breaker trips, maintenance personnel need to check the leakage points on the branch or user side one by one, which takes a long time (usually several hours) and affects the reliability of power supply.

[0003] 2. Severe data silos: The data of terminal equipment in the transformer area is not effectively integrated. Instead, it relies solely on the total residual current value of a single transformer area for judgment, which is highly susceptible to interference from harmonics and load fluctuations, leading to false tripping or failure to trip.

[0004] 3. Communication depends on the main station: Existing solutions mostly upload data to the main station for analysis, resulting in high response latency and making it impossible to make rapid decisions locally.

[0005] 4. The early warning design is too lax: the main protection only activates when the total leakage current of the transformer area (i.e., the total residual current of the transformer area) exceeds the action threshold (set according to the line capacity, generally set to 300mA), which cannot achieve early intervention. Summary of the Invention

[0006] Therefore, it is necessary to provide a method and product for locating and warning of leakage faults in low-voltage distribution substations based on edge computing, which addresses the problems of difficulty in locating faults, frequent false trips, and slow response in existing low-voltage distribution substation leakage protection methods that use a central protection system.

[0007] This invention is achieved using the following technical solution: In a first aspect, the present invention discloses a method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing, comprising the following steps: Step 1: Deploy a data collection unit with integrated edge computing module on the low-voltage side of the transformer, deploy corresponding branch detection units on each branch line to collect environmental and electrical data of the branch lines, and deploy corresponding smart circuit breakers on each user's electricity meter to support remote opening and closing and leakage detection. Step 2: The data aggregation unit issues a topology discovery command and receives the device ID and parent node information returned by each branch detection unit and each smart circuit breaker to generate a tree topology that represents the circuit connection status of the low-voltage distribution substation. In the tree topology, the transformer is the first-level node, each branch monitoring unit is the second-level node, and each smart circuit breaker is the third-level node. Step 3: The data collection unit monitors the total residual current of the transformer area in real time according to the preset sampling period; If the total residual current value of the transformer area meets the early warning conditions, the early warning pre-judgment is triggered; otherwise, step three is repeated. The early warning pre-judgment is based on the multi-source data fusion analysis algorithm to identify interference in the multi-source data of each node in order to eliminate false signals. If the pre-warning judgment passes, it is determined that a fault has occurred, and an early warning and leakage current location are initiated to the corresponding third-level node; otherwise, it is determined to be a false signal, and the early warning is canceled.

[0008] This edge computing-based method for locating and warning of leakage faults in low-voltage distribution substations implements the method or process according to embodiments of this disclosure.

[0009] Secondly, this invention discloses a computer program product, including a computer program. When executed by a processor, this computer program implements the steps of the edge computing-based low-voltage distribution substation leakage fault location and early warning method disclosed in the first aspect.

[0010] This type of computer program product implements the methods or processes according to embodiments of the present disclosure.

[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. On the one hand, the present invention deploys a data collection unit with an integrated edge computing module on the low-voltage side of the transformer to support data analysis and processing at the edge, effectively reducing latency and improving response; on the other hand, it deploys corresponding branch detection units on each branch line and corresponding smart circuit breakers on each user meter, and through multi-point synchronous sampling and analysis, leakage faults can be located to specific user meter boxes.

[0012] 2. This invention constructs a tree-like topology for the entire transformer area and monitors the total residual current of the transformer area in real time. By setting early warning conditions and combining early warning pre-judgment, it reduces false alarms of pseudo signals while achieving effective early warning.

[0013] 3. The warning threshold used in the warning conditions of the present invention is a dynamic threshold that is set based on sensitivity and is lower than the action threshold, which can achieve early intervention.

[0014] 4. The early warning judgment of the present invention is based on multi-source data composed of power data and corresponding environmental data, and performs corresponding feature extraction and weighted fusion to obtain a comprehensive risk index for threshold judgment, so as to avoid misjudgment caused by single data.

[0015] 5. The leakage current location of the present invention is designed based on the correlation between the residual current of each branch line and the total residual current of the transformer area. The overdetermined equation is solved by the least squares method to find the third-level node corresponding to the largest leakage current state coefficient as the leakage point, so as to achieve a clear fault location.

[0016] 6. This invention supports individual power disconnection for users with leakage after leakage location, achieving precise fault isolation, avoiding power outages for non-faulty users, directly improving power supply reliability, and significantly reducing manual inspection costs. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of the low-voltage distribution substation leakage fault location and early warning method based on edge computing provided in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of a low-voltage distribution substation after step one, as provided in Embodiment 1 of the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] It should be noted that when a component is said to be "installed on" another component, it can be directly on the other component or it may be in a component that is centered on it. When a component is said to be "set on" another component, it can be directly set on the other component or it may also be in a component that is centered on it. When a component is said to be "fixed to" another component, it can be directly fixed to the other component or it may also be in a component that is centered on it.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "or / and" as used herein includes any and all combinations of one or more of the associated listed items.

[0022] First, it should be noted that a low-voltage distribution substation typically follows a three-level tree topology, which includes: a transformer connected to the main substation (located on the grid side) on the high-voltage side; several branch boxes connected to the low-voltage side of the transformer; and several user meter boxes connected to each branch box (forming several branch lines). The user meter box serves as the connection point on the user side and contains several user meters—each user meter corresponding to a specific user.

[0023] See Figure 1 This embodiment 1 provides a method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing, including the following steps: Step 1: Deploy a data collection unit with integrated edge computing module on the low-voltage side of the transformer, deploy corresponding branch detection units on each branch line to collect environmental and electrical data of the branch lines, and deploy corresponding smart circuit breakers on each user's electricity meter to support remote opening and closing and leakage current detection.

[0024] Therefore, for the data aggregation unit, the address of each branch detection unit and each smart circuit breaker is known.

[0025] See Figure 2 This is a schematic diagram of the low-voltage distribution substation after step one.

[0026] It should be noted that: As an edge intelligent analysis platform, the edge computing module does not rely on the main station on the power grid side. It can perform data analysis and processing at the edge, effectively reducing latency and improving response. The edge computing module can use an MCU (such as an ARM architecture chip, a RISC-V architecture chip, etc.) or an FPGA acceleration card. In this embodiment 1, it is recommended to use an ARM Cortex-A55 processor for the edge computing module to achieve a balance between cost and performance.

[0027] In addition to integrating the edge computing module, the data aggregation unit also integrates a communication module (supporting multiple communication protocols, such as 4G, 5G, NB-IoT, LoRa, HPLC, etc.) to achieve communication between upstream and downstream devices, and a storage unit (on-chip eMMC is recommended) to ensure data computation and storage.

[0028] The branch detection unit is adapted for low-voltage power distribution network monitoring. It integrates not only environmental temperature and humidity sensors, but also Rogowski coils (or other types of current transformers), and can synchronously collect environmental data (such as temperature, humidity, etc.) and electrical data (such as voltage, current, residual current, etc.) of the branch lines.

[0029] Intelligent circuit breakers are devices that control and protect low-voltage power distribution networks. They integrate voltage phase angle detection modules, residual current transformers, and support remote opening and closing as well as leakage current detection.

[0030] Step 2: The data aggregation unit issues a topology discovery command and receives the device ID and parent node information returned by each branch detection unit and each smart circuit breaker to generate a tree topology that represents the circuit connection status of the low-voltage distribution substation.

[0031] It should be noted that: for a branch monitoring unit, its parent node is the transformer; for a smart circuit breaker, its parent node is the branch monitoring unit to which it belongs.

[0032] In the tree topology, the transformer is the first-level node, each branch monitoring unit is the second-level node, and each smart circuit breaker is the third-level node, and each is assigned a unique logical number.

[0033] Step 3: The data collection unit monitors the total residual current of the transformer area in real time according to the preset sampling period; If the total residual current value of the transformer area meets the early warning conditions, the early warning pre-judgment is triggered; otherwise, step three is repeated. If the pre-warning judgment passes, it is determined that a fault has occurred, and an early warning and leakage current location are initiated to the corresponding third-level node; otherwise, it is determined to be a false signal, and the early warning is canceled. It should be noted that: 1. The warning condition is set as follows: the total residual current in the transformer area continuously exceeds the warning threshold I0 for a preset duration, and the overall trend is upward. The preset duration t0 is generally set to 5 minutes (but can be fine-tuned according to the actual situation, for example, fluctuating by 2 minutes); the warning threshold I0 is not a fixed value, but a dynamic threshold set according to sensitivity.

[0034] Existing designs using total protection typically set the action threshold to 300mA. This method, aiming for early intervention, sets the warning threshold I0 below the action threshold. Specifically: in high-sensitivity mode, the warning threshold I0 is set to 1 / 6 to 1 / 3 of the action threshold (70mA recommended); in low-sensitivity mode, the warning threshold I0 is set to 65% to 75% of the action threshold (210mA recommended).

[0035] 2. The early warning judgment is mainly based on the multi-source data fusion analysis algorithm to identify interference in the multi-source data of each node in order to eliminate false signals. The process is as follows: S301, Perform data synchronization call: The data collection unit initiates a high-priority data call to all nodes to obtain multi-source data (specifically including: power data and corresponding environmental data) that are accurately synchronized by each node.

[0036] In this embodiment 1, the power data type is selected as residual current and three-phase voltage; the environmental data type is selected as ambient humidity.

[0037] S302, perform spurious signal elimination—the specific discrimination steps are as follows: S3021, Feature extraction based on multi-source data of a certain node: ① First, denoise the residual current I_rms of the node (wavelet denoising is recommended), and calculate the linear regression slope I_rms_trend of the denoised signal within the most recent preset time window (i.e., starting from the current time and working backwards to the interval ΔT) - which reflects the changing trend of the residual current.

[0038] It should be noted that ΔT is recommended to be set to 10s~30s. This is because leakage faults (especially those caused by insulation aging or moisture) usually have a gradual development trend. A window of 10s-30s can capture this gradual trend and effectively smooth out millisecond-level pulse interference.

[0039] ② First, the three-phase voltage U_abc of this node is denoised. Then, the fundamental phase angle θ of the voltage is extracted from the denoised signal. Next, the deviation sequence Δθ between θ and the rated phase angle θ_base is calculated. 3ph Then, we obtain its standard deviation normalized value U_flu, which reflects the degree of abnormality of voltage fluctuation.

[0040] It should be noted that, since this is a three-phase electrical system, θ actually includes the fundamental phase angles of the voltages for phases A, B, and C (which can be represented as θa, θb, and θc, respectively). Similarly, θ_base also includes the rated phase angles of phases A, B, and C (which can be represented as θa,rated, θb,,rated, and θc,rated, respectively) – which can be directly calculated using the theoretical standard values ​​of power grid design: if the rated phase angle of phase A is 0°, the rated phase angle of phase B is -120° or 240°, and the rated phase angle of phase C is 120°, then Δθ 3ph The specific calculation formula is as follows: ; It should be noted that θ is a numerical sequence that varies with time t, Δθ 3ph It is also a numerical sequence that changes with time t, so for Δθ 3ph Calculate the standard deviation and normalize it to obtain U_flu.

[0041] ③ Compare the ambient humidity H of this node with the lower limit of the humidity threshold H_set1 (generally set to 40%) and the upper limit of the humidity threshold H_set2 (generally set to 70%). If H does not exceed H_set1, it indicates that the environment is dry. In this case, the humidity correction factor H_hum is lowered to 0.9 to appropriately reduce the risk assessment in a dry environment. The judgment is mainly based on current and voltage characteristics to avoid interference from environmental factors. If H exceeds H_set1 but does not exceed H_set2, it indicates that the environment is in a normal humidity environment. The humidity correction factor H_hum is kept at the default value of 1. If H exceeds H_set2, it indicates that the environment is humid. In this case, the humidity correction factor H_hum is increased (it is recommended to use a linear growth function for calculation) to increase the role of humidity in risk assessment.

[0042] Specifically, the linear growth function is designed as follows: H_hum = 1 + k × (H - Hset); In the formula, k represents the growth slope (ranging from 0.01 to 0.05), which is used to adjust the weight of the influence of humidity on the comprehensive risk index.

[0043] However, it should be noted that H_hum cannot be increased indefinitely; its upper limit is set to 2.

[0044] S3022, I_rms_trend, U_flu, and H_hum are weighted and fused to obtain the comprehensive risk index R_risk of the node.

[0045] This is to avoid misjudgment caused by a single data source, and to introduce corresponding weights for multi-source data to calculate R_risk.

[0046] Specifically, the formula for calculating R_risk is: R_risk=W1×I_rms_trend+W2×U_flu+W3×H_hum; In the formula, W1, W2, and W3 represent the corresponding weight coefficients, and satisfy: W1 + W2 + W3 = 1.

[0047] In this embodiment 1, the recommended values ​​are: W1=0.5, W2=0.3, W3=0.2.

[0048] S3023, compare R_risk with the leakage current threshold T_dynamic: If R_risk≥T_dynamic, it is determined to be a leakage signal, not a false signal. The pre-warning judgment is passed, and the subsequent leakage location is entered.

[0049] If R_risk < T_dynamic, it is determined to be a spurious signal (which may be caused by instantaneous harmonic interference or sensor noise). The pre-warning judgment fails, and the spurious signal needs to be excluded and no warning is issued.

[0050] Additionally, it should be noted that T_dynamic is recommended to adaptively adjust based on ambient humidity, and its mathematical formula is as follows: T_dynamic=T_base×[1+α×(H_hum-1)]; In the formula, T_base represents the baseline threshold (usually taken as 0.5); α represents the adjustment coefficient (usually taken as 0.2).

[0051] However, it should be noted that although it is a false signal, there is indeed an anomaly in the area. Therefore, it is still recommended to send a reminder to the maintenance personnel so that they can intervene in advance.

[0052] S3024, traverse all nodes, repeating S3021~S3023.

[0053] 3. The residual current variation trend of the branch where the leakage point is located is highly consistent with the total residual current variation trend of the transformer area - this is the core basis for leakage location.

[0054] Leakage current location supports locating the corresponding third-level node. Specifically, the leakage current location process is as follows: S303, Perform preliminary correlation screening: The correlation between the residual current of each branch line and the total residual current of the transformer area was calculated using the weighted correlation coefficient method.

[0055] If R i Let represent the correlation between the i-th branch line and the total residual current of the transformer area. Its calculation formula is: ; In the formula, N represents the total number of sampling periods; I total (t) represents the measured value of the total residual current in the transformer area during the t-th sampling period; I total_avg I represents the average value of the total residual current in the transformer area over N sampling periods; branch_i (t) represents the measured value of the residual current of the i-th branch line in the t-th sampling period; I branch_i_avg This represents the average value of the residual current of the i-th branch line over N sampling periods.

[0056] In other words, R i Characterizing the residual current of the i-th branch line and the total residual current of the transformer area in history N Correlation of rates of change within each sampling period.

[0057] Next, mark the top K (K is recommended to be 3 or less) branches with the highest correlation values ​​as suspected faulty branches.

[0058] S304, for precise impedance positioning: It is recommended to use the least squares method based on line impedance to locate the leakage point in the suspected faulty branch.

[0059] Specifically, S304 includes: S3041, for the xth suspected faulty branch, obtain the ground voltage of its third-level node (when leakage occurs, the leakage point will generate a ground voltage - which is the "source" driving the leakage current, which can be obtained through the smart circuit breaker), and combine it with its line impedance parameters (generally entered during the construction of the transformer area or obtained through online impedance identification) to construct the overdetermined equation: ; In the formula, I branch_x This represents the measured value of the residual current of the line in the x-th suspected fault branch; M represents the total number of third-level nodes in the x-th suspected fault branch (i.e., the total number of users under the x-th suspected fault branch). This represents the voltage to ground of the j-th third-level node within the x-th suspected faulty branch; This represents the line impedance between the second-level node corresponding to the x-th suspected faulty branch and the j-th third-level node within it; This represents the leakage current status coefficient of the j-th third-level node within the x-th suspected fault branch: if the j-th user within the x-th suspected fault branch is a normal user and has no leakage current, Close to 0; otherwise Significantly greater than 0.

[0060] S3042, the above overdetermined equations are solved using the least squares method, and the calculation is performed to make... The coefficient vector with the smallest residual value. So, let's take K. x The third-level node corresponding to the leakage current state coefficient with the largest median value is taken as the leakage point.

[0061] S3043, traverse K suspected fault branches until all leakage points are located.

[0062] The unique logical number corresponding to each leakage point serves as the location result—which can be used to locate the specific user's meter box. The location result can be simultaneously sent to maintenance personnel, effectively shortening the repair and location time.

[0063] It should be noted that the above method for locating leakage current can be replaced by other equivalent methods, such as using machine learning models (such as random forest, LSTM) or graph neural networks (GNN) for leakage current location.

[0064] It should also be emphasized that the calculation and analysis process in step three above is undertaken by the edge computing module, and all analysis and decision-making can be completed on the edge side without relying on the main station.

[0065] Of course, further fault isolation can be performed: S305 sends a tripping command to the corresponding smart circuit breaker based on the positioning result, isolating it from the power grid.

[0066] The S305 can also be configured so that the operation and maintenance personnel decide whether to send a trip command after receiving the location results.

[0067] Traditional methods using a main circuit breaker can cause a power outage in the entire distribution area due to leakage current tripping. This method, however, achieves fault isolation—since the power supply to non-faulty areas is not interrupted (the main circuit breaker does not trip), there is no need for restoration. Instead, it achieves a self-healing effect of precise fault isolation and zero-impact on non-faulty areas. This approach avoids the economic losses and decreased user experience caused by a power outage in the entire distribution area.

[0068] In summary, this method has the following outstanding advantages: 1. High positioning accuracy: Through multi-point synchronous sampling and analysis, leakage faults can be located to specific user meter boxes with a positioning accuracy of 95%, and the troubleshooting time is shortened from hours to within 5 minutes.

[0069] 2. Strong early warning capability: Supports different sensitivity designs and provides corresponding early warnings to effectively prevent accidents.

[0070] 3. Good anti-interference performance: It integrates multi-dimensional data such as electrical data and environmental data to effectively distinguish between real leakage current and false signals such as harmonic interference and load change, reducing the false trip rate by more than 80%.

[0071] 4. High autonomy: All analysis and decision-making are completed at the edge, without relying on the main station, and the response latency is less than 1 second.

[0072] 5. Significant economic benefits: Reduces the number of households affected by power outages (avoiding power outages for non-faulty users), improves power supply reliability (i.e., improves SAIDI index), and reduces manual inspection costs.

[0073] Example 2 This embodiment 2 provides a specific application example of the low-voltage distribution transformer area leakage fault location and early warning method based on edge computing provided in embodiment 1: This method was deployed in the JD-09019 transformer substation on the 10kV Haiwang 11 line in a certain urban residential area, and set to high-sensitivity mode (warning threshold set to 70mA). The data collection unit communicates with the main station via 4G, connecting to 12 branch monitoring units (one per building) and 200 miniature circuit breakers. One day, the total residual current in the substation was detected to slowly rise from 20mA to 80mA and remain there for 5 minutes, triggering a warning. After recalling the data, it was found that the residual current in the branch of Building #3 had synchronously risen to 75mA, while the other branches were normal. Impedance localization determined that the problem was due to aging wiring in a user's unit 2 of Building #3. Maintenance personnel went directly to that user to handle the issue, preventing a power outage in the entire substation area.

[0074] Among the detected incidents, the residual current of the main protection device rose to 80mA and remained there for 5 minutes—although this did not reach the 300mA activation threshold, it still triggered an early warning under the high-sensitivity model. Through subsequent precise location of the fault, maintenance personnel were able to address potential problems before they escalated, demonstrating the value of the early warning mechanism.

[0075] The edge computing module calculated the correlation coefficient R between the 12 branches and the total protection. i The R3 of the branch at building #3 was found to be 0.98, significantly higher than other branches (all <0.3), therefore it was listed as the only suspected faulty branch. (The last sentence appears to be incomplete and possibly refers to an error in the original text.) i This significantly reduced the scope of complex impedance calculations required, from 200 households across the entire district to just 10 households, greatly improving positioning speed and efficiency.

[0076] Since there are only a few users (let's say 10) under Building #3, impedance-based precise location was applied directly to these 10 users. Using the least squares method, it was found that a user in Unit 2 had the highest leakage current state coefficient, thus identifying this user as the leakage point.

[0077] Example 3 This embodiment 3 discloses a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the steps of the low-voltage distribution substation leakage fault location and early warning method based on edge computing disclosed in this embodiment 3.

[0078] The computer equipment can be either a mobile terminal or a fixed terminal. Examples of the former include mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (such as in-vehicle navigation terminals); examples of the latter include digital TVs and desktop computers.

[0079] This embodiment 3 also discloses a readable storage medium that stores computer program instructions. When the computer program instructions are read and executed by a processor, the steps of the low-voltage distribution substation leakage fault location and early warning method based on edge computing disclosed in embodiment 3 are performed.

[0080] The readable storage medium may include, but is not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination of the above.

[0081] This embodiment 3 also discloses a computer program product, including a computer program. When executed by a processor, the computer program implements the steps of the edge computing-based low-voltage distribution substation leakage fault location and early warning method disclosed in embodiment 3.

[0082] It should be noted that the computer program used to execute the above can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as C or similar languages. The computer program can be executed entirely on the user's computer, partially on the user's computer, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer through any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN).

[0083] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0084] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for locating and warning of leakage current faults in low-voltage distribution transformer areas based on edge computing, characterized in that, include: Step 1: Deploy a data collection unit with integrated edge computing module on the low-voltage side of the transformer, deploy corresponding branch detection units on each branch line to collect environmental and electrical data of the branch lines, and deploy corresponding smart circuit breakers on each user's electricity meter to support remote opening and closing and leakage detection. Step 2: The data aggregation unit issues a topology discovery command and receives the device ID and parent node information returned by each branch detection unit and each smart circuit breaker to generate a tree topology that represents the circuit connection status of the low-voltage distribution substation. In the tree topology, the transformer is the first-level node, each branch monitoring unit is the second-level node, and each smart circuit breaker is the third-level node. Step 3: The data collection unit monitors the total residual current of the transformer area in real time according to the preset sampling period; If the total residual current value of the transformer area meets the early warning conditions, the early warning pre-judgment is triggered; otherwise, step three is repeated. The early warning pre-judgment is based on the multi-source data fusion analysis algorithm to identify interference in the multi-source data of each node in order to eliminate false signals. If the pre-warning judgment passes, it is determined that a fault has occurred, and an early warning and leakage current location are initiated to the corresponding third-level node; otherwise, it is determined to be a false signal, and the early warning is canceled.

2. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 1, characterized in that, In step two, the parent node of the branch monitoring unit is the transformer; the parent node of the smart circuit breaker is the branch monitoring unit to which it belongs.

3. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 1, characterized in that, In step three, the warning condition is set as follows: the total residual current in the transformer area continues to exceed the warning threshold I0 for a preset duration T0, and the overall trend is upward. Where I0 is the dynamic threshold set based on sensitivity.

4. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 3, characterized in that, T0 is set to 5 minutes; Or / and, in high sensitivity mode, I0 is 1 / 6 to 1 / 3 of the action threshold; In low sensitivity mode, I0 is 65% to 75% of the action threshold.

5. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 1, characterized in that, In step three, the pre-warning judgment includes: S301, the data aggregation unit initiates a high-priority data call to all nodes to obtain multi-source data that is accurately synchronized by each node; among which, the multi-source data includes: residual current, three-phase voltage, and ambient humidity; S302, perform spurious signal elimination; S302 includes: S3021, feature extraction is performed based on multi-source data of a certain node; the feature extraction process includes: first, denoising the residual current I_rms of the node, and then calculating the linear regression slope I_rms_trend of the denoised signal within the most recent preset time window; second, denoising the three-phase voltage U_abc of the node, then extracting the fundamental phase angle θ from the denoised signal, and then calculating the deviation sequence Δθ between θ and the rated phase angle θ_base. 3ph Then, its standard deviation normalized value U_flu is obtained; the ambient humidity H of this node is compared with the lower limit of the humidity threshold H_set1 and the upper limit of the humidity threshold H_set2; if H does not exceed H_set1, the humidity correction factor H_hum is reduced to 0.9; if H exceeds H_set1 but does not exceed H_set2, H_hum remains at 1; if H exceeds H_set2, H_hum is increased. S3022, I_rms_trend, U_flu, and H_hum are weighted and fused to obtain the comprehensive risk index R_risk of the node; Where R_risk = W1 × I_rms_trend + W2 × U_flu + W3 × H_hum; In the formula, W1, W2, and W3 represent the corresponding weight coefficients, and satisfy: W1 + W2 + W3 = 1; S3023, if R_risk≥T_dynamic, then it is determined to be a leakage signal, not a false signal, and the pre-warning judgment is passed; If R_risk < T_dynamic, it is determined to be a false signal, and the pre-warning judgment fails; Where T_dynamic represents the leakage current threshold, and its mathematical formula is: T_dynamic=T_base×[1+α×(H_hum−1)]; In the formula, T_base represents the baseline threshold; α represents the adjustment coefficient; S3024, traverse all nodes, repeating S3021~S3023.

6. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 5, characterized in that, Δθ 3ph The calculation formula is: ； In the formula, θa, θb, and θc represent the fundamental phase angles of the voltages in phases A, B, and C, respectively; θa,rated, θb,,rated, and θc,rated represent the rated phase angles of phases A, B, and C, respectively. Or / and, when H exceeds H_set2, H_hum is calculated using a linear growth function; The linear growth function is: H_hum = 1 + k × (H - Hset); In the formula, k represents the growth slope; Or / and, the upper limit of H_hum is 2.

7. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 6, characterized in that, Step three includes methods for locating leakage current: S303, the weighted correlation coefficient method is used to calculate the correlation between the residual current of each branch line and the total residual current of the transformer area. Then, the K branch lines with the largest correlation values ​​are marked as suspected fault branches. S304 uses the least squares method based on line impedance to locate the leakage point in the suspected faulty branch.

8. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 7, characterized in that, In S303, the correlation R between the i-th branch line and the total residual current of the transformer area is... i The calculation formula is: ； In the formula, N represents the total number of sampling periods; I total (t) represents the measured value of the total residual current in the transformer area during the t-th sampling period; I total_avg I represents the average value of the total residual current in the transformer area over N sampling periods; branch_i (t) represents the measured value of the residual current of the i-th branch line in the t-th sampling period; I branch_i_avg This represents the average value of the residual current of the i-th branch line over N sampling periods; Or / and, S304 includes: S3041, for the xth suspected fault branch, obtain the ground voltage of its third-level node, and combine it with its line impedance parameters to construct an overdetermined equation; The overdetermined equations are as follows: ； In the formula, I branch_x This represents the measured value of the residual current of the line in the x-th suspected fault branch; M represents the total number of third-level nodes in the x-th suspected fault branch; This represents the voltage to ground of the j-th third-level node within the x-th suspected faulty branch; This represents the line impedance between the second-level node corresponding to the x-th suspected faulty branch and the j-th third-level node within it; This represents the leakage current state coefficient of the j-th third-level node within the x-th suspected fault branch: S3042, using the least squares method to solve the overdetermined equations, calculates... Minimum coefficient vector ; Take K x The third-level node corresponding to the leakage current state coefficient with the largest median value is taken as the leakage point. S3043, traverse K suspected fault branches until all leakage points are located.

9. The method for locating and warning of leakage faults in low-voltage distribution transformer areas based on edge computing as described in claim 7, characterized in that, Step three, the method for locating leakage current also includes: S305 sends a tripping command to the corresponding smart circuit breaker based on the positioning result, isolating it from the power grid.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the low-voltage distribution substation leakage fault location and early warning method based on edge computing as described in any one of claims 1-9.

Images