Data processing method for different types of power utilization safety at client side
Through the customer-side power safety data processing method, including data preprocessing and feature selection, the problem of massive electricity consumption information occupying memory is solved, the in-depth analysis of the operation of power enterprises and the accurate extraction of electricity consumption information is achieved, and data processing is improved Efficiency and accuracy.
Patent Information
- Application Number
- CN202411051393.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The massive amount of power consumption information accumulated by power companies during the management process occupies a large amount of computer storage system memory, reducing computer operation efficiency and increasing the operating burden of the power grid. It is difficult for the existing technology to effectively extract and analyze this data to obtain guiding power consumption information.
The data processing methods for different types of electricity use safety on the customer side include building a data model, extracting users with abnormal electricity use behavior, performing data preprocessing, feature selection and dimensionality reduction, classifying power users through the European distance algorithm, and using pull The Grangian interpolation formula handles missing values and minimum-maximum normalization methods for data normalization.
Through in-depth data analysis, electricity consumption information that is instructive to the operation and management of power enterprises is obtained, data processing efficiency and accuracy are improved, data missing and duplicate information are eliminated, and model analysis results are ensured.
Smart Images

Figure CN119090670A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of data processing methods, and in particular to a method for processing data for different types of electricity safety at a client side. Background Art
[0002] With the continuous improvement of information technology, power companies have accumulated a huge amount of electricity consumption information in their daily management process, such as electricity consumption information in different equipment such as power information systems, relay protection systems, and power grid operation systems. The massive amount of electricity data stored in these metering devices occupies a large amount of memory in the power grid computer storage system, reducing the computer's operating efficiency and increasing the overall operating burden of the power grid.
[0003] Therefore, it is urgent to use data extraction technology to conduct in-depth extraction and analysis of these huge amounts of power data. It is extremely important and meaningful to obtain electricity consumption information that is instructive and helpful for the operation and management of power companies. Summary of the invention
[0004] To this end, the present invention provides a method for processing different types of power safety data on the client side to solve the above problems in the prior art. In order to achieve the above purpose, the present invention provides the following technical solutions: According to a first aspect of the present invention, a method for processing different types of power safety data on the client side includes:
[0005] Step 1: Build a data model to extract users with abnormal electricity usage behavior;
[0006] Step 2: Preprocess the electricity consumption data, clean up the empty, zero, and identical data in the data set, and eliminate the power users with serious missing electricity consumption data, process the missing values of the electricity consumption data, and standardize the electricity consumption data;
[0007] Step 3: Classify power users according to data similarity. The specific classification algorithm is as follows:
[0008]
[0009] Among them, ||x i -c j ‖ is the Euclidean distance between the data object xi and the cluster center cj;
[0010] Step 4: Feature selection and dimension reduction of electricity consumption data.
[0011] Furthermore, a data model is constructed to extract users with abnormal electricity usage behaviors, including:
[0012] 101. Before carrying out the data extraction task, it is necessary to first analyze the business needs, that is, to figure out what to do and what effect to achieve, and to determine that the extraction goal is to use the load data of power users to extract users with abnormal power consumption behavior;
[0013] 102. Collecting the required extraction data set from the data storage device according to the determined extraction target, and performing data cleaning, data transformation, and data reduction processing on the collected data set;
[0014] 103. After analyzing the category of the extracted object and the desired extraction purpose, preprocess the data of the extracted object, consider which appropriate extraction method to choose, and conduct modeling analysis on the selected method;
[0015] 104. Select appropriate extraction methods to construct extraction models according to the characteristics and rules of the extracted data and the purpose of extraction;
[0016] 105. To evaluate the performance of the constructed model, the model evaluation indicators can be used to analyze the adopted extraction method.
[0017] Furthermore, the electricity consumption data is preprocessed, including:
[0018] For electricity consumption data cleaning, the electricity consumption data of some electricity users may contain a large number of missing values, continuous null values, or a large number of identical values. The judgment rule for eliminating electricity users with serious missing values in electricity consumption data is as follows:
[0019] For the user's daily load data, one quarter or more of the load data is missing;
[0020] One-third of electricity users are missing from monthly load data;
[0021] Electricity users whose monthly load data is zero for more than half;
[0022] Users with the same daily and monthly load data.
[0023] Furthermore, the electricity consumption data is preprocessed, and the following steps are also included:
[0024] The specific processing formula for missing values in electricity consumption data is as follows:
[0025] Lagrange interpolation formula:
[0026] For a given n+1 data points (x_0, y_0), (x_1, y_1), ..., (x_n, y_n), the Lagrange interpolation polynomial L(x) can be expressed as:
[0027] L(x)=Σ[i=0to n]y_i*l_i(x)
[0028] Where l_i(x) is the Lagrangian basis function, expressed as:
[0029] l_i(x)=Π[j=0to n,j! =i](x-x_j) / (x_i-x_j).
[0030] Furthermore, the electricity consumption data is preprocessed, and the following steps are also included:
[0031] The electricity consumption data is standardized by using the minimum-maximum standardization method to normalize the collected electricity consumption data, that is, the electricity consumption data after missing data interpolation is linearly transformed so that the processed electricity consumption data is between 0 and 1. The specific calculation formula is as follows:
[0032]
[0033] Among them, xnew is the electricity consumption data after standardization, x is the user's original electricity consumption data, and xmin and xmax are the minimum and maximum values of the original electricity consumption data before standardization.
[0034] Furthermore, the electricity consumption data feature selection includes:
[0035] Assume that the electricity consumption data of each type of user includes (l i ,i=1,2,…,n|l i ∈N) users’ electricity consumption data for H days, each user’s electricity consumption data sequence can be represented as an H-dimensional vector The electricity consumption data sequence set of each type of user is s l =(s l ,l=1,2,…,l i ), each type of extracted feature set can be expressed as V = (V l ,l=1,2,…,l i ).
[0036] The present invention has the following advantages:
[0037] By establishing an analysis model, staff can conduct in-depth extraction and analysis of massive power data, and obtain electricity consumption information that is instructive and helpful to the operation and management of power companies. This is extremely important and meaningful. The data is converted into a dimensionless and unified format through a certain format conversion, so that it can better adapt to the requirements of the extraction model for data characteristics, ensure the accuracy of the model analysis results, pre-process the data, clean up the empty, zero, and identical data in the data set, and eliminate power users with serious missing electricity consumption data, so as to improve the efficiency of subsequent data processing. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 A flowchart of a method for processing data of different types of electricity safety at the client side is provided for some embodiments of the present invention.
[0039] Figure 2 A data extraction flow chart of a method for processing data for different types of electricity safety at the client side provided in some embodiments of the present invention.
[0040] Figure 3 A structural diagram of an electricity usage information collection object for a method for processing data for different types of electricity safety at the client side provided in some embodiments of the present invention. DETAILED DESCRIPTION
[0041] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0042] like Figures 1 to 3 As shown, a method for processing different types of power safety data on the client side in an embodiment of the first aspect of the present invention includes:
[0043] Step 1: Build a data model to extract users with abnormal electricity usage behavior;
[0044] Step 2: Preprocess the electricity consumption data, clean up the empty, zero, and identical data in the data set, and eliminate the power users with serious missing electricity consumption data, process the missing values of the electricity consumption data, and standardize the electricity consumption data;
[0045] Step 3: Classify power users according to data similarity. The specific classification algorithm is as follows:
[0046]
[0047] Among them, ||x i -c j || is the Euclidean distance between the data object xi and the cluster center cj;
[0048] Step 4: Feature selection and dimension reduction of electricity consumption data.
[0049] In this embodiment, by establishing an analysis model, the staff can conduct in-depth extraction and analysis of the huge amount of power data, and obtain electricity consumption information that is instructive and helpful to the operation and management of power companies, which is extremely important and meaningful. The data is converted into a dimensionless and unified format through a certain format conversion, so that it can better adapt to the requirements of the extraction model for data characteristics, ensure the accuracy of the model analysis results, pre-process the data, clean up the empty, zero, and identical data in the data set, and eliminate power users with serious missing electricity consumption data, so as to improve the efficiency of data processing in the later stage.
[0050] Build a data model to extract users with abnormal electricity usage behavior, including:
[0051] 101. Before carrying out the data extraction task, it is necessary to first analyze the business needs, that is, to figure out what to do and what effect to achieve, and to determine that the extraction goal is to use the load data of power users to extract users with abnormal power consumption behavior;
[0052] 102. Collecting the required extraction data set from the data storage device according to the determined extraction target, and performing data cleaning, data transformation, and data reduction processing on the collected data set;
[0053] 103. After analyzing the category of the extracted object and the desired extraction purpose, preprocess the data of the extracted object, consider which appropriate extraction method to choose, and conduct modeling analysis on the selected method;
[0054] 104. Select appropriate extraction methods to construct extraction models according to the characteristics and rules of the extracted data and the purpose of extraction;
[0055] 105. To evaluate the performance of the constructed model, the model evaluation indicators can be used to analyze the adopted extraction method.
[0056] In this embodiment, before performing the data extraction task, it is necessary to first analyze the business needs, that is, to figure out what to do and what effect to achieve. The extraction goal in this article is to use the load data of power users to extract users with abnormal power consumption behavior. The extraction goal is the first and key step in data extraction. The determination of the extraction goal lays the foundation for the smooth progress of data extraction.
[0057] Data collection mainly involves collecting the required extraction data set from the data storage device according to the determined extraction target. This is an important step for the smooth progress of data extraction. Data preprocessing mainly involves data cleaning, data transformation, data reduction and other processing on the collected data set. Data cleaning mainly involves deleting various numerical anomalies such as irrelevant data, duplicate data and smooth noise data in the original data set, and cleaning up according to the numerical anomalies. The purpose of data cleaning is mainly to solve data quality problems and make the data more suitable for data extraction and analysis. Common data cleaning methods include missing value processing, outlier processing, and noise value processing. Data transformation mainly involves dimensionless and unified processing of data through certain format conversion, so that it can better adapt to the requirements of the extraction model for data characteristics and ensure the accuracy of the model analysis results.
[0058] The model obtained by using data extraction technology may not reach the preset extraction target, or may not accurately represent the true meaning of the extracted object, and may even be contrary to the facts in some cases. Therefore, it is necessary to evaluate the performance of the constructed model. The model evaluation index can be used to analyze the adopted extraction method. In order to further determine the accuracy of the method used, it can also be compared and verified with other methods, so as to find the best model suitable for the extraction target.
[0059] Preprocess the electricity consumption data, including:
[0060] For electricity consumption data cleaning, the electricity consumption data of some electricity users may contain a large number of missing values, continuous null values, or a large number of identical values. The judgment rule for eliminating electricity users with serious missing values in electricity consumption data is as follows:
[0061] For the user's daily load data, one quarter or more of the load data is missing;
[0062] One-third of electricity users are missing from monthly load data;
[0063] Electricity users whose monthly load data is zero for more than half;
[0064] Users with the same daily and monthly load data.
[0065] Preprocessing of electricity consumption data also includes:
[0066] The specific processing formula for missing values in electricity consumption data is as follows:
[0067] Lagrange interpolation formula:
[0068] For a given n+1 data points (x_0, y_0), (x_1, y_1), ..., (x_n, y_n), the Lagrange interpolation polynomial L(x) can be expressed as:
[0069] L(x)=Σ[i=0to n]y_i*l_i(x)
[0070] Where l_i(x) is the Lagrangian basis function, expressed as:
[0071] l_i(x)=Π[j=0to n,j! =i](x-x_j) / (x_i-x_j).
[0072] Preprocessing of electricity consumption data also includes:
[0073] The electricity consumption data is standardized by using the minimum-maximum standardization method to normalize the collected electricity consumption data, that is, the electricity consumption data after missing data interpolation is linearly transformed so that the processed electricity consumption data is between 0 and 1. The specific calculation formula is as follows:
[0074]
[0075] Among them, xnew is the electricity consumption data after standardization, x is the user's original electricity consumption data, and xmin and xmax are the minimum and maximum values of the original electricity consumption data before standardization.
[0076] In this embodiment, before analyzing the electrical status data, the data must be preprocessed to clean up empty, zero, or identical data in the data set and eliminate power users with serious missing electricity data. The small amount of missing data in the data set is interpolated, and the processed electricity data is further normalized. Data preprocessing mainly includes data cleaning, data integration, data transformation, and data reduction. In the process of power companies extracting user electricity data from the power user electricity information collection system, human or natural factors may cause missing values in the collected data. If these missing data are deleted directly, the user's electricity information will be incomplete, which will directly affect the abnormal detection model's judgment of the user's electricity behavior, and may even result in missed judgments, misjudgments, etc., so it is necessary to process the missing data in the collected data. Since the Lagrange interpolation method has a compact structure and is easy to calculate;
[0077] Since different electricity consumption indicators usually have different dimensions and orders of magnitude, when the levels of the indicators differ greatly, if the original data is directly used for electricity consumption anomaly detection and analysis, the indicators with higher values will be highlighted and the indicators with lower values will be relatively weakened, which will seriously affect the final results of anomaly detection. Therefore, in order to ensure the reliability of the data extraction results, the original electricity consumption data needs to be standardized. Data standardization mainly involves scaling the data by a certain ratio to reduce the data to a specified smaller range.
[0078] The selection of electricity consumption data features includes:
[0079] Assume that the electricity consumption data of each type of user includes (l i ,i=1,2,...,n|l i ∈N) users’ electricity consumption data for H days, each user’s electricity consumption data sequence can be represented as an H-dimensional vector The electricity consumption data sequence set of each type of user is s l =(s l ,l=1,2,…,l i ), each type of extracted feature set can be expressed as V = (V l ,l=1,2,…,l i ).
[0080] In this embodiment, the clustering algorithm is a fuzzy clustering algorithm based on an objective function. Its idea is to maximize the similarity between data divided into the same group and minimize the similarity between different groups of data according to a given rule, that is, to follow the principle of "high cohesion within the cluster and low coupling outside the cluster". Since the load data has periodic characteristics in terms of days, weeks, months, etc., it is necessary to extract features from the time series according to the period. Therefore, this patent analyzes the characteristics of the load data in time series and extracts features from four aspects: statistical indicators, trend indicators, variability indicators and load characteristic indicators.
[0081] Since there are too many features in the extracted electricity consumption data, feature matching will be too complicated, and these features will overlap, so it is necessary to reduce the dimension of the features, that is, map the high-dimensional features to low-dimensional features through a certain function as new features. It is proposed to use principal component analysis to reduce the dimension of the extracted electricity consumption data feature set of different categories of users.
Claims
1. A method for processing data of different types of electricity safety at the client side, characterized in that: include: Step 1: Build a data model to extract users with abnormal electricity usage behavior; Step 2: Preprocess the electricity consumption data, clean up the empty, zero, and identical data in the data set, and eliminate the power users with serious missing electricity consumption data, process the missing values of the electricity consumption data, and standardize the electricity consumption data; Step 3: Classify power users according to data similarity. The specific classification algorithm is as follows: Among them, ||x i -c j || is the Euclidean distance between the data object xi and the cluster center cj; Step 4: Feature selection and dimension reduction of electricity consumption data.
2. A method for processing data for different types of electricity safety at the client side according to claim 1, characterized in that: Build a data model to extract users with abnormal electricity usage behavior, including:
101. Before carrying out the data extraction task, it is necessary to first analyze the business needs, that is, to figure out what to do and what effect to achieve, and to determine that the extraction goal is to use the load data of power users to extract users with abnormal power consumption behavior; 102. Collecting the required extraction data set from the data storage device according to the determined extraction target, and performing data cleaning, data transformation, and data reduction processing on the collected data set; 103. After analyzing the category of the extracted object and the desired extraction purpose, preprocess the data of the extracted object, consider which appropriate extraction method to choose, and conduct modeling analysis on the selected method; 104. Select appropriate extraction methods to construct extraction models according to the characteristics and rules of the extracted data and the purpose of extraction; 105. To evaluate the performance of the constructed model, the model evaluation indicators can be used to analyze the adopted extraction method.
3. A method for processing data for different types of electricity safety at the client side according to claim 1, characterized in that: Preprocess the electricity consumption data, including: For electricity consumption data cleaning, the electricity consumption data of some electricity users may contain a large number of missing values, continuous null values, or a large number of identical values. The judgment rule for eliminating electricity users with serious missing values in electricity consumption data is as follows: For the user's daily load data, one quarter or more of the load data is missing; One-third of electricity users are missing from monthly load data; Electricity users whose monthly load data is zero for more than half; Users with the same daily and monthly load data.
4. A method for processing data for different types of electricity safety at the client side according to claim 1, characterized in that: Preprocessing of electricity consumption data also includes: The specific processing formula for missing values in electricity consumption data is as follows: Lagrange interpolation formula: For a given n+1 data points (x_0, y_0), (x_1, y_1), ..., (x_n, y_n), the Lagrange interpolation polynomial L(x) can be expressed as: L(x)=Σ[i=0to n]y_i*l_i(x) Where l_i(x) is the Lagrangian basis function, expressed as: l_i(x)=Π[j=0to n,j! =i](x-x_j) / (x_i-x_j).
5. A method for processing data for different types of electricity safety at the client side according to claim 1, characterized in that: Preprocessing of electricity consumption data also includes: The electricity consumption data is standardized by using the minimum-maximum standardization method to normalize the collected electricity consumption data, that is, the electricity consumption data after missing data interpolation is linearly transformed so that the processed electricity consumption data is between 0 and 1. The specific calculation formula is as follows: Among them, xnew is the electricity consumption data after standardization, x is the user's original electricity consumption data, and xmin and xmax are the minimum and maximum values of the original electricity consumption data before standardization.
6. A method for processing data for different types of electricity safety at the client side according to claim 1, characterized in that: The selection of electricity consumption data features includes: Assume that the electricity consumption data of each type of user includes (l i ,i=1,2,…,n|l i ∈N) users’ electricity consumption data for H days, each user’s electricity consumption data sequence can be represented as an H-dimensional vector The electricity consumption data sequence set of each type of user is s l =(s l ,l=1,2,…,l i ), each type of extracted feature set can be expressed as V = (V l ,l=1,2,…,l i ).
