Feedback method and system for line loss abnormity of transformer area

Abstract

The invention discloses a transformer area line loss abnormity feedback method and system, and belongs to the technical field of line loss data analysis, and the method comprises the steps: obtaining a multi-dimensional transformer area data set and a meter data set, and inputting the data into a preset line loss trend analysis model, obtaining a line loss value and a line loss prediction interval according to the season trend factor, the user feature correction factor and the whole line temperature compensation coefficient, carrying out line loss abnormal triggering judgment, and generating an abnormal triggering condition instruction; and performing preset value fault mode matching detection according to the abnormal triggering condition instruction, calculating a propagation probability, positioning a line loss abnormal source and a involved user area, and generating a transformer area line loss abnormal feedback. According to the technical scheme, temperature, season and user factors are used for the line loss trend analysis model, and accurate line loss abnormity judgment is achieved; user monitoring is carried out by acquiring an abnormal triggering condition instruction, the fault propagation probability is calculated, and line loss abnormal positioning and feedback generation are realized.

Description

Technical Field

[0001] This application relates to the field of line loss data analysis technology, and in particular to a feedback method and system for abnormal line loss in transformer substations. Background Technology

[0002] In power systems, transformer substation line loss is a crucial indicator for measuring the management level of power supply companies and the efficiency of power grid operation. Accurate calculation of transformer substation line loss based on edge computing utilizes edge computing technology to process and analyze transformer substation power data at the network edge, close to the data source, to achieve accurate calculation of transformer substation line loss. Smart meters and sensor devices with communication capabilities are deployed at key locations in the transformer substation (such as transformers and branch lines) to collect various power data such as voltage, current, power, and energy consumption in real time. Dedicated data processing and analysis programs are pre-deployed at the edge nodes to perform preliminary preprocessing on the collected raw data and then perform real-time analysis and processing.

[0003] Currently, traditional transformer substation line loss calculations rely on electrical parameters as the data source. However, in reality, transformer substation line loss is also affected by various other factors, including the environment. Relying solely on a single electrical parameter data source results in an incomplete analysis of transformer substation line loss. When multiple substations experience abnormal line loss simultaneously, traditional methods cannot distinguish between independent faults and cascading reactions, making it difficult to accurately capture the complex and ever-changing line loss situation in actual operation. This leads to a high false alarm rate in fault diagnosis results and an inability to respond promptly to anomaly diagnoses and take appropriate measures. Therefore, the limitation of existing technologies lies in their inability to accurately and quickly estimate transformer substation line loss anomalies and formulate fault repair plans. Summary of the Invention

[0004] This application provides a feedback method and system for abnormal line loss in distribution transformer areas, which can solve the problem of accurate judgment and feedback of abnormal line loss in distribution transformer areas based on multi-source data.

[0005] This application provides a feedback method for abnormal line loss in a transformer substation, including: Obtain the real-time transformer area dataset and the real-time meter dataset, and use the real-time transformer area dataset and the real-time meter dataset as the first dataset; The first dataset is input into a preset line loss trend analysis model. Based on the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient, the actual line loss value and line loss prediction range for each first user are calculated. The line loss trend analysis model is a network model trained based on historical transformer area datasets and full line temperature compensation coefficients. The seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient are obtained based on the first dataset. When the actual line loss value corresponding to any first user meets the line loss prediction interval under a first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated. Based on the abnormal trigger condition instruction, fault mode matching detection is performed to obtain the abnormal line loss status corresponding to each first user. The fault mode matching detection is obtained by dynamically adding interference to each first user. Based on the abnormal line loss status of each first user and the first dataset, the transmission area propagation probability is calculated, and the area with the highest sum of propagation probabilities within the transmission area is taken as the abnormal propagation source, generating an abnormal line loss feedback result.

[0006] This application utilizes a line loss trend analysis model to calculate the theoretical line loss value, actual line loss value, and predicted line loss range for each user from a first dataset. This achieves efficient transformation from multi-source transformer area datasets to precise line loss characteristics, establishing a comprehensive data foundation reflecting user power consumption status and line loss conditions, and providing a mathematical basis for subsequent line loss anomaly judgment. The line loss trend analysis model is trained using historical transformer area datasets and full-line temperature compensation coefficients, giving the model output environmental significance and improving its accuracy. By adding dynamic interference, a comparison of user impedance spectral rate and fault modes under the same scenario is constructed, achieving highly robust judgment of user line loss status. By calculating the propagation probability of anomalies in each zone of the transformer area, the area with the highest propagation probability is identified as the propagation source of line loss anomalies, effectively distinguishing between local faults and cascading anomalies, thereby locating the source of line loss anomalies and providing decision support for rapid fault location and targeted handling. Compared to existing technologies, this application can perform rapid line loss anomaly diagnosis based on multi-source electrical data, combined with actual and predicted values, achieving the technical effect of rapid and accurate line loss anomaly location and feedback.

[0007] Further, the step of inputting the first dataset into a preset line loss trend analysis model, and calculating the actual line loss value and line loss prediction interval for each first user based on the seasonal trend factor, user characteristic correction factor, and full-line temperature compensation coefficient, includes: Calculate the theoretical line loss value based on the resistance value, current value, and user characteristic correction factor of the first dataset; The actual line loss value is obtained based on the theoretical line loss value, the temperature compensation coefficient of the entire line, the resistance value of the first dataset, and the statistical line loss value of the first dataset. Based on the line loss characteristics and the seasonal trend factor, the line loss prediction range is obtained.

[0008] By inputting real-time data from the transformer substations into the line loss trend analysis model, the theoretical line loss value, incorporating the influence of user behavior data, is first calculated. Through correlation analysis between user behavior data and sudden changes in line loss, abnormal power consumption patterns are accurately identified. By introducing a temperature compensation coefficient based on the theoretical line loss value, dynamic corrections are made for resistance changes caused by ambient temperature. This ensures that the output actual line loss value, in addition to basic power data, also reflects user behavior data and temperature data, significantly improving the accuracy and scientific rigor of line loss calculation. Furthermore, a seasonal trend factor, reflecting the systematic fluctuations of power consumption data within a fixed period, is introduced. By adding this factor, a line loss prediction range considering seasonal factors is generated, providing a dynamic and reasonable threshold range for subsequent line loss anomaly detection, enhancing the adaptability and reliability of anomaly detection.

[0009] Furthermore, the line loss trend analysis model is obtained by training a network model based on historical transformer area datasets and the temperature compensation coefficient of the entire line, including: Set the physical constraints of the network model; Based on the temperature compensation coefficient of the entire line, construct the model loss function; Based on the historical transformer area dataset, the physical constraints, and the model loss function, a line loss trend analysis model is trained.

[0010] By constructing a line loss trend analysis model, the line loss range for the next cycle can be predicted, enabling proactive maintenance of distribution lines, avoiding misjudgments based on fixed thresholds, and improving the flexibility of early warning. By building a model loss function based on the temperature compensation coefficient of the entire line and training this loss function using historical distribution area datasets, the line loss trend analysis model can actively learn and internalize the impact of temperature changes on line losses during the optimization process. This significantly improves the accuracy of the model's output results and ensures that the calculation results naturally include the physical meaning of temperature changes, enhancing the model's interpretability and environmental adaptability. By setting physical constraints on the model during training, it is further ensured that the model output conforms to the actual operating laws of the power system, improving the scientific nature of the model structure and the practical reference value of the output results. This makes the prediction results not only accurate but also possess realistic physical meaning and interpretability.

[0011] Furthermore, the step of performing fault mode matching detection based on the abnormal triggering condition command to obtain the abnormal line loss status corresponding to each first user is specifically as follows: According to the abnormal triggering condition command, dynamic interference is added to the corresponding first user. When the impedance spectral rate decreases, similarity matching detection is performed with the preset fault mode to obtain the similarity after matching. When the similarity after matching is greater than a preset first threshold, the line loss is determined to be abnormal, and the user is marked as an abnormal power consumption user.

[0012] By receiving abnormal trigger condition commands, the impedance spectral rate of users suspected of having abnormal line loss is compared with the impedance spectral rate after dynamic interference is added. The impedance spectral rate under the fault mode can be used to further verify and determine the user's line loss status more accurately. By recording each abnormal user, the generation of subsequent abnormal monitoring information sheets for users in the transformer area can be realized.

[0013] Further, the step of calculating the transmission area propagation probability based on the abnormal line loss status corresponding to each first user, and taking the area with the highest sum of propagation probabilities within the transmission area as the abnormal propagation source, and generating an abnormal line loss feedback result, includes: Based on the abnormal line loss status of each first user, the regional propagation probability of the area where each abnormal point is located is calculated according to the impedance between cables in the first data set, the time difference of the abnormal event, and the preset impedance attenuation factor. The sum of the propagation probabilities of several areas within the transformer area is calculated, and the area with the highest sum of propagation probabilities is taken as the source of abnormal propagation, generating line loss feedback results.

[0014] By calculating and accumulating the propagation probability of anomalies in each region according to the division of the transformer area, the sum of the propagation probabilities of that node pointing to other nodes is obtained. If the sum of the propagation probabilities of one region is higher than that of other regions, then the anomaly point here can be identified as the source of the anomaly propagation. This method can effectively identify the key starting point of anomaly propagation in space, and achieve accurate location of the source of line loss faults, thereby providing clear target guidance for subsequent operation and maintenance decisions.

[0015] Furthermore, when the actual line loss value corresponding to any first user and the line loss prediction interval meet a first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated, specifically as follows: Based on the actual line loss value and the first dataset, impedance spectral rate analysis is performed. When the decrease in impedance spectral rate is greater than the second preset threshold, an abnormal trigger condition command is generated. When the actual line loss value continuously exceeds the predicted range value, an abnormal trigger condition instruction is generated; When the actual line loss value exceeds the preset standard line loss threshold and the real-time load used by the user increases, an abnormal trigger condition command is generated. When the actual line loss value exceeds the third preset threshold and the duration exceeds the first preset time, an abnormal trigger condition command is generated.

[0016] By establishing a proactive anomaly diagnosis mechanism through multiple line loss anomaly judgment conditions such as abnormal time and load, the system can intelligently trigger the anomaly judgment mechanism when the actual line loss value meets any condition, and generate corresponding anomaly trigger condition commands. This shortens the fault response cycle and completes the shift from "passive repair" to "proactive early warning." The design of the anomaly trigger condition commands ensures the timely detection of potential abnormal power users, avoiding the waste of computing resources caused by directly applying dynamic interference to all users, and achieving an effective combination of precise triggering and efficient diagnosis.

[0017] Furthermore, the overall line temperature compensation coefficient is obtained based on the first dataset, specifically as follows: Based on the temperature data of the first dataset and the reference temperature resistance value, the line resistance is dynamically corrected to obtain the corrected resistance value. The interval temperature compensation coefficient is calculated based on the corrected resistance value and the reference temperature resistance value; wherein, the interval temperature compensation coefficient can also be obtained from the standard temperature coefficient of the wire material and the difference between the reference temperature and the actual temperature. Based on the temperature compensation coefficient, and according to the first dataset, the update time point is set to obtain the temperature compensation coefficient for the entire line.

[0018] By constructing a temperature compensation coefficient, temperature data is scientifically integrated into the line loss calculation process, ensuring that the line loss value not only reflects electrical characteristics but also possesses a clear climatological and physical meaning. By setting update time points, the temperature compensation coefficient for the entire line can be dynamically acquired, ensuring that the line loss calculation is updated in real time with changes in ambient temperature, effectively improving the timeliness and environmental adaptability of the line loss analysis results.

[0019] Furthermore, the acquisition of the real-time distribution area dataset and the real-time meter dataset specifically involves: The real-time transformer area dataset includes electrical data generated by the lines, environmental datasets, and user behavior datasets. The data of the transformer substations is spatially divided, and each subdivided area is labeled and coded, and the spatial boundary coordinates are marked. The construction of the electrical dataset includes electrical data generated by the lines; The environmental dataset includes data on the external environment of the line body; The user behavior dataset includes electricity data used by users.

[0020] By fusing line electrical data, environmental data, and user behavior data to construct a multi-source real-time dataset, the operating status of the transformer substation can be comprehensively and accurately reflected, eliminating the errors of traditional single-dimensional line loss estimation, and precisely quantifying actual line losses, thereby improving the accuracy and reliability of line loss calculation results. By spatially dividing and marking the coordinates of the transformer substation, local faults and cascading anomalies can be effectively distinguished during subsequent anomaly location, thus accurately pinpointing the propagation source of line loss anomalies.

[0021] Furthermore, after generating the line loss feedback results, it also includes: Based on the transformer area line paths in the first dataset, a transformer area association map is established, and the real-time status mapping of transformer area lines is performed. The location of transformer area lines is fused using positioning technology. An anomaly heatmap is generated based on the frequency of users with abnormal power consumption. Based on the anomaly heatmap, locations prone to anomalies are marked. The anomaly level of the locations prone to anomalies is marked by the number of anomalies and the cause of the fault. Based on the line loss feedback results, the fault repair time is sorted according to the anomaly marking level.

[0022] By establishing a transformer area association map for real-time mapping of line status, the system can intuitively and quickly locate areas where anomalies occur. Feedback results are used to visualize heatmaps of users with abnormal power consumption, marking locations prone to line loss and assigning them a classification. This allows the system to automatically prioritize fault repair tasks, helping maintenance personnel quickly identify key anomalies, determine the order of handling, and effectively improve fault response efficiency and the scientific nature of maintenance decisions.

[0023] This application provides a feedback system for abnormal line loss in a transformer area, including: a data acquisition module, a model calculation module, a line loss judgment module, and a result feedback module; The data acquisition module is used to acquire real-time transformer area dataset and real-time meter dataset, and to use the real-time transformer area dataset and the real-time meter dataset as the first dataset. The model calculation module is used to input the first dataset into a preset line loss trend analysis model, and calculate the actual line loss value and line loss prediction range for each first user based on the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient; wherein, the line loss trend analysis model is obtained by training a network model based on historical transformer area dataset and full line temperature compensation coefficient; wherein, the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient are obtained based on the first dataset; The line loss judgment module is used to generate an abnormal trigger condition instruction for any first user when the actual line loss value and the line loss prediction interval meet a first preset condition, and to perform fault mode matching detection based on the abnormal trigger condition instruction to obtain the abnormal line loss status of each first user; wherein, the fault mode matching detection is obtained by dynamically adding interference to each first user. The result feedback module is used to calculate the transmission area propagation probability based on the abnormal line loss status of each first user and the first dataset, and to take the area with the highest sum of propagation probabilities within the transmission area as the abnormal propagation source, and generate the abnormal line loss feedback result. Attached Figure Description

[0024] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating an embodiment of the feedback method for abnormal line loss in the transformer area provided in this application.

[0026] Figure 2 This is a schematic diagram of one embodiment of a feedback system for abnormal line loss in a transformer substation, provided in this application. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0029] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0030] In the description of the embodiments of this application, the size of the interval and the threshold are set for the purpose of facilitating comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0031] In the description of the embodiments of this application, all formulas are dimensionless and numerical calculations. The formulas are derived from the most recent real-world situation obtained by software simulation using a large amount of collected data. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0032] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0033] Example 1 See Figure 1 To address the issues of abnormal line loss monitoring and location feedback in existing distribution transformer substations, this application provides an embodiment of a method for feedback on abnormal line loss in distribution transformer substations, comprising steps 11 to 14, the specific steps of which are as follows: Step 11: Obtain the real-time transformer area dataset and the real-time meter dataset, and use the real-time transformer area dataset and the real-time meter dataset as the first dataset.

[0034] Furthermore, the acquisition of the real-time transformer area dataset and the real-time meter dataset specifically involves: the real-time transformer area dataset including electrical data generated by the lines, environmental dataset, and user behavior dataset; spatially dividing the transformer area data, labeling and encoding each divided area, and marking the spatial boundary coordinates; wherein, the construction of the electrical dataset includes electrical data generated by the lines; wherein, the environmental dataset includes external environmental data of the lines; and wherein, the user behavior dataset includes electricity data used by users.

[0035] In one embodiment, obtaining a real-time transformer area dataset and a real-time meter dataset, and using the real-time transformer area dataset and the real-time meter dataset as a first dataset, includes step 1101, which is as follows: Step 1101: Obtain the real-time transformer area dataset and the real-time meter dataset.

[0036] The monitoring area is spatially divided according to the continuity and the integrity of power generation, transmission and distribution functions. The N partitions are labeled, coded and marked with spatial boundary coordinates.

[0037] Based on the power parameter sensors and environmental parameter sensors installed at each key node of the transformer substation, the relevant power parameters and environmental parameters generated by the transformer substation lines are collected, and the coordinate points of each key node are marked. The electrical data generated by the lines are collected to construct an electrical dataset, and the external environmental data of the lines are used to construct an environmental dataset. The electrical dataset and the environmental dataset include node coordinate point markings.

[0038] Specifically, based on the user's smart meter reading the user's load curve, a load data sheet for the user is generated, and peak and valley periods and sudden events are marked. A user behavior dataset is constructed based on the user's electricity usage data.

[0039] The real-time electrical data, environmental dataset, and user behavior dataset generated by the line are used as the real-time transformer area dataset; real-time labeled data collected by the line monitoring equipment are acquired to construct the real-time meter dataset; finally, the real-time transformer area dataset and the real-time meter dataset are used as the first dataset.

[0040] By partitioning the transformer substation data and recording coordinate points, a precise correspondence between substation data and spatial location is established. This lays a geographic information foundation for subsequent calculation of propagation probability based on partitions, thereby enabling anomaly location of line loss. Furthermore, by integrating electrical data, environmental data, and user behavior data to create a multi-source real-time dataset, the complex operational status of the transformer substations is more realistically reflected. This provides more comprehensive and dynamic data support for subsequent model building and line loss prediction, improving the scientific rigor and accuracy of the feedback results.

[0041] Step 12: Input the first dataset into the preset line loss trend analysis model, and calculate the actual line loss value and line loss prediction range for each first user based on the seasonal trend factor, user characteristic correction factor and full line temperature compensation coefficient; wherein, the line loss trend analysis model is obtained by training the network model based on the historical transformer area dataset and the full line temperature compensation coefficient; wherein, the seasonal trend factor, user characteristic correction factor and full line temperature compensation coefficient are obtained based on the first dataset.

[0042] Further, the step of inputting the first dataset into a preset line loss trend analysis model and calculating the actual line loss value and line loss prediction interval for each first user based on the seasonal trend factor, user characteristic correction factor, and full-line temperature compensation coefficient includes: calculating the theoretical line loss value based on the resistance value, current value, and user characteristic correction factor of the first dataset; obtaining the actual line loss value based on the theoretical line loss value, the full-line temperature compensation coefficient, the resistance value of the first dataset, and the statistical line loss value of the first dataset; and obtaining the line loss prediction interval based on the line loss characteristics and the seasonal trend factor.

[0043] Furthermore, the line loss trend analysis model is obtained by training a network model based on historical transformer area datasets and the temperature compensation coefficients of the entire line, including: setting physical constraints for the network model; constructing a model loss function based on the temperature compensation coefficients of the entire line; and training the line loss trend analysis model based on the historical transformer area datasets, the physical constraints, and the model loss function.

[0044] Furthermore, the full-line temperature compensation coefficient is obtained based on the first dataset, specifically by: dynamically correcting the line resistance based on the temperature data and reference temperature resistance value in the first dataset to obtain the corrected resistance value; calculating the interval temperature compensation coefficient based on the corrected resistance value and the reference temperature resistance value; wherein, the interval temperature compensation coefficient can also be obtained from the standard temperature coefficient of the line material and the difference between the reference temperature and the actual temperature; based on the temperature compensation coefficient, setting the update time point according to the first dataset to obtain the full-line temperature compensation coefficient.

[0045] In one embodiment, the first dataset is input into a preset line loss trend analysis model. Based on the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient, the actual line loss value and line loss prediction range for each first user are calculated, including steps 1201 to 1203, each step of which is as follows: Step 1201: Calculate the temperature compensation coefficient for the entire line based on the first dataset.

[0046] Based on the environmental dataset in the first dataset, temperature data under sunny conditions is obtained. By comparing the temperature data with the line loss values ​​at different times, a curve analysis and observation graph is established to observe the trend of line loss values ​​under the influence of temperature data. The curve analysis and observation graph uses environmental parameters as the horizontal axis and temperature data as the vertical axis.

[0047] Based on the electrical data set in the first dataset, the resistance value at the reference temperature is obtained. By combining real-time temperature data, the line resistance is dynamically corrected to obtain the corrected resistance value. Through the corrected resistance value Resistance value at reference temperature The temperature compensation coefficient W for the interval is calculated using the following formula: Where β is a standard temperature coefficient set according to the material of the yarn. This represents the difference between the real-time temperature and the reference temperature.

[0048] Based on the above temperature compensation coefficient formula, the update time point is intelligently set for the data to generate real-time temperature compensation coefficients for the entire line.

[0049] By constructing a temperature compensation coefficient, temperature data is scientifically integrated into the line loss calculation process, ensuring that the line loss results not only reflect electrical characteristics but also possess clear climatological and physical meaning. The dynamic correction of resistance values ​​accurately calibrates the line resistance based on real-time temperature, reducing the impact of resistance drift caused by temperature changes on line loss calculations and ensuring that line loss values ​​more closely reflect actual operating conditions. By setting update time points, the system can dynamically acquire the temperature compensation coefficient for the entire line, ensuring that line loss calculations are updated in real time with changes in ambient temperature, effectively improving the timeliness and environmental adaptability of line loss analysis results.

[0050] Step 1202: Train the line loss trend analysis model based on the temperature compensation coefficient of the entire line and the historical transformer area dataset.

[0051] The historical transformer area dataset includes historical electrical datasets and historical environmental datasets.

[0052] Specifically, based on the historical electrical dataset and historical environmental dataset, an initial network model based on a fully connected neural network architecture is set as a line loss trend analysis model. Key data is obtained from the historical electrical dataset and historical environmental dataset. Physical constraints and loss functions of the network model are set, and the temperature compensation coefficient is embedded into the loss function of the initial network model to train and obtain the line loss trend analysis model.

[0053] By constructing a line loss trend analysis model, the line loss range for the next cycle can be predicted, enabling proactive maintenance of distribution lines, avoiding misjudgments based on fixed thresholds, and improving the flexibility of early warning. By embedding the temperature compensation coefficient of the entire line into the loss function and training the model based on historical distribution area datasets, the line loss trend analysis model internalizes the influence of temperature on line loss during the optimization process, thereby improving prediction accuracy and enhancing the physical interpretability of the results. By setting physical constraints on the model during training, it is further ensured that the model output conforms to the actual operating laws of the power system, improving the scientific nature of the model structure and giving the prediction results realistic physical meaning and interpretability.

[0054] Step 1203: Extract correlation features from the first dataset, input the correlation features into the trained line loss trend analysis model, and obtain the theoretical line loss value, the actual line loss value, and the line loss prediction range.

[0055] Specifically, data for each first user within the prediction period is extracted from the first dataset and used as input to the line loss trend analysis model for that first user. Correlation features are extracted from the electrical dataset and the environmental dataset. The theoretical line loss value is calculated by setting the processing layer of the line loss trend analysis model. By adding a seasonal trend factor, the line loss prediction interval within the period is obtained. Each user in the transformer area dataset is considered as the first user.

[0056] The specific formula for calculating the theoretical line loss value is as follows: in, This is the theoretical line loss value. Let be the current value at time D. D is the user feature correction factor, N is the index factor representing the number of times, and N is the total number of current data.

[0057] The specific formula for calculating the actual line loss value is as follows: in, Actual line loss value after temperature compensation This is the theoretical line loss value. The statistical line loss value is the difference between the power supply line loss value and the sales line loss value, where R is the resistance value. The dynamic coefficient value is automatically set according to the working scenario when the line voltage is 10kV distribution network. When the line voltage is 35kV, the line voltage is 2. For 4,110kV lines It is 8.

[0058] The specific calculation formula for the line loss prediction interval [Y-2d, Y+2d] within the period is as follows: Where Y is the predicted value, K is the number of line loss features obtained, I is the collected line loss features, and X is the set seasonal trend factor.

[0059] The seasonal trend factor is obtained by classifying electricity consumption data for the four quarters according to seasonal cycles, calculating the average value of each cycle, and finally calculating the mean of all cycle averages. The specific calculation method is: Seasonal Trend Factor = *100% reflects the systematic fluctuations in electricity consumption data over a fixed period.

[0060] The user feature correction factor is derived from the user's electricity consumption data. The specific calculation method is: User feature correction factor = 1 + Number of abnormal electricity consumption * 0.3; where the number of abnormal electricity consumption is obtained through user behavior dataset.

[0061] By setting seasonal trend factors, user characteristic correction factors, and introducing a temperature compensation coefficient, multi-source data is embedded into the line loss trend analysis model, improving the accuracy and scientific rigor of line loss calculation. This ultimately yields a precise line loss value that incorporates the influence of temperature and user factors. Furthermore, by calculating the line loss prediction interval, accurate data range constraints are provided for subsequent anomaly detection in line loss.

[0062] Step 13: When the actual line loss value corresponding to any first user and the line loss prediction interval meet the first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated, and fault mode matching detection is performed based on the abnormal trigger condition instruction to obtain the abnormal line loss status corresponding to each first user; wherein, the fault mode matching detection is obtained by dynamically adding interference to each first user.

[0063] Furthermore, the step of performing fault mode matching detection based on the abnormal triggering condition command to obtain the abnormal line loss status corresponding to each first user specifically involves: according to the abnormal triggering condition command, dynamic interference is added to the corresponding first user; when the impedance spectral rate decreases, similarity matching detection is performed with a preset fault mode to obtain the similarity after matching; when the similarity after matching is greater than a preset first preset threshold, the line loss is determined to be an abnormal state, and the user is marked as an abnormal power consumption state user.

[0064] Further, the step of generating an abnormal trigger condition instruction for the first user when the actual line loss value and the predicted line loss interval for any first user meet a first preset condition specifically involves: performing impedance spectral rate analysis based on the actual line loss value and the first dataset; generating an abnormal trigger condition instruction when the decrease in the impedance spectral rate is greater than a second preset threshold; generating an abnormal trigger condition instruction when the actual line loss value continuously exceeds the predicted interval value; generating an abnormal trigger condition instruction when the actual line loss value exceeds a preset standard line loss threshold and the real-time load used by the user increases; and generating an abnormal trigger condition instruction when the actual line loss value exceeds a third preset threshold and the duration exceeds a first preset time.

[0065] In one embodiment, when the actual line loss value corresponding to any first user and the line loss prediction interval meet a first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated. Based on the abnormal trigger condition instruction, fault mode matching detection is performed to obtain the abnormal line loss status corresponding to each first user, including steps 1301 to 1302, each step of which is as follows: Step 1301: Based on the first preset condition, determine the abnormal line loss trigger and generate an abnormal trigger condition instruction.

[0066] Specifically, the associated feature dataset is generated by extracting associated features from the electrical dataset, environmental dataset, and real-time meter dataset of the real-time lines in the first dataset.

[0067] The first preset condition includes: performing impedance spectral rate analysis based on the associated feature dataset and the actual line loss value; generating an abnormal trigger condition command when the impedance spectral rate shows an abnormal decrease, i.e., exceeds a second preset threshold; generating an abnormal trigger condition command when the actual line loss value continuously exceeds the predicted interval value; generating an abnormal trigger condition command when the actual line loss value exceeds a preset standard line loss threshold and the real-time load used by the user increases; and generating an abnormal trigger condition command when the actual line loss value exceeds the threshold of the standard line loss value, i.e., exceeds a third preset threshold, and the duration exceeds a first preset time.

[0068] By establishing multiple anomaly detection conditions, such as abnormal time and load, the system can intelligently trigger the anomaly detection mechanism when the actual line loss value meets any of these conditions, generating corresponding anomaly trigger condition commands to achieve accurate detection of line loss anomalies. The design of these anomaly trigger condition commands ensures the timely detection of potential abnormal power users, providing trigger conditions for further assessment of user line loss status. This avoids the waste of computational resources caused by directly applying dynamic interference to all users, achieving an effective combination of accurate triggering and efficient diagnosis.

[0069] Step 1302: Perform the abnormal feature assistance mechanism operation according to the abnormal trigger condition instruction to obtain the assistance mechanism operation monitoring form. Based on the assistance mechanism operation monitoring form and the preset value fault mode matching detection, obtain the line loss abnormal matching point and generate the transformer area user abnormal monitoring information form.

[0070] Obtain the abnormal trigger condition command, and dynamically add interference to the corresponding user according to the abnormal trigger condition command. When the impedance spectral rate decreases after adding interference, perform matching detection according to the preset fault mode to obtain the similarity after matching. When the similarity after matching is greater than the standard matching state, that is, greater than the first preset threshold, it means that the line loss consumption is abnormal and the line loss is judged to be in an abnormal state.

[0071] The first user whose line loss is judged to be abnormal is marked as having an abnormal power consumption status. The line loss status of all first users in the transformer area is obtained, the abnormal line loss matching point of the user with abnormal power consumption status is obtained, and the abnormal monitoring information sheet of the transformer area user is generated.

[0072] By receiving abnormal trigger condition commands, the impedance spectral rate of users suspected of having abnormal line loss is compared with that of the dynamically added interference. This allows for a more accurate verification and determination of the user's line loss status. By recording each abnormal user, a data foundation is provided for calculating the propagation source of abnormal line loss in subsequent steps.

[0073] Step 14: Calculate the transmission area propagation probability based on the abnormal line loss status of each first user and the first dataset, and take the area with the highest sum of propagation probabilities within the transmission area as the abnormal propagation source to generate abnormal line loss feedback results.

[0074] Further, the step of calculating the distribution area propagation probability based on the abnormal line loss status corresponding to each first user and the first dataset, and taking the area with the highest sum of propagation probabilities within the distribution area as the abnormal propagation source, and generating a line loss abnormality feedback result, includes: based on the abnormal line loss status corresponding to each first user, calculating the regional propagation probability of the area where each abnormal point is located within the distribution area according to the impedance between cables, the time difference of the abnormal event, and a preset impedance attenuation factor in the first dataset; calculating the sum of regional propagation probabilities of several areas within the distribution area according to the regional propagation probabilities, and taking the area with the highest sum of regional propagation probabilities as the abnormal propagation source, and generating a line loss feedback result.

[0075] Furthermore, after generating the line loss feedback result, the method further includes: establishing a transformer area association map based on the transformer area line paths in the first dataset, performing real-time status mapping of the transformer area lines, and performing location fusion of the transformer area lines using positioning technology; forming an anomaly point heatmap based on the frequency of users with abnormal power consumption status, marking locations prone to anomalies based on the anomaly point heatmap, and marking the anomaly marking level of the locations prone to anomalies based on the number of anomalies and the cause of the fault; and sorting the fault repair time based on the line loss feedback result and the anomaly marking level.

[0076] In one embodiment, based on the abnormal line loss status corresponding to each first user and the first dataset, the transmission area propagation probability is calculated, and the area with the highest sum of propagation probabilities within the transmission area is taken as the abnormal propagation source, generating an abnormal line loss feedback result, including steps 1401 to 1402, each step as follows: Step 1401: Extract the time-series characteristics of line loss mutations in each transformer area, calculate the fault propagation probability based on the line path of the transformer area, and generate a line loss anomaly feedback information sheet.

[0077] Generate anomaly monitoring information sheets for users in the transformer area. When multiple anomalies appear simultaneously in the anomaly monitoring information sheets for users in the transformer area, perform fault propagation analysis of the line path in the transformer area.

[0078] By constructing a physical connection diagram and treating each transformer substation as a node, if substations I and J are directly connected by a cable, the probability of propagation from substation I to J is calculated. Based on the regions divided within the substation, the propagation probability of each region containing an anomaly is analyzed, yielding the sum P of the propagation probabilities from that node to other nodes. If the sum of the propagation probabilities of one region is higher than that of other regions, then that anomaly is identified as the source of the anomaly propagation. In this way, based on substation line topology analysis and historical fault mode database matching, fault source tracing and propagation analysis capabilities are built. Through time-series feature extraction and fault propagation probability calculation, the root cause of the anomaly can be located and the scope of impact assessed.

[0079] The specific formula for calculating the propagation probability L is as follows: The specific formula for calculating the sum of propagation probabilities P from each node to other nodes is as follows: Where Z is the impedance of the cable between transformer substations I and J. is the time difference between abnormal events in transformer areas I and J, and s is the industry standard reference automatically set impedance attenuation factor.

[0080] By combining the fault propagation probability with the actual line loss value, an abnormal line loss feedback information sheet is generated.

[0081] Step 1402: Based on the abnormal line loss feedback information form, perform an abnormal diagnosis feedback operation.

[0082] A transformer area association map is established based on the transformer area line path. The real-time status of the transformer area lines is mapped through the transformer area association map, and the location of the transformer area lines is fused by combining positioning technology.

[0083] An anomaly heatmap is created by analyzing the frequency of anomalies. Locations in the transformer substation prone to anomalies are marked, and the anomaly frequency and cause of failure are used to classify the areas into different levels and label them accordingly.

[0084] Based on the level of the abnormal points in the transformer area, when an abnormal point fault is received, the fault repair time is sorted according to the level.

[0085] By establishing a transformer area association map for real-time mapping of line status, the system can intuitively and quickly locate areas where anomalies occur. Feedback results are used to visualize heatmaps of users with abnormal power consumption, marking locations prone to line loss and assigning them a classification. This allows the system to automatically prioritize fault repair tasks, helping maintenance personnel quickly identify key anomalies, determine the order of handling, and effectively improve fault response efficiency and the scientific nature of maintenance decisions.

[0086] See Figure 2Another embodiment of this application also provides a feedback system for abnormal line loss in a transformer area, including: a data acquisition module 201, a model calculation module 202, a line loss judgment module 203, and a result feedback module 204.

[0087] The data acquisition module 201 is used to acquire real-time transformer area dataset and real-time meter dataset, and to use the real-time transformer area dataset and the real-time meter dataset as the first dataset.

[0088] The model calculation module 202 is used to input the first dataset into a preset line loss trend analysis model, and calculate the actual line loss value and line loss prediction range for each first user based on the seasonal trend factor, user characteristic correction factor and full line temperature compensation coefficient; wherein, the line loss trend analysis model is obtained by training a network model based on historical transformer area dataset and full line temperature compensation coefficient; wherein, the seasonal trend factor, user characteristic correction factor and full line temperature compensation coefficient are obtained based on the first dataset.

[0089] The line loss judgment module 203 is used to generate an abnormal trigger condition instruction for the first user when the actual line loss value corresponding to any first user and the line loss prediction interval meet a first preset condition, and to perform fault mode matching detection based on the abnormal trigger condition instruction to obtain the abnormal line loss status corresponding to each first user; wherein, the fault mode matching detection is obtained by dynamically adding interference to each first user.

[0090] The result feedback module 204 is used to calculate the transmission area propagation probability based on the abnormal line loss status of each first user and the first dataset, and to take the area with the highest sum of propagation probabilities within the transmission area as the abnormal propagation source, and generate the abnormal line loss feedback result.

[0091] This approach, through a line loss trend analysis model, calculates the theoretical line loss value, actual line loss value, and predicted line loss range for each user from the first dataset. This achieves efficient transformation from multi-source transformer area datasets to precise line loss characteristics, establishing a comprehensive data foundation reflecting user power consumption status and line loss conditions, providing a mathematical basis for subsequent line loss anomaly judgment. Training the line loss trend analysis model with historical transformer area datasets and full-line temperature compensation coefficients imbues the model output with environmental significance, improving its accuracy. By adding dynamic interference, comparing user impedance spectral rates and fault modes under the same scenario, highly robust judgment of user line loss status is achieved. By calculating the propagation probability of anomalies within each zone of the transformer area, the area with the highest propagation probability is identified as the propagation source of line loss anomalies, effectively distinguishing between local faults and cascading anomalies, thus locating the source of line loss anomalies and providing decision support for rapid fault location and targeted handling. Compared to existing technologies, this application enables rapid line loss anomaly diagnosis based on multi-source electrical data, combined with actual and predicted values, achieving rapid and accurate line loss anomaly location and feedback.

[0092] It is understood that the above system embodiments correspond to the method embodiments of this application, and can implement the feedback method for abnormal line loss in the transformer area provided by any of the above method embodiments of this application.

[0093] It should be noted that the system embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided in this application, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0094] The above description is the preferred embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications are also considered to be within the scope of protection of this application.

Claims

1. A feedback method for abnormal line loss in a transformer substation, characterized in that, include: Obtain the real-time transformer area dataset and the real-time meter dataset, and use the real-time transformer area dataset and the real-time meter dataset as the first dataset; The first dataset is input into a preset line loss trend analysis model. Based on the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient, the actual line loss value and line loss prediction range for each first user are calculated. The line loss trend analysis model is a network model trained based on historical transformer area datasets and full line temperature compensation coefficients. The seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient are obtained based on the first dataset. When the actual line loss value corresponding to any first user meets the line loss prediction interval under a first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated. Based on the abnormal trigger condition instruction, fault mode matching detection is performed to obtain the abnormal line loss status corresponding to each first user. The fault mode matching detection is obtained by dynamically adding interference to each first user. Based on the abnormal line loss status of each first user and the first dataset, the transmission area propagation probability is calculated, and the area with the highest sum of propagation probabilities within the transmission area is taken as the abnormal propagation source, generating an abnormal line loss feedback result.

2. The feedback method for abnormal line loss in transformer substations according to claim 1, characterized in that, The step of inputting the first dataset into a preset line loss trend analysis model, and calculating the actual line loss value and line loss prediction range for each first user based on seasonal trend factors, user characteristic correction factors, and full-line temperature compensation coefficients, includes: Calculate the theoretical line loss value based on the resistance value, current value, and user characteristic correction factor of the first dataset; The actual line loss value is obtained based on the theoretical line loss value, the temperature compensation coefficient of the entire line, the resistance value of the first dataset, and the statistical line loss value of the first dataset. Based on the line loss characteristics and the seasonal trend factor, the line loss prediction range is obtained.

3. The feedback method for abnormal line loss in transformer substations according to claim 1, characterized in that, The line loss trend analysis model is a network model trained based on historical transformer area datasets and the temperature compensation coefficients of the entire line, including: Set the physical constraints of the network model; Based on the temperature compensation coefficient of the entire line, construct the model loss function; Based on the historical transformer area dataset, the physical constraints, and the model loss function, a line loss trend analysis model is trained.

4. The feedback method for abnormal line loss in transformer substations according to claim 1, characterized in that, The fault mode matching detection based on the abnormal triggering condition command, to obtain the abnormal line loss status corresponding to each first user, specifically includes: According to the abnormal trigger condition command, dynamic interference is added to the corresponding first user. When the impedance spectral rate decreases, similarity matching detection is performed with the preset fault mode to obtain the similarity after matching. When the similarity after matching is greater than a preset first threshold, the line loss is determined to be abnormal, and the user is marked as an abnormal power consumption user.

5. The feedback method for abnormal line loss in transformer substations according to claim 1, characterized in that, The step of calculating the transmission line loss propagation probability based on the abnormal line loss status of each first user and the first dataset, and taking the area with the highest sum of propagation probabilities within the transmission line area as the abnormal propagation source, and generating an abnormal line loss feedback result includes: Based on the abnormal line loss status of each first user, the regional propagation probability of the area where each abnormal point is located is calculated according to the impedance between cables in the first data set, the time difference of the abnormal event, and the preset impedance attenuation factor. Based on the regional propagation probability, the sum of regional propagation probabilities for several areas within the transformer area is calculated. The area with the highest sum of regional propagation probabilities is taken as the source of abnormal propagation, and line loss feedback results are generated.

6. The feedback method for abnormal line loss in transformer substations according to claim 1, characterized in that, When the actual line loss value corresponding to any first user and the line loss prediction interval meet a first preset condition, an abnormal trigger condition instruction corresponding to the first user is generated, specifically as follows: Based on the actual line loss value and the first dataset, impedance spectral rate analysis is performed. When the decrease in impedance spectral rate is greater than the second preset threshold, an abnormal trigger condition command is generated. When the actual line loss value continuously exceeds the predicted range value, an abnormal trigger condition instruction is generated; When the actual line loss value exceeds the preset standard line loss threshold and the real-time load used by the user increases, an abnormal trigger condition command is generated. When the actual line loss value exceeds the third preset threshold and the duration exceeds the first preset time, an abnormal trigger condition command is generated.

7. The feedback method for abnormal line loss in the transformer area according to claim 2, characterized in that, The overall line temperature compensation coefficient is obtained based on the first dataset, specifically as follows: Based on the temperature data of the first dataset and the reference temperature resistance value, the line resistance is dynamically corrected to obtain the corrected resistance value. The interval temperature compensation coefficient is calculated based on the corrected resistance value and the reference temperature resistance value; wherein, the interval temperature compensation coefficient can also be obtained from the standard temperature coefficient of the wire material and the difference between the reference temperature and the actual temperature. Based on the temperature compensation coefficient, and according to the first dataset, the update time point is set to obtain the temperature compensation coefficient for the entire line.

8. The feedback method for abnormal line loss in a transformer area according to claim 1, characterized in that, The acquisition of the real-time transformer area dataset and the real-time meter dataset specifically involves: The real-time transformer area dataset includes electrical data generated by the lines, environmental datasets, and user behavior datasets. The data of the transformer substations is spatially divided, and each subdivided area is labeled and coded, and the spatial boundary coordinates are marked. The construction of the electrical dataset includes electrical data generated by the lines; The environmental dataset includes data on the external environment of the line body; The user behavior dataset includes electricity data used by users.

9. The feedback method for abnormal line loss in transformer substations according to claims 1 to 8, characterized in that, After generating the line loss feedback result, the following is also included: Based on the transformer area line paths in the first dataset, a transformer area association map is established, and the real-time status mapping of transformer area lines is performed. The location of transformer area lines is fused using positioning technology. An anomaly heatmap is generated based on the frequency of users with abnormal power consumption. Based on the anomaly heatmap, locations prone to anomalies are marked. The anomaly level of the locations prone to anomalies is marked by the number of anomalies and the cause of the fault. Based on the line loss feedback results, the fault repair time is sorted according to the anomaly marking level.

10. A feedback system for abnormal line loss in a transformer substation, characterized in that, The system is used to implement a feedback method for abnormal line loss in a transformer area as described in any one of claims 1 to 9, comprising: a data acquisition module, a model calculation module, a line loss judgment module, and a result feedback module; The data acquisition module is used to acquire real-time transformer area dataset and real-time meter dataset, and to use the real-time transformer area dataset and the real-time meter dataset as the first dataset. The model calculation module is used to input the first dataset into a preset line loss trend analysis model, and calculate the actual line loss value and line loss prediction range for each first user based on the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient; wherein, the line loss trend analysis model is obtained by training a network model based on historical transformer area dataset and full line temperature compensation coefficient; wherein, the seasonal trend factor, user characteristic correction factor, and full line temperature compensation coefficient are obtained based on the first dataset; The line loss judgment module is used to generate an abnormal trigger condition instruction for any first user when the actual line loss value and the line loss prediction interval meet a first preset condition, and to perform fault mode matching detection based on the abnormal trigger condition instruction to obtain the abnormal line loss status of each first user; wherein, the fault mode matching detection is obtained by dynamically adding interference to each first user. The result feedback module is used to calculate the transmission area propagation probability based on the abnormal line loss status of each first user and the first dataset, and to take the area with the highest sum of propagation probabilities within the transmission area as the abnormal propagation source, and generate the abnormal line loss feedback result.

Images