Transformer area cable equivalent length measuring and calculating method combining topology and line loss constraint

Abstract

The invention belongs to the technical field of cable equivalent length measurement and calculation, and particularly relates to a transformer area cable equivalent length measurement and calculation method combining topology and line loss constraint, which comprises the following steps: S1, data aggregation and preprocessing; s2, carrying out topology analysis on the power grid of the transformer area; s3, determining a line loss constraint condition; s4, constructing an equivalent length measuring and calculating model; and S5, model verification and application. According to the transformer area cable equivalent length measuring and calculating method combining the topology and the line loss constraint, data aggregation and preprocessing are carried out by using an RPA technology, multi-source data can be automatically captured, analyzed and integrated, cross-system data integration is realized, the data collection efficiency is greatly improved, and the labor burden is reduced; and meanwhile, the data is subjected to cleaning, denoising, missing value filling and other operations, so that the accuracy and integrity of the data are ensured, high-quality data support is provided for subsequent analysis and model training, measurement and calculation errors caused by data problems are avoided, measurement and calculation results are more reliable, and the overall efficiency and quality of data processing are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cable equivalent length measurement, and in particular to a method for measuring the equivalent length of cables in a substation area by combining topology and line loss constraints. Background Art

[0002] In power systems, calculating the equivalent length of substation cables is crucial for accurately assessing line losses, optimizing grid layout, and improving power system efficiency. Traditional methods for calculating substation cable length often only consider physical length, ignoring factors such as grid topology and line losses. This leads to significant deviations from actual results. As the scale and complexity of power grids continue to expand, traditional methods are no longer sufficient for the refined management of substation line losses.

[0003] Currently, substation line loss management faces numerous challenges. On the one hand, the low-voltage power grid's complex topology, the wide variety of equipment, and its widespread distribution make data collection and management challenging. This makes obtaining accurate grid topology information difficult, which in turn affects the accuracy of calculating equivalent cable lengths. On the other hand, line loss is affected by a combination of factors, such as resistance, magnetic fields, power consumption, and equipment operating status. Traditional measurement methods struggle to fully account for these factors, resulting in inaccurate line loss calculations and a failure to provide a reliable basis for developing loss reduction measures. Traditional manual inspections and simple data analysis methods are unable to promptly detect and address line loss anomalies, making it difficult to effectively control substation line losses. Therefore, there is an urgent need for a new method for calculating equivalent cable length in substations that comprehensively considers topology and line loss constraints, improves measurement accuracy and efficiency, and provides strong support for substation line loss management. Summary of the Invention

[0004] The main purpose of the present invention is to provide a method for calculating the equivalent length of cables in a substation area by combining topology and line loss constraints, which can solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention proposes a method for calculating the equivalent length of cables in a substation area by combining topology and line loss constraints, comprising the following steps:

[0006] S1. Data aggregation and preprocessing: Utilize RPA technology to automatically capture, parse, and integrate data from various data sources, and perform preprocessing operations such as cleaning, denoising, and filling missing values ​​on the collected data.

[0007] S2. Substation grid topology analysis: Construct a substation grid topology model, analyze its topological structure characteristics, and determine the electrical connection paths of each line;

[0008] S3. Determine line loss constraints: Based on the distribution network line loss theory, analyze the causes of line losses and determine line loss constraints based on historical line loss data;

[0009] S4. Equivalent length calculation model construction: Combine the topology structure and line loss constraints to build a calculation model, and use machine learning algorithms for modeling and prediction;

[0010] S5. Model verification and application: Use actual substation operation data to verify the model, optimize the model based on the verification results, and apply the optimized model to substation line loss management and power grid planning.

[0011] Preferably, in the data aggregation and preprocessing steps, RPA technology is used to seamlessly connect multiple business systems, call applications and various web pages in the desktop system, realize cross-system data integration, and perform preliminary cleaning and format conversion on the captured data.

[0012] Preferably, in the substation power grid topology analysis step, a graph theory method is used to convert the power grid topology structure into a graph model, in which nodes represent power grid components and edges represent connection lines between components, and attributes such as line length, resistance, and reactance are assigned to the edges. The line topology information is obtained by traversing the topology graph through a depth-first search or breadth-first search algorithm.

[0013] Preferably, in the step of determining the line loss constraint condition, the average current method is used to calculate the theoretical line loss. The power distribution network to be calculated is regarded as a resistance element with a resistance value of R. When the current passing through the element is i, the power loss ΔA of the element during the calculation period is:

[0014]

[0015] Among them, I eff is the root mean square current, which is used to integrate the actual current over time. The ratio K of the root mean square current to the average current is the shape factor. The coefficient value is related to the load rate of the station area. Substituting the shape factor K into the above formula, we get:

[0016]

[0017] Where, AP and AQ are the active and reactive power in the calculation period; U av The average voltage value during the calculation period is used; at the same time, the variation range and trend of line loss are determined in combination with historical line loss data, and the reasonable range of line loss rate and the value range of related factors are set as constraints.

[0018] Preferably, in the step of constructing the equivalent length measurement model, 21 parameters are selected as model training features, including the average load rate, the maximum load rate, the proportion of data with a load rate less than 20%, the maximum and minimum values ​​of the transformer outlet voltage, the proportion of data with a power factor greater than 0.9, the transformer capacity, the number of electricity sales meters, the number of access points, the maximum power supply radius, the minimum power supply radius, the number of outgoing cables, the number of branch nodes, the maximum distance between the branch node and the transformer, the minimum distance between the branch node and the transformer, the total line length, the cable line length, the overhead line length, the maximum three-phase unbalance rate, and the proportion of data with an unbalance rate greater than 25%.

[0019] Preferably, in the equivalent length measurement model construction step, the LightGBM algorithm is used for modeling and prediction, the selected features are standardized, the data are divided into training set, validation set and test set, the k-fold validation and early stopping method are used to optimize the model parameters, and the grid search method is used to determine the optimal hyperparameter configuration.

[0020] Preferably, in the model validation and application steps, the accuracy and reliability of the model are evaluated using evaluation indicators such as mean absolute error (MAE), mean square error (MSE), and coefficient of determination (R2), where: 2

[0021]

[0022] Among them, n is the number of test samples, y i is the actual value, is the predicted value, is the average value of the actual value; analyze the cause of the error and optimize the model based on the evaluation results.

[0023] Preferably, if the model error is caused by unreasonable feature selection, re-evaluate the feature importance, add, delete or combine transformation features; if it is due to improper hyperparameter settings, re-search the hyperparameters; if it is due to low quality of training data, further clean and preprocess the data or increase the amount of training data.

[0024] Preferably, when the optimized model is applied to the line loss management of the substation, the operation data is monitored in real time, the equivalent length of the cable and the line loss are calculated, an automatic warning is issued when an abnormality occurs, and loss reduction measures are formulated according to the calculation results.

[0025] Preferably, in terms of power grid planning, the optimized model is used to predict the equivalent length of cables and line losses under different planning schemes, evaluate the feasibility and economy of the schemes, and select the optimal planning scheme.

[0026] The present invention provides a method for calculating the equivalent length of cables in a substation area by combining topology and line loss constraints. It has the following beneficial effects:

[0027] (1) This method for calculating the equivalent length of cables in substations, which combines topology and line loss constraints, uses RPA technology for data aggregation and preprocessing. It can automatically capture, parse and integrate multi-source data, realize cross-system data integration, greatly improve data collection efficiency, and reduce manual burden. At the same time, it cleans, denoises and fills missing values ​​on the data to ensure the accuracy and completeness of the data, provide high-quality data support for subsequent analysis and model training, avoid measurement errors caused by data problems, make the measurement results more reliable, and effectively improve the overall efficiency and quality of data processing.

[0028] (2) This method for calculating the equivalent length of cables in a substation area, which combines topology and line loss constraints, uses graph theory to construct a substation grid topology model, and combines depth-first or breadth-first search algorithms to obtain line topology information, which can accurately grasp the connection relationship of grid components; uses the average current method to calculate the theoretical line loss, and combines historical data to determine the line loss constraint conditions, comprehensively considering the effects of resistance, magnetic field and various influencing factors, and making the analysis of grid topology and line loss more accurate. It provides a solid theoretical and data foundation for the calculation of equivalent cable length, helps to gain a deeper understanding of the grid operation status, and provides a strong basis for loss reduction and planning.

[0029] (3) This method combines topology and line loss constraints to calculate the equivalent length of cables in the substation area. It uses the LightGBM algorithm to build a model. It combines standardization, k-fold validation, early stopping, and grid search techniques to optimize model parameters and improve the generalization ability and prediction accuracy of the model. The model is evaluated using indicators such as mean absolute error, mean square error, and determination coefficient. The model is optimized based on the results to further improve the accuracy of the model. The accurate model can more accurately calculate the equivalent length of cables, providing reliable support for substation line loss management and power grid planning.

[0030] (4) This method combines topology and line loss constraints to calculate the equivalent length of cables in the substation area. The optimized model is applied to the line loss management of the substation area, and the operation data is monitored in real time. Anomalies are detected and early warnings are issued in a timely manner. Loss reduction measures are formulated based on the measurement results, effectively reducing line losses and improving the operating efficiency and economic benefits of the power system. In terms of power grid planning, by predicting the equivalent length and line loss of cables under different schemes, the feasibility and economic efficiency of the schemes are evaluated, and the optimal planning scheme is selected to achieve the rational allocation of power grid resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0032] Figure 1 It is the overall flow chart of the present invention;

[0033] Figure 2 This is a flow chart of data aggregation and preprocessing of the present invention;

[0034] Figure 3 This is a flow chart of the power grid topology analysis in the substation area of ​​the present invention;

[0035] Figure 4 A flow chart for determining line loss constraint conditions of the present invention;

[0036] Figure 5 Construct a flow chart for the equivalent length calculation model of the present invention;

[0037] Figure 6 This is a flow chart for model verification and application of the present invention.

[0038] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] See also Figures 1-6 The present invention proposes a method for calculating the equivalent length of cables in a substation area by combining topology and line loss constraints, comprising the following steps:

[0041] S1. Data aggregation and preprocessing: Utilize RPA technology to automatically capture, parse, and integrate data from various data sources, and perform preprocessing operations such as cleaning, denoising, and filling missing values ​​on the collected data.

[0042] S2. Substation grid topology analysis: Construct a substation grid topology model, analyze its topological structure characteristics, and determine the electrical connection paths of each line;

[0043] S3. Determine line loss constraints: Based on the distribution network line loss theory, analyze the causes of line losses and determine line loss constraints based on historical line loss data;

[0044] S4. Equivalent length calculation model construction: Combine the topology structure and line loss constraints to build a calculation model, and use machine learning algorithms for modeling and prediction;

[0045] S5. Model verification and application: Use actual substation operation data to verify the model, optimize the model based on the verification results, and apply the optimized model to substation line loss management and power grid planning.

[0046] In step S1 of the present invention, RPA robots are first configured with corresponding RPA robot tasks for different data sources, such as the power company's internal meter data management system, the equipment operation monitoring platform, external meteorological data websites, and the power grid topology database. The RPA robots simulate manual operations, automatically log into each system, and capture data according to preset rules. For example, they obtain real-time data such as power consumption, current, and voltage from the meter data management system; status data such as transformer load factor and oil temperature from the equipment operation monitoring platform; meteorological data that may affect line loss, such as temperature, humidity, and wind speed, from the meteorological data website; and topological data such as line connectivity and equipment location from the power grid topology database. During the data capture process, the RPA robots perform preliminary data cleansing and format conversion. For example, they unify the data's time format and standardize date representation across different systems; and remove invalid characters from the data to ensure data consistency and availability. After the data is captured, it is stored in a unified data warehouse. Professional data processing tools and algorithms are then used to deeply cleanse, denoise, and fill in missing values. For cleansing, data records with obvious errors are removed by setting reasonable data thresholds. For example, when the current value recorded by the meter is detected to be negative or far outside the normal range, it is marked as abnormal data and corrected or deleted. For denoising, data smoothing algorithms, such as the moving average method, are used to process data with large fluctuations so that the data better reflects the actual situation. For filling missing values, an appropriate method is selected based on the characteristics and relevance of the data. If the data has time series characteristics, linear interpolation can be used to estimate missing values ​​based on the data values ​​at adjacent time points. For data with strong correlation with other variables, machine learning-based methods, such as decision tree regression models, are used to predict missing values ​​based on the values ​​of other related variables.

[0047] Furthermore, in step S2 of the present invention, detailed information on all grid components within the substation is collected, including the location, model, and connection relationships of transformers, transmission lines, switches, user terminals, and the like. A topological map is constructed, with the transformer as the key node and the transmission lines connecting the transformer to other components as edges. Each edge is assigned attributes such as line length, resistance, and reactance. These attributes can be obtained from grid design drawings, equipment parameter manuals, or actual measurement data. For example, line length can be measured using a geographic information system (GIS) combined with actual line alignment; resistance and reactance can be determined based on the line's material, specifications, and relevant electrical calculation formulas. After constructing the topological model, the topological map is traversed using a depth-first search (DFS) or breadth-first search (BFS) algorithm. Taking the DFS algorithm as an example, starting from a selected starting node (such as the main transformer node), the map explores nodes as deeply as possible along a path until it is no longer possible to continue or the target node is reached. The map then backtracks to the last unexplored node and continues exploring other paths. During the traversal process, topological information such as the starting point, end point, nodes passed through, and branching of each line is recorded. This information can be used to clearly determine the electrical connection path of each line, providing a basis for subsequent analysis of the line's electrical characteristics and equivalent length calculation.

[0048] Furthermore, in step S3 of the present invention, in terms of resistance, heat is generated when current passes through the line resistance, resulting in power loss, and the magnitude of the loss is proportional to the square of the current, the resistance, and the power-on time. In terms of magnetic field effects, when equipment such as transformers are in operation, the alternating magnetic field in the iron core will induce hysteresis and eddy current effects, resulting in excitation loss. The average current method is used to calculate the theoretical line loss. The distribution network to be calculated is equivalent to a resistance element with a resistance value of R. When the current passing through the element is i, the power loss ΔA of the element during the calculation period is:

[0049]

[0050] Among them, I eff is the root mean square current, which is used to integrate the actual current over time. The ratio K of the root mean square current to the average current is the shape factor. The coefficient value is related to the load rate of the station area. Substituting the shape factor K into the above formula, we get:

[0051]

[0052] Where, AP and AQ are the active and reactive power in the calculation period; U av is the average voltage during the calculation period.

[0053] Collect and organize historical line loss data, analyzing trends in line loss over time, load variations, and other factors. Through statistical analysis, determine a reasonable range for line loss rates. For example, based on data from the past year, the line loss rate for this substation is typically between 3% and 8%, which can be used as a reference standard. Furthermore, set value ranges for relevant factors, such as line current, voltage fluctuation range, and power factor, as line loss constraints. If the current of a particular line exceeds the normal range for a long period of time, this may lead to an abnormal increase in line loss. In this case, this current range can be used as a constraint for subsequent equivalent length measurement and line loss analysis.

[0054] Furthermore, in step S4 of the present invention, 21 parameters are selected as model training features based on the relationship between the line loss in the substation and the power supply, load rate, average voltage, power factor, equipment parameters and grid structure. The average load rate, maximum load rate and the proportion of data with a load rate less than 20% are used to characterize the shape factor K, reflecting the change in the substation load. The substation voltage level is measured by the maximum and minimum values ​​of the transformer outlet voltage. Voltage fluctuations will affect the line loss and thus affect the equivalent length of the cable. The power factor level of the substation is evaluated by the proportion of data with a power factor greater than 0.9. The lower the power factor, the greater the reactive power loss, and the greater the impact on the line loss and equivalent length. Parameters such as transformer capacity, number of electricity meters, number of access points, maximum power supply radius, minimum power supply radius, number of outgoing cables, number of branch nodes, maximum distance between branch nodes and transformers, minimum distance between branch nodes and transformers, total line length, cable line length and overhead line length describe the substation grid structure. These structural features are closely related to line loss and equivalent length. For example, an excessively large power supply radius may increase line resistance, increase line losses, and change the equivalent length accordingly. Considering the impact of neutral current on line losses caused by three-phase load imbalance, the maximum three-phase imbalance rate and the proportion of data with an imbalance rate greater than 25% were introduced to assess the degree of three-phase load imbalance in the substation. The 21 selected features were standardized to make features of different dimensions comparable. The LightGBM algorithm was used for modeling and prediction. The selected features were standardized, and the data was divided into training, validation, and test sets. The k-fold validation and early stopping method were used to optimize the model parameters, and the grid search method was used to determine the optimal hyperparameter configuration.

[0055] Furthermore, in step S5 of the present invention, the accuracy and reliability of the model are evaluated using evaluation indicators such as mean absolute error (MAE), mean square error (MSE), and coefficient of determination (R2), where: 2

[0056]

[0057] Among them, n is the number of test samples, y i is the actual value, is the predicted value, is the average value of the actual value. Analyze the cause of model error based on the evaluation index results. If MAE and MSE are large, R 2 If the value is low, it indicates that the model has a large error. If it is caused by unreasonable feature selection, re-evaluate the importance of the feature. By calculating the correlation coefficient between the feature and the target variable (equivalent cable length), such as the Pearson correlation coefficient, filter out features with strong correlation, delete features with weak correlation or redundant features, or try to combine and transform existing features to create new features. If the hyperparameter setting is improper, re-search the hyperparameters, expand the search range or use more sophisticated search methods, such as random search or Bayesian optimization. If the quality of the training data is not high, further clean and preprocess the data to remove outliers and noise data, or increase the amount of training data to improve the generalization ability of the model.

[0058] The optimized model is applied to substation line loss management. A real-time monitoring system is established, connected to various monitoring devices in the substation, to obtain real-time operational data. The model calculates the equivalent cable length and line loss based on real-time data, sets reasonable thresholds, and automatically issues warnings when the calculated results exceed the thresholds. For example, when the line loss rate of a certain line exceeds the set normal range of 8%, the system issues an alarm. Loss reduction measures are formulated based on the measurement results. For example, for lines with excessive equivalent lengths resulting in high line losses, optimization of the grid layout and shortening of line lengths can be considered. For substations with unbalanced three-phase loads, load distribution is adjusted to achieve a balanced load on the three phases.

[0059] In grid planning, the optimized model predicts equivalent cable lengths and line losses under different planning scenarios. Parameters for different grid planning scenarios, such as the location and capacity of new substations, and the construction and renovation of new lines, are input into the model for simulation and prediction. By comparing line loss indicators and equivalent lengths under different scenarios, the feasibility and economic viability of each scenario are evaluated. The planning solution with the lowest line loss and the lowest combined construction and operating costs is selected to achieve optimal allocation of grid resources and sustainable development.

[0060] By utilizing RPA technology for data aggregation and preprocessing, this invention can automatically capture, parse, and integrate multi-source data, achieve cross-system data integration, significantly improve data collection efficiency, and reduce manual workload. At the same time, it cleans, denoises, and fills missing values ​​in the data to ensure data accuracy and completeness, providing high-quality data support for subsequent analysis and model training, avoiding measurement errors caused by data problems, making measurement results more reliable, and effectively improving the overall efficiency and quality of data processing.

[0061] At the same time, a graph theory method is used to construct a substation power grid topology model, combined with a depth-first or breadth-first search algorithm to obtain line topology information, which can accurately grasp the connection relationship between power grid components. The average current method is used to calculate theoretical line loss, and the line loss constraints are determined in combination with historical data. The comprehensive consideration of resistance, magnetic field effects, and various influencing factors makes the analysis of power grid topology and line loss more accurate, providing a solid theoretical and data foundation for the calculation of equivalent cable length, helping to gain a deeper understanding of power grid operation status and providing a strong basis for loss reduction and planning.

[0062] At the same time, the LightGBM algorithm was used for modeling, combined with standardization, k-fold validation, early stopping, and grid search techniques to optimize model parameters and improve the model's generalization and prediction accuracy. The model was evaluated using metrics such as mean absolute error, mean square error, and coefficient of determination, and the model was optimized based on the results to further improve its accuracy. This precise model can more accurately calculate equivalent cable length, providing reliable support for substation line loss management and power grid planning.

[0063] The above description is only a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural transformation made by using the contents of the present description and drawings under the inventive concept of the present invention, or directly / indirectly applied in other related technical fields, is included in the patent protection scope of the present invention.

Claims

1. A method for calculating the equivalent length of cables in a substation area by combining topology and line loss constraints, characterized in that: The following steps are involved: S1. Data aggregation and preprocessing: Utilize RPA technology to automatically capture, parse, and integrate data from various data sources, and perform preprocessing operations such as cleaning, denoising, and filling missing values ​​on the collected data. S2. Substation grid topology analysis: Construct a substation grid topology model, analyze its topological structure characteristics, and determine the electrical connection paths of each line; S3. Determine line loss constraints: Based on the distribution network line loss theory, analyze the causes of line losses and determine line loss constraints based on historical line loss data; S4. Equivalent length calculation model construction: Combine the topology structure and line loss constraints to build a calculation model, and use machine learning algorithms for modeling and prediction; S5. Model verification and application: Use actual substation operation data to verify the model, optimize the model based on the verification results, and apply the optimized model to substation line loss management and power grid planning.

2. The method for calculating the equivalent cable length in a substation area by combining topology and line loss constraints according to claim 1, characterized in that: In the data aggregation and preprocessing steps, RPA technology is used to seamlessly connect multiple business systems, call applications and various web pages in the desktop system, realize cross-system data integration, and perform preliminary cleaning and format conversion on the captured data.

3. The method for calculating the equivalent cable length in a substation area by combining topology and line loss constraints according to claim 1, characterized in that: In the substation grid topology analysis step, a graph theory method is used to convert the grid topology structure into a graph model, where nodes represent grid components and edges represent connection lines between components. Edges are assigned attributes such as line length, resistance, and reactance, and the line topology information is obtained by traversing the topology graph using a depth-first search or breadth-first search algorithm.

4. The method for calculating the equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 1, characterized in that: In the step of determining the line loss constraint condition, the average current method is used to calculate the theoretical line loss. The distribution network to be calculated is regarded as a resistance element with a resistance value of R. When the current passing through the element is i, the energy loss ΔA of the element during the calculation period is: Among them, I eff is the root mean square current, which is used to integrate the actual current over time. The ratio K of the root mean square current to the average current is the shape factor. The coefficient value is related to the load rate of the station area. Substituting the shape factor K into the above formula, we get: Where, AP and AQ are the active and reactive power in the calculation period; U av The average voltage value during the calculation period is used; at the same time, the variation range and trend of line loss are determined in combination with historical line loss data, and the reasonable range of line loss rate and the value range of related factors are set as constraints.

5. The method for calculating the equivalent cable length in a substation area by combining topology and line loss constraints according to claim 1, characterized in that: In the step of constructing the equivalent length measurement model, 21 parameters are selected as model training features, including average load rate, maximum load rate, data proportion of load rate less than 20%, maximum and minimum values ​​of transformer outlet voltage, data proportion of power factor greater than 0.9, transformer capacity, number of electricity sales meters, number of access points, maximum power supply radius, minimum power supply radius, number of outgoing cables, number of branch nodes, maximum distance between branch node and transformer, minimum distance between branch node and transformer, total line length, cable line length, overhead line length, maximum three-phase unbalance rate and data proportion of unbalance rate greater than 25%.

6. A method for calculating equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 5, characterized in that: In the equivalent length measurement model construction step, the LightGBM algorithm is used for modeling and prediction, the selected features are standardized, the data are divided into training set, validation set and test set, the k-fold validation and early stopping method are used to optimize the model parameters, and the grid search method is used to determine the optimal hyperparameter configuration.

7. The method for calculating equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 1, characterized in that: In the model verification and application steps, the accuracy and reliability of the model are evaluated using evaluation indicators such as mean absolute error (MAE), mean square error (MSE), and coefficient of determination (R2), among which: Among them, n is the number of test samples, y i is the actual value, is the predicted value, is the average value of the actual value; analyze the cause of the error and optimize the model based on the evaluation results.

8. The method for calculating equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 7, characterized in that: If the model error is caused by unreasonable feature selection, re-evaluate the importance of features and add, delete, or combine transformation features; If the hyperparameters are not set correctly, re-search the hyperparameters; If the quality of the training data is not high, further clean and preprocess the data or increase the amount of training data.

9. The method for calculating equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 8, characterized in that: When the optimized model is applied to line loss management in the substation area, the operating data is monitored in real time, the equivalent cable length and line loss are calculated, an automatic warning is issued when an abnormality occurs, and loss reduction measures are formulated based on the measurement results.

10. The method for calculating equivalent cable length in a substation area in combination with topology and line loss constraints according to claim 1, characterized in that: In terms of power grid planning, the optimized model is used to predict the equivalent cable length and line loss under different planning schemes, evaluate the feasibility and economy of the schemes, and select the optimal planning scheme.

Images