Electricity larceny prevention analysis method and system based on electricity utilization big data and intelligent algorithm model

Abstract

The invention discloses an electricity larceny prevention analysis method and system based on electricity utilization big data and an intelligent algorithm model. The method comprises the following steps: analyzing a first electricity larceny analysis result for line loss data in a region; acquiring power consumption data and adjoint factor data of a user, wherein the power consumption data comprises power consumption time sequence data; analyzing a second electricity stealing analysis result by using an intelligent algorithm model based on the electricity utilization data and the adjoint factor data of the user; based on the first electricity larceny analysis result and the second electricity larceny analysis result, determining an electricity larceny condition; and analyzing and positioning an electricity stealing position and an electricity stealing time period based on the line loss data in the region, the electricity utilization data of the user and the accompanying factor data under the condition of determining that electricity stealing exists. According to the method, the electricity stealing analysis efficiency is effectively improved through a mode of firstly analyzing the line loss of the transformer area and then analyzing the users of the transformer area.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity theft analysis, and in particular to an anti-electricity theft analysis method and system based on electricity consumption big data and an intelligent algorithm model. Background Art

[0002] Electricity theft is an important component of non-technical losses in the power system. A series of electricity thefts on the distribution network side will not only cause huge economic losses, but also bring immeasurable safety hazards. Although the existing technology has made certain progress in electricity theft detection, there are still some shortcomings: First, traditional detection methods such as line tapping or manual inspection of meter tampering are inefficient and costly. Second, with the popularization of smart grids, the amount of user electricity consumption data has increased dramatically, and the huge amount of data has increased the difficulty of data processing and model building; third, there are various types of electricity theft, and conventional two-classification electricity theft models often cannot accurately capture the subtle differences between different electricity theft behaviors, and it is difficult to meet the needs of power companies. Summary of the invention

[0003] In view of the problems existing in the above-mentioned prior art, the present invention provides an anti-electricity theft analysis method and system based on electricity consumption big data and intelligent algorithm model, which effectively improves the efficiency and accuracy of anti-electricity theft analysis. The technical solution is as follows:

[0004] In the first aspect, an anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model is provided, comprising the following steps:

[0005] Analyze the first electricity theft analysis result for the line loss data in the area;

[0006] Acquire the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes the time series data of electricity consumption; the accompanying factor data includes the weather, calendar, and family attribute data accompanying the user's electricity consumption at the same time;

[0007] Based on the user's electricity consumption data and accompanying factor data, the second electricity theft analysis results are analyzed using an intelligent algorithm model;

[0008] Determining the existence of electricity theft based on the first electricity theft analysis result and the second electricity theft analysis result;

[0009] When it is determined that electricity theft has occurred, the location and time period of electricity theft are analyzed and located based on line loss data in the area, user electricity consumption data and accompanying factor data.

[0010] In some implementations, analyzing the second electricity theft analysis result using an intelligent algorithm model based on the user's electricity usage data and accompanying factor data includes:

[0011] Perform trend similarity analysis based on the daily, weekly and monthly electricity consumption time series data of all users, and classify users with similar trends into one category. If the trends of the user's daily electricity consumption time series data, weekly electricity consumption time series data and monthly electricity consumption time series data are all similar, the user is classified into one category. If one of the three trends is not similar, the user is not classified into one category and is determined to be a candidate electricity theft user.

[0012] A class is formed based on users who are not classified into the class, and users in the class whose number of users is less than a first preset value are determined as candidate electricity theft users;

[0013] For a candidate electricity theft user, obtain the user's historical electricity consumption time series data and historical accompanying factor time series data;

[0014] Based on the user's historical electricity consumption time series data and historical accompanying factor time series data, the user's electricity consumption pattern under the influence of accompanying factors is mined.

[0015] Based on the power consumption time series data and the accompanying factor time series data of the candidate power stealing users in the current time period, the user rules are utilized to analyze whether the candidate power stealing users in the current time period have committed power stealing behavior.

[0016] In some embodiments, the trend similarity analysis comprises:

[0017] Calculate the growth rate value of the later moment of the daily electricity consumption time series data of the first user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the first user; calculate the growth rate value of the later moment of the daily electricity consumption time series data of the second user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the second user;

[0018] Based on the first user's daily electricity consumption time series growth rate value sequence and the second user's daily electricity consumption time series growth rate value sequence, the difference between the first user's daily electricity consumption time series growth rate value and the second user's daily electricity consumption time series growth rate value at the same moment is calculated, and the sum of the differences corresponding to all moments in the daily electricity consumption time series data is obtained as the similarity between the first user's daily electricity consumption time series data trend and the second user's daily electricity consumption time series data trend.

[0019] In some implementations, mining the user's electricity usage pattern under the influence of accompanying factors based on the user's historical electricity usage time series data and historical accompanying factor time series data includes:

[0020] Based on the historical electricity consumption time series data and the historical accompanying factor time series data of different users as the first data set, after training the neural network model, the electricity consumption pattern of the user under the action of the accompanying factors is mined, recorded as the first electricity consumption pattern, and the first electricity consumption analysis model representing the first electricity consumption pattern is obtained;

[0021] Taking the first electricity consumption analysis model that characterizes the first electricity consumption pattern as the basic model, based on the historical electricity consumption time series data and the historical accompanying factor time series data of the same user as the second data set, the basic model is transformed to mine the electricity consumption pattern of the same user under the influence of the accompanying factors, which is recorded as the second electricity consumption pattern, and obtain the second electricity consumption analysis model that characterizes the second electricity consumption pattern.

[0022] In some implementations, transforming the basic model to obtain a second power consumption analysis model representing a second power consumption pattern includes:

[0023] Based on the second sample of the second data set, the basic model is input to obtain the final encoding result of the basic model encoding the power consumption time series data and the accompanying factor time series data, that is, the result before the output layer of the basic model;

[0024] Outputting a first coding transformation result based on the final coding result of the basic model through a first linear transformation matrix;

[0025] Based on the first coding transformation result, a nonlinear transformation is performed through a nonlinear activation function layer to obtain a nonlinear coding result;

[0026] Based on the nonlinear coding result, a first linear inverse transformation matrix opposite to the first linear transformation matrix is ​​used to obtain a first decoding result, the final coding result of the basic model and the first decoding result are combined as a fused first decoding result, and the fused first decoding result is input into the output layer of the basic model for full connection processing to obtain the electricity theft identification result of the second sample;

[0027] Based on the error between the electricity theft labeling data of the second sample and the actual electricity theft identification result, the first linear transformation matrix, the nonlinear activation function layer, and the first linear inverse transformation matrix are trained to obtain the second electricity consumption analysis model.

[0028] In some implementations, forming a separate class based on users not classified into the class, and determining users in the class whose number of users is less than a first preset value as candidate electricity theft users, further includes:

[0029] Based on users in a class whose number of users is greater than a first preset value, determine the height distance between each user's power consumption time series data curve and the power consumption time series data curves of other users in the class, and obtain the sum of the height distances between each user's power consumption time series data curve and the power consumption time series data curves of all other users in the class;

[0030] The sum of the height distances corresponding to each user's electricity consumption time series data curve is sorted from large to small, and the user's electricity consumption time series data curve ranked at a front preset value is obtained as a candidate electricity theft user.

[0031] In some embodiments, mining the user's electricity usage pattern under the influence of accompanying factors based on the user's historical electricity usage time series data and historical accompanying factor time series data includes: using the user's historical electricity usage time series data and historical accompanying factor time series data as input, training an intelligent algorithm model based on a neural network to mine the user's electricity usage pattern, and characterizing the user's electricity usage pattern based on the trained neural network model;

[0032] The neural network-based intelligent algorithm includes:

[0033] The power consumption time series data and accompanying factor time series data of a preset long period of time are used as the first input data.

[0034] Obtaining a growth value of a later moment of the power consumption time series data and the accompanying factor time series data over a preset long period of time compared to a previous moment, obtaining the power consumption time series growth rate data and the accompanying factor growth rate data of the preset long period of time, and using the power consumption time series growth rate data and the accompanying factor growth rate data as the second input data;

[0035] Inputting the data sequence of each time series data of the first input data within a preset long period of time into the sequence feature extraction unit, extracting the time series feature of each time series data, and characterizing the time relationship feature of each time series data;

[0036] Inputting the electricity consumption time series growth rate data and the accompanying factor growth rate data at the same time of the second input data into the convolutional layer unit to obtain the spatial relationship characteristics of the electricity consumption time series growth rate data and the accompanying factor growth rate data;

[0037] Based on the temporal relationship characteristics and spatial relationship characteristics of the input data, it is input into the fully connected layer for analysis to determine whether the input data is characteristic data of electricity theft.

[0038] In the second aspect, an anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model is provided, and the system includes:

[0039] A first electricity theft analysis unit, used for analyzing a first electricity theft analysis result based on line loss data in a region;

[0040] An electricity consumption data acquisition unit is used to acquire the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes time series data of electricity consumption; the accompanying factor data includes meteorological, calendar, and family attribute data accompanying the user's electricity consumption at the same time;

[0041] A second electricity theft analysis unit is used to analyze the second electricity theft analysis result using an intelligent algorithm model based on the user's electricity consumption data and accompanying factor data;

[0042] A third electricity theft analysis unit, configured to determine whether electricity theft exists based on the first electricity theft analysis result and the second electricity theft analysis result;

[0043] The electricity theft locating unit is used to analyze and locate the electricity theft location and electricity theft time period based on the line loss data in the area, the user's electricity consumption data and the accompanying factor data when it is determined that electricity theft exists.

[0044] According to a third aspect, an electronic device is provided, the electronic device comprising:

[0045] processor;

[0046] a memory for storing processor-executable instructions;

[0047] Among them, the processor implements the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model as described in the first aspect above by running the executable instructions.

[0048] In a fourth aspect, a computer-readable storage medium is provided, on which computer instructions are stored. When the instructions are executed by a processor, the steps of the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model as described in the first aspect are implemented.

[0049] The present invention provides an anti-electricity theft analysis method and system based on electricity consumption big data and intelligent algorithm model, which has the following beneficial effects: the present invention first determines whether there is abnormal line loss based on the calculation results of the line loss in the substation area, and if there is abnormal line loss, further analyzes the users in the substation area as the target object, and analyzes whether the user's electricity consumption conforms to the normal law of electricity consumption changes by considering factors such as weather, temperature, calendar, time, and season. The present invention first analyzes the line loss in the substation area and then analyzes the users in the substation area, which effectively improves the efficiency of electricity theft analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a flowchart of an anti-electricity theft analysis method based on electricity consumption big data and an intelligent algorithm model in an embodiment of the present application;

[0051] Figure 2 It is a flowchart of a method for analyzing a second electricity theft analysis result using an intelligent algorithm model based on the user's electricity consumption data and accompanying factor data in an embodiment of the present application;

[0052] Figure 3 It is a flowchart of a method for analyzing trend similarity of time series data in an embodiment of the present application;

[0053] Figure 4 It is a flowchart of a method for mining the user's electricity consumption pattern under the influence of accompanying factors based on the user's historical electricity consumption time series data and historical accompanying factor time series data in an embodiment of the present application;

[0054] Figure 5It is a structural diagram of an anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model in an embodiment of the present application. DETAILED DESCRIPTION

[0055] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0056] The present application provides an anti-electricity theft analysis method based on electricity consumption big data and an intelligent algorithm model, comprising the following steps:

[0057] Step 1, analyzing the line loss data in the area to obtain a first electricity theft analysis result;

[0058] Step 2, obtaining the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes the time series data of electricity consumption; the accompanying factor data includes the weather, calendar, and family attribute data (family attribute data such as the composition structure of family members, income, etc.) accompanying the user's electricity consumption at the same time;

[0059] Step 3, based on the user's electricity consumption data and accompanying factor data, using an intelligent algorithm model to analyze the second electricity theft analysis result;

[0060] Step 4, determining whether electricity theft exists based on the first electricity theft analysis result and the second electricity theft analysis result;

[0061] Step 5: When it is determined that electricity theft exists, the electricity theft location and the electricity theft time period are analyzed and located based on the line loss data in the area, the user's electricity consumption data and the accompanying factor data.

[0062] In the embodiment of the present application, based on the calculation results of the line loss in the substation area, it is first determined whether there is an abnormal line loss. In the case of abnormal line loss, it is considered that the theft of electricity will lead to an increase in the line loss of the power grid. In this application, the substation users are further analyzed as the target objects, and the user's electricity consumption is analyzed to see whether it conforms to the normal law of electricity consumption changes, taking into account factors such as weather, temperature, calendar, time, and season. The present application first analyzes the substation line loss and then analyzes the substation users, which effectively improves the efficiency of the electricity theft analysis.

[0063] In one embodiment, the above step 3, based on the user's electricity consumption data and accompanying factor data, uses an intelligent algorithm model to analyze the second electricity theft analysis result, including the following steps:

[0064] Step 31, based on the daily, weekly and monthly electricity consumption time series data of all users, trend similarity analysis is performed, and users with similar trends are divided into one category, wherein if the trend of the user's daily electricity consumption time series data, weekly electricity consumption time series data and monthly electricity consumption time series data are all similar, the user is divided into one category; if there is one trend among the three that is not similar, the user is not divided into one category and the user is determined to be a candidate electricity theft user;

[0065] Step 32, forming a separate class based on users not classified into the class, and determining the users in the class whose number of users is less than a first preset value as candidate electricity theft users;

[0066] Step 33: For a candidate electricity stealing user, obtain the user's historical electricity consumption time series data and historical accompanying factor time series data;

[0067] Step 34, mining the user's electricity consumption pattern under the influence of the accompanying factors based on the user's historical electricity consumption time series data and historical accompanying factor time series data,

[0068] Step 35 , based on the power consumption time series data and the accompanying factor time series data of the candidate power stealing users in the current time period, and using the user rules, analyze whether the candidate power stealing users in the current time period have committed power stealing behavior.

[0069] It is understandable that different users have different electricity consumption patterns. For example, some users increase their electricity consumption on weekends. For office workers, electricity consumption is less during the day, and electricity consumption is higher in the morning and evening after get off work. For commercial users, the daily electricity consumption is basically the same, with little fluctuation. There is a clear difference between the load curve fluctuation pattern of electricity theft users and normal users. Normal users have obvious characteristic fluctuation patterns in electricity consumption. Based on the similarity analysis of the electricity consumption trends of different users, the users who are classified into one category have similar patterns. For those who are not classified into one category or the number of users in the category is too small, the electricity consumption pattern of the users does not conform to the electricity consumption pattern of most users, and the users may be electricity theft users.

[0070] It can be understood that the daily electricity consumption time series data trend, weekly electricity consumption time series data trend, and monthly electricity consumption time series data trend of normal electricity users will be similar to the electricity consumption time series data trend of other users with similar electricity consumption patterns. For electricity theft users, one of their daily electricity consumption time series data trend, weekly electricity consumption time series data trend, and monthly electricity consumption time series data trend may be disguised to be similar to the electricity consumption time series trend of other users with similar electricity consumption patterns, but the electricity consumption change characteristics extracted from multiple observation dimensions will still show abnormalities from the electricity consumption change characteristics of normal users.

[0071] In one embodiment, the trend similarity analysis in step 31 above includes the following steps:

[0072] Step 311, calculate the growth rate value of the later moment of the daily electricity consumption time series data of the first user compared with the previous moment, and obtain the daily electricity consumption time series growth rate value sequence of the first user; calculate the growth rate value of the later moment of the daily electricity consumption time series data of the second user compared with the previous moment, and obtain the daily electricity consumption time series growth rate value sequence of the second user;

[0073] Step 312, based on the first user's daily electricity consumption time series growth rate value sequence and the second user's daily electricity consumption time series growth rate value sequence, calculate the difference between the first user's daily electricity consumption time series growth rate value and the second user's daily electricity consumption time series growth rate value sequence at the same time, and obtain the sum of the differences corresponding to all moments in the daily electricity consumption time series data as the similarity between the first user's daily electricity consumption time series data trend and the second user's daily electricity consumption time series data trend.

[0074] The method for calculating the similarity of the trend of time series data in the embodiment of the present application includes:

[0075] Calculate the growth rate value of the later moment of the daily electricity consumption time series data of the first user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the first user; calculate the growth rate value of the later moment of the daily electricity consumption time series data of the second user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the second user;

[0076] Based on the first user's daily electricity consumption time series growth rate value sequence and the second user's daily electricity consumption time series growth rate value sequence, the difference between the first user's daily electricity consumption time series growth rate value and the second user's daily electricity consumption time series growth rate value at the same moment is calculated, and the sum of the differences corresponding to all moments in the daily electricity consumption time series data is obtained as the similarity between the first user's daily electricity consumption time series data trend and the second user's daily electricity consumption time series data trend.

[0077] Similarly, the similarity between the weekly electricity consumption time series data trend of the first user and the weekly electricity consumption time series data trend of the second user is calculated.

[0078] It can be understood that in the embodiment of the present application, considering the similarity of the trend of the two daily power consumption time series data, it is not analyzed by the correlation calculation method such as Euclidean distance or Pearson coefficient. The Euclidean distance measures the difference in the numerical distribution of the two daily power consumption time series data, but cannot analyze whether the change trend of the two daily power consumption time series data is similar; the Pearson coefficient analyzes the change correlation of the two daily power consumption time series data, but cannot accurately measure the similarity or difference of the trend of the two daily power consumption time series data. In the embodiment of the present application, the difference in the growth rate value of the two daily power consumption time series data curves at the same time is used to measure the difference in the trend of the two daily power consumption time series data. It can be understood that in the embodiment of the present application, if the growth rate value on one of the daily power consumption time series data curves at the same time is greater than 0 (for example, 4), and the growth rate value on the other daily power consumption time series data curve is less than 0 (for example, -3), it indicates that the difference in the trend of the two daily power consumption time series data at this moment is very large (the trend difference is 7), one is an increasing trend, and the other is a decreasing trend. The difference in the growth rate value of the two daily power consumption time series data curves at the same time can be used to measure the difference in the trend of the two daily power consumption time series data.

[0079] In one embodiment, the above step 34 mines the user's electricity consumption pattern under the influence of the accompanying factors based on the user's historical electricity consumption time series data and historical accompanying factor time series data, and includes the following steps:

[0080] Step 3401, based on the historical power consumption time series data and the historical accompanying factor time series data of different users as the first data set, after training the neural network model, the power consumption pattern of the user under the action of the accompanying factors is mined, recorded as the first power consumption pattern, and a first power consumption analysis model representing the first power consumption pattern is obtained;

[0081] Step 3402, taking the first electricity consumption analysis model that characterizes the first electricity consumption pattern as the basic model, based on the historical electricity consumption time series data and the historical accompanying factor time series data of the same user as the second data set, transforming the basic model, mining the electricity consumption pattern of the same user under the influence of the accompanying factors, recorded as the second electricity consumption pattern, and obtaining the second electricity consumption analysis model that characterizes the second electricity consumption pattern.

[0082] In an embodiment of the present application, the power consumption patterns of users are mined from two data sets, wherein the historical power consumption time series data and the historical accompanying factor time series data of different users are used as the first data set, and the historical power consumption time series data and the historical accompanying factor time series data of the same user are used as the second data set; it can be understood that the first power consumption rule mined based on the first data set characterizes the power consumption of users when the accompanying factors take different values, and the correlation between the changes of different accompanying factors and the power consumption of users is mined based on the first data set, but the first power consumption rule mined based on the first data set cannot be finely matched with the power consumption habits of a single user.

[0083] In this application, a second data set is used to mine a second electricity usage pattern that is finely adapted to the electricity usage habits of a single user. In the second data set, certain accompanying factors will remain basically unchanged for a long period of time (for example, family attribute data such as the composition structure of family members and income situation), and then will change at a certain time, resulting in changes in the user's electricity usage behavior. The diversity of the sample data in the second data set is insufficient, and it is difficult to fully mine the user's second electricity usage pattern.

[0084] In an embodiment of the present application, based on a first electricity usage rule that conforms to different users and guided by a second electricity usage rule mined from a second data set with a small number of samples, the first electricity usage rule is continuously optimized to form a second electricity usage rule that conforms to the data characteristics of the second data set.

[0085] In one embodiment, the bootstrapping process may be as follows: training a neural network intelligent algorithm based on a first data set to obtain a first intelligent analysis network model for predicting user electricity consumption under the influence of various related factors; on the basis of the first intelligent analysis network model, fine-tuning the model parameters of the original first intelligent analysis network model with a second data set with less sample data to obtain an optimized second intelligent analysis network model that meets the data characteristics of the second data set, and the second intelligent analysis network model is used to characterize the second electricity consumption pattern.

[0086] In one implementation, the above step 3402, transforming the basic model to obtain a second power consumption analysis model representing the second power consumption pattern, includes:

[0087] Step 34021, based on the second sample of the second data set (the sample in the second data set is recorded as the second sample), the basic model is input to obtain the final encoding result of the basic model encoding the power consumption time series data and the accompanying factor time series data, that is, the result before the output layer of the basic model;

[0088] Step 34022, outputting a first coding transformation result based on the final coding result of the basic model through a first linear transformation matrix;

[0089] Step 34023, performing nonlinear transformation through a nonlinear activation function layer based on the first coding transformation result to obtain a nonlinear coding result;

[0090] Step 34023, based on the nonlinear coding result, a first linear inverse transformation matrix opposite to the first linear transformation matrix is ​​used to obtain a first decoding result, the final coding result of the basic model and the first decoding result are combined as a fused first decoding result, and the fused first decoding result is input into the output layer of the basic model for full connection processing to obtain the electricity theft identification result of the second sample;

[0091] Step 34024: Based on the error between the electricity theft labeling data of the second sample and the actual electricity theft identification result, the first linear transformation matrix, the nonlinear activation function layer, and the first linear inverse transformation matrix are trained to obtain a second electricity consumption analysis model.

[0092] In the embodiment of the present application, the first power consumption analysis model that characterizes the first power consumption law is used as the basic model. When mining the second power consumption law to obtain the second power consumption analysis model, the model parameters of the basic model are guaranteed to remain unchanged. On the basis of the basic model, the existing first power consumption analysis model is partially improved. In the embodiment of the present application, the encoding result of the basic model for the input data is further subjected to linear transformation, nonlinear transformation, and linear inverse transformation. The linear transformation and nonlinear transformation are combined to improve the ability to express the characteristic differences of the historical power consumption time series data and historical accompanying factor time series data of the same user compared with the historical power consumption time series data and historical accompanying factor time series data of different users. Furthermore, in the embodiment of the present application, it is only necessary to obtain the parameters of the first linear transformation matrix and the first linear inverse transformation matrix. The final encoding result of the basic model and the first decoding result are combined as the fused first decoding result to ensure that the data input to the output layer of the basic model is a fine-tuning of the final encoding result of the basic model.

[0093] In one embodiment, the above step 32, forming a class based on users not classified into the class, and determining users in the class whose number of users is less than a first preset value as candidate electricity theft users, further includes:

[0094] Step 321, based on the users in the class whose number of users is greater than the first preset value, determine the height distance between the power consumption time series data curve of each user and the power consumption time series data curves of the other users in the class, and obtain the sum of the height distances between the power consumption time series data curve of each user and the power consumption time series data curves of all other users in the class;

[0095] Step 322 , sort the sum values ​​of the height distances corresponding to each user's power consumption time series data curve from large to small, and obtain the user's power consumption time series data curve with a ranking at a front preset value as a candidate power stealing user.

[0096] In the embodiment of the present application, it is considered that the trend of the change of the electricity consumption time series data of users in a class is similar, but there are electricity stealing users who achieve the purpose of stealing electricity by reducing the actual electricity consumption by a fixed ratio based on normal electricity consumption behavior. At this time, the change trend of the electricity consumption time series curve of the electricity stealing user is the same as that of the normal user, and the only difference is the amount of electricity consumption. In the embodiment of the present application, by calculating the size of the sum of the height distances corresponding to the electricity consumption time series data curve of each user, it is determined whether the distribution position of the user's electricity consumption time series data curve is in an area where all the users' electricity consumption time series data curves are densely distributed. If the sum of the height distances corresponding to the electricity consumption time series data curve of a certain user is smaller, the distribution position of the electricity consumption time series data curve of the user is in an area where all the users' electricity consumption time series data curves are densely distributed. The height distance is determined based on the difference in the electricity consumption values ​​corresponding to the same time on the electricity consumption time series data curve of two users, and further based on the sum of the differences corresponding to all the times on the electricity consumption time series data curve as the height distance of the two users' electricity consumption time series data curves.

[0097] In one embodiment, the above step 34, mining the user's electricity usage pattern under the influence of accompanying factors based on the user's historical electricity usage time series data and historical accompanying factor time series data, includes: using the user's historical electricity usage time series data and historical accompanying factor time series data as input, training an intelligent algorithm model based on a neural network to mine the user's electricity usage pattern, and characterizing the user's electricity usage pattern based on the trained neural network model;

[0098] The neural network-based intelligent algorithm includes:

[0099] Step 3411, using the power consumption time series data and the accompanying factor time series data of a preset long period of time as the first input data,

[0100] Step 3412, obtaining the growth value of the power consumption time series data and the accompanying factor time series data in a preset long time period compared with the previous moment, obtaining the power consumption time series growth rate data and the accompanying factor growth rate data in the preset long time period, and using the power consumption time series growth rate data and the accompanying factor growth rate data as the second input data;

[0101] Step 3413, inputting the data sequence of each time series data of the first input data within a preset long period of time into a sequence feature extraction unit, extracting the time series feature of each time series data, and characterizing the time relationship feature of each time series data;

[0102] Step 3414, inputting the power consumption time series growth rate data and the accompanying factor growth rate data at the same time of the second input data into the convolutional layer unit to obtain the spatial relationship features of the power consumption time series growth rate data and the accompanying factor growth rate data;

[0103] Step 3415: Based on the temporal relationship features and spatial relationship features of the input data, the input data is input to the fully connected layer for analysis to determine whether the input data is electricity theft feature data.

[0104] In the embodiment of the present application, based on the concept of channel attention mechanism and spatial attention mechanism, the time relationship characteristics and spatial relationship characteristics of electricity consumption time series data and accompanying factor time series data are analyzed. The relationship between electricity consumption time series data and accompanying factor time series data is mined from multiple dimensions. Based on the data sequence input of each time series data in the input multi-dimensional sequence data within a preset long time period, a sequence feature extraction unit is input to extract the time series characteristics of each time series data, characterize the time relationship characteristics of each time series data, and mine the spatial relationship between electricity consumption time series growth rate data and accompanying factor growth rate data through convolution kernels based on multi-dimensional time series growth rate sequence data, and perform binary classification analysis based on time relationship characteristics and spatial relationship characteristics. In the embodiment of the present application, the sequence feature extraction unit can be implemented based on the LSTM model, and there are multiple convolution layers in the convolution layer unit, and each convolution layer includes multiple convolution kernels of different scales to extract multi-scale spatial relationship characteristics.

[0105] In another embodiment, the neural network-based intelligent algorithm further includes: converting the power consumption time series data curve and the accompanying factor time series data curve based on a preset long time period into a two-dimensional image with time as the horizontal axis, wherein the two-dimensional image includes the power consumption time series data curve and the accompanying factor time series data curve having the same horizontal axis in the same coordinate system area;

[0106] Based on the multi-dimensional time series data curve image input, the convolutional neural network recognizes and mines the change characteristics and change correlation characteristics of the electricity consumption time series data curve and the accompanying factor time series data curve, and identifies whether the distribution characteristics of the electricity consumption time series data and the accompanying factor time series data within a preset long time period in the image conform to the normal electricity consumption change characteristics; the convolutional layer in the convolutional neural network includes multiple layers, and each convolutional layer includes multiple non-one-dimensional convolution kernels.

[0107] In an embodiment of the present application, for a convolutional layer unit that obtains spatial relationship features, when the input is the electricity consumption time series growth rate data and the accompanying factor growth rate data for a preset long time period, in order to mine the spatial relationship features, in an embodiment of the present application, the electricity consumption time series growth rate data and the accompanying factor growth rate data at the same time are input into the convolutional layer unit, and the convolution kernel uses a one-dimensional convolution sliding to perform a convolution operation on the input data.

[0108] For the convolutional neural network used in the multi-dimensional time series data curve image, in order to identify whether the distribution characteristics of the power consumption time series data and the accompanying factor time series data in the preset long time period in the image meet the normal power consumption change characteristics, the convolution kernel in the convolution layer uses multiple non-one-dimensional convolution kernels, such as 3*3 convolution kernels, 5*5 convolution kernels, etc. In the embodiment of the present application, considering the time change trend of the power consumption time series data and the accompanying factor time series data for different users, they can be classified into multiple types according to the power consumption habits of different users. For example, the power consumption of office workers is relatively large in the three time periods of morning, noon, and evening after get off work during the working day, and the power consumption in the rest of the time will be relatively small (in the non-working time period, it will show a trend of first rising and then falling, and it will decrease when resting at night), while the power consumption of office workers during the weekend rest day is small. Therefore, the use of convolutional neural networks to identify the distribution characteristics of user power consumption time series data and accompanying factor time series data can identify whether it is a normal power consumption change characteristic, if not, it is power theft.

[0109] Furthermore, in an embodiment of the present application, the above-mentioned time relationship characteristics and spatial relationship characteristics based on the input data are input into the fully connected layer for analysis to determine whether the input data is characteristic data of electricity theft, and the analysis result is recorded as the first analysis result. The above-mentioned analysis result based on the multi-dimensional time series data curve image is input into the convolutional neural network to identify whether the distribution characteristics of the electricity consumption time series data and the accompanying factor time series data in a preset long time period in the image conform to the normal electricity consumption change characteristics. The analysis result is recorded as the second analysis result, and the final electricity theft analysis result is determined by a comprehensive decision based on the first analysis result and the second analysis result.

[0110] The embodiment of the present application provides an anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model, including:

[0111] A first electricity theft analysis unit, used for analyzing a first electricity theft analysis result based on line loss data in a region;

[0112] An electricity consumption data acquisition unit is used to acquire the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes time series data of electricity consumption; the accompanying factor data includes meteorological, calendar, and family attribute data accompanying the user's electricity consumption at the same time;

[0113] A second electricity theft analysis unit is used to analyze the second electricity theft analysis result using an intelligent algorithm model based on the user's electricity consumption data and accompanying factor data;

[0114] A third electricity theft analysis unit, configured to determine whether electricity theft exists based on the first electricity theft analysis result and the second electricity theft analysis result;

[0115] The electricity theft locating unit is used to analyze and locate the electricity theft location and electricity theft time period based on the line loss data in the area, the user's electricity consumption data and the accompanying factor data when it is determined that electricity theft exists.

[0116] For the specific definition of the anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model, please refer to the definition of the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model in the above text, which will not be repeated here. Each unit in the above-mentioned anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned units can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned units.

[0117] An embodiment of the present application provides an electronic device, the electronic device comprising:

[0118] processor;

[0119] a memory for storing processor-executable instructions;

[0120] The processor implements the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model as described above by running the executable instructions.

[0121] Among them, the processor may include one or more processing cores, such as a 4-core processor, an 8-core processor, etc. The processor may be implemented in at least one hardware form of digital signal processing (DSP), field programmable gate array (FPGA), and programmable logic array (PLA). The memory may include one or more computer-readable storage media, which may be non-transient. The memory may also include a high-speed random access memory, and a non-volatile memory, such as one or more disk storage devices and flash memory storage devices. In some embodiments, the electronic device may also optionally include: an input interface and an output interface. The processor, the memory, and the input interface and the output interface may be connected via a bus or a signal line. Each peripheral device may be connected to the input interface and the output interface via a bus, a signal line, or a circuit board. The input interface and the output interface may be used to connect at least one peripheral device related to input / output to the processor and the memory.

[0122] The embodiment of the present application provides a computer-readable storage medium on which computer instructions are stored, and when the instructions are executed by a processor, the steps of the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model are implemented. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage node, etc.

[0123] The present invention is not limited to the above-mentioned specific implementation modes. Various changes made by ordinary technicians in this field based on the above-mentioned concepts without creative work are all within the protection scope of the present invention.

Claims

1. An anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model, characterized in that: The steps include: Analyze the first electricity theft analysis result for the line loss data in the area; Acquire the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes the time series data of electricity consumption; the accompanying factor data includes the weather, calendar, and family attribute data accompanying the user's electricity consumption at the same time; Based on the user's electricity consumption data and accompanying factor data, the second electricity theft analysis results are analyzed using an intelligent algorithm model; Determining the existence of electricity theft based on the first electricity theft analysis result and the second electricity theft analysis result; When it is determined that electricity theft has occurred, the location and time period of electricity theft are analyzed and located based on line loss data in the area, user electricity consumption data and accompanying factor data.

2. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 1 is characterized in that: The method of analyzing the second electricity theft analysis result based on the user's electricity consumption data and accompanying factor data using an intelligent algorithm model includes: Perform trend similarity analysis based on the daily, weekly and monthly electricity consumption time series data of all users, and classify users with similar trends into one category. If the trends of the user's daily electricity consumption time series data, weekly electricity consumption time series data and monthly electricity consumption time series data are all similar, the user is classified into one category. If one of the three trends is not similar, the user is not classified into one category and is determined to be a candidate electricity theft user. A class is formed based on users who are not classified into the class, and users in the class whose number of users is less than a first preset value are determined as candidate electricity theft users; For a candidate electricity theft user, obtain the user's historical electricity consumption time series data and historical accompanying factor time series data; Based on the user's historical electricity consumption time series data and historical accompanying factor time series data, the user's electricity consumption pattern under the influence of accompanying factors is mined. Based on the power consumption time series data and the accompanying factor time series data of the candidate power stealing users in the current time period, the user rules are utilized to analyze whether the candidate power stealing users in the current time period have committed power stealing behavior.

3. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 2 is characterized in that: The trend similarity analysis includes: Calculate the growth rate value of the later moment of the daily electricity consumption time series data of the first user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the first user; calculate the growth rate value of the later moment of the daily electricity consumption time series data of the second user compared with the previous moment, and obtain the time series growth rate value sequence of the daily electricity consumption of the second user; Based on the first user's daily electricity consumption time series growth rate value sequence and the second user's daily electricity consumption time series growth rate value sequence, the difference between the first user's daily electricity consumption time series growth rate value and the second user's daily electricity consumption time series growth rate value at the same moment is calculated, and the sum of the differences corresponding to all moments in the daily electricity consumption time series data is obtained as the similarity between the first user's daily electricity consumption time series data trend and the second user's daily electricity consumption time series data trend.

4. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 2 is characterized in that: The method of mining the user's electricity consumption pattern under the influence of accompanying factors based on the user's historical electricity consumption time series data and historical accompanying factor time series data includes: Based on the historical electricity consumption time series data and the historical accompanying factor time series data of different users as the first data set, after training the neural network model, the electricity consumption pattern of the user under the action of the accompanying factors is mined, recorded as the first electricity consumption pattern, and the first electricity consumption analysis model representing the first electricity consumption pattern is obtained; Taking the first electricity consumption analysis model that characterizes the first electricity consumption pattern as the basic model, based on the historical electricity consumption time series data and the historical accompanying factor time series data of the same user as the second data set, the basic model is transformed to mine the electricity consumption pattern of the same user under the influence of the accompanying factors, which is recorded as the second electricity consumption pattern, and obtain the second electricity consumption analysis model that characterizes the second electricity consumption pattern.

5. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 4 is characterized in that: The basic model is transformed to obtain a second power consumption analysis model representing a second power consumption law, including: Based on the second sample of the second data set, the basic model is input to obtain the final encoding result of the basic model encoding the power consumption time series data and the accompanying factor time series data, that is, the result before the output layer of the basic model; Outputting a first coding transformation result based on the final coding result of the basic model through a first linear transformation matrix; Based on the first coding transformation result, a nonlinear transformation is performed through a nonlinear activation function layer to obtain a nonlinear coding result; Based on the nonlinear coding result, a first linear inverse transformation matrix opposite to the first linear transformation matrix is ​​used to obtain a first decoding result, the final coding result of the basic model and the first decoding result are combined as a fused first decoding result, and the fused first decoding result is input into the output layer of the basic model for full connection processing to obtain the electricity theft identification result of the second sample; Based on the error between the electricity theft labeling data of the second sample and the actual electricity theft identification result, the first linear transformation matrix, the nonlinear activation function layer, and the first linear inverse transformation matrix are trained to obtain the second electricity consumption analysis model.

6. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 2 is characterized in that: The method of forming a class based on users not classified into the class and determining users in the class whose number of users is less than a first preset value as candidate electricity stealing users also includes: Based on users in a class whose number of users is greater than a first preset value, determine the height distance between each user's power consumption time series data curve and the power consumption time series data curves of other users in the class, and obtain the sum of the height distances between each user's power consumption time series data curve and the power consumption time series data curves of all other users in the class; The sum of the height distances corresponding to each user's electricity consumption time series data curve is sorted from large to small, and the user's electricity consumption time series data curve ranked at a front preset value is obtained as a candidate electricity theft user.

7. The anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model according to claim 2 is characterized in that: The mining of the user's electricity usage pattern under the influence of accompanying factors based on the user's historical electricity usage time series data and historical accompanying factor time series data includes: using the user's historical electricity usage time series data and historical accompanying factor time series data as input, training an intelligent algorithm model based on a neural network to mine the user's electricity usage pattern, and characterizing the user's electricity usage pattern based on the trained neural network model; The neural network-based intelligent algorithm includes: The power consumption time series data and accompanying factor time series data of a preset long period of time are used as the first input data. Obtaining a growth value of a later moment of the power consumption time series data and the accompanying factor time series data over a preset long period of time compared to a previous moment, obtaining the power consumption time series growth rate data and the accompanying factor growth rate data of the preset long period of time, and using the power consumption time series growth rate data and the accompanying factor growth rate data as the second input data; Inputting the data sequence of each time series data of the first input data within a preset long period of time into the sequence feature extraction unit, extracting the time series feature of each time series data, and characterizing the time relationship feature of each time series data; Inputting the electricity consumption time series growth rate data and the accompanying factor growth rate data at the same time of the second input data into the convolutional layer unit to obtain the spatial relationship characteristics of the electricity consumption time series growth rate data and the accompanying factor growth rate data; Based on the temporal and spatial relationship features of the input data, the input data is input into the fully connected layer for analysis to determine whether the input data is electricity theft feature data. 8.An anti-electricity theft analysis system based on electricity consumption big data and intelligent algorithm model, characterized in that: include: A first electricity theft analysis unit, used for analyzing a first electricity theft analysis result based on line loss data in a region; An electricity consumption data acquisition unit, used to acquire the user's electricity consumption data and accompanying factor data, wherein the electricity consumption data includes time series data of electricity consumption; The accompanying factor data include weather, calendar, and household attribute data accompanying the user's electricity use; A second electricity theft analysis unit, configured to analyze a second electricity theft analysis result using an intelligent algorithm model based on the user's electricity consumption data and accompanying factor data; A third electricity theft analysis unit, configured to determine whether electricity theft exists based on the first electricity theft analysis result and the second electricity theft analysis result; The electricity theft locating unit is used to analyze and locate the electricity theft location and electricity theft time period based on the line loss data, user electricity consumption data and accompanying factor data in the area when it is determined that electricity theft exists.

9. An electronic device, characterized in that: The electronic device comprises: processor; a memory for storing processor-executable instructions; The processor implements the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model as described in any one of claims 1 to 7 by running the executable instructions.

10. A computer-readable storage medium having computer instructions stored thereon, characterized in that: When the instructions are executed by the processor, the steps of the anti-electricity theft analysis method based on electricity consumption big data and intelligent algorithm model as described in any one of claims 1 to 7 are implemented.

Images