Power system line loss assessment method and system based on big data analysis

Abstract

The invention provides an electric power system line loss assessment method and system based on big data analysis, and relates to the technical field of line loss assessment, and the method comprises the steps: obtaining a multi-source electric power data set of an electric power system, carrying out the feature extraction of the preprocessed multi-source electric power data set, and obtaining a multi-dimensional electric power feature vector; the method comprises the following steps: acquiring line loss rate data of a power system, preprocessing the line loss rate data to obtain a line loss rate sequence, and performing sudden change detection on the line loss rate sequence to obtain a line loss sudden change significant quantity; according to the line loss abrupt change significant quantity and the multi-dimensional power feature vector, constructing an abrupt change correlation sequence, and converting the multi-dimensional power feature vector into a core power feature space based on the abrupt change correlation sequence; and determining a line loss prediction coefficient according to the core electric power feature space and the line loss prediction model, and judging the line loss evaluation grade of the electric power system through the line loss prediction coefficient, so that the features associated with the line loss abnormity of the electric power system can be extracted from the multi-source data to improve the prediction precision of the line loss prediction model.

Description

Technical Field

[0001] The present application relates to the technical field of line loss assessment, and more specifically, to a method and system for assessing line loss in an electric power system based on big data analysis. Background Art

[0002] Power system line loss assessment based on big data analysis is a method that uses modern data processing technology and advanced algorithms to comprehensively diagnose and optimize energy loss in power systems. By collecting multi-source data of the power system, including real-time load data, voltage and current curves, equipment operating status, historical line loss records, environmental parameters, etc., a power system feature database is formed. Combined with the distributed storage and computing capabilities of the big data platform, machine learning, statistical analysis, deep learning and other algorithms are used to clean, model and mine data, which can identify the key factors affecting line loss, predict line loss trends, and put forward optimization suggestions. At the same time, it can also support multiple tasks such as power grid dispatch optimization, equipment maintenance management, and energy efficiency improvement, providing technical support for the realization of a green and low-carbon power system.

[0003] However, the factors that affect the line loss of the power system are numerous and complex. Line loss is not only directly affected by the operating conditions of the power grid (such as load distribution, power factor, network topology, etc.), but may also be closely related to environmental conditions (such as temperature and humidity, seasonal changes) and equipment status (such as aging degree, operating efficiency). In the prior art, the interaction of such multi-dimensional factors leads to low prediction accuracy of the line loss prediction model and inaccurate assessment of the line loss of the power system. Therefore, how to extract features associated with abnormal line loss in the power system from multi-source data to improve the prediction accuracy of the line loss prediction model is a difficult problem faced by the industry. Summary of the invention

[0004] The present application provides a method and system for evaluating power system line loss based on big data analysis, which can extract features associated with power system line loss anomalies from multi-source data to improve the prediction accuracy of the line loss prediction model.

[0005] In a first aspect, the present application provides a method for evaluating line loss in a power system based on big data analysis, the evaluation method comprising the following steps:

[0006] Acquire a multi-source power data set of the power system, then preprocess the multi-source power data set, perform feature extraction on the preprocessed multi-source power data set, and obtain a multi-dimensional power feature vector of the power system;

[0007] Acquire line loss rate data of the power system, preprocess the line loss rate data to obtain a line loss rate sequence, perform mutation detection on the line loss rate sequence, and obtain a significant amount of line loss mutation of the power system;

[0008] Constructing a mutation association sequence of the power system according to the line loss mutation significant amount and the multidimensional power feature vector, and converting the multidimensional power feature vector into a core power feature space of the power system based on the mutation association sequence;

[0009] A pre-trained line loss prediction model is obtained, a line loss prediction coefficient of the power system is determined according to the core power feature space and the line loss prediction model, and a line loss assessment level of the power system is judged by the line loss prediction coefficient.

[0010] In this embodiment, preprocessing the multi-source power data set is to clean and standardize the multi-source power data set.

[0011] In this embodiment, feature extraction is performed on the preprocessed multi-source power data set to obtain the multi-dimensional power feature vector of the power system, specifically including:

[0012] Extracting power parameter data, power time data and power environment data of the power system from the preprocessed multi-source power data set;

[0013] Extracting features from the power parameter data, the power time data, and the power environment data respectively to obtain a plurality of power features;

[0014] A multi-dimensional power feature vector of the power system is constructed according to all the extracted power features.

[0015] In this embodiment, the preprocessing of the line loss rate data is to use a wavelet denoising algorithm to perform smoothing preprocessing on the line loss rate data.

[0016] In this embodiment, performing mutation detection on the line loss rate sequence to obtain the significant amount of line loss mutation of the power system specifically includes:

[0017] Obtain a given reference line loss rate sequence;

[0018] respectively determining the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence;

[0019] Determining a line loss fluctuation significance coefficient of the power system according to the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence;

[0020] The significant amount of line loss mutation of the power system is determined based on the line loss fluctuation significant coefficient, the line loss rate sequence and the reference line loss rate sequence.

[0021] In this embodiment, constructing the mutation association sequence of the power system according to the line loss mutation significant amount and the multi-dimensional power feature vector specifically includes:

[0022] Determine the mutation correlation between the line loss mutation significant amount and each power feature in the multidimensional power feature vector;

[0023] A mutation correlation sequence of the power system is constructed according to all mutation correlation degrees.

[0024] In this embodiment, converting the multidimensional power feature vector into the core power feature space of the power system based on the mutation association sequence specifically includes:

[0025] Obtain a preset mutation correlation threshold;

[0026] Performing core screening on each power feature in the multidimensional power feature vector according to the mutation correlation threshold and the mutation correlation sequence to obtain a plurality of core power features;

[0027] A core power feature space of the power system is constructed based on all the screened core power features.

[0028] In this embodiment, the line loss prediction coefficient of the power system is determined according to the core power feature space and the line loss prediction model by inputting all core power features in the core power feature space into the line loss prediction model for line loss prediction, thereby obtaining the line loss prediction coefficient of the power system.

[0029] In this embodiment, judging the line loss assessment level of the power system by the line loss prediction coefficient specifically includes:

[0030] Obtaining a line loss level assessment table for the power system;

[0031] The line loss assessment level of the power system is determined according to the line loss level assessment table and the line loss prediction coefficient.

[0032] In a second aspect, the present application provides a power system line loss assessment system based on big data analysis, which is used to perform a power system line loss assessment method based on big data analysis, and the assessment system includes:

[0033] A multi-source data acquisition module is used to acquire a multi-source power data set of a power system, and then pre-process the multi-source power data set, perform feature extraction on the pre-processed multi-source power data set, and obtain a multi-dimensional power feature vector of the power system;

[0034] A mutation detection module is used to obtain line loss rate data of the power system, pre-process the line loss rate data to obtain a line loss rate sequence, perform mutation detection on the line loss rate sequence, and obtain a significant amount of line loss mutation of the power system;

[0035] A feature space construction module, used to construct a mutation association sequence of the power system according to the line loss mutation significant amount and the multidimensional power feature vector, and convert the multidimensional power feature vector into a core power feature space of the power system based on the mutation association sequence;

[0036] The line loss assessment module is used to obtain a pre-trained line loss prediction model, determine the line loss prediction coefficient of the power system according to the core power feature space and the line loss prediction model, and judge the line loss assessment level of the power system by the line loss prediction coefficient.

[0037] The technical solution provided by the embodiments disclosed in this application has the following beneficial effects:

[0038] A multi-source power data set of the power system is obtained, and then the multi-source power data set is preprocessed, and feature extraction is performed on the preprocessed multi-source power data set to obtain a multi-dimensional power feature vector of the power system; line loss rate data of the power system is obtained, and the line loss rate data is preprocessed to obtain a line loss rate sequence, and mutation detection is performed on the line loss rate sequence to obtain a significant amount of line loss mutations in the power system; a mutation association sequence of the power system is constructed according to the significant amount of line loss mutations and the multi-dimensional power feature vector, and the multi-dimensional power feature vector is converted into a core power feature space of the power system based on the mutation association sequence; a pre-trained line loss prediction model is obtained, and a line loss prediction coefficient of the power system is determined according to the core power feature space and the line loss prediction model, and a line loss assessment level of the power system is judged by the line loss prediction coefficient.

[0039] It can be seen that in this application, firstly, the preprocessing step can ensure that the subsequent model is based on higher quality data for analysis, thereby improving the prediction accuracy. By constructing a multi-dimensional power feature vector, the multiple dimensional characteristics of the power system can be captured at the same time to help the model understand the complex dynamic behavior of the power system; then, by preprocessing the line loss rate data, unnecessary interference can be effectively removed, and based on mutation detection and calculation of the significant amount of line loss mutations, it can reflect the important changes in the line loss rate in the power system, and by combining the characteristics of the significant amount of line loss mutations, the prediction model can better identify abnormal patterns; secondly, by constructing mutation association sequences and extracting core power features, the relevant information in multi-source power data can be effectively focused on the features directly related to line loss changes, reducing the interference of noise on prediction accuracy; finally, based on the core power feature space, the line loss prediction model can focus on key information closely related to line loss, making line loss prediction more accurate.

[0040] In summary, the technical solution adopted in this application can extract features associated with power system line loss anomalies from multi-source data to improve the prediction accuracy of the line loss prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:

[0042] Figure 1 It is a flow chart of a method for evaluating power system line loss based on big data analysis provided in this application;

[0043] Figure 2 is an exemplary flow chart for determining a significant amount of line loss mutation in a power system according to the present application;

[0044] Figure 3 is an exemplary flow chart for determining the core power feature space of a power system according to the present application;

[0045] Figure 4 This is a module structure diagram of the power system line loss assessment system based on big data analysis provided in this application. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments and drawings. The schematic implementation modes and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention. It should be noted that the present invention is already in the actual development and use stage.

[0047] Embodiment 1

[0048] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods. Figure 1 As shown in the figure, this figure is an exemplary flow chart of a power system line loss assessment method based on big data analysis according to this embodiment of the present application, and the assessment method includes the following steps:

[0049] In step S1, a multi-source power data set of a power system is acquired, and then the multi-source power data set is preprocessed, and feature extraction is performed on the preprocessed multi-source power data set to obtain a multi-dimensional power feature vector of the power system.

[0050] It should be noted that in the present application, the multi-source power data set is a data set composed of power parameter data, power time data and power environment data of the power system, wherein the power parameter data mainly includes the voltage, current, power and power factor of the power equipment in the power system, the power time data mainly includes the load time series, operation cycle characteristics, equipment start and stop time and load change time, and the power environment data mainly includes external influencing factors of the power system, such as temperature, humidity and atmospheric pressure; in specific implementation, the power parameter data, power time data and power environment data can be obtained through the intelligent sensor array in the power system.

[0051] In this embodiment, preprocessing the multi-source power data set is to clean and standardize the multi-source power data set; it should be noted that preprocessing the multi-source power data set is to ensure the quality, consistency and accuracy of the data, so as to provide a reliable basis for subsequent analysis and modeling. In this application, the main tasks of preprocessing include data cleaning and standardization, wherein the purpose of data cleaning is to remove or correct errors, anomalies and inconsistencies in the multi-source power data set to ensure the integrity and accuracy of the multi-source power data set, and data standardization is to convert data of different dimensions and scales into a unified standard scale to make them comparable, so as to avoid affecting the performance of the line loss prediction model due to large differences in the range of characteristic values.

[0052] In this embodiment, feature extraction is performed on the preprocessed multi-source power data set to obtain the multi-dimensional power feature vector of the power system in the following manner, namely:

[0053] Extracting power parameter data, power time data and power environment data of the power system from the preprocessed multi-source power data set;

[0054] Extracting features from the power parameter data, the power time data, and the power environment data respectively to obtain a plurality of power features;

[0055] A multi-dimensional power feature vector of the power system is constructed according to all the extracted power features.

[0056] In specific implementation, first, the power parameter data, power time data and power environment data of the power system can be extracted from the preprocessed multi-source power data set by data traversal; then, the power parameter data, power time data and power environment data can be statistically extracted respectively to obtain multiple power features. The multiple power features in this application mainly include the power factor, average voltage, average current, periodicity, seasonality, trend, temperature fluctuation, humidity fluctuation and light intensity of the power system. The extraction of power parameter data reflects the operating status of the power system. The extraction of power time data is to timely understand the time dynamics of power data. The extraction of power environment data is to understand the operating influencing factors of the power system. Finally, a multidimensional power feature vector of the power system can be constructed based on all the extracted power features, wherein the multidimensional power feature vector is a multidimensional feature vector used to represent the operating status of the power system. All the extracted power features can be directly spliced ​​together to obtain a multidimensional feature vector, that is, the multidimensional power feature vector of the power system.

[0057] It should be noted that the preprocessing step can ensure the integrity, consistency and accuracy of the data set, ensure that the subsequent model is analyzed based on higher quality data, thereby improving the prediction accuracy. By constructing a multi-dimensional power feature vector, the multiple dimensional characteristics of the power system can be captured at the same time to help the model understand the complex dynamic behavior of the power system.

[0058] In step S2, line loss rate data of the power system is obtained, the line loss rate data is preprocessed to obtain a line loss rate sequence, mutation detection is performed on the line loss rate sequence to obtain a significant amount of line loss mutation of the power system.

[0059] In specific implementation, the line loss rate data of the power system can be calculated based on the operating data of the power system. The data usually comes from the operating data of the power company, that is, the ratio of the loss power to the total power generation is multiplied by 100%, and the result is used as the line loss rate. The line loss rate data of the power system can be obtained in the above way.

[0060] In this embodiment, the line loss rate data is preprocessed by smoothing the line loss rate data using a wavelet denoising algorithm, and then the preprocessed line loss rate data is sorted in chronological order, and the sorting result is used as a line loss rate sequence; it should be noted that the line loss rate data is preprocessed by smoothing using a wavelet denoising algorithm, which can effectively remove noise, retain the main trends and features of the data, improve the quality of the line loss rate data, and provide more accurate data support for subsequent mutation detection and anomaly analysis.

[0061] Preferably, in this embodiment, reference Figure 2As shown, this figure is an exemplary flow chart for determining a significant amount of line loss mutation in a power system in an embodiment of the present application. In this embodiment, mutation detection is performed on the line loss rate sequence to obtain a significant amount of line loss mutation in the power system, which can be specifically achieved by the following steps:

[0062] First, in step S21, a given reference line loss rate sequence is obtained;

[0063] Then, in step S22, the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence are determined respectively;

[0064] Secondly, in step S23, the line loss fluctuation significance coefficient of the power system is determined according to the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence;

[0065] Finally, in step S24, the significant amount of line loss mutation of the power system is determined based on the line loss fluctuation significant coefficient, the line loss rate sequence and the reference line loss rate sequence.

[0066] In the specific implementation, first, a given reference line loss rate sequence can be obtained, which is the historical line loss rate data of the power system under normal operating conditions, and the reference line loss rate sequence reflects the typical line loss situation of the power system under no abnormal conditions; then, the line loss volatility of the line loss rate sequence and the line loss volatility of the reference line loss rate sequence are determined respectively, wherein the line loss volatility represents the degree of fluctuation of the line loss of the power system, the variance of all data in the line loss rate sequence can be used as the line loss volatility of the line loss rate sequence, and the variance of all data in the reference line loss rate sequence can be used as the line loss volatility of the reference line loss rate sequence.

[0067] In addition, in specific implementation, the line loss fluctuation significance coefficient of the power system can be determined according to the line loss fluctuation degree of the line loss rate sequence and the line loss fluctuation degree of the reference line loss rate sequence, wherein the line loss fluctuation significance coefficient is an indicator used to measure the significance of the fluctuation difference between the current line loss rate sequence and the reference line loss rate sequence. In actual implementation, the line loss fluctuation significance coefficient can be determined by the following formula:

[0068]

[0069] Among them, γ represents the significant coefficient of line loss fluctuation of the power system, n represents the length of the line loss rate sequence, α represents the line loss fluctuation degree of the line loss rate sequence, m represents the length of the reference line loss rate sequence, and β represents the line loss fluctuation degree of the reference line loss rate sequence; finally, the significant amount of line loss mutation of the power system can be determined based on the significant coefficient of line loss fluctuation, the line loss rate sequence and the reference line loss rate sequence, wherein the significant amount of line loss mutation represents the significance of the line loss rate mutation of the power system. The larger the significant amount of line loss mutation, the more serious line loss mutation may occur in the current power system. In actual implementation, the significant amount of line loss mutation can be determined by the following formula:

[0070]

[0071] Wherein, λ represents the significant amount of line loss mutation in the power system, n represents the length of the line loss rate sequence, μ(n) represents the mean line loss rate in the line loss rate sequence, m represents the length of the reference line loss rate sequence, and μ(m) represents the mean line loss rate in the reference line loss rate sequence.

[0072] It should be noted that by preprocessing the line loss rate data, unnecessary interference can be effectively removed. Based on mutation detection and calculation of the significant amount of line loss mutation, it can reflect the important changes in the line loss rate in the power system, indicating that abnormal events have occurred in the power system, such as equipment failure, abnormal load fluctuation, line damage or other emergencies. In addition, by combining the characteristics of the significant amount of line loss mutation, the prediction model can better identify abnormal patterns and predict line loss situations that do not conform to conventional rules, thereby improving the model's adaptability to abnormal situations.

[0073] In step S3, a mutation association sequence of the power system is constructed according to the line loss mutation significance and the multidimensional power feature vector, and the multidimensional power feature vector is converted into a core power feature space of the power system based on the mutation association sequence.

[0074] In this embodiment, the following method can be used to construct the mutation association sequence of the power system according to the line loss mutation significant amount and the multi-dimensional power feature vector, namely:

[0075] Determine the mutation correlation between the line loss mutation significant amount and each power feature in the multidimensional power feature vector;

[0076] A mutation correlation sequence of the power system is constructed according to all mutation correlation degrees.

[0077] In the specific implementation, first, the mutation correlation between the line loss mutation significance amount and each power feature in the multidimensional power feature vector can be determined, wherein the mutation correlation is an indicator used to measure the degree of linear relationship between the line loss mutation significance amount and the power feature, and the Pearson correlation coefficient between the line loss mutation significance amount and the power feature can be calculated, and the result can be used as the mutation correlation between the line loss mutation significance amount and the power feature. The mutation correlation between the line loss mutation significance amount and each power feature in the multidimensional power feature vector can be obtained in the above manner; then, the mutation correlation sequence of the power system can be constructed according to all the mutation correlations, that is, the sequence composed of all mutation correlations is used as the mutation correlation sequence of the power system.

[0078] Preferably, in this embodiment, reference Figure 3 As shown, this figure is an exemplary flow chart of determining the core power feature space of the power system in an embodiment of the present application. In this embodiment, converting the multidimensional power feature vector into the core power feature space of the power system based on the mutation association sequence can be implemented by the following steps:

[0079] First, in step S31, a preset mutation correlation threshold is obtained;

[0080] Then, in step S32, core screening is performed on each power feature in the multidimensional power feature vector according to the mutation correlation threshold and the mutation correlation sequence to obtain a plurality of core power features;

[0081] Finally, in step S33, the core power feature space of the power system is constructed according to all the screened core power features.

[0082] In specific implementation, first, a pre-set mutation correlation threshold can be obtained, which can help screen out power features that are significantly correlated with line loss mutations. The mutation correlation threshold can be statistically analyzed based on the correlation between historical line loss rate data and feature data to calculate the mutation correlation threshold that can best distinguish between normal and mutation states, which will not be repeated here; then, each power feature in the multidimensional power feature vector can be core screened based on the mutation correlation threshold and the mutation correlation sequence, that is, for each power feature in the multidimensional power feature vector, the power feature whose corresponding mutation correlation in the mutation correlation sequence is greater than the mutation correlation threshold is used as the core power feature. In the above manner, multiple core power features can be obtained. It should be noted that the core power feature is a feature that can effectively characterize the key operating state in the power system, and the core power feature has a high correlation with the significant amount of line loss mutation; finally, the core power feature space of the power system can be constructed based on all the screened core power features, wherein the core power feature space is a feature space composed of multiple core power features, and all the screened core power features can be mapped into the feature space to obtain the core power feature space of the power system.

[0083] It should be noted that by constructing mutation association sequences and extracting core power features, the relevant information in multi-source power data can be effectively focused on the features directly related to line loss changes, and those features that are irrelevant or have little impact on line loss prediction can be effectively removed, reducing the interference of noise on prediction accuracy.

[0084] In step S4, a pre-trained line loss prediction model is obtained, a line loss prediction coefficient of the power system is determined according to the core power feature space and the line loss prediction model, and a line loss assessment level of the power system is judged by the line loss prediction coefficient.

[0085] In specific implementation, a pre-trained line loss prediction model can be obtained. The line loss prediction model selected in this application is a multivariate regression linear model based on machine learning. In actual implementation, other models can also be selected as line loss prediction models. There is no limitation here. The line loss prediction model is trained by using the prepared training data to obtain the line loss prediction model.

[0086] In this embodiment, determining the line loss prediction coefficient of the power system according to the core power feature space and the line loss prediction model is to input all the core power features in the core power feature space into the line loss prediction model for line loss prediction, and then obtain the line loss prediction coefficient of the power system; in specific implementation, after all the core power features in the core power feature space are input into the line loss prediction model, the core power features are calculated from the input layer to the output layer, and the line loss prediction model will generate a prediction result, i.e., the line loss prediction coefficient, according to weights, activation functions, etc.; it should be noted that in this application, the line loss prediction coefficient is an indicator obtained by prediction to represent the degree of line loss in the power system. The larger the line loss prediction coefficient, the greater the loss in the power system, which may be due to equipment aging, uneven load, line problems, etc.

[0087] In this embodiment, the line loss assessment level of the power system may be determined by the line loss prediction coefficient in the following manner, namely:

[0088] Obtaining a line loss level assessment table for the power system;

[0089] The line loss assessment level of the power system is determined according to the line loss level assessment table and the line loss prediction coefficient.

[0090] In the specific implementation, first, a line loss level assessment table for the power system can be obtained. The line loss level assessment table is a data table used to define different line loss assessment levels corresponding to different line loss prediction coefficients. The line loss level assessment table can be preset according to the historical data, standards and industry regulations of the power system; then, the line loss assessment level of the power system can be judged according to the line loss level assessment table and the line loss prediction coefficient, that is, the line loss level assessment table and the line loss prediction coefficient are compared, so that the line loss assessment level of the power system can be obtained from the line loss level assessment table.

[0091] It should be noted that in this application, the line loss assessment level of the power system can be divided into low loss level, medium loss level and high loss level. Among them, the low loss level indicates that the power system operates efficiently and no large-scale adjustments are required. At this time, monitoring can continue to ensure its stability; the medium loss level indicates that the power operation team needs to conduct inspections to check for potential equipment aging, line losses or unreasonable operations, and take optimization measures (such as equipment maintenance, load adjustment, etc.) to reduce losses: the high loss level indicates that there are serious problems in the power system and emergency measures need to be taken immediately, such as replacing damaged power equipment, optimizing line design, adjusting power dispatching, etc., to reduce energy losses and ensure the stability and economy of power supply.

[0092] In addition, it should be noted that based on the core power feature space, the line loss prediction model can focus on key information closely related to line loss, making the line loss prediction more accurate. It can quantify the loss level of the power system according to different line loss prediction coefficients and line loss assessment levels, accurately reflect the system operation status, and avoid the occurrence of sudden failures.

[0093] It can be seen that in this application, firstly, the preprocessing step can ensure that the subsequent model is based on higher quality data for analysis, thereby improving the prediction accuracy. By constructing a multi-dimensional power feature vector, the multiple dimensional characteristics of the power system can be captured at the same time to help the model understand the complex dynamic behavior of the power system; then, by preprocessing the line loss rate data, unnecessary interference can be effectively removed, and based on mutation detection and calculation of the significant amount of line loss mutations, it can reflect the important changes in the line loss rate in the power system, and by combining the characteristics of the significant amount of line loss mutations, the prediction model can better identify abnormal patterns; secondly, by constructing mutation association sequences and extracting core power features, the relevant information in multi-source power data can be effectively focused on the features directly related to line loss changes, reducing the interference of noise on prediction accuracy; finally, based on the core power feature space, the line loss prediction model can focus on key information closely related to line loss, making line loss prediction more accurate.

[0094] In summary, the technical solution adopted in this application can extract features associated with power system line loss anomalies from multi-source data to improve the prediction accuracy of the line loss prediction model.

[0095] Embodiment 2

[0096] This application provides a power system line loss assessment system based on big data analysis, referring to Figure 4 As shown, this figure is a schematic diagram of an evaluation system according to this embodiment of the present application, and the evaluation system includes:

[0097] The multi-source data acquisition module 100 is used to acquire a multi-source power data set of the power system, and then pre-process the multi-source power data set, perform feature extraction on the pre-processed multi-source power data set, and obtain a multi-dimensional power feature vector of the power system;

[0098] A mutation detection module 200 is used to obtain line loss rate data of the power system, pre-process the line loss rate data to obtain a line loss rate sequence, perform mutation detection on the line loss rate sequence, and obtain a significant amount of line loss mutation of the power system;

[0099] A feature space construction module 300 is used to construct a mutation association sequence of the power system according to the line loss mutation significant amount and the multidimensional power feature vector, and convert the multidimensional power feature vector into a core power feature space of the power system based on the mutation association sequence;

[0100] The line loss assessment module 400 is used to obtain a pre-trained line loss prediction model, determine the line loss prediction coefficient of the power system according to the core power feature space and the line loss prediction model, and judge the line loss assessment level of the power system by the line loss prediction coefficient.

[0101] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for evaluating power system line loss based on big data analysis, characterized in that: The evaluation method comprises the following steps: Acquire a multi-source power data set of the power system, then preprocess the multi-source power data set, perform feature extraction on the preprocessed multi-source power data set, and obtain a multi-dimensional power feature vector of the power system; Acquire line loss rate data of the power system, preprocess the line loss rate data to obtain a line loss rate sequence, perform mutation detection on the line loss rate sequence, and obtain a significant amount of line loss mutation of the power system; Constructing a mutation association sequence of the power system according to the line loss mutation significant amount and the multidimensional power feature vector, and converting the multidimensional power feature vector into a core power feature space of the power system based on the mutation association sequence; A pre-trained line loss prediction model is obtained, a line loss prediction coefficient of the power system is determined according to the core power feature space and the line loss prediction model, and a line loss assessment level of the power system is judged by the line loss prediction coefficient.

2. A method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Preprocessing the multi-source power data set is to clean and standardize the multi-source power data set.

3. A method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Feature extraction is performed on the preprocessed multi-source power data set to obtain the multi-dimensional power feature vector of the power system, specifically including: Extracting power parameter data, power time data and power environment data of the power system from the preprocessed multi-source power data set; Extracting features from the power parameter data, the power time data, and the power environment data respectively to obtain a plurality of power features; A multi-dimensional power feature vector of the power system is constructed according to all the extracted power features.

4. A method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: The preprocessing of the line loss rate data is to use a wavelet denoising algorithm to perform smoothing preprocessing on the line loss rate data.

5. The method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Performing mutation detection on the line loss rate sequence to obtain a significant amount of line loss mutation of the power system specifically includes: Obtain a given reference line loss rate sequence; respectively determining the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence; Determining a line loss fluctuation significance coefficient of the power system according to the line loss fluctuation of the line loss rate sequence and the line loss fluctuation of the reference line loss rate sequence; The significant amount of line loss mutation of the power system is determined based on the line loss fluctuation significant coefficient, the line loss rate sequence and the reference line loss rate sequence.

6. A method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Constructing the mutation association sequence of the power system according to the line loss mutation significant amount and the multi-dimensional power feature vector specifically includes: Determine the mutation correlation between the line loss mutation significant amount and each power feature in the multidimensional power feature vector; A mutation correlation sequence of the power system is constructed according to all mutation correlation degrees.

7. The method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Converting the multidimensional power feature vector into the core power feature space of the power system based on the mutation association sequence specifically includes: Obtain a preset mutation correlation threshold; Performing core screening on each power feature in the multidimensional power feature vector according to the mutation correlation threshold and the mutation correlation sequence to obtain a plurality of core power features; A core power feature space of the power system is constructed based on all the screened core power features.

8. The method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Determining the line loss prediction coefficient of the power system according to the core power feature space and the line loss prediction model is to input all the core power features in the core power feature space into the line loss prediction model to perform line loss prediction, and then obtain the line loss prediction coefficient of the power system.

9. The method for evaluating power system line loss based on big data analysis according to claim 1, characterized in that: Determining the line loss assessment level of the power system by using the line loss prediction coefficient specifically includes: Obtaining a line loss level assessment table for the power system; The line loss assessment level of the power system is determined according to the line loss level assessment table and the line loss prediction coefficient.

10. A power system line loss assessment system based on big data analysis, used to execute a power system line loss assessment method based on big data analysis as claimed in any one of claims 1 to 9, characterized in that: The evaluation system comprises: A multi-source data acquisition module is used to acquire a multi-source power data set of a power system, and then pre-process the multi-source power data set, perform feature extraction on the pre-processed multi-source power data set, and obtain a multi-dimensional power feature vector of the power system; A mutation detection module is used to obtain line loss rate data of the power system, pre-process the line loss rate data to obtain a line loss rate sequence, perform mutation detection on the line loss rate sequence, and obtain a significant amount of line loss mutation of the power system; A feature space construction module, used to construct a mutation association sequence of the power system according to the line loss mutation significant amount and the multidimensional power feature vector, and convert the multidimensional power feature vector into a core power feature space of the power system based on the mutation association sequence; The line loss assessment module is used to obtain a pre-trained line loss prediction model, determine the line loss prediction coefficient of the power system according to the core power feature space and the line loss prediction model, and judge the line loss assessment level of the power system by the line loss prediction coefficient.

Images