Line loss analysis method and device, electronic equipment and storage medium

Through the combination of robust anomaly point detection, clustering and least squares support vector machine algorithms, a line loss factor positioning model is constructed, which solves the problem of factor analysis deviation in the existing technology, and improves the accuracy and efficiency of line loss analysis. Improved user experience.

CN119046716APending Publication Date: 2024-11-29BEIJING CHINA POWER INFORMATION TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410891659.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-04
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The existing linear loss factor analysis method is based on a simplified circuit loss model and cannot fully and objectively reflect the influence characteristics of all factors. The data quality and actual circuit structure depend on high, resulting in a large deviation from the analysis results.

Method used

Using robust anomaly point detection algorithm, clustering algorithm and least squares support vector machine algorithm, a factor positioning model for line loss anomalies is constructed, and the model training function is fitted through the filtering sample data correlation. Whether the comprehensive factors exceed the threshold, positioning and sorting The influence of different data on line loss.

Benefits of technology

It realizes the characteristics of more objectively reflecting the impact of all factors on line loss, improves the accuracy and efficiency of analysis, provides more accurate quantitative directions, and improves user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119046716A_ABST
    Figure CN119046716A_ABST
Patent Text Reader

Abstract

The invention provides a line loss analysis method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining at least one group of basic data of a power grid, and determining at least one line loss rate corresponding to the at least one group of basic data; performing abnormal line loss rate screening on the at least one line loss rate by using a robust abnormal point detection algorithm, taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, and otherwise, taking the basic data as sample data; performing clustering and grouping processing on the sample data to obtain at least one classification group, determining any specific data of the basic data corresponding to any classification group as a first factor, and determining whether the first factor exceeds a set threshold; responding to the condition that the first factor does not exceed a set threshold value, taking the line loss rate corresponding to any classification group as a second factor, and determining an analysis formula by utilizing a least square support vector machine algorithm; and substituting specific data corresponding to the abnormal data into an analysis formula, and determining and outputting the influence of different specific data on the line loss.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of power grid operation management, and in particular to a line loss analysis method, device, electronic equipment and storage medium. Background Art

[0002] With the rapid development of the world economy and industry, the power consumption of enterprises is constantly increasing. While the power supply department supplies electricity to enterprises, there is a large amount of power loss. From a macro perspective, the actual line loss can be divided into two categories: technical line loss and management line loss. Technical line loss is usually considered to be the power loss of normal lines, which is inevitable; while management line loss is man-made power loss caused by poor equipment and management, and due to inadequate management, the power loss formed accounts for a large proportion of the entire line loss.

[0003] Traditional line loss influencing factor analysis methods are mainly based on the equivalent resistance method and power flow algorithm, etc., through simplified circuit loss models, formula derivation is used to qualitatively obtain the influencing characteristics of each factor. Although this method can obtain strong theoretical support, during the construction of the model, on the one hand, certain simplified conditions are added, which cannot fully and objectively reflect the influencing characteristics of all factors; on the other hand, it is highly dependent on data quality and actual circuit structure, and the influencing characteristics derived based on the mechanism are often greatly deviated from the actual ones. Summary of the invention

[0004] In view of this, the present application proposes a line loss analysis method, device, electronic device and storage medium to solve or partially solve the above problems.

[0005] Based on the above purpose, the present application provides a line loss analysis method, including:

[0006] Acquire at least one set of basic data of the power grid, and respectively determine at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a plurality of specific data;

[0007] Using a robust outlier detection algorithm to screen the at least one line loss rate for abnormal line loss rates, and taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data;

[0008] Performing clustering processing on the sample data to obtain at least one classification group, determining any specific data of the basic data corresponding to any classification group as a first factor, and determining whether the first factor exceeds a set threshold;

[0009] In response to the first factor not exceeding the set threshold, taking the line loss rate corresponding to any one of the classification groups as the second factor, and determining an analysis formula by using the first factor and the second factor using a least squares support vector machine algorithm;

[0010] The specific data corresponding to the first factor in the abnormal data is brought into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

[0011] In some exemplary embodiments, the screening of abnormal line loss rates for the at least one line loss rate using a robust outlier detection algorithm includes:

[0012] Fitting the at least one line loss rate into a Gaussian mixture model using an expectation maximization algorithm, and determining a first parameter according to means and variances of different Gaussian components of the Gaussian mixture model;

[0013] The first parameter is screened for errors according to the Laida criterion, and the error exceeding the limit is taken as the abnormal line loss rate.

[0014] In some exemplary embodiments, fitting the at least one line loss rate into a Gaussian mixture model using an expectation maximization algorithm includes:

[0015] The at least one line loss rate is centered and standardized, and the processed result is fitted into the Gaussian mixture model.

[0016] In some exemplary embodiments, the clustering process of the sample data includes:

[0017] The K-means clustering algorithm is used to cluster the sample data in the spatial dimension.

[0018] In some exemplary embodiments, after determining whether the first factor exceeds a set threshold, the method further includes:

[0019] In response to the first factor exceeding a set threshold, the first factor is output as an abnormal factor with the highest priority.

[0020] In some exemplary embodiments, determining the analysis formula using the first factor and the second factor by using a least squares support vector machine algorithm includes:

[0021] Using the first factor and the second factor as training data, and fitting the training model function using the correlation relationship of the least squares support vector machine algorithm;

[0022] The trained model function is differentiated to obtain the analytical formula.

[0023] In some exemplary embodiments, the specific data includes at least one of power supply, power sales, substation name, transformer equipment classification, transformer status, transformer model, usage nature, operation and maintenance unit, maintenance team, voltage level, rated capacity, short-circuit loss, no-load loss, low-voltage side rated current, number of households, power supply radius, line length and load rate;

[0024] The line loss rate is calculated based on the power supply and the power sales amount.

[0025] Based on the same concept, the present application also provides a line loss analysis device, including:

[0026] An acquisition module, used to acquire at least one set of basic data of the power grid, and respectively determine at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a variety of specific data;

[0027] A screening module, used for screening the at least one line loss rate for abnormal line loss rate by using a robust outlier detection algorithm, and taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data;

[0028] A clustering module, used to perform clustering processing on the sample data to obtain at least one classification group, determine any specific data of the basic data corresponding to any classification group as a first factor, and determine whether the first factor exceeds a set threshold;

[0029] a determination module, configured to, in response to the first factor not exceeding a set threshold, use the line loss rate corresponding to any one of the classification groups as a second factor, and determine an analysis formula using a least squares support vector machine algorithm through the first factor and the second factor;

[0030] The output module is used to bring the specific data corresponding to the first factor in the abnormal data into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

[0031] Based on the same concept, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the methods described above when executing the program.

[0032] Based on the same concept, the present application also provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement any of the methods described above.

[0033] As can be seen from the above, the present application provides a line loss analysis method, device, electronic device and storage medium. The present application comprehensively uses a robust outlier detection algorithm, a clustering algorithm and a least squares support vector machine algorithm to construct a factor location model for line loss anomalies. The model training function is fitted by associating the screened sample data, and the judgment of whether the comprehensive factors exceed the threshold is used to locate and sort the influence of different data on line loss, providing a more accurate quantitative guidance basis for line loss anomaly management. By applying the solution provided by the present application, the influence characteristics of all factors on line loss can be more objectively reflected, and the accuracy is greatly improved compared with the original method. Ultimately, it can provide more accurate quantitative direction of line loss influencing factors, improve the efficiency of line loss impact confirmation, and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the related technologies, the drawings required for use in the embodiments or the related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 A flow chart of an exemplary method provided in an embodiment of the present application is shown.

[0036] Figure 2 A schematic diagram of the process of determining the Gaussian mixture density of the exemplary method provided in an embodiment of the present application is shown.

[0037] Figure 3 A schematic diagram showing the results of the K-means clustering algorithm of the exemplary method provided in an embodiment of the present application is shown.

[0038] Figure 4 A schematic diagram of the implementation flow of the exemplary method provided in an embodiment of the present application in a specific scenario is shown.

[0039] Figure 5 A schematic diagram of the structure of an exemplary device provided in an embodiment of the present application is shown.

[0040] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of this specification more clear, this specification is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0042] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements, objects or method steps appearing before the word cover the elements, objects or method steps listed after the word and their equivalents, without excluding other elements, objects or method steps. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0043] As mentioned in the background technology section, the analysis and research on line loss is mainly about the management line loss of power supply enterprises. Line loss management is an important measure for power grid companies to plug leaks, increase revenue, reduce costs and increase efficiency, and is also an important yardstick for evaluating their management level. At present, the energy-saving development and planning evaluation of low-voltage substations are inseparable from the guidance of theoretical line losses. How to calculate the theoretical line loss of the target substation and analyze the effect of various influencing factors on this basis have always been breakthroughs in substation loss reduction research.

[0044] Line power loss is an important factor affecting the economic operation of the power system. In order to reduce line loss, it is necessary to count the actual line loss and compare the statistical line loss with the theoretical line loss in order to conduct line loss analysis and strengthen the management of line loss. The power loss of the distribution network is not only related to the structure and load properties of the power grid, but also to the management level of the enterprise. In the line loss management of the power supply station, the management of line loss is an important factor affecting the statistical line loss. Since this kind of loss has no rules to follow and is not easy to calculate, it is usually called unknown loss. The power loss caused by inadequate management accounts for a large proportion of the actual line loss. In some places, some links are even quite serious.

[0045] Through line loss analysis, we can gain in-depth understanding of the causes, properties, proportions of each component and other factors of the unit's line loss, identify the main factors affecting the loss, and take corresponding measures in a targeted manner to achieve greater loss reduction effects and economic benefits with less investment.

[0046] At present, there are discussions on the main factors affecting line loss, mainly including the following methods: (1) Expert analysis method. Expert analysis method is a method based on subjective judgment. Its advantage is that when the data is incomplete, it can make judgments with the help of expert experience. Its disadvantage is that it is difficult to standardize and quantify, it is difficult to compare the credit status of multiple customers horizontally and vertically, and it is impossible to evaluate and effectively monitor expert decisions. In particular, over-reliance on expert experience will result in high labor costs. If there is a lack of experienced experts, it will bring great difficulties to the credit evaluation work. (2) Decision tree model. The decision tree model method is a method for classifying samples. The space is divided according to a simple classification rule, and only one feature of the sample is separated at a time. Therefore, the final result of the space division is non-overlapping units. (3) Power forward and backward method. The power forward and backward method is to define the nodes of each link on the trunk line of the distribution network, determine the number of the node closest to the power source one by one, and determine the order of forward and backward through the hierarchical numbering of the distribution network, so as to solve the power loss value of the distribution network. (4) Association analysis, also known as association mining, is to find frequent patterns, associations, correlations or causal structures between sets of items or sets of objects in transaction data, relational data or other information carriers. In other words, association analysis is to discover the connection between different commodities (items) in the transaction database. Association analysis is a simple and practical analysis technique that discovers the associations or correlations in a large number of data sets, thereby describing the laws and patterns of the simultaneous occurrence of certain attributes in a thing.

[0047] However, in practice, it is found that the above traditional theoretical line loss influencing factor analysis method mainly uses simplified circuit loss models and formula derivation to qualitatively obtain the influencing characteristics of each factor. Although this method can obtain strong theoretical support, during the construction of the model, on the one hand, certain simplified conditions are added, which cannot fully and objectively reflect the influencing characteristics of all factors; on the other hand, it is highly dependent on data quality and actual circuit structure, and the influencing characteristics derived based on the mechanism often deviate greatly from the actual ones.

[0048] In view of the above-mentioned actual situation, an embodiment of the present application provides a line loss analysis method. The present application comprehensively uses a robust outlier detection algorithm, a clustering algorithm, and a least squares support vector machine algorithm to construct a factor location model for line loss anomalies. The model training function is fitted using the selected sample data association, and the judgment of whether the comprehensive factors exceed the threshold is used to locate and sort the influence of different data on line loss, providing a more accurate quantitative guidance basis for line loss anomaly management. By applying the solution provided by the present application, the influence characteristics of all factors on line loss can be more objectively reflected, and the accuracy is greatly improved compared with the original method. Ultimately, it can provide a more accurate quantitative direction of line loss influencing factors, improve the efficiency of line loss impact confirmation, and improve user experience.

[0049] Figure 1 A flow chart of an exemplary method provided in an embodiment of the present application is shown.

[0050] like Figure 1 As shown, the line loss analysis method exemplarily proposed in the embodiment of the present application specifically includes the following steps.

[0051] Step 102, obtaining at least one set of basic data of the power grid, and determining at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a variety of specific data.

[0052] In this step, the main purpose is to collect the corresponding basic data from the electricity information collection system and user files, and calculate the line loss rate index. Among them, based on the electricity information collection system, the electricity customer's electricity consumption history data can be collected, including: power supply, daily electricity sales, etc.; based on the marketing business application system, the electricity customer file can be collected, which may include the name of the substation, transformer equipment classification, transformer status, transformer model, usage nature, operation and maintenance unit, maintenance team, voltage level, rated capacity, short-circuit loss, no-load loss, low-voltage side rated current, number of households, power supply radius, line length, load rate, etc.

[0053] First of all, line loss is the line loss caused by power supply companies in the transmission, distribution, and management of electric energy, including power loss in the power grid, power loss in electrical equipment, and unknown damage. Theoretically, only part of the line loss can be accurately calculated by formulas and measuring instruments. Therefore, the total line loss can only be obtained by subtracting the total power sales from the total power supply measured by the meter. Its accuracy depends on the accuracy of the power metering system that measures the power supply and sales, as well as the scientific and reasonable recording and statistical system for the user's power sales. The so-called power supply refers to the amount of electricity supplied by a power plant, power supply area or power grid to a user, including line loss in the process of transmitting and distributing electric energy. The calculation formula is:

[0054] Ag =A f -A y -A ch +A r

[0055] Among them, A g A is the total power supply in the power supply area or power grid, f A is the power generation of power plants in this region or grid, y A is the power consumption of the power plant, ch A is the amount of electricity exported to other power grids. r The amount of electricity sold refers to the amount of electricity sold by power companies to users and the amount of electricity supplied by power companies to their own non-power production departments. Because the non-power production departments of the company should be regarded as users. Therefore, the amount of electricity sold in the power supply area or power grid is equal to the sum of the electricity meters of users.

[0056] Line loss rate: The percentage of line loss in power supply is called line loss rate, or line loss rate for short. The calculation formula is: Line loss rate = (power supply - power sales) / power supply * 100%. Line loss rate is a comprehensive technical indicator of the power system and an important economic indicator for assessing the power sector. In the operation and management of the power grid, the line loss obtained by subtracting the total power sales from the total power supply is called statistical line loss, and the corresponding line loss rate is called statistical line loss rate. Among the statistical line loss, a part is unavoidable in the process of transmitting and distributing electric energy, which is determined by the load conditions of the power grid and the parameters of the power supply equipment at that time. This part of the loss is called technical loss, which can be obtained through theoretical calculation, so it is also called theoretical line loss, and the corresponding line loss rate is called theoretical line loss rate. Among the statistical line loss, another part of the loss is unknown loss, also called management loss. This part of the loss can and should be avoided or reduced by taking necessary measures.

[0057] In this embodiment, each set of basic data includes the corresponding specific data collected from the power consumption information collection system and user files, among which the power supply and power sales are generally used to calculate the line loss rate, and other specific data can be used as various factors affecting the line loss to make the final influence judgment. Multiple sets of basic data can correspond to different time periods, such as days, one set of basic data per day, or regions or areas, etc.

[0058] Step 104, using a robust outlier detection algorithm to screen the at least one line loss rate for abnormal line loss rates, and taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data.

[0059] In this step, the statistical line loss value is mainly used as the analysis indicator, and the robust outlier point detection algorithm is used to screen the abnormal line loss rate. The basic data corresponding to the identified line loss rate anomaly is used as abnormal data, and the basic data corresponding to the unidentified line loss rate anomaly is used as sample data.

[0060] The traditional method of determining abnormal line loss mainly relies on the experience of business personnel. Based on the proficient business knowledge, the power consumption information collection system is used to realize the real-time monitoring of line loss data. The power consumption information collection system is the previous automatic meter reading system. It is a system that uses modern digital communication technology, computer software and hardware and other high-tech technologies to wirelessly transmit the power consumption data of field equipment and classify and collect and summarize it. The application of the power consumption information collection system also provides power supply companies, especially grassroots companies, with more advanced technical means to manage important indicators. The management mode has gradually changed from the "extensive" manual management to the "information-based and intelligent" management mode.

[0061] Among them, the application of line loss management in power supply companies is the most important example. Line loss indicators involve various links of power supply companies, such as power transmission, transformation, distribution, and power consumption. Loss reduction technology and line loss management are the main manifestations of the comprehensive level of power supply companies in power grid planning, production regulation, marketing and operation. However, in grassroots enterprises, a large amount of data can only be summarized and formed into reports within a cycle before the problems can be analyzed. The real-time performance is poor, and the line loss indicator control work is in a passive state.

[0062] In this embodiment, the data mining and analysis method of robust outlier detection is used for data preprocessing, and the data collected by the power consumption information system is screened to identify the data with abnormal line loss rate of lines, substations, and large users. The data with abnormal line loss rate is regarded as abnormal data, and the data with unrecognized abnormal line loss rate is regarded as sample data. This algorithm realizes the real-time, comprehensive and accurate data collection.

[0063] For the robust outlier detection algorithm, it first uses the Gaussian Mixture Model (GMM), which can be regarded as a combination of multiple Gaussian distributions with different weights. A GMM is represented as follows: It means that the above GMM is composed of M Gaussian components. denote the weight, mean and covariance of the jth Gaussian component respectively. Furthermore, a D-dimensional feature vector The Gaussian mixture density is Figure 2 As shown. Furthermore, the D-dimensional feature vector The likelihood probability is:

[0064]

[0065] here, Represents the probability of the j-th Gaussian component, which can be specifically expressed as:

[0066]

[0067] Afterwards, in this embodiment, based on the simulation capability of GMM for any data, GMM can be used to fit the statistical line loss rate, and then the 3σ criterion can be used to detect abnormal points. The 3σ criterion, also known as the three sigma criterion or the Laida criterion, is a rule commonly used in statistics to measure the variability of data and identify abnormal values. This criterion is based on the properties of the normal distribution (or approximate normal distribution) and provides a standard to measure the discreteness of a data set.

[0068] The 3σ criterion is based on repeated measurements of equal precision under normal distribution, and the interference or noise that causes singular data is difficult to satisfy the normal distribution. If the absolute value of the residual error of a measurement value in a set of measurement data is greater than ±3σ, the measurement value is a bad value and should be eliminated. Taking the error equal to ±3σ as the limit error, for the random error of the normal distribution, the probability of falling outside ±3σ is only 0.27%, and the probability of it occurring in a finite number of measurements is very small, that is, it can be eliminated. In specific application scenarios, the value of σ can be set according to the specific scenario, and it can be set to 1 for general scenarios.

[0069] In some embodiments, before screening, the line loss rate may be preprocessed to avoid the influence of abnormal data on the detection. The specific preprocessing process may be to centralize and standardize the line loss rate, and the method is:

[0070]

[0071] Among them, x i and i are the statistical line loss rates before and after centralization and standardization, respectively. m The line loss rate {x i}, s m For {x i}The median of the absolute values ​​of the median deviations, calculated as: s m =umedian{|x i -x m |}, where u is a preset constant, which can be 1.4826, in order to make s m becomes an unbiased estimator of the standard deviation of normally distributed data.

[0072] After that, the expectation-maximization (EM) algorithm can be used to convert the centralized and standardized statistical line loss rate {y i} Fit it into GMM to obtain a Gaussian mixture model containing M Gaussian components

[0073] Recalculate where μ j and j are the mean and variance of the jth Gaussian component in the Gaussian mixture model. Make |d i |≥3, then x i is an outlier point; and if any Gaussian component, |d i | are all less than 3, then x i Here, σ is set to 1 in the 3σ criterion for judgment.

[0074] That is, in some embodiments, the use of a robust outlier detection algorithm to screen the at least one line loss rate for abnormal line loss rates includes: using an expectation maximization algorithm to fit the at least one line loss rate into a Gaussian mixture model, and determining a first parameter according to the mean and variance of different Gaussian components of the Gaussian mixture model; and performing error screening on the first parameter according to the Laida criterion, and taking the error exceeding the limit as the abnormal line loss rate. The first parameter corresponds to the aforementioned d i .

[0075] Afterwards, in some embodiments, fitting the at least one line loss rate into a Gaussian mixture model using an expectation-maximization algorithm includes: performing centering and standardization processing on the at least one line loss rate, and fitting the processed result into the Gaussian mixture model.

[0076] Step 106, clustering the sample data to obtain at least one classification group, determining any specific data of the basic data corresponding to any classification group as a first factor, and determining whether the first factor exceeds a set threshold.

[0077] In this step, the sample data is first clustered, and clustering can be mainly performed using clustering algorithms, such as K-means clustering algorithm (K-MEANS), K-center point clustering algorithm (K-MEDOIDS), Clara algorithm, Clarans algorithm, etc. After the clustering is completed, it can be judged whether all the basic data corresponding to any classification group exceeds the set threshold. Here, it is mainly aimed at judging other specific data in the basic data except for the calculation of line loss rate. For example, if a classification group includes 10 groups of sample data, the specific data corresponding to these 10 groups of sample data are judged respectively. The specific process of the judgment can be to extract some specific data of these 10 groups of sample data (such as 10 rated capacity data, etc.) for judgment. Afterwards, a threshold is set as the basis for judgment, which can be determined according to the user's pre-settings. For example, in a specific scenario, it can be set according to the requirements in the substation operation management specifications. That is, any specific data of the basic data corresponding to the classification group is used as the first factor to determine whether the first factor exceeds the set threshold. The first factor here is a certain specific data (such as rated capacity data, etc.).

[0078] In some embodiments, in order to make each cluster as compact as possible and to separate each cluster as much as possible, a K-means clustering algorithm may be used for clustering. That is, in some embodiments, the clustering of the sample data includes: clustering the sample data in the spatial dimension using a K-means clustering algorithm.

[0079] The K-means clustering algorithm first randomly selects K objects as the initial cluster centers. Then the distance between each object and each seed cluster center is calculated, and each object is assigned to the cluster center closest to it. The cluster centers and the objects assigned to them represent a cluster. Once all objects are assigned, the cluster center of each cluster will be recalculated based on the existing objects in the cluster. This process will be repeated until a certain termination condition is met. The termination condition can be that no (or a minimum number) of objects are reassigned to different clusters, no (or a minimum number) of cluster centers change again, and the sum of squared errors is locally minimized. Finally, the clustering results using the K-means clustering algorithm are as follows: Figure 3 As shown, x n These are the targets that need to be clustered one by one.

[0080] In a specific scenario, clustering is a process of classifying and organizing data members in a data set that are similar in some aspects. Clustering is a technique for discovering this internal structure. Clustering technology is often referred to as unsupervised learning. The K-means clustering algorithm is the most famous partitioning clustering algorithm. Due to its simplicity and efficiency, it has become the most widely used of all clustering algorithms. Given a set of data points and the required number of clusters k, k is specified by the user. The K-means clustering algorithm repeatedly divides the data into k clusters according to a certain distance function. The specific K-means clustering algorithm first randomly selects K objects as the initial cluster centers. Then calculate the distance between each object and each seed cluster center, and assign each object to the cluster center closest to it. The cluster centers and the objects assigned to them represent a cluster. Once all objects are assigned, the cluster center of each cluster will be recalculated based on the existing objects in the cluster. This process will be repeated until a certain termination condition is met. The termination condition can be any of the following: (1) no (or a minimum number) of objects are reassigned to different clusters; (2) no (or a minimum number) of cluster centers change again; (3) the sum of squared errors is locally minimized, etc.

[0081] Afterwards, in some embodiments, for each classification group, the sample data indicators are divided into influencing factor indicators (i.e., the first factor, other specific data except for the calculation of line loss rate) and statistical line loss indicators (i.e., the second factor, line loss rate). Thus, according to the foregoing, the influencing factor indicators may include: substation name, transformer equipment classification, transformer status, transformer model, usage nature, operation and maintenance unit, maintenance team, voltage level, rated capacity, short-circuit loss, no-load loss, low-voltage side rated current, number of households, power supply radius, line length, load rate, etc. The statistical line loss indicator is the line loss rate calculated using the power supply and power sales. The judgment of whether the first factor exceeds the set threshold is to judge whether the influencing factor indicator exceeds the provisions of the substation operation management specifications.

[0082] Finally, for the judgment result, if the first factor is judged to be within the set threshold, then continue to execute step 108. If the first factor is judged to be outside the set threshold, it means that the first factor has obviously affected the line loss and is a more obvious factor causing the line loss. Its influence and priority should be relatively high. Then, for such a first factor, the subsequent steps can be directly omitted, and it can be directly used as a factor with a higher or highest priority to determine the influence and locate it. That is, in some embodiments, after determining whether the first factor exceeds the set threshold, the method also includes: in response to the first factor exceeding the set threshold, outputting the first factor as an abnormal factor with the highest priority.

[0083] Here we first explain the role of influence ranking. According to the influence ranking, we can give priority to the factors with greater influence in the abnormal management of line loss. The first factor outside the set threshold can be regarded as the most concerned factor, so as to focus on and adjust it in line loss management.

[0084] Step 108, in response to the first factor not exceeding the set threshold, the line loss rate corresponding to any one of the classification groups is used as the second factor, and the analysis formula is determined by using the least squares support vector machine algorithm through the first factor and the second factor.

[0085] In this step, the first factor judged to be reasonable (within the set threshold) in step 106 can be used as an influencing factor index set, that is, X i ={x1,x2,…,x n}, where {x1,x2,…,x n} represents the data corresponding to the first factor in different groups of basic data. For example, if a classification group has 10 groups of basic data, then here is the rated capacity data in the 10 groups of basic data. After that, the line loss rate corresponding to the corresponding classification group can be used as the second factor z i For example, if a classification group has 10 sets of basic data, then these 10 sets of basic data correspond to 10 line loss rates. These two factors can be used as training data, and the least squares support vector machine algorithm can be used for fitting, and finally an analysis formula is generated.

[0086] Among them, the least squares support vector machine algorithm (LSSVM) is a variant of the traditional support vector machine (SVM). It changes the inequality constraints of the slack variables in the support vector machine into equality constraints, so that the LSSVM can be solved by solving a system of linear equations. The traditional SVM improves its generalization ability through the principle of structural risk minimization, and can better solve practical problems such as nonlinearity and high dimensions. As an improvement of the support vector machine (SVM), the least squares support vector machine (LSSVM) transforms the quadratic programming problem of the SVM into solving linear equations by selecting different loss functions, which speeds up the convergence speed and reduces the difficulty of solving.

[0087] Specifically, given a sample P = {(X1,z1),(X2,z2),…,(X p ,z p )}, where X i ={x1,x2,…,x n}∈R nis the model input, and n is the number of features. In order to accurately characterize the influence of theoretical line loss, it is assumed that the number of households, power supply radius, line length, load rate and other factors are selected and normalized as model input; i ∈R is the model output, i.e. the normalized line loss in the substation area; p is the number of samples. The mathematical model of LSSVM is

[0088]

[0089] Among them, C is the penalty parameter, C∈R; ε is the weight vector; ξ i is the model error; b is the deviation. Introduce the Lagrangian function: Where: α i is the Lagrange multiplier. According to The optimization conditions are

[0090] By defining the kernel function The optimization problem is transformed into solving a system of linear equations, namely

[0091]

[0092] Finally, the regression coefficient α is solved based on the least squares method i And deviation b, we can get the final influencing factor mapping relationship model:

[0093]

[0094] After that, the kernel function can use a polynomial kernel function:

[0095] K(X i ,X j )=(γX T X j +r) d , γ>0, where: d is the order of the kernel function; γ is the width of the kernel function; r is the bias coefficient.

[0096] Based on the mapping relationship f(X), through the first-order Taylor expansion, as shown in the following formula:

[0097]

[0098] Right now: For any factor x j Its impact is: By keeping the other factors unchanged (i.e. the corresponding Δx = 0, and fixing the step length Δx j Adjustment factor x j , send the factor sequence before and after adjustment to z i =f(X i ) to calculate the corresponding zj and z′ j Finally, by fitting Δz j With x j The correlation relationship between the factors is obtained, and the influence f′(x j ).

[0099] That is, for the first factor judged to be reasonable (within the set threshold) in step 106, the correlation relationship fitting training model function z=f(X)=f(x1,x2,…,x n ). Then, for any factor x j Its influence is Therefore, the influencing factor analysis model is Δz=f ′ (x1)Δx1+…+f ′ (x n )Δx n , which is the final analysis formula, that is, the sensitivity calculation formula is z ′ =f ′ (x1,x2,…,x n ), which can be understood as taking the derivative of the trained model function to obtain the final analytical formula z ′ =f ′ (x1,x2,…,x n That is, in some embodiments, the determining of the analysis formula by using the first factor and the second factor and the least squares support vector machine algorithm includes: using the first factor and the second factor as training data, and fitting the training model function using the correlation relationship of the least squares support vector machine algorithm; and deriving the trained model function to obtain the analysis formula.

[0100] Step 110: bring the specific data corresponding to the first factor in the abnormal data into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

[0101] In this step, after obtaining the specific analysis formula in step 108, the specific data corresponding to the abnormal data obtained in step 104 can be brought into the analysis formula to perform sensitivity (influence) calculation. In a specific scenario, the specific data here are also other data except the power supply and power sales for calculating the line loss rate, that is, the influencing factor data.

[0102] Finally, for these first factors that are determined not to exceed the set threshold in step 106, the specific data corresponding to the abnormal data also contains the same type of data, and then these data can be brought into the analysis formula obtained from the same type of data of the sample data, so as to obtain the corresponding z' value, which can express the influence of the corresponding specific data. After that, the influence values ​​corresponding to different specific data (here, those determined by step 106, as for those that have not passed step 106, as introduced above, they will be directly defined as the highest priority influence and directly output) are calculated, and can be sorted to express the influence of different influencing factors, so that in the actual abnormal management work, the factors with greater influence can be given priority, providing a quantitative guidance basis for abnormal management, supporting line loss diagnosis, and assisting substation management. For example, according to the final influence ranking, the factors with greater influence are managed first, focused on, and adjusted quickly and directly, while those with less influence can be managed later.

[0103] In specific application scenarios, with the intelligent and digital development of large-scale power grid operation and management, as well as the continuous maturity of artificial intelligence and big data technologies, big data algorithms are gradually being applied to the theoretical line loss calculation of low-voltage substations. By building a complete influencing factor index system to train the calculation model, the requirements for data quality are relatively low, and it does not depend on the actual physical circuit structure of the substation. The calculation performance is also much higher than other traditional algorithms. Therefore, the influence characteristics obtained based on the analysis of the big data algorithm model are also closer to the objective laws of the actual substation.

[0104] Furthermore, Figure 4 As shown, the main process of an embodiment of the present application during specific execution is as follows: the first step is basic data collection, and the data collected from the power consumption information collection system and the user files are used for index calculation; the second step is to use the statistical line loss value as the analysis index, and adopt the robust abnormal point detection algorithm to screen the abnormal line loss rate, and use the data with abnormal line loss rate as abnormal data, and the data with unrecognized abnormal line loss rate as sample data; the third step is to classify and group the sample data index in the spatial dimension using the K-means clustering algorithm, and for different classification groups, for each classification group, the data except the power supply and power sales for line loss rate calculation are the influencing factor index (i.e., the first factor), and the line loss rate is the statistical line loss index (i.e., the second factor); the fourth step is to identify and judge the influencing factor index in combination with the substation operation management specification to determine whether the influencing factor index is reasonable; the fifth step is to, if reasonable, the reasonable influencing factor index X i ={x1, x2, ..., x n} and the corresponding statistical line loss index z iAs training data, the least squares support vector machine algorithm is used to fit the correlation relationship of the training model function z = f (X) = f (x1, x2, ..., x n ); Step 6: For any factor x j Its influence is Therefore, the influencing factor analysis model is Δz=f′(x1)Δx1+…+f′(x n )Δx n , that is, the sensitivity calculation formula is z′=f′(x1,x2,…,x n ); Step 7: Substitute the factor sequence data of the abnormal data into the sensitivity calculation formula z′=f′(x1,x2,…,x n ), calculate the actual influence of each factor, locate and sort the abnormal factors; in the eighth step, if it is judged to be unreasonable in the fourth step, the abnormal factors can be directly located. Combined with the abnormal factor location results of the seventh and eighth steps, in the actual abnormal management work, priority can be given to the factors with greater influence, which provides a quantitative guidance basis for abnormal management, supports line loss diagnosis, and assists in substation management.

[0105] It can be seen from the above embodiments that a line loss analysis method provided by the embodiments of the present application includes: obtaining at least one set of basic data of the power grid, and determining at least one line loss rate corresponding to at least one set of basic data; wherein any set of basic data includes multiple specific data; using a robust outlier detection algorithm to screen at least one line loss rate for abnormal line loss rate, and using the basic data corresponding to at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data; clustering and grouping the sample data to obtain at least one classification group, determining any specific data of the basic data corresponding to any classification group as the first factor, and determining whether the first factor exceeds the set threshold; in response to the first factor not exceeding the set threshold, using the line loss rate corresponding to any classification group as the second factor, and determining the analysis formula using the least squares support vector machine algorithm through the first factor and the second factor; bringing the specific data corresponding to the first factor in the abnormal data into the analysis formula, so as to determine and output the influence of different specific data on the line loss according to the calculation results. The present application comprehensively uses the robust outlier detection algorithm, the clustering algorithm and the least squares support vector machine algorithm to construct a factor positioning model for line loss anomalies. The model training function is fitted by using the selected sample data association, and the judgment of whether the comprehensive factors exceed the threshold is used to locate and sort the influence of different data on line loss, providing more accurate quantitative guidance for abnormal line loss management. By applying the solution provided by this application, it is possible to more objectively reflect the influence characteristics of all factors on line loss, and compared with the original method, it is also greatly improved in accuracy. Ultimately, it can provide more accurate quantitative direction of line loss influencing factors, improve the efficiency of line loss impact confirmation, and improve user experience.

[0106] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of the embodiment of the present application can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0107] It should be noted that the above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0108] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a line loss analysis device.

[0109] refer to Figure 5 , the line loss analysis device comprises:

[0110] The acquisition module 210 is used to acquire at least one set of basic data of the power grid, and respectively determine at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a variety of specific data.

[0111] The screening module 220 is used to screen the at least one line loss rate for abnormal line loss rate using a robust outlier detection algorithm, and use the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise use it as sample data.

[0112] The clustering module 230 is used to perform clustering processing on the sample data to obtain at least one classification group, determine any specific data of the basic data corresponding to any classification group as a first factor, and determine whether the first factor exceeds a set threshold.

[0113] The determination module 240 is used to determine the analysis formula using the least squares support vector machine algorithm through the first factor and the second factor, in response to the first factor not exceeding the set threshold, using the line loss rate corresponding to any classification group as the second factor.

[0114] The output module 250 is used to bring the specific data corresponding to the first factor in the abnormal data into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

[0115] In some exemplary embodiments, the screening module 220 is further configured to:

[0116] Fitting the at least one line loss rate into a Gaussian mixture model using an expectation maximization algorithm, and determining a first parameter according to means and variances of different Gaussian components of the Gaussian mixture model;

[0117] The first parameter is screened for errors according to the Laida criterion, and the error exceeding the limit is taken as the abnormal line loss rate.

[0118] In some exemplary embodiments, the screening module 220 is further configured to:

[0119] The at least one line loss rate is centered and standardized, and the processed result is fitted into the Gaussian mixture model.

[0120] In some exemplary embodiments, the clustering module 230 is further configured to:

[0121] The K-means clustering algorithm is used to cluster the sample data in the spatial dimension.

[0122] In some exemplary embodiments, the clustering module 230 is further configured to:

[0123] In response to the first factor exceeding a set threshold, the first factor is output as an abnormal factor with the highest priority.

[0124] In some exemplary embodiments, the determining module 240 is further configured to:

[0125] Using the first factor and the second factor as training data, and fitting the training model function using the correlation relationship of the least squares support vector machine algorithm;

[0126] The trained model function is differentiated to obtain the analytical formula.

[0127] In some exemplary embodiments, the specific data includes at least one of power supply, power sales, substation name, transformer equipment classification, transformer status, transformer model, usage nature, operation and maintenance unit, maintenance team, voltage level, rated capacity, short-circuit loss, no-load loss, low-voltage side rated current, number of households, power supply radius, line length and load rate;

[0128] The line loss rate is calculated based on the power supply amount and the power sales amount.

[0129] For the convenience of description, the above devices are described in terms of functions divided into various modules. Of course, when implementing the embodiments of the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0130] The device of the above embodiment is used to implement the corresponding line loss analysis method in the above embodiment, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0131] Based on the same concept, corresponding to any of the above-mentioned embodiments, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the line loss analysis method as described in any of the above embodiments is implemented.

[0132] Figure 6 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0133] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0134] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0135] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0136] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0137] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0138] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0139] The electronic device of the above embodiment is used to implement the corresponding line loss analysis method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0140] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the line loss analysis method described in any of the above embodiments.

[0141] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be achieved by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0142] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the line loss analysis method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0143] Based on the same concept, corresponding to any of the above-mentioned embodiments, the present application also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of a computer so that the computer and / or the processor execute the line loss analysis method. Corresponding to the execution subject corresponding to each step in each embodiment of the line loss analysis method, the processor that executes the corresponding step can belong to the corresponding execution subject.

[0144] The computer program product of the above embodiment is used to enable the computer and / or the processor to execute the line loss analysis method described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0145] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0146] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0147] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0148] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.

Claims

1. A line loss analysis method, It is characterized in that include: Acquire at least one set of basic data of the power grid, and respectively determine at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a plurality of specific data; Using a robust outlier detection algorithm to screen the at least one line loss rate for abnormal line loss rates, and taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data; Performing clustering processing on the sample data to obtain at least one classification group, determining any specific data of the basic data corresponding to any classification group as a first factor, and determining whether the first factor exceeds a set threshold; In response to the first factor not exceeding the set threshold, taking the line loss rate corresponding to any one of the classification groups as the second factor, and determining an analysis formula by using the first factor and the second factor using a least squares support vector machine algorithm; The specific data corresponding to the first factor in the abnormal data is brought into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

2. The method according to claim 1, It is characterized in that The screening of abnormal line loss rates for the at least one line loss rate by using a robust outlier detection algorithm includes: Fitting the at least one line loss rate into a Gaussian mixture model using an expectation maximization algorithm, and determining a first parameter according to means and variances of different Gaussian components of the Gaussian mixture model; The first parameter is screened for errors according to the Laida criterion, and the error exceeding the limit is taken as the abnormal line loss rate.

3. The method according to claim 2, It is characterized in that The step of fitting the at least one line loss rate into a Gaussian mixture model by using an expectation maximization algorithm includes: The at least one line loss rate is centered and standardized, and the processed result is fitted into the Gaussian mixture model.

4. The method according to claim 1, It is characterized in that The clustering process of the sample data includes: The K-means clustering algorithm is used to cluster the sample data in the spatial dimension.

5. The method according to claim 1, It is characterized in that After determining whether the first factor exceeds a set threshold, the method further includes: In response to the first factor exceeding a set threshold, the first factor is output as an abnormal factor with the highest priority.

6. The method according to claim 1, It is characterized in that The method of determining the analysis formula by using the least squares support vector machine algorithm based on the first factor and the second factor includes: Using the first factor and the second factor as training data, and fitting the training model function using the correlation relationship of the least squares support vector machine algorithm; The trained model function is differentiated to obtain the analytical formula.

7. The method according to claim 1, It is characterized in that The specific data include at least one of power supply, power sales, substation name, transformer equipment classification, transformer status, transformer model, usage nature, operation and maintenance unit, maintenance team, voltage level, rated capacity, short-circuit loss, no-load loss, low-voltage side rated current, number of households, power supply radius, line length and load rate; The line loss rate is calculated based on the power supply amount and the power sales amount.

8. A line loss analysis device, It is characterized in that include: An acquisition module, used to acquire at least one set of basic data of the power grid, and respectively determine at least one line loss rate corresponding to the at least one set of basic data; wherein any set of basic data includes a variety of specific data; A screening module, used for screening the at least one line loss rate for abnormal line loss rate by using a robust outlier detection algorithm, and taking the basic data corresponding to the at least one line loss rate belonging to the abnormal line loss rate as abnormal data, otherwise as sample data; A clustering module, used to perform clustering processing on the sample data to obtain at least one classification group, determine any specific data of the basic data corresponding to any classification group as a first factor, and determine whether the first factor exceeds a set threshold; a determination module, configured to, in response to the first factor not exceeding a set threshold, use the line loss rate corresponding to any one of the classification groups as a second factor, and determine an analysis formula using a least squares support vector machine algorithm through the first factor and the second factor; The output module is used to bring the specific data corresponding to the first factor in the abnormal data into the analysis formula to determine and output the influence of different specific data on the line loss according to the calculation result.

9. An electronic device, It is characterized in that The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium, It is characterized in that The non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to implement the method according to any one of claims 1 to 7.