Power engineering design power distribution network planning method based on three-dimensional modeling

Abstract

The invention discloses an electric power engineering design power distribution network planning method based on three-dimensional modeling, and relates to the technical field of electric power engineering and information, and the method comprises the following steps: collecting the current effective value of a power distribution network and the axial temperature data of equipment, extracting features after preprocessing, combining three-dimensional coordinates, analyzing the time sequence correlation by using a gated cycle unit network, and obtaining the current effective value of the power distribution network. And constructing a random forest model to divide risk areas, fusing the two to generate an equipment health score, performing classified early warning, positioning abnormal equipment, and triggering a control instruction. Through combination of three-dimensional modeling and the advanced information technology, the intelligent level of power distribution network planning is remarkably improved, the method not only can accurately predict equipment faults and reduce the operation and maintenance cost, but also can quickly position and early warn abnormal equipment and guarantee stable operation of a power grid, in addition, resource configuration is optimized, the operation efficiency is improved, and the method is suitable for popularization and application. The method injects new vitality to sustainable development of the power industry, and has wide social and economic benefits and application prospects.

Description

Technical Field

[0001] The present invention relates to the fields of power engineering and information technology, and particularly to a power engineering design and distribution network planning method based on three-dimensional modeling. Background Art

[0002] A distribution network refers to a power network that receives electric energy from a transmission network or a regional power plant and distributes it locally or step by step according to voltage levels to various users through distribution facilities. It mainly consists of distribution equipment and ancillary facilities such as overhead lines, cables, poles, distribution transformers, switchgear, and reactive compensation capacitors. It is a network that plays an important role in distributing electric energy in the power grid. The structure of the distribution network is usually radial, and a closed-loop design is adopted to improve the operation flexibility and power supply reliability. However, it operates in an open-loop state during actual operation to limit the short-circuit fault current and control the scope of fault propagation. In addition, the distribution network has characteristics such as multiple voltage levels, complex network structures, and diverse equipment types. Therefore, its safety risk factors are relatively numerous. The main task of the distribution network is to further transmit electric power from the transmission network of the power system to users, realizing the distribution and supply of electric energy. It is widely used in various power supply areas such as cities, rural areas, and factories and is the power supply foundation for modern urbanized life. At the same time, the planning and development of the distribution network need to formulate system expansion and renovation plans according to future load growth and the current situation of the urban distribution network.

[0003] In order to solve the problems of fault prediction of distribution equipment and abnormal distribution equipment location in the distribution network, the existing technology mainly uses the method of traditional data analysis and two-dimensional drawing comparison for processing. This method relies on historical fault data and manual inspections, and predicts possible fault points through comparison and analysis. However, due to the complex structure of the distribution network and the wide distribution of equipment, the traditional method often has the situation of untimely data update and low fault prediction accuracy, resulting in low efficiency in locating abnormal distribution equipment and being difficult to quickly and effectively eliminate faults. In addition, the traditional method has limitations in dealing with the equipment relationship in three-dimensional space and is difficult to accurately reflect the actual layout and mutual influence between equipment, thus increasing the difficulty of fault troubleshooting. Therefore, in order to improve the accuracy of distribution equipment fault prediction and the efficiency of abnormal location, a power engineering design and distribution network planning method based on three-dimensional modeling is proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a power engineering design and distribution network planning method based on three-dimensional modeling to solve the problems raised in the above background art.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is: A power engineering design and distribution network planning method based on three-dimensional modeling, including the following steps:

[0006] S1. Collect the effective current value data of the distribution network line and the axial temperature data of the equipment surface;

[0007] S2. Based on the collected effective current value data and the axial temperature data of the equipment surface, call the three-dimensional coordinate data to form an input feature vector, use a gated recurrent unit network to construct a time series correlation model, and explore the time series correlation between current harmonics and temperature gradients;

[0008] S3. Combine the three-dimensional coordinates and the random forest algorithm to construct a risk zoning model, and divide the risk areas of the distribution equipment in the distribution network;

[0009] S4. Synthesize the analysis results of the time series correlation model and the risk zoning model, construct a health assessment model, and output the health score of the distribution equipment;

[0010] S5. Divide the threshold range of the equipment health score, classify the health status of the equipment, locate the abnormal equipment in combination with the three-dimensional coordinates, issue an equipment warning according to the health score, and trigger the corresponding control instruction.

[0011] A further improvement of the technical solution of the present invention is that in S1, the process of collecting the effective current value data of the distribution network line and the axial temperature data of the equipment surface includes:

[0012] Deploy high-precision current sensors at the incoming and outgoing ends of the transformer and the cable joints of the distribution network line. Determine the spacing of the high-precision current sensors according to the equipment coordinate density in the three-dimensional modeling. Adopt a through-core deployment structure, sleeve the high-precision current sensor outside the outer insulation layer of the wire, and calculate the effective current value through an integral circuit;

[0013] Lay a distributed optical fiber temperature sensor along the axial center line of the surface of the distribution equipment in the distribution network. The distribution equipment includes a circuit breaker and a disconnecting switch. The optical fiber is closely attached to the outer shell of the distribution equipment. Adopt an armored protection structure, arrange temperature measurement points at intervals of 0.5 meters, form a continuous monitoring network through series connection of the optical fiber links, and calculate the absolute temperature value of the distribution equipment by using the light intensity ratio.

[0014] A further improvement of the technical solution of the present invention is that in S1, the process of preprocessing the collected effective current value data and the axial temperature data of the equipment surface includes:

[0015] Perform wavelet transform decomposition on the original current waveform signal in the collection process, decompose the current signal into fundamental wave and 2 - 50th harmonic components through fast Fourier transform, and retain the harmonic amplitude;

[0016] The original temperature data of the fiber optic sensor is identified for abnormal points using the sliding window standard deviation method. If the deviation of a temperature point from the average value of the adjacent 5 points within the window exceeds 3 times the standard deviation, it is determined as an outlier and removed. For the temperature sequence after removing the outliers, the sliding window averaging method is used for processing. The window width is set to 7 sampling points, and the arithmetic average of the temperature data within the window is taken to replace the original value. Along the axial center line of the power distribution equipment, in the order of the temperature measurement points at intervals of 0.5 meters, the ratio of the temperature difference between adjacent points to the distance is calculated, and the axial temperature gradient sequence is output;

[0017] The effective current data and the axial temperature data on the surface of the equipment are injected into a unified clock source to align the time tags. For data segments with a deviation exceeding ±2ms, interpolation compensation is started. When the temperature monitoring points continuously have missing data for more than 5 sampling periods, the average temperature of the two adjacent monitoring points along the axis of the same equipment is taken to replace the missing value.

[0018] A further improvement of the technical solution of the present invention lies in that: in the S1, the process of feature extraction from the preprocessed effective current data and the axial temperature data on the surface of the equipment includes:

[0019] Based on the amplitudes of the 2 - 50th harmonic currents retained after preprocessing, the sum of the proportions of the energies of each harmonic component to the fundamental wave energy is calculated to obtain the harmonic energy concentration. Based on the effective current data of the continuous period sequence, the relative change rate of the effective current between adjacent periods is calculated;

[0020] Along the axial temperature gradient sequence on the surface of the equipment, the axial temperature gradient extreme values are selected, and the algebraic sum of the axial temperature gradients on the surface of the equipment within a single - day monitoring period is statistically calculated to obtain the temperature cumulative change amount.

[0021] A further improvement of the technical solution of the present invention lies in that: in the S2, the process of calling the three - dimensional coordinate data to form the input feature vector includes:

[0022] A three - dimensional model database pre - set in the power engineering design stage for storing the spatial position information of the power distribution equipment is introduced. The harmonic energy concentration and the axial temperature gradient extreme value data are automatically associated with the equipment coordinates in the three - dimensional model database through the equipment number. The corresponding three - dimensional coordinates are extracted from the three - dimensional model database according to the equipment number, and the three - dimensional coordinates are verified with the actual position of the equipment;

[0023] The current harmonic energy concentration, the axial temperature gradient extreme value data, the time stamp, the equipment number, and the three - dimensional coordinates are bound. If the equipment number cannot match the database coordinates, the data stream is frozen and an alarm is triggered. If the equipment number matches the database coordinates, an input feature vector is formed.

[0024] A further improvement of the technical solution of the present invention lies in that: in the S2, the process of constructing the time - series correlation model includes:

[0025] The gated recurrent unit network receives an input feature vector containing current harmonic energy concentration, axial temperature gradient extreme value data, timestamp, device number, and three-dimensional coordinates. The device number and three-dimensional coordinates are stored separately as static identification data. The timestamp, current harmonic energy concentration, and temperature gradient extreme value constitute dynamic time-series data, which are input into the gated recurrent unit network at time steps to construct a time-series correlation model. The gating mechanism of the gated recurrent unit network includes a reset gate and an update gate;

[0026] The reset gate calculates the forgetting ratio based on the current current harmonic energy concentration and the previous hidden state. When the current harmonic energy concentration suddenly increases by more than 20% of the historical mean, the reset gate value approaches 0, discarding irrelevant historical states. The update gate determines the information update intensity according to the change trend of the temperature gradient extreme value. If the temperature gradient extreme value rises continuously for 3 time steps, the update gate value approaches 1, strengthening the current temperature rise feature;

[0027] Fuse the historical state filtered by the reset gate with the current input to capture the time-series causal relationship between the sudden increase in current harmonic energy concentration and the lagging influence intensity of the temperature gradient extreme value. Fuse the candidate state and the historical state through the update gate to retain the harmonic cumulative effect during the daily load peak period and at the same time respond to the instantaneous harmonics caused by lightning strikes;

[0028] Rebind the hidden state with the static three-dimensional coordinates to generate a time-series feature vector with spatial labels, and transmit the time-series feature vector with spatial labels to the random forest model. Combining the spatial distribution characteristics of the device coordinates, divide the device risk area.

[0029] A further improvement of the technical solution of the present invention lies in: in S3, the process of constructing the risk zoning model includes:

[0030] Use the random forest model to receive the time-series feature vector with spatial labels output by the time-series correlation model to construct a risk zoning model. Discretize the distribution area of the distribution network into cube grids with a side length of 0.5 meters according to the three-dimensional coordinates (x, y, z). Each grid corresponds to a risk analysis unit. Calculate the mean value of the time-series correlation intensity of the distribution equipment in the grid. Take the historical failure rate of the distribution equipment in the grid as the classification target, divide the actual failure level into normal, faulty, and warning according to the historical failure data of the distribution equipment, calculate the decrease in Gini impurity at the candidate splitting point, and select the three-dimensional coordinate axis and splitting point that maximize the purity of the decision tree sub-nodes in the risk zoning model. Calculate the total splitting gain ratio of the three-dimensional coordinate axis in the decision tree. If the importance of an axis is higher than 60%, then determine that the direction of this axis is the risk-dominant dimension;

[0031] The hidden state output by the time-series correlation model represents the time-series correlation strength of the sudden increase in the current harmonic energy concentration of the distribution equipment and the lagging effect of the extreme temperature gradient. The failure probability per unit time of the distribution equipment in the grid over the past three years is calculated using the Poisson distribution. It is set that the real-time state contributes 60% to the short-term risk. Based on the contribution degree of the time-series correlation strength and the historical failure probability, the risk score R of each grid is calculated. Based on the risk score, the risk levels are divided into high risk, medium risk, and low risk. The 95th percentile of the R-value distribution 30 days before the failure in the historical data is set as the high-risk threshold, the 75th to 95th percentiles as the medium-risk threshold, and the 75th percentile as the low-risk threshold. Every quarter, according to the newly generated failure data, the quantile thresholds are recalculated;

[0032] The input data is the feature vector containing the hidden state and three-dimensional coordinates corresponding to the historical failure records, and the label is the actual failure level. The input data is stratified and sampled according to the three-dimensional coordinates, and the number of decision trees and the range of the maximum tree depth are determined through grid search;

[0033] The high-risk area coordinate set is sent to the inspection robot, and the inspection path is planned and marked as a red flashing icon in the three-dimensional model. The medium-risk area coordinate set is pushed to the handheld terminal of the operation and maintenance personnel and marked as an orange flashing icon in the three-dimensional model. The low-risk area starts the parameter adaptive adjustment operation and is marked as a green flashing icon in the three-dimensional model.

[0034] A further improvement of the technical solution of the present invention lies in: in S4, the process of constructing the health evaluation model includes:

[0035] Bind the dynamically acquired time-series feature vector containing the timestamp, current harmonic energy concentration, and extreme axial temperature gradient with the device number and three-dimensional coordinates, and input it into the trained time-series correlation model. The time-series correlation model outputs the time-series correlation strength and quantifies the real-time correlation between the sudden increase in the current harmonic energy concentration of the distribution equipment and the extreme temperature gradient. Input the time-series feature vector containing the hidden state output by the time-series correlation model and the three-dimensional coordinates of the distribution equipment into the trained risk zoning model. The risk zoning model outputs the grid risk score, which represents the historical and real-time comprehensive risks of the grid area where the distribution equipment belongs;

[0036] Allocate time-series weights and space weights according to the device type and historical failure data, and calculate the health score of the distribution equipment based on the hidden state and the grid risk score. When the hidden state of the distribution equipment is higher than 0.8 and the grid risk score is higher than 0.7 for three consecutive days, start the adaptive adjustment operation. Compare the health score with the actual failure records and calculate the prediction accuracy. If the prediction accuracy is lower than 90%, trigger the retraining of the time-series correlation model and the risk zoning model.

[0037] A further improvement of the technical solution of the present invention lies in: in S5, the process of dividing the equipment health score threshold range and classifying the equipment health status includes:

[0038] Select the equipment health score data without fault records in the past three years, calculate the 95th percentile of its distribution as the upper limit of the health score threshold for normal power distribution equipment, extract the health score data of the power distribution equipment within 24 hours before the fault occurs, calculate its mean value as the critical value for early warning and fault, and based on the equipment health score threshold range, divide the power distribution equipment into normal state, early warning state and fault state;

[0039] Draw the working characteristic curve of the power distribution equipment of the health score against the fault event, calculate the area under the curve, set the area threshold under the curve, and if the area reduction under the curve in the same area within a single day exceeds 20%, recalculate the equipment health score threshold range.

[0040] A further improvement of the technical solution of the present invention lies in: in S5, the process of locating abnormal equipment, giving equipment early warning, and triggering corresponding control instructions includes:

[0041] Uniquely bind the health score with the three-dimensional coordinates of the power distribution equipment through the equipment number, mark the equipment status icon according to the coordinates in the three-dimensional model according to the threshold range to which the health score belongs, calculate the mean value of the health scores of the power distribution equipment in the same grid, set the grid risk threshold, and if the mean value of the health score is lower than the grid risk threshold, mark the coordinates of the power distribution equipment in the grid as an abnormal area;

[0042] If it is determined that the power distribution equipment is in the normal state, mark the corresponding three-dimensional model area as green and maintain routine monitoring. If it is determined that the power distribution equipment is in the early warning state, push the corresponding coordinate set to the inspection robot, mark the corresponding three-dimensional model area as orange, generate an inspection work order and associate it with the equipment maintenance record. If it is determined that the power distribution equipment is in the fault state, trigger the circuit breaker tripping instruction, mark the corresponding three-dimensional model area as red, and push the fault coordinates to the emergency repair terminal.

[0043] Due to the adoption of the above technical solution, the technical progress achieved by the present invention compared with the prior art is:

[0044] 1. The present invention provides a power engineering design and distribution network planning method based on three-dimensional modeling, which can significantly improve the accuracy and timeliness of power distribution equipment fault prediction. By collecting and analyzing key data such as the effective value of current and the axial temperature of the equipment surface, combined with three-dimensional coordinate information, using an advanced gated recurrent unit network to explore temporal correlations, it realizes the accurate assessment of the equipment health status, effectively prevents potential faults, reduces unplanned power outages, and improves power supply reliability.

[0045] 2. The present invention provides a distribution network planning method for power engineering design based on 3D modeling, which innovatively integrates 3D modeling and machine learning technologies, realizes the intelligent division of equipment risk areas, and based on the random forest model, combined with the 3D coordinates of equipment, can accurately identify high-risk areas, providing a scientific basis for distribution network planning and maintenance, optimizing resource allocation, reducing operation and maintenance costs, and improving the overall operation efficiency.

[0046] 3. The present invention provides a distribution network planning method for power engineering design based on 3D modeling. By setting the threshold range of equipment health scores, it realizes the dynamic classification and early warning of equipment health status. Combined with 3D coordinate positioning technology, it can quickly lock the location of abnormal equipment, trigger corresponding control instructions according to the health scores, take intervention measures in a timely manner, effectively prevent the expansion of faults, ensure the safe and stable operation of the power grid, and improve user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0048] Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0049] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0050] Embodiment, as Figure 1 shown, the present invention provides a distribution network planning method for power engineering design based on 3D modeling, including the following steps:

[0051] S1. Collect the effective current data of the distribution network line and the axial temperature data of the equipment surface. Deploy high-precision current sensors at the incoming and outgoing line ends of the transformers and cable joints of the distribution network line. Determine the high-precision current sensor spacing according to the equipment coordinate density in the 3D modeling. Adopt a through-hole deployment structure, sleeve the high-precision current sensor on the outer insulation layer of the wire, and calculate the effective current value through an integrating circuit. Lay distributed fiber optic temperature sensors along the axial centerline of the surface of the distribution equipment in the distribution network. The distribution equipment includes circuit breakers and disconnect switches. The optical fiber is closely attached to the outer shell of the distribution equipment, and a armored protection structure is adopted. Temperature measurement points are arranged at intervals of 0.5 meters, and a continuous monitoring network is formed by connecting in series through the optical fiber link. Calculate the absolute temperature value of the distribution equipment by using the light intensity ratio. Perform wavelet transform decomposition on the original current waveform signal during the acquisition process. Decompose the current signal into the fundamental wave and the 2nd - 50th harmonic components through fast Fourier transform, and retain the harmonic amplitudes. Use the sliding window standard deviation method to identify abnormal points in the original temperature data of the fiber optic sensor. If the deviation between a temperature point and the average value of the adjacent 5 points within the window exceeds 3 times the standard deviation, it is determined as an abnormal value and excluded. For the temperature sequence after excluding abnormal values, use the sliding window average method for processing. Set the window width to 7 sampling points, and take the arithmetic average of the temperature data within the window to replace the original value. Along the axial centerline of the distribution equipment, calculate the ratio of the temperature difference between adjacent points to the spacing according to the order of the temperature measurement points at intervals of 0.5 meters, and output the axial temperature gradient sequence. Inject the effective current data and the axial temperature data of the equipment surface into a unified clock source to align the time tags. For data segments with a deviation exceeding ±2ms, start interpolation compensation. When the temperature monitoring points continuously lack data for more than 5 sampling periods, take the temperature average value of two adjacent monitoring points in the same equipment axis to replace the missing value. Based on the 2nd - 50th harmonic current amplitudes retained in the preprocessing, calculate the sum of the proportions of the energies of each harmonic component to the fundamental wave energy to obtain the harmonic energy concentration. Based on the effective current data of the continuous cycle sequence, calculate the relative change rate of the effective current between adjacent cycles. Along the axial temperature gradient sequence of the equipment surface, select the axial temperature gradient extreme values, and statistically calculate the algebraic sum of the axial temperature gradients on the equipment surface within a single-day monitoring cycle to obtain the temperature cumulative change amount;

[0052] S2. Based on the collected effective current data and the axial temperature data on the device surface, call the three-dimensional coordinate data to form an input feature vector. Use a gated recurrent unit network to construct a temporal correlation model to explore the temporal correlation between current harmonics and temperature gradients. Introduce a three-dimensional model database that pre-stores the spatial location information of power distribution equipment in the power engineering design stage. Automatically associate the harmonic energy concentration and temperature gradient extreme value data with the device coordinates in the three-dimensional model database through the device number. Extract the corresponding three-dimensional coordinates from the three-dimensional model database according to the device number, and verify the three-dimensional coordinates with the actual location of the device. Bind the current harmonic energy concentration, axial temperature gradient extreme value data, timestamp, device number, and three-dimensional coordinates. If the device number cannot match the database coordinates, freeze the data stream and trigger an alarm. If the device number matches the database coordinates, form an input feature vector. The gated recurrent unit network receives the input feature vector containing the current harmonic energy concentration, axial temperature gradient extreme value data, timestamp, device number, and three-dimensional coordinates, and stores the device number and three-dimensional coordinates separately as static identification data. The timestamp, current harmonic energy concentration, and temperature gradient extreme value constitute dynamic temporal data, which are input into the gated recurrent unit network at time steps to construct a temporal correlation model. The gating mechanism of the gated recurrent unit network includes a reset gate and an update gate. The reset gate calculates the forgetting ratio based on the current harmonic energy concentration and the previous hidden state. When the current harmonic energy concentration suddenly increases by more than 20% of the historical average, the reset gate value approaches 0, discarding irrelevant historical states. The update gate determines the information update intensity according to the change trend of the temperature gradient extreme value. If the temperature gradient extreme value rises continuously for 3 time steps, the update gate value approaches 1, strengthening the current temperature rise feature. Integrate the historical state filtered by the reset gate with the current input to capture the temporal causal relationship between the sudden increase in current harmonic energy concentration and the lagging influence intensity of the temperature gradient extreme value. Integrate the candidate state and the historical state through the update gate, retain the harmonic cumulative effect during the daily load peak period, and at the same time respond to the instantaneous harmonics caused by lightning strikes. Rebind the hidden state with the static three-dimensional coordinates to generate a temporal feature vector with spatial tags, and transmit the temporal feature vector with spatial tags to a random forest model to divide the device risk area according to the spatial distribution characteristics of the device coordinates;

[0053] S3. Combine the three-dimensional coordinates and the random forest algorithm to construct a risk zoning model, divide the risk areas of distribution equipment in the distribution network, use the random forest model to receive the time-series feature vectors with spatial labels output by the time-series correlation model, construct the risk zoning model, discretize the area covered by the distribution network into cube grids with a side length of 0.5 meters according to the three-dimensional coordinates (x, y, z), each grid corresponds to a risk analysis unit, calculate the average value of the time-series correlation intensity of the distribution equipment in the grid, take the historical failure rate of the distribution equipment in the grid as the classification target, divide the actual failure levels into normal, faulty, and warning according to the historical failure data of the distribution equipment, calculate the decrease in Gini impurity at the candidate splitting points, select the three-dimensional coordinate axis and the splitting point that maximize the purity of the decision tree sub-nodes in the risk zoning model, count the proportion of the total splitting gain of the three-dimensional coordinate axis in the decision tree, if there is an axis with an importance higher than 60%, then determine that the direction of this axis is the risk-dominant dimension, the time-series correlation intensity affected by the sudden increase in the current harmonic energy concentration and the lag of the temperature gradient extreme value of the distribution equipment is characterized by the hidden state output by the time-series correlation model, use the Poisson distribution to calculate the unit-time failure probability of the distribution equipment in the grid in the past three years, set the contribution of the real-time state to the short-term risk to account for 60%, calculate the risk score R of each grid based on the contribution degree of the time-series correlation intensity and the historical failure probability, divide the risk levels into high risk, medium risk, and low risk based on the risk score, set the 95th percentile of the R-value distribution in the 30 days before the failure in the historical data as the high-risk threshold, the 75th to 95th percentiles as the medium-risk threshold, and the 75th percentile as the low-risk threshold, recalculate the percentile thresholds every quarter according to the newly generated failure data, the input data is the feature vector containing the hidden state and three-dimensional coordinates corresponding to the historical failure records, the label is the actual failure level, input the data by stratified sampling according to the three-dimensional coordinates, determine the number of decision trees and the range of the maximum tree depth through grid search, send the high-risk area coordinate set to the inspection robot, and plan the inspection path, mark it as a red flashing icon in the three-dimensional model, push the medium-risk area coordinate set to the handheld terminal of the operation and maintenance personnel, mark it as an orange flashing icon in the three-dimensional model, and start the parameter adaptive adjustment operation for the low-risk area, mark it as a green flashing icon in the three-dimensional model;

[0054] S4. Integrate the analysis results of the comprehensive time-series correlation model and the risk zoning model to construct a health evaluation model, output the health score of the power distribution equipment, bind the dynamically acquired time-series feature vector containing the timestamp, current harmonic energy concentration, and axial temperature gradient extreme value with the equipment number and three-dimensional coordinates, input it into the trained time-series correlation model. The time-series correlation model outputs the time-series correlation intensity, quantifies the real-time correlation between the sudden increase in the current harmonic energy concentration of the power distribution equipment and the extreme value of the temperature gradient. Input the time-series feature vector containing the hidden state output by the time-series correlation model and the three-dimensional coordinates of the power distribution equipment into the trained risk zoning model. The risk zoning model outputs the grid risk score, which represents the historical and real-time comprehensive risks of the grid area where the power distribution equipment is located. Allocate time-series weights and spatial weights according to the equipment type and historical fault data, and calculate the health score of the power distribution equipment based on the hidden state and the grid risk score. When the hidden state of the power distribution equipment is higher than 0.8 and the grid risk score is higher than 0.7 for three consecutive days, start the adaptive adjustment operation. Compare the health score with the actual fault record, calculate the prediction accuracy rate. If the prediction accuracy rate is lower than 90%, trigger the retraining of the time-series correlation model and the risk zoning model;

[0055] S5. Divide the threshold range of the equipment health score, classify the health status of the equipment, locate the abnormal equipment in combination with the three-dimensional coordinates, issue equipment warnings according to the health score, and trigger corresponding control instructions. Select the equipment health score data without fault records in the past three years, calculate the 95th percentile of its distribution as the upper limit of the health score threshold for normal power distribution equipment. Extract the health score data of the power distribution equipment within 24 hours before the fault occurs, and calculate its mean value as the critical value for warning and fault. Based on the threshold range of the power distribution equipment health score, divide the power distribution equipment into normal state, warning state, and fault state. Draw the operating characteristic curve of the power distribution equipment with the health score against the fault event, calculate the area under the curve, set the threshold of the area under the curve. If the decrease in the area under the curve within the same area on a single day exceeds 20%, recalculate the threshold range of the health score. Uniquely bind the health score with the three-dimensional coordinates of the power distribution equipment through the equipment number. According to the threshold range to which the health score belongs, mark the status icon of the power distribution equipment in the three-dimensional model according to the coordinates. Calculate the mean value of the health scores of the power distribution equipment within the same grid, set the grid risk threshold. If the mean value of the health score is lower than the grid risk threshold, mark the coordinates of the power distribution equipment within the grid as an abnormal area. If it is determined that the power distribution equipment is in the normal state, mark the corresponding three-dimensional model area as green and maintain regular monitoring. If it is determined that the power distribution equipment is in the warning state, push the corresponding coordinate set to the inspection robot, mark the corresponding three-dimensional model area as orange, generate an inspection work order and associate it with the equipment maintenance record. If it is determined that the power distribution equipment is in the fault state, trigger the circuit breaker tripping instruction, mark the corresponding three-dimensional model area as red, and push the fault coordinates to the emergency repair terminal.

[0056] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A distribution network planning method for power engineering design based on 3D modeling, characterized in that, It includes the following steps: S1. Collect the effective current value data of the distribution network line and the axial temperature data of the equipment surface; S2. Based on the collected effective current value data and the axial temperature data of the equipment surface, call the three-dimensional coordinate data to form an input feature vector, use the gated recurrent unit network to construct a time-series correlation model, and explore the time-series correlation between current harmonics and temperature gradients; S3. Combine the three-dimensional coordinates and the random forest algorithm to construct a risk zoning model, and divide the risk areas of the distribution equipment in the distribution network; S4. Integrate the analysis results of the time-series correlation model and the risk zoning model to construct a health assessment model, and output the health score of the distribution equipment; S5. Divide the threshold range of the equipment health score, classify the health status of the equipment, locate the abnormal equipment in combination with the three-dimensional coordinates, give early warnings to the equipment according to the health score, and trigger corresponding control instructions.

2. A power engineering design and distribution network planning method based on 3D modeling according to claim 1, characterized in that: In the above S1, the process of collecting the effective current value data of the distribution network line and the axial temperature data of the equipment surface includes: Deploy high-precision current sensors at the incoming and outgoing line ends of the transformer and the cable joints of the distribution network line. Determine the spacing of the high-precision current sensors according to the equipment coordinate density in the three-dimensional modeling, and calculate the effective current value through the integration circuit; Lay distributed fiber optic temperature sensors along the axial center line of the surface of the distribution equipment in the distribution network. The distribution equipment includes circuit breakers and disconnectors. Arrange the temperature measurement points at intervals of 0.5 m, form a continuous monitoring network through series connection of the optical fiber links, and calculate the absolute temperature value of the distribution equipment by using the light intensity ratio.

3. A power engineering design and distribution network planning method based on 3D modeling according to claim 2, characterized in that: In the above S1, the process of preprocessing the collected effective current value data and the axial temperature data of the equipment surface includes: Perform wavelet transform decomposition on the original current waveform signal in the collection process, decompose the current signal into fundamental wave and 2-50th harmonic components through fast Fourier transform, and retain the harmonic amplitude; Use the sliding window standard deviation method to identify abnormal points in the original temperature data of the fiber optic sensor. If the deviation of a temperature point from the average value of the adjacent 5 points in the window exceeds 3 times the standard deviation, it is determined as an abnormal value and excluded. For the temperature sequence after excluding the abnormal values, use the sliding window average method for processing. Set the window width to 7 sampling points, take the arithmetic average of the temperature data in the window to replace the original value, and calculate the ratio of the temperature difference between adjacent points to the spacing along the axial center line of the distribution equipment at intervals of 0.5 m, and output the axial temperature gradient sequence; Inject the effective current value data and the axial temperature data of the equipment surface into a unified clock source, align the time tags, and start interpolation compensation for the data segment with a deviation exceeding ±2 ms. When the temperature monitoring points are continuously missing data for more than 5 sampling periods, take the average temperature of the two adjacent monitoring points along the axis of the same equipment to replace the missing value.

4. A power engineering design and distribution network planning method based on 3D modeling according to claim 3, characterized in that: In the above S1, the process of feature extraction from the preprocessed effective current value data and the axial temperature data of the equipment surface includes: Based on the 2-50th harmonic current amplitudes retained in the preprocessing, calculate the sum of the proportions of the energies of each harmonic component to the fundamental wave energy to obtain the harmonic energy concentration. Based on the effective current value data of the continuous periodic sequence, calculate the relative change rate of the effective current values of adjacent periods; Along the axial temperature gradient sequence on the equipment surface, select the extreme values of the axial temperature gradient, and statistically calculate the algebraic sum of the axial temperature gradients on the equipment surface within a single-day monitoring period to obtain the cumulative temperature change.

5. A power engineering design distribution network planning method based on 3D modeling according to claim 4, characterized in that: In S2, the process of calling three-dimensional coordinate data to form an input feature vector includes: Introduce a three-dimensional model database that pre-stores the spatial location information of power distribution equipment in the power engineering design stage. Automatically associate the harmonic energy concentration and temperature gradient extreme value data with the equipment coordinates in the three-dimensional model database through the equipment number. Extract the corresponding three-dimensional coordinates from the three-dimensional model database according to the equipment number, and verify the three-dimensional coordinates with the actual location of the equipment. Bind the current harmonic energy concentration, axial temperature gradient extreme value data, time stamp, equipment number, and three-dimensional coordinates. If the equipment number cannot match the database coordinates, freeze the data stream and trigger an alarm. If the equipment number matches the database coordinates, form an input feature vector.

6. A power engineering design distribution network planning method based on 3D modeling according to claim 5, characterized in that: In S2, the process of constructing a time-series correlation model includes: The gated recurrent unit network receives an input feature vector containing the current harmonic energy concentration, axial temperature gradient extreme value data, time stamp, equipment number, and three-dimensional coordinates. Store the equipment number and three-dimensional coordinates separately as static identification data. The time stamp, current harmonic energy concentration, and temperature gradient extreme value constitute dynamic time-series data, which are input into the gated recurrent unit network in time steps to construct a time-series correlation model. The gating mechanism of the gated recurrent unit network includes a reset gate and an update gate. The reset gate calculates the forgetting ratio based on the current harmonic energy concentration and the previous hidden state. When the sudden increase in the harmonic energy concentration exceeds 20% of the historical average, the reset gate value approaches 0, discarding irrelevant historical states. The update gate determines the information update intensity according to the change trend of the temperature gradient extreme value. If the temperature gradient extreme value rises continuously for three time steps, the update gate value approaches 1, strengthening the current temperature rise feature. Fuse the historical state filtered by the reset gate with the current input to capture the time-series causal relationship between the sudden increase in the harmonic energy concentration and the lagging influence intensity of the temperature gradient extreme value. Fuse the candidate state and the historical state through the update gate to retain the harmonic cumulative effect during the daily load peak period and at the same time respond to the instantaneous harmonics caused by lightning strikes. Rebind the hidden state with the static three-dimensional coordinates to generate a time-series feature vector with spatial labels, and transmit the time-series feature vector with spatial labels to the random forest model. Combine the spatial distribution characteristics of the equipment coordinates to divide the equipment risk area.

7. A power engineering design and distribution network planning method based on 3D modeling according to claim 6, characterized in that: In S3, the process of constructing a risk zoning model includes: The random forest model is used to receive the time-series feature vectors with spatial labels output by the time-series correlation model, construct a risk zoning model, discretize the distribution network coverage area into cubic grids with a side length of 0.5 meters according to the three-dimensional coordinates (x, y, z), each grid corresponds to a risk analysis unit, calculate the average value of the time-series correlation intensity of the power distribution equipment in the grid, take the historical failure rate of the power distribution equipment in the grid as the classification target, divide the actual failure levels into normal, faulty and warning according to the historical failure data of the power distribution equipment, calculate the decrease in Gini impurity at the candidate splitting points, select the three-dimensional coordinate axis and splitting point that maximize the purity of the decision tree sub-nodes in the risk zoning model, and count the proportion of the total splitting gain of the three-dimensional coordinate axis in the decision tree. If the importance of an axis is higher than 60%, it is determined that the direction of this axis is the risk-dominant dimension; The time-series correlation intensity of the sudden increase in the current harmonic energy concentration and the lag effect of the temperature gradient extreme value of the power distribution equipment is characterized by the hidden state output by the time-series correlation model. The failure probability per unit time of the power distribution equipment in the grid in the past three years is calculated using the Poisson distribution. It is set that the real-time state contributes 60% to the short-term risk. Based on the contribution degree of the time-series correlation intensity and the historical failure probability, the risk score R of each grid is calculated. Based on the risk score, the risk levels are divided into high risk, medium risk and low risk. The 95th percentile of the R value distribution 30 days before the failure in the historical data is set as the high-risk threshold, the 75th to 95th percentiles are set as the medium-risk threshold, and the 75th percentile is set as the low-risk threshold. Every quarter, the quantile thresholds are recalculated according to the newly generated failure data; The input data is the feature vector containing the hidden state and three-dimensional coordinates corresponding to the historical failure records, and the label is the actual failure level. The input data is sampled by layer according to the three-dimensional coordinates, and the number of decision trees and the range of the maximum tree depth are determined through grid search; The high-risk area coordinate set is sent to the inspection robot, and the inspection path is planned and marked as a red flashing icon in the three-dimensional model. The medium-risk area coordinate set is pushed to the handheld terminal of the operation and maintenance personnel and marked as an orange flashing icon in the three-dimensional model. The low-risk area starts the parameter adaptive adjustment operation and is marked as a green flashing icon in the three-dimensional model.

8. A power engineering design distribution network planning method based on 3D modeling according to claim 7, characterized in that: In S4, the process of constructing the health evaluation model includes: Bind the newly collected dynamic time-series feature vectors containing the timestamp, current harmonic energy concentration, and axial temperature gradient extreme value with the equipment number and three-dimensional coordinates, input them into the trained time-series correlation model. The time-series correlation model outputs the time-series correlation intensity and quantifies the real-time correlation between the sudden increase in the current harmonic energy concentration and the temperature gradient extreme value of the power distribution equipment. Input the time-series feature vector containing the hidden state output by the time-series correlation model and the three-dimensional coordinates of the power distribution equipment into the trained risk zoning model. The risk zoning model outputs the grid risk score, which represents the historical and real-time comprehensive risks of the grid area where the power distribution equipment is located; Allocate temporal weights and spatial weights according to the device type and historical fault data, calculate the health score of the distribution equipment based on the hidden state and grid risk score. When the hidden state of the distribution equipment is higher than 0.8 and the grid risk score is higher than 0.7 for three consecutive days, start the adaptive adjustment operation, compare the health score with the actual fault record, calculate the prediction accuracy rate. If the prediction accuracy rate is lower than 90%, trigger the retraining of the temporal correlation model and the risk zoning model.

9. A power engineering design and distribution network planning method based on three-dimensional modeling according to claim 8, characterized in that: In step S5, the process of dividing the threshold range of the equipment health score and classifying the health status of the equipment includes: Select the health score data of the equipment without fault records in the past three years, calculate the 95th percentile of its distribution as the upper limit of the health score threshold of the normal distribution equipment, extract the health score data of the distribution equipment within 24 hours before the fault occurs, calculate its mean value as the critical value of early warning and fault. Based on the threshold range of the distribution equipment health score, divide the distribution equipment into normal state, early warning state and fault state; Draw the working characteristic curve of the distribution equipment with the health score against the fault event, calculate the area under the curve, set the threshold of the area under the curve. If the decrease in the area under the curve within the same region on a single day exceeds 20%, recalculate the threshold range of the health score.

10. A power engineering design and distribution network planning method based on 3D modeling according to claim 9, characterized in that: In step S5, the process of locating abnormal equipment, giving equipment early warning, and triggering corresponding control instructions includes: Bind the health score and the three-dimensional coordinates of the distribution equipment uniquely through the equipment number. According to the threshold range to which the health score belongs, mark the status icon of the distribution equipment according to the coordinates in the three-dimensional model. Calculate the mean value of the health scores of the distribution equipment within the same grid, set the grid risk threshold. If the mean value of the health score is lower than the grid risk threshold, mark the coordinates of the distribution equipment within the grid as an abnormal area; If it is determined that the distribution equipment is in the normal state, mark the corresponding three-dimensional model area as green and maintain routine monitoring. If it is determined that the distribution equipment is in the early warning state, push the corresponding coordinate set to the inspection robot, mark the corresponding three-dimensional model area as orange, generate an inspection work order and associate it with the equipment maintenance record. If it is determined that the distribution equipment is in the fault state, trigger the breaker tripping instruction, mark the corresponding three-dimensional model area as red, and push the fault coordinates to the emergency repair terminal.

Images