Power distribution network maintenance method and device, computer equipment, readable storage medium and program product

Abstract

The invention relates to a power distribution network maintenance method and device, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a voltage value, acquired by a voltage sensor, of each detection node in a power distribution network, and acquiring a potential difference between every two detection nodes based on the voltage values; under the condition that the potential difference is not in the corresponding target potential difference range, inputting a potential difference signal corresponding to the potential difference into the trained K adjacent model to obtain an abnormal state category corresponding to the potential difference signal; and determining a fault category and a fault position of the power distribution network based on the abnormal state category, and maintaining the power distribution network based on the fault category and the fault position. According to the method provided by the invention, by maintaining the power distribution network, the power supply quality of the power distribution network can be improved, and the risk and cost of system operation and maintenance can be reduced.

Description

Technical Field

[0001] The present application relates to the technical field of power grid data processing, and in particular to a distribution network maintenance method, apparatus, computer equipment, computer-readable storage medium, and computer program product. Background Art

[0002] With the transformation and upgrading of modern distribution networks towards automation, informatization and intelligence, the requirements for voltage stability and power supply reliability are constantly increasing, and the environment faced by distribution networks is becoming more complex and changeable. In the daily operation and maintenance of distribution networks, it is extremely important to maintain stable and reliable power supply quality. In order to achieve this goal, it is necessary to perceive the distribution network in real time and respond quickly to potential abnormal situations. Existing technologies usually monitor the status of distribution networks through voltage fluctuations, but existing technologies are difficult to quickly and accurately locate problems and take effective countermeasures when encountering abnormal voltage fluctuations, which not only affects the power supply quality, but also increases the risk and cost of system operation and maintenance. Summary of the invention

[0003] Based on this, it is necessary to provide a distribution network maintenance method, device, computer equipment, computer-readable storage medium and computer program product that can effectively improve the power supply quality of the distribution network in response to the above technical problems.

[0004] In a first aspect, the present application provides a distribution network maintenance method, the method comprising:

[0005] Acquire the voltage value of each detection node in the distribution network collected by the voltage sensor, and acquire the potential difference between every two detection nodes based on the voltage value;

[0006] When the potential difference is not within the corresponding target potential difference range, inputting the potential difference signal corresponding to the potential difference into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal;

[0007] A fault category and a fault location of the power distribution network are determined based on the abnormal state category, and maintenance is performed on the power distribution network based on the fault category and the fault location.

[0008] In one embodiment, the process of determining the target potential difference range includes:

[0009] Based on the load condition and network topology of the distribution network, obtaining an initial potential difference range between different detection nodes in the distribution network;

[0010] For every two detection nodes, based on the node types, node positions and network topology of the two detection nodes, the initial potential difference range is adjusted to obtain a corresponding intermediate potential difference range of the two detection nodes;

[0011] The current temperature and the current humidity of the environment in which the two detection nodes are located are obtained, and the intermediate potential difference range is adjusted based on the current temperature and the current humidity to obtain the target potential difference range corresponding to the two detection nodes.

[0012] In one embodiment, the training process of the K-neighbor model includes:

[0013] Acquire sample potential difference signals and preset distance measurement criteria corresponding to the potential difference signals, each sample potential difference signal is marked with a corresponding state category;

[0014] Preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, wherein the optimal K value is used to predict the category of the sample in the K neighbor model, wherein the preprocessing includes noise reduction processing and missing value interpolation;

[0015] Based on the optimal K value and the preset distance metric criterion, construct a pre-trained K nearest neighbor model;

[0016] The sample potential difference signal is input into the pre-trained K-neighbor model to obtain the trained K-neighbor model.

[0017] In one embodiment, determining the optimal K value based on the preprocessed sample potential difference signal includes:

[0018] Extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform;

[0019] Based on the amplitude, the frequency domain characteristics and the time domain characteristics, a plurality of initial statistical characteristics of the sample potential difference signal are obtained;

[0020] Determining at least one target statistical feature from the multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process;

[0021] The optimal K value is determined based on the target statistical characteristics.

[0022] In one embodiment, inputting the sample potential difference signal into the pre-trained K-neighbor model to obtain the trained K-neighbor model includes:

[0023] Obtaining at least one performance evaluation indicator output by the pre-trained K-neighbor model;

[0024] When at least one performance evaluation indicator does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion;

[0025] Based on the adjusted optimal K value and the preset distance metric criterion, return to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance metric criterion, and continue to execute until each performance evaluation indicator meets the corresponding preset indicator range.

[0026] In one embodiment, the potential difference signal is transmitted in a transmission area of ​​the power distribution network via a target radio signal; and the process of determining the target radio signal includes:

[0027] Obtaining initial frequencies and initial bands of all radio signals in the transmission area, and obtaining the number of times each initial frequency and each initial band is used in the transmission area;

[0028] Determine the frequency and the band with the least number of uses among the initial frequencies and the initial bands as the target frequency and the target band respectively;

[0029] Determining transmission power based on the area and terrain characteristics of the transmission area;

[0030] The target radio signal is determined based on the target frequency, the target band, the transmission power, a preset antenna gain, and a preset channel bandwidth.

[0031] In a second aspect, the present application further provides a distribution network maintenance device, the device comprising:

[0032] A first acquisition module is used to acquire a voltage value of each detection node in the distribution network collected by a voltage sensor, and acquire a potential difference between every two detection nodes based on the voltage value;

[0033] A first input module is used for inputting a potential difference signal corresponding to the potential difference into a trained K-neighbor model to obtain an abnormal state category corresponding to the potential difference signal when the potential difference is not within a corresponding target potential difference range;

[0034] A determination module is used to determine a fault category and a fault location of the distribution network based on the abnormal state category, and to maintain the distribution network based on the fault category and the fault location.

[0035] In a third aspect, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in any one of the above embodiments when executing the computer program.

[0036] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the method in any one of the above embodiments are implemented.

[0037] In a fifth aspect, the present application further provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the method in any one of the above embodiments are implemented.

[0038] The distribution network maintenance method, device, computer equipment, computer-readable storage medium and computer program product described above obtain the voltage value of each detection node in the distribution network collected by the voltage sensor, and obtain the potential difference between every two detection nodes based on the voltage value; when the potential difference is not within the corresponding target potential difference range, the potential difference signal corresponding to the potential difference is input into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal; the fault category and fault location of the distribution network are determined based on the abnormal state category, and the distribution network is maintained based on the fault category and fault location. The method provided in the present application can improve the power supply quality of the distribution network and reduce the risk and cost of system operation and maintenance by maintaining the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0040] Figure 1 A schematic diagram of a flow chart of a distribution network maintenance method in one embodiment;

[0041] Figure 2 Schematic diagram of a flow chart of a method for determining a target potential difference range in one embodiment;

[0042] Figure 3 A schematic flow chart of a distribution network maintenance method according to another embodiment;

[0043] Figure 4 is a structural block diagram of a power distribution network maintenance device in one embodiment;

[0044] Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0045] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0046] In one embodiment, Figure 1 As shown, a distribution network maintenance method is provided. This embodiment takes the method applied to a terminal as an example. It can be understood that the method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0047] S102: Obtain a voltage value of each detection node in the distribution network collected by a voltage sensor, and obtain a potential difference between every two detection nodes based on the voltage value.

[0048] Among them, the corresponding voltage range can be obtained according to the voltage level of the distribution network, and a voltage sensor adapted to the corresponding voltage range can be selected to measure the voltage data of the detection node. For example, the voltage level of the distribution network is 10kV, and the selected voltage sensor can operate in the range of 9kV to 11kV, and its accuracy needs to be within ±1%; the power system analysis software PSS / E or MATLAB can be used to calculate the real-time potential difference.

[0049] For example, voltage sensors are installed at two key detection nodes A and B in the distribution network to monitor and compare the voltage levels of A and B. From the data collected by the sensors, it can be seen that the voltage value of node A is 15kV, while the voltage value of node B is 12kV. The potential difference between the two nodes is calculated as 15kV-12kV=3kV, indicating that there is a voltage drop between the two detection points. For the voltage drop, the operator must evaluate whether it exceeds the normal operating parameters. If the design standard stipulates that the maximum potential difference between any two nodes of the distribution network shall not exceed 5kV, then this 3kV potential difference is acceptable. For the continuously monitored data, if the potential difference between nodes A and B often exceeds 5kV during the peak load period every afternoon, this indicates that there may be grid congestion or local overload problems. By analyzing this data, the operation and maintenance team can take measures to optimize load distribution or increase transmission capacity.

[0050] For the voltage values ​​detected at different detection nodes, these voltage values ​​need to be time synchronized and synchronized according to a unified time standard. GPS (Global Positioning System) or NTP (Network Time Protocol) can be used for time synchronization to ensure that the collected voltage data has an accurate timestamp. For example, all timestamps can be calibrated to Greenwich Mean Time.

[0051] For voltage sensors located outdoors, since the ambient temperature is very high in the hot summer, it is necessary to continuously monitor whether the ambient temperature exceeds the operating temperature range of the voltage sensor. If it exceeds, the sensor settings need to be adjusted. For example, the ambient temperature can reach 40°C, while the specified operating temperature range of the equipment is -20°C to 60°C. The ambient temperature data shows that the critical working conditions have not been reached, so there is no need to adjust the sensor settings, but attention should be paid to continuous monitoring to ensure that the temperature does not exceed the range and make adjustments if necessary.

[0052] S104. When the potential difference is not within the corresponding target potential difference range, input the potential difference signal corresponding to the potential difference into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal.

[0053] Among them, if the potential difference is not within the corresponding target potential difference range, it means that there is an abnormal condition in the distribution network, which will trigger the alarm and event recording mechanism, and record the timestamp of the out-of-range event, the duration of the event, the potential difference, the load condition and related environmental parameters, so as to facilitate subsequent analysis and take necessary maintenance or adjustment measures.

[0054] S106: Determine a fault category and a fault location of the distribution network based on the abnormal state category, and perform maintenance on the distribution network based on the fault category and the fault location.

[0055] Among them, after determining the abnormal state category, the cause of the potential difference signal abnormality can be determined by comparing historical abnormal cases, and the fault category and fault location can be identified based on the abnormal cause; after determining the fault category and fault location, query the plan database to select the corresponding emergency plan number and start-up conditions. If the fault detection information meets the emergency plan start-up conditions, the emergency plan is activated, and the isolation strategy and emergency measures are extracted from it; according to the isolation strategy, the isolation device is controlled to perform operations to isolate the faulty node. For example, the power grid monitoring system detects that the voltage of a substation suddenly drops to 80% of the normal value, accompanied by a short-term power jump. Preliminary analysis shows that it may be due to an internal fault in a main transformer of the substation. At this time, the system automatically confirms that the fault type is a transformer short-circuit fault through the action information of the fault indicator and the protection device, and determines that the fault location is on the 35kV main transformer side of the substation through fault waveform analysis. Then, the system queries the plan database and finds the corresponding emergency plan number "EA-035". The startup condition is "the main transformer voltage drops below 80% of the normal value and is accompanied by protection action". Since the current fault detection information meets the startup condition, the fault monitoring system activates the emergency plan "EA-035". The isolation strategy extracted from the emergency plan "EA-035" includes first cutting off the high-voltage side power supply of the faulty transformer, that is, opening 3 5kV high-voltage circuit breaker, and then cut off the low-voltage side load of the faulty transformer, that is, open the 10kV low-voltage circuit breaker. The emergency measures include activating the spare transformer and reorganizing the load distribution to reduce the scope of the power outage; notifying relevant personnel and departments, preparing emergency repair equipment and materials, and promptly releasing power outage information and estimated restoration time to the public; according to the isolation strategy, the control center operator issued instructions through the remote control system and performed the following operations. Within 3 minutes after the accident, the 35kV high-voltage circuit breaker was successfully opened to isolate the high-voltage side of the faulty transformer with a fault voltage of 28kV. Subsequently, the 10kV low-voltage circuit breaker was opened within 5 minutes to isolate the low-voltage side load of the faulty transformer with a fault current of 0A; the above operations successfully isolated the faulty node, avoided the fault spread, and ensured the stable operation of the rest of the power grid. At the same time, the emergency repair team was quickly dispatched to the scene to start fault repair work.

[0056] According to the location of the fault point and the predicted load demand, an improved algorithm based on the Newton-Raphson method is used to simulate several reconstruction schemes, and the power supply capacity of the scheme and its impact on the stability of the power grid are calculated. If a simulated reconstruction scheme meets the load demand and does not affect the stability of the power grid, the areas under the reconstruction scheme are prioritized. According to the priority and protection coordination principles, a specific control strategy is generated, including adjustment instructions for the switchgear and transformer taps that need to be opened and closed. Fault simulation is performed on the generated control strategy to verify the response of the protection device and the stability of the reconstructed power supply path. If the simulation verification results are qualified, the power supply path is automatically reconstructed. The reconstruction execution process and the operating status of the new power supply path after reconstruction are monitored by the state monitoring system to ensure a smooth reconstruction process and reliable power supply. For example, a power grid system fails and the fault point is located on a transmission line. According to the predicted load demand, the load peak is 100MW. Three different reconstruction schemes are simulated using an improved algorithm based on the Newton-Raphson method. Scheme 1 transfers the load to adjacent areas by opening and closing certain switchgear. The simulation results show that this scheme can supply up to 95MW of power, but has a greater impact on the stability of the power grid, with frequency fluctuations reaching ±5Hz, exceeding the allowable range. In addition to opening and closing switchgear, scheme 2 also adjusts the transformer taps to optimize the voltage profile. The simulation results show that the power supply capacity of this scheme is 100MW, and the frequency fluctuation is ±1Hz, which is within the allowable range. Scheme 3 further adds a backup transmission line on the basis of scheme 2. The simulation results show that the power supply capacity of this scheme can reach 105MW, and the stability of the power grid has been enhanced, with the frequency fluctuation reduced to ±05Hz. Among these three schemes, Among the schemes, Scheme 2 meets the load demand and has little impact on the stability of the power grid. Therefore, the areas under this scheme are further prioritized. Area A, which is adjacent to the fault area, has the highest priority, followed by Area B and then Area C. According to the priority and protection coordination principles, a specific control strategy is generated, which requires closing switches S1 and S2, opening switch S3, and adjusting the tap of transformer T1 from gear 5 to gear 4; performing fault simulation to verify the response of the protection device. For example, the simulation found that the protection device successfully isolated the fault within 150ms, and the stability of the reconstructed power supply path was good, without overvoltage or frequency anomalies. Once the simulation verification results are qualified, the power supply path is automatically reconstructed, and the reconstruction execution process is continuously monitored by the status monitoring system. The monitoring data shows that after the reconstruction, the voltage in area A is restored to 220kV, and the frequency is stabilized at 50Hz, without abnormal fluctuations, indicating that the reconstruction process is smooth and the power supply is reliable.

[0057] GIS technology (Geographic Information System) is used to obtain the location of power grid faults, and combined with the identification of faulty equipment, the precise location of the affected equipment and the surrounding environment map are automatically drawn to obtain the equipment model and identify the fault type. According to the equipment model and the nature of the fault, the corresponding standard operating procedures and repair plans are extracted from the maintenance knowledge base, and a fault report is provided to the operation and maintenance team, including the location, type and repair plan of the power grid fault. The recording method of the fault handling progress is set, and the frequency of reporting and updating is determined according to the severity of the fault and the complexity of the repair. For example, through GIS technical analysis, it can be learned that a specific power transformer T-1234 has a fault. The system automatically locates the fault point in the industrial area in the north of the city. By analyzing the terrain and building data of the area, GIS generates a map containing the precise location of the faulty transformer and the surrounding environment. The model of the transformer is identified as "XYZ-500kV". The system preliminarily determines the fault type as "abnormal insulation loss" through model identification; then, the system retrieves the standard operating procedures for this model and fault type from the maintenance knowledge base as follows: cut off the power supply of transformer T-1234, perform an insulation resistance test, and record a test value of 5 megohms, which is far below the normal threshold of 2 megohms. According to the test results, replace the insulation material and redo the test to ensure recovery to normal; for operation and maintenance The fault report provided by the maintenance team contains the above information and a specific repair plan, lists the tools and materials that need to be prepared, and the estimated completion time is 3 hours; the recording method of the fault handling progress is decided to update the progress every hour, and record the start and completion time of each key step in the system. At 10:00AM, the fault diagnosis is completed and the repair plan is confirmed. At 11:00AM, the power is cut off and the insulation test is carried out. At 12:00PM, the insulation material is replaced and ready for retesting; in view of the medium severity of the fault and the low complexity of the repair, the report update frequency is determined to be once an hour to ensure that the progress is provided to key personnel in a timely manner. If any problems are encountered during the process, such as the insulation test value does not meet the expected results, the report will be updated immediately and the maintenance plan will be adjusted.

[0058] In the above distribution network maintenance method, the voltage value of each detection node in the distribution network collected by the voltage sensor is obtained, and the potential difference between every two detection nodes is obtained based on the voltage value; when the potential difference is not within the corresponding target potential difference range, the potential difference signal corresponding to the potential difference is input into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal; the fault category and fault location of the distribution network are determined based on the abnormal state category, and the distribution network is maintained based on the fault category and fault location. The method provided in the present application can improve the power supply quality of the distribution network and reduce the risk and cost of system operation and maintenance by maintaining the distribution network.

[0059] In some embodiments, Figure 2 As shown in FIG. 1 , the process of determining the target potential difference range includes:

[0060] S202: Based on the load condition and network topology of the distribution network, an initial potential difference range between different detection nodes in the distribution network is obtained.

[0061] S204 . For every two detection nodes, based on the node types, node positions and network topology of the two detection nodes, the initial potential difference range is adjusted to obtain a corresponding intermediate potential difference range of the two detection nodes.

[0062] S206 , obtaining the current temperature and the current humidity of the environment where the two detection nodes are located, and adjusting the intermediate potential difference range based on the current temperature and the current humidity to obtain the corresponding target potential difference ranges of the two detection nodes.

[0063] Specifically, based on the load conditions and network topology of the distribution network, the normal fluctuation range of the potential difference between every two detection nodes in the distribution network can be determined, and then the initial potential difference range can be comprehensively determined based on the normal fluctuation range of the potential difference, the maximum safe potential difference threshold and the minimum induced potential difference threshold. After determining the initial potential difference range, the corresponding maximum safe potential difference threshold and the minimum induced potential difference threshold of the two specific detection nodes can be adjusted according to the node type, node location and network topology, so as to adjust the initial potential difference range of the two detection nodes and obtain the corresponding intermediate potential difference range. Since environmental factors and weather factors will affect the potential difference, the intermediate potential difference range is adjusted again based on the current temperature and current humidity to obtain the target potential difference range. For example, the normal fluctuation range of the potential difference of the distribution network is 100kV to 110kV, the maximum safe potential difference threshold is set to 120kV, and the minimum induced potential difference threshold is set to 90kV, then the initial potential difference range can be determined to be 90kV to 120kV; the distribution network includes a detection node C and a detection node D, wherein the load of the detection node C is 150MW, the load of the detection node D is 100MW, the detection node C is a power generation node, and the detection node D is a distribution node. The initial potential difference range can be adjusted to the corresponding intermediate potential difference range of the two nodes based on the node type, node location and network topology of the detection node C and the detection node D; in the environment where the two detection nodes are located, the temperature on that day is 35°C, which is 25°C higher than the average temperature, and the humidity is high, reaching 80%. Since high temperature and high humidity will reduce the efficiency of the transmission line, which may cause the potential difference to increase, therefore, the maximum safe potential difference threshold is adjusted to 115kV, and then the intermediate potential difference range is adjusted based on the adjusted maximum safe potential difference threshold to provide more operating room.

[0064] In this embodiment, based on the relevant information of the distribution network, the relevant information of the two specific detection nodes and the environmental conditions, the target potential difference range is finally determined, so that the judgment of the abnormal state category based on the target potential difference range is more accurate.

[0065] In some embodiments, the training process of the K neighbor model includes: obtaining a sample potential difference signal and a preset distance measurement criterion corresponding to the potential difference signal, each sample potential difference signal is marked with a corresponding state category; preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, the optimal K value is used to predict the category of the sample in the K neighbor model, and the preprocessing includes noise reduction and missing value interpolation; based on the optimal K value and the preset distance measurement criterion, constructing a pre-trained K neighbor model; inputting the sample potential difference signal into the pre-trained K neighbor model to obtain a trained K neighbor model.

[0066] Among them, the preset distance measurement criterion can be selected between Euclidean distance and Manhattan distance. Through verification of the sample potential difference signal, it can be seen that the accuracy of the K-nearest neighbor model using Euclidean distance is 94%, and the recall rate is 93%. The accuracy of the K-nearest neighbor model using Manhattan distance is 92%, and the recall rate is 90%. Therefore, Euclidean distance is selected as the preset distance measurement criterion; the sample potential difference signal can be the potential difference value obtained by the voltage sensor every minute in the past week. These sample potential difference signals may be missing due to sensor failure. Therefore, linear interpolation can be used to fill these missing values. For noise reduction processing, a moving average filter can be used, in which the window size is 5 data points to smooth short-term fluctuations.

[0067] In this embodiment, the sample potential difference signal is preprocessed to make the sample potential difference signal more accurate, thereby making the performance of the K-neighbor model finally constructed more excellent.

[0068] In some embodiments, determining the optimal K value based on the preprocessed sample potential difference signal includes: extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; acquiring multiple initial statistical features of the sample potential difference signal based on amplitude, frequency domain features and time domain features; determining at least one target statistical feature from multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process; and determining the optimal K value based on the target statistical feature.

[0069] Specifically, in one embodiment, the initial statistical features may include the mean and standard deviation of the sample potential difference corresponding to the sample potential difference signal. The main energy of the sample potential difference signal is concentrated near 60 Hz. Since the information gain of the standard deviation is 5 and the mutual information of the peak amplitude at 60 Hz is 7, it means that the standard deviation is very effective in distinguishing normal states from abnormal states. Therefore, the standard deviation is determined as the target statistical feature. After determining the target statistical feature, the optimal K value of the K neighbor model can be determined by using a grid search combined with a cross-validation method. For example, the average accuracy of the cross-validation is determined when K=3, 5, 7 or 9, respectively. Since the average accuracy of the cross-validation is the highest when K=5, which is 95%, 5 is determined as the optimal K value.

[0070] In this embodiment, the optimal K value is determined based on Fourier transform, information gain and mutual information indicators, so that the determined optimal K value is more accurate.

[0071] In some embodiments, the sample potential difference signal is input into a pre-trained K neighbor model to obtain a trained K neighbor model, including: obtaining at least one performance evaluation indicator output by the pre-trained K neighbor model; in the case that at least one performance evaluation indicator does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion; based on the adjusted optimal K value and the preset distance measurement criterion, returning to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance measurement criterion, and continuing to execute until each performance evaluation indicator meets the corresponding preset indicator range.

[0072] Among them, the performance evaluation index is used to evaluate the performance of the K nearest neighbor model. The performance evaluation index may include accuracy, recall rate and F1 score. When all performance evaluation indicators meet the corresponding preset indicator ranges, the K nearest neighbor model at this time is determined as the trained K nearest neighbor model. For example, in a round of training, the accuracy of the model output is 95%, the recall rate is 93%, and the F1 score is 94%, all of which meet the corresponding preset indicator ranges. The model at this time is determined as the trained K nearest neighbor model; the optimal K value can be adjusted by reselecting the K value or reselecting the target statistical features.

[0073] In this embodiment, when there is at least one performance evaluation indicator that does not meet the corresponding preset indicator range, by adjusting the optimal K value and the preset distance measurement criterion, the performance of the finally trained K neighbor model can be made better.

[0074] In some embodiments, the potential difference signal is transmitted within the transmission area of ​​the distribution network via a target radio signal; the process of determining the target radio signal includes: obtaining the initial frequency and initial band of all radio signals within the transmission area, and obtaining the number of times each initial frequency and each initial band are used within the transmission area; determining the frequency and band with the least number of uses among the initial frequencies and initial bands as the target frequency and target band, respectively; determining the transmission power based on the area and terrain characteristics of the transmission area; and determining the target radio signal based on the target frequency, target band, transmission power, preset antenna gain, and preset channel bandwidth.

[0075] Specifically, in the same area, the more times the radio signals of the same frequency and the same band are used, the greater the interference between the radio signals of the same frequency and the same band. Therefore, it is necessary to obtain the frequency information and band information of all radio signals in the distribution network transmission area, and determine the frequency and band with the least number of uses as the frequency and band of the target radio signal. The target radio signal determined in this way will be subject to the least interference in the process of transmitting the potential difference signal. For example, in the transmission area, there are a large number of radio signals in the 4GHz Wi-Fi band, and the model density of the 5GHz Wi-Fi band is relatively low. In order to avoid interference, the 5GHz band is determined as the corresponding frequency band of the target radio signal; the transmission power determined according to the area and terrain characteristics of the transmission area can ensure that the target radio signal covers the entire transmission area. For example, the transmission area is an urban area with a radius of 2 kilometers. The urban area is a roughly flat terrain. After calculation, in order to meet To achieve a received signal strength of at least -75dBm, the transmission power needs to be set to 20dBm; in order to help the target radio signal better focus on the transmission area and reduce unnecessary propagation in other directions, the antenna gain needs to be adjusted to an appropriate value. For example, a directional antenna with a gain of 9dBi can be used; by simulating the multipath propagation effect of the radio signal, the signal transmission path of the target radio signal is designed to reduce the signal quality degradation caused by signal reflection, refraction or diffraction. For example, in order to reduce the impact of multipath propagation problems that radio signals may encounter in urban environments, when designing the signal transmission path, avoid directly aiming at large metal buildings, because these buildings may cause strong reflection of the signal; in order to meet the rate requirements of voltage data transmission, an appropriate channel bandwidth needs to be set. For example, since a data transmission rate of up to 150Mbps can theoretically be provided within a 20MHz channel bandwidth, the preset channel bandwidth is set to 20MHz.

[0076] In this embodiment, by setting the target frequency and target band, the interference of other radio signals on the target radio signal can be reduced. By setting the appropriate transmission power, it can be ensured that the target radio signal covers the entire transmission area. By setting the appropriate preset antenna gain, the directionality of the signal can be enhanced and the coverage quality of the signal can be improved. By setting the appropriate preset channel bandwidth, the rate requirements of voltage data transmission can be met.

[0077] In one embodiment, Figure 3 As shown, another distribution network maintenance method is provided, which includes the following contents:

[0078] (1) Deploy voltage sensors at the detection nodes of the distribution network to obtain the voltage data of the detection nodes and calculate the potential difference.

[0079] (2) Use radio signal transmission and establish a communication network to ensure stable voltage data transmission.

[0080] (3) Set the potential difference threshold, compare the voltage data with the potential difference threshold in real time, and obtain the potential difference abnormality signal.

[0081] (4) Obtain abnormal signals of potential difference, use the KNN algorithm to automatically classify and identify the abnormal types, and determine the cause of the power grid abnormality based on the fluctuation characteristics of the potential difference.

[0082] (5) Based on the determined cause of the grid abnormality, trigger the relevant emergency plan and isolate the faulty node.

[0083] (6) Generate adjustment instructions and automatically reconstruct the distribution network power supply path to ensure that unaffected areas continue to have stable power supply.

[0084] (7) Provide fault reports to the operation and maintenance team, including the location, type and repair plan of the power grid fault, and assist in on-site processing.

[0085] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0086] Based on the same inventive concept, the embodiment of the present application also provides a distribution network maintenance device for implementing the distribution network maintenance method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more distribution network maintenance device embodiments provided below can refer to the limitations on the distribution network maintenance method above, and will not be repeated here.

[0087] In an exemplary embodiment, Figure 4 As shown, a distribution network maintenance device 400 is provided, comprising: a first acquisition module 401, a first input module 402 and a determination module 403, wherein:

[0088] The first acquisition module 401 is used to acquire the voltage value of each detection node in the power distribution network collected by the voltage sensor, and acquire the potential difference between every two detection nodes based on the voltage value.

[0089] The first input module 402 is used to input the potential difference signal corresponding to the potential difference into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal when the potential difference is not within the corresponding target potential difference range.

[0090] The determination module 403 is configured to determine a fault category and a fault location of the distribution network based on the abnormal state category, and perform maintenance on the distribution network based on the fault category and the fault location.

[0091] In some embodiments, the distribution network maintenance device 400 is specifically used to obtain an initial potential difference range between different detection nodes in the distribution network based on the load conditions and network topology of the distribution network; for every two detection nodes, based on the node type, node position and network topology of the two detection nodes, adjust the initial potential difference range to obtain the corresponding intermediate potential difference range of the two detection nodes; obtain the current temperature and current humidity of the environment in which the two detection nodes are located, adjust the intermediate potential difference range based on the current temperature and the current humidity, and obtain the corresponding target potential difference range of the two detection nodes.

[0092] In some embodiments, the distribution network maintenance device 400 further includes:

[0093] The second acquisition module is used to acquire the sample potential difference signal and the preset distance measurement criterion corresponding to the potential difference signal, and each sample potential difference signal is marked with a corresponding state category.

[0094] A preprocessing module is used to preprocess the sample potential difference signal and determine the optimal K value based on the preprocessed sample potential difference signal. The optimal K value is used to predict the category of the sample in the K neighbor model. The preprocessing includes noise reduction and missing value interpolation.

[0095] A construction module is used to construct a pre-trained K nearest neighbor model based on the optimal K value and the preset distance metric criterion.

[0096] The second input module is used to input the sample potential difference signal into the pre-trained K-neighbor model to obtain the trained K-neighbor model.

[0097] In some embodiments, the preprocessing module is also used to extract amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; based on the amplitude, the frequency domain features and the time domain features, obtain multiple initial statistical features of the sample potential difference signal; based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process, determine at least one target statistical feature from the multiple initial statistical features; and determine the optimal K value based on the target statistical feature.

[0098] In some embodiments, the second input module is also used to obtain at least one performance evaluation indicator output by the pre-trained K neighbor model; when there is at least one performance evaluation indicator that does not meet the corresponding preset indicator range, the optimal K value and the preset distance measurement criterion are adjusted; based on the adjusted optimal K value and the preset distance measurement criterion, the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance measurement criterion is returned, and the execution is continued until each performance evaluation indicator meets the corresponding preset indicator range.

[0099] In some embodiments, the potential difference signal is transmitted within the transmission area of ​​the distribution network via a target radio signal; the distribution network maintenance device 400 is further used to obtain the initial frequency and initial band of all radio signals in the transmission area, and obtain the number of times each initial frequency and each initial band are used in the transmission area; the frequency and band with the least number of times used among the initial frequencies and the initial bands are respectively determined as the target frequency and the target band; the transmission power is determined based on the area and terrain characteristics of the transmission area; the target radio signal is determined based on the target frequency, the target band, the transmission power, the preset antenna gain and the preset channel bandwidth.

[0100] Each module in the above-mentioned distribution network maintenance device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0101] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 5 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC) or other technologies. When the computer program is executed by the processor, a distribution network maintenance method is implemented.

[0102] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0103] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the following steps when executing the computer program: obtaining a voltage value of each detection node in a distribution network collected by a voltage sensor, and obtaining a potential difference between every two detection nodes based on the voltage value; when the potential difference is not within a corresponding target potential difference range, inputting a potential difference signal corresponding to the potential difference into a trained K-neighbor model to obtain an abnormal state category corresponding to the potential difference signal; determining a fault category and a fault location of the distribution network based on the abnormal state category, and maintaining the distribution network based on the fault category and the fault location.

[0104] In one embodiment, the process of determining the target potential difference range implemented when the processor executes a computer program includes: obtaining an initial potential difference range between different detection nodes in the distribution network based on the load condition and network topology of the distribution network; for every two detection nodes, adjusting the initial potential difference range based on the node type, node position and network topology of the two detection nodes to obtain an intermediate potential difference range corresponding to the two detection nodes; obtaining the current temperature and current humidity of the environment in which the two detection nodes are located, and adjusting the intermediate potential difference range based on the current temperature and the current humidity to obtain a target potential difference range corresponding to the two detection nodes.

[0105] In one embodiment, the training process of the K neighbor model implemented when the processor executes a computer program includes: obtaining a sample potential difference signal and a preset distance measurement criterion corresponding to the potential difference signal, each sample potential difference signal is marked with a corresponding state category; preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, the optimal K value is used to predict the category of the sample in the K neighbor model, the preprocessing includes noise reduction and missing value interpolation; based on the optimal K value and the preset distance measurement criterion, constructing a pre-trained K neighbor model; inputting the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model.

[0106] In one embodiment, the method of determining the optimal K value based on the preprocessed sample potential difference signal implemented when the processor executes a computer program includes: extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; obtaining multiple initial statistical features of the sample potential difference signal based on the amplitude, the frequency domain features and the time domain features; determining at least one target statistical feature from the multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process; and determining the optimal K value based on the target statistical feature.

[0107] In one embodiment, the processor executes a computer program to input the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model, including: obtaining at least one performance evaluation indicator output by the pre-trained K neighbor model; if there is at least one performance evaluation indicator that does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion; based on the adjusted optimal K value and the preset distance measurement criterion, returning to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance measurement criterion, and continuing to execute until each performance evaluation indicator meets the corresponding preset indicator range.

[0108] In one embodiment, the potential difference signal implemented when the processor executes the computer program is transmitted through the target radio signal within the transmission area of ​​the distribution network; the process of determining the target radio signal includes: obtaining the initial frequency and initial band of all radio signals in the transmission area, and obtaining the number of times each initial frequency and each initial band are used in the transmission area; determining the frequency and band with the least number of times used among the initial frequencies and the initial bands as the target frequency and the target band respectively; determining the transmission power based on the area and terrain characteristics of the transmission area; and determining the target radio signal based on the target frequency, the target band, the transmission power, the preset antenna gain and the preset channel bandwidth.

[0109] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: obtaining a voltage value of each detection node in a distribution network collected by a voltage sensor, and obtaining a potential difference between every two detection nodes based on the voltage value; when the potential difference is not within a corresponding target potential difference range, inputting a potential difference signal corresponding to the potential difference into a trained K-neighbor model to obtain an abnormal state category corresponding to the potential difference signal; determining a fault category and a fault location of the distribution network based on the abnormal state category, and maintaining the distribution network based on the fault category and the fault location.

[0110] In one embodiment, the process of determining the target potential difference range implemented when the computer program is executed by the processor includes: obtaining the initial potential difference range between different detection nodes in the distribution network based on the load condition and network topology of the distribution network; for every two detection nodes, adjusting the initial potential difference range based on the node type, node position and network topology of the two detection nodes to obtain the intermediate potential difference range corresponding to the two detection nodes; obtaining the current temperature and current humidity of the environment in which the two detection nodes are located, and adjusting the intermediate potential difference range based on the current temperature and the current humidity to obtain the target potential difference range corresponding to the two detection nodes.

[0111] In one embodiment, the training process of the K neighbor model implemented when the computer program is executed by the processor includes: obtaining a sample potential difference signal and a preset distance measurement criterion corresponding to the potential difference signal, each sample potential difference signal is marked with a corresponding state category; preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, the optimal K value is used to predict the category of the sample in the K neighbor model, the preprocessing includes noise reduction and missing value interpolation; based on the optimal K value and the preset distance measurement criterion, constructing a pre-trained K neighbor model; inputting the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model.

[0112] In one embodiment, the computer program implemented when executed by the processor determines the optimal K value based on the preprocessed sample potential difference signal, including: extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; obtaining multiple initial statistical features of the sample potential difference signal based on the amplitude, the frequency domain features and the time domain features; determining at least one target statistical feature from the multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process; and determining the optimal K value based on the target statistical feature.

[0113] In one embodiment, the computer program implemented when executed by the processor inputs the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model, including: obtaining at least one performance evaluation indicator output by the pre-trained K neighbor model; in the case that at least one performance evaluation indicator does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion; based on the adjusted optimal K value and the preset distance measurement criterion, returning to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance measurement criterion, and continuing to execute until each performance evaluation indicator meets the corresponding preset indicator range.

[0114] In one embodiment, the potential difference signal implemented when the computer program is executed by the processor is transmitted through the target radio signal within the transmission area of ​​the distribution network; the process of determining the target radio signal includes: obtaining the initial frequency and initial band of all radio signals in the transmission area, and obtaining the number of times each initial frequency and each initial band are used in the transmission area; determining the frequency and band with the least number of times used among the initial frequencies and the initial bands as the target frequency and target band respectively; determining the transmission power based on the area and terrain characteristics of the transmission area; determining the target radio signal based on the target frequency, the target band, the transmission power, the preset antenna gain and the preset channel bandwidth.

[0115] In one embodiment, a computer program product is provided, comprising a computer program, which implements the following steps when executed by a processor: obtaining a voltage value of each detection node in a distribution network collected by a voltage sensor, and obtaining a potential difference between every two detection nodes based on the voltage value; when the potential difference is not within a corresponding target potential difference range, inputting a potential difference signal corresponding to the potential difference into a trained K-neighbor model to obtain an abnormal state category corresponding to the potential difference signal; determining a fault category and a fault location of the distribution network based on the abnormal state category, and maintaining the distribution network based on the fault category and the fault location.

[0116] In one embodiment, the process of determining the target potential difference range implemented when the computer program is executed by the processor includes: obtaining the initial potential difference range between different detection nodes in the distribution network based on the load condition and network topology of the distribution network; for every two detection nodes, adjusting the initial potential difference range based on the node type, node position and network topology of the two detection nodes to obtain the intermediate potential difference range corresponding to the two detection nodes; obtaining the current temperature and current humidity of the environment in which the two detection nodes are located, and adjusting the intermediate potential difference range based on the current temperature and the current humidity to obtain the target potential difference range corresponding to the two detection nodes.

[0117] In one embodiment, the training process of the K neighbor model implemented when the computer program is executed by the processor includes: obtaining a sample potential difference signal and a preset distance measurement criterion corresponding to the potential difference signal, each sample potential difference signal is marked with a corresponding state category; preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, the optimal K value is used to predict the category of the sample in the K neighbor model, the preprocessing includes noise reduction and missing value interpolation; based on the optimal K value and the preset distance measurement criterion, constructing a pre-trained K neighbor model; inputting the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model.

[0118] In one embodiment, the computer program implemented when executed by the processor determines the optimal K value based on the preprocessed sample potential difference signal, including: extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; obtaining multiple initial statistical features of the sample potential difference signal based on the amplitude, the frequency domain features and the time domain features; determining at least one target statistical feature from the multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process; and determining the optimal K value based on the target statistical feature.

[0119] In one embodiment, the computer program implemented when executed by the processor inputs the sample potential difference signal into the pre-trained K neighbor model to obtain the trained K neighbor model, including: obtaining at least one performance evaluation indicator output by the pre-trained K neighbor model; in the case that at least one performance evaluation indicator does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion; based on the adjusted optimal K value and the preset distance measurement criterion, returning to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance measurement criterion, and continuing to execute until each performance evaluation indicator meets the corresponding preset indicator range.

[0120] In one embodiment, the potential difference signal implemented when the computer program is executed by the processor is transmitted through the target radio signal within the transmission area of ​​the distribution network; the process of determining the target radio signal includes: obtaining the initial frequency and initial band of all radio signals in the transmission area, and obtaining the number of times each initial frequency and each initial band are used in the transmission area; determining the frequency and band with the least number of times used among the initial frequencies and the initial bands as the target frequency and target band respectively; determining the transmission power based on the area and terrain characteristics of the transmission area; determining the target radio signal based on the target frequency, the target band, the transmission power, the preset antenna gain and the preset channel bandwidth.

[0121] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0122] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0123] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0124] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be construed as limiting the scope of the present application. It should be noted that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A distribution network maintenance method, characterized in that: The method comprises: Acquire the voltage value of each detection node in the distribution network collected by the voltage sensor, and acquire the potential difference between every two detection nodes based on the voltage value; When the potential difference is not within the corresponding target potential difference range, inputting the potential difference signal corresponding to the potential difference into the trained K-neighbor model to obtain the abnormal state category corresponding to the potential difference signal; A fault category and a fault location of the power distribution network are determined based on the abnormal state category, and maintenance is performed on the power distribution network based on the fault category and the fault location.

2. The method according to claim 1, characterized in that The process of determining the target potential difference range includes: Based on the load condition and network topology of the distribution network, obtaining an initial potential difference range between different detection nodes in the distribution network; For every two detection nodes, based on the node types, node positions and network topology of the two detection nodes, the initial potential difference range is adjusted to obtain a corresponding intermediate potential difference range of the two detection nodes; The current temperature and the current humidity of the environment in which the two detection nodes are located are obtained, and the intermediate potential difference range is adjusted based on the current temperature and the current humidity to obtain the target potential difference range corresponding to the two detection nodes.

3. The method according to claim 1, characterized in that The training process of the K-neighbor model includes: Acquire sample potential difference signals and preset distance measurement criteria corresponding to the potential difference signals, each sample potential difference signal is marked with a corresponding state category; Preprocessing the sample potential difference signal, and determining an optimal K value based on the preprocessed sample potential difference signal, wherein the optimal K value is used to predict the category of the sample in the K neighbor model, wherein the preprocessing includes noise reduction processing and missing value interpolation; Based on the optimal K value and the preset distance metric criterion, construct a pre-trained K nearest neighbor model; The sample potential difference signal is input into the pre-trained K-neighbor model to obtain the trained K-neighbor model.

4. The method according to claim 3, characterized in that The determining of the optimal K value based on the preprocessed sample potential difference signal comprises: Extracting amplitude, frequency domain features and time domain features from the preprocessed sample potential difference signal based on Fourier transform; Based on the amplitude, the frequency domain characteristics and the time domain characteristics, a plurality of initial statistical characteristics of the sample potential difference signal are obtained; Determining at least one target statistical feature from the multiple initial statistical features based on information gain, mutual information index, and the degree of influence of each initial statistical feature on the abnormal state category identification process; The optimal K value is determined based on the target statistical characteristics.

5. The method according to claim 3, characterized in that: The step of inputting the sample potential difference signal into the pre-trained K-neighbor model to obtain the trained K-neighbor model comprises: Obtaining at least one performance evaluation indicator output by the pre-trained K-neighbor model; When at least one performance evaluation indicator does not meet the corresponding preset indicator range, adjusting the optimal K value and the preset distance measurement criterion; Based on the adjusted optimal K value and the preset distance metric criterion, return to the step of constructing the pre-trained K neighbor model based on the optimal K value and the preset distance metric criterion, and continue to execute until each performance evaluation indicator meets the corresponding preset indicator range.

6. The method according to claim 1, characterized in that The potential difference signal is transmitted in the transmission area of ​​the power distribution network through a target radio signal; the determination process of the target radio signal includes: Obtaining initial frequencies and initial bands of all radio signals in the transmission area, and obtaining the number of times each initial frequency and each initial band is used in the transmission area; Determine the frequency and the band with the least number of uses among the initial frequencies and the initial bands as the target frequency and the target band respectively; Determining transmission power based on the area and terrain characteristics of the transmission area; The target radio signal is determined based on the target frequency, the target band, the transmission power, a preset antenna gain, and a preset channel bandwidth.

7. A distribution network maintenance device, characterized in that: The device comprises: A first acquisition module is used to acquire a voltage value of each detection node in the distribution network collected by a voltage sensor, and acquire a potential difference between every two detection nodes based on the voltage value; A first input module is used for inputting a potential difference signal corresponding to the potential difference into a trained K-neighbor model to obtain an abnormal state category corresponding to the potential difference signal when the potential difference is not within a corresponding target potential difference range; A determination module is used to determine a fault category and a fault location of the distribution network based on the abnormal state category, and to maintain the distribution network based on the fault category and the fault location.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images