Dynamic analysis of electricity theft and electricity violation behavior based on load curve

Abstract

The present invention relates to a method for dynamically analyzing electricity theft and electricity use violations based on load curves, including: establishing a basic database, preliminarily establishing a database based on the load curve of the user collection system, and classifying and storing according to industry type and user type; establishing a typical load curve library, and forming an industry typical load curve F1(t) based on the basic database; electricity consumption data analysis, electricity use violation analysis: according to the industry typical load curve F1(t), comparing the corresponding industry user power load real-time curve F(t) 实时 , by calculating the actual value F(t) 实时 The linear correlation degree with the estimated value F1(t) curve can be used to determine whether the user has violated the electricity usage regulations. It can more accurately screen out violations such as changes in the nature of electricity use, and provide detailed information on behaviors other than voltage loss and phase failure, such as loose screws in metering devices and electricity theft by high-tech covert means, so as to accurately investigate and deal with them and achieve proactive operation and maintenance.

Description

Technical Field

[0001] The present invention relates to the technical field of electricity consumption analysis, and in particular to a method for dynamically analyzing electricity theft and electricity violation behaviors based on load curves. Background Art

[0002] At present, it is difficult to actively track the problems of electricity use in violation of regulations and electricity theft. The only ways to investigate and deal with abnormalities are through big data screening of the electricity user collection system and manual on-site verification. The former is only applicable to simple problems such as loss of voltage, phase failure, and three-phase imbalance based on available meter data. The latter requires a lot of manpower and is time-consuming and labor-intensive. Since the amount of data carried by the electricity user collection system is too large, the analysis function is relatively weak. Active monitoring cannot be achieved for extreme situations such as changes in the nature of user electricity use and electricity use, and high-tech electricity theft. Summary of the invention

[0003] In order to solve the technical problems existing in the background technology, the present invention provides a dynamic analysis of electricity theft and breach of contract electricity use based on load curve, which can more accurately screen out breach of contract electricity use such as changes in the nature of electricity use, and provide detailed information on behaviors other than voltage loss and phase failure, such as loose screws in metering devices, and electricity theft by high-tech covert means, so as to accurately investigate and deal with them and achieve proactive operation and maintenance.

[0004] The technical solution adopted by the present invention to solve its technical problem is:

[0005] The method of dynamically analyzing electricity theft and electricity violation based on load curve includes:

[0006] (1) Establishing a basic database

[0007] Initially establish a database based on the user collection system load curve, and classify and store it according to industry type and user type;

[0008] (2) Establish a typical load curve library

[0009] Based on the basic database, the industry typical load curve F1(t) is formed;

[0010] (3) Electricity consumption data analysis

[0011] Analysis of power consumption in default: Based on the typical load curve F1(t) of the industry, compare the real-time power load curve F(t) of the corresponding industry users 实时 , by calculating the actual value F(t) 实时 The degree of linear correlation with the estimated value F1(t) curve is used to determine whether the user has violated the electricity usage regulations.

[0012] Furthermore, based on the basic database, a typical load curve F2(t) of the user is formed;

[0013] The typical user load curve F2(t) is established based on the basic data. The correlation between the real-time power load and the typical load value in the new year is calculated to obtain the correction coefficient α2. By comparing the estimated value α2F2(t) with the user load curve F(t) of the year 实时 , calculate the actual value F(t) 实时 The linear correlation degree with the estimated value α2F2(t) is used to determine whether the user has committed electricity theft.

[0014] Furthermore, the industry typical load curve F1(t) needs to be subdivided into types of electricity consumption.

[0015] Furthermore, the user's typical load curve F2(t) needs to be fitted with the user's load curve at each time point over several years, and the user's typical load curve F2(t) is formed based on the data.

[0016] Furthermore, in each subsequent year, the user's typical load curve F2(t) is corrected according to actual electricity consumption to obtain α2F2(t).

[0017] Furthermore, the user's typical load curve F2(t) needs to fit the user's load curve at each time point within 3-5 years.

[0018] Beneficial effects of the present invention:

[0019] The present invention can preliminarily establish a user load curve database, and by forming a typical load curve, it can more accurately screen out electricity use violations such as changes in the nature of electricity use, and details of behaviors other than voltage loss and phase failure, such as loose screws in metering devices, and electricity theft by high-tech covert means, so as to accurately investigate and deal with them and achieve proactive operation and maintenance. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0021] Figure 1 is a typical load curve;

[0022] Figure 2 is the load curve of school 1;

[0023] Figure 3 is the load curve of school 2;

[0024] Figure 4 is a typical load estimation curve;

[0025] Figure 5 It is the actual power load curve. DETAILED DESCRIPTION

[0026] The present invention is further described in detail below with reference to the accompanying drawings.

[0027] (1) Establish a basic database: Preliminarily establish a database based on the user's collection system load curve, and classify and store it according to industry type and electricity consumption nature;

[0028] (2) Establish a typical load curve library:

[0029] Based on the basic database, the first step is to form a typical industry load curve F1(t), such as the electricity load curve for agricultural irrigation, the electricity load curve for school education, the electricity load curve for product processing plants, etc. The load curve needs to be subdivided into small categories of electricity consumption. For example, school electricity consumption can be divided into kindergartens, elementary schools, junior high schools, high schools, and universities. Kindergartens can be further divided into boarding and day schools. The more targeted the subdivision is, the more accurate the typical load value curve library of the industry will be.

[0030] The second is to form the user's typical load curve F2(t). This curve needs to fit the user's load curve at each time point within 3-5 years, and form the user's typical load curve F2(t) based on this data. In each subsequent year, the typical load curve is corrected according to actual electricity consumption to obtain α2F2(t).

[0031] (3) Electricity consumption data analysis

[0032] a Analysis of power consumption in violation of contract: Based on the typical load curve F1(t) of the industry, compare the real-time power load curve F(t) of the corresponding industry users within the jurisdiction of the power supply station. 实时 , by calculating the actual value F(t) 实时 The degree of linear correlation with the estimated value F1(t) reflects whether the user has violated the electricity use rules. Example: Assuming that the typical load curve of the Jinqueshan District Day Primary School on April 23 and the actual load curves of the two day primary schools in the street are shown in the attached figure, by calculating the correlation between the load curves of School 1 and School 2 and the typical load curve of the area at that time, it is found that the load curve of School 1 is highly linearly related to the typical load curve, and the load curve of School 2 is not correlated with the typical load curve of the area, that is, the difference is large.

[0033] When calculating the correlation between the power load curve and the typical load curve, it is generally necessary to find a coefficient or formula to determine the fluctuation.

[0034] In actual operation, the correlation coefficient is generally used. The correlation coefficient is a common indicator for judging the strength of the linear relationship between two variables, and can also be used to judge the degree of correlation between two curves. The calculation formula is as follows: r = Cov(X,Y) / (sd(X)*sd(Y)) where Cov(X,Y) represents the covariance between X and Y, and sd(X) and sd(Y) represent the standard deviations of X and Y, respectively. When the value of r is close to 1 or -1, it means that there is a strong positive or negative correlation between the two curves; when the value of r is close to 0, it means that there is no linear correlation between the two curves.

[0035] Neural network methods can also be used. Neural network is an algorithm based on artificial neural network, which can be used to build nonlinear models and predict future results. In power load forecasting, neural network can be used to build complex load forecasting models and improve forecasting accuracy.

[0036] In the specific implementation, the historical load data must first be processed and divided into a training set and a test set. The training set is used to train the neural network, and the test set is used to verify the prediction ability of the model.

[0037] When building a neural network model, it is necessary to determine parameters such as network structure, activation function, loss function, etc. Different network structures and parameters can be selected for adjustment and optimization according to actual conditions.

[0038] When training a neural network, a back-propagation algorithm is usually used to continuously adjust the weights and bias terms in the network so that the objective function (such as the mean square error) is gradually reduced, thereby optimizing the model's predictive ability.

[0039] After training, the neural network can be used for load forecasting. By inputting the current external factors, the network will produce a predicted value to predict future load.

[0040] Through the above calculations, it can be determined that the electricity usage of School 1 is normal, while the electricity usage of School 2 is abnormal. It is necessary to investigate on site whether there is any violation of the electricity usage regulations.

[0041] b. Electricity theft analysis: Based on the basic data, the typical load curve F2(t) of the user is established. The correlation between the real-time power load and the typical load value in the new year is calculated to obtain the correction coefficient α2. By comparing the estimated value α2F2(t) with the user load curve F(t) of the year, 实时 , calculate the actual value F(t) 实时The degree of linear correlation with the estimated value α2F2(t) reflects whether the user has electricity theft or abnormal electricity use behavior. Example: Assuming that the typical load estimation curve of Jinqueshan Primary School on April 23 and the actual load curve of the primary school on that day are shown in the attached figure, through the correlation between the actual load curve and the typical load estimation curve, it is found that the actual load curve of the user is highly different from the typical load estimation curve. It is necessary to confirm on site whether there is any violation of electricity use.

[0042] Here, it is also necessary to find a coefficient or formula to determine the deviation. The method and formula used to calculate the correlation between the power load curve and the typical load curve can be referred to.

[0043] Based on the above ideal embodiments of the present invention, the relevant staff can make various changes and modifications without departing from the technical concept of the present invention through the above description. The technical scope of the present invention is not limited to the contents of the specification, and its technical scope must be determined according to the scope of the claims.

Claims

1. A method for dynamically analyzing electricity theft and electricity violation behaviors based on load curves, characterized in that: include: (1) Establishing a basic database Initially establish a database based on the user collection system load curve, and classify and store it according to industry type and user type; (2) Establish a typical load curve library Based on the basic database, the industry typical load curve F1(t) is formed; (3) Electricity consumption data analysis Analysis of power consumption in default: Based on the typical load curve F1(t) of the industry, compare the real-time power load curve F(t) of the corresponding industry users 实时 , by calculating the actual value F(t) 实时 The linear correlation degree with the estimated value F1(t) curve is used to judge whether the user has violated the electricity usage contract. Based on the basic database, the user's typical load curve F2(t) is formed; The typical user load curve F2(t) is established based on the basic data. The correlation between the real-time power load and the typical load value in the new year is calculated to obtain the correction coefficient α2. By comparing the estimated value α2F2(t) with the user load curve F(t) of the year 实时 , calculate the actual value F(t) 实时 The linear correlation degree with the estimated value α2F2(t) is used to determine whether the user has committed electricity theft.

2. The method for dynamically analyzing electricity theft and electricity violation based on load curve according to claim 1 is characterized in that: The industry's typical load curve F1(t) needs to be subdivided into different types of electricity consumption.

3. The method for dynamically analyzing electricity theft and electricity violation based on load curve according to claim 1 is characterized in that: The user's typical load curve F2(t) needs to fit the user's load curve at each time point in a few years, and the user's typical load curve F2(t) is formed based on the data.

4. The method for dynamically analyzing electricity theft and electricity violation based on load curve according to claim 3 is characterized in that: In each subsequent year, the user's typical load curve F2(t) will be corrected according to actual electricity consumption to obtain α2F2(t).

5. The method for dynamically analyzing electricity theft and electricity violation based on load curve according to claim 3 is characterized in that: The user's typical load curve F2(t) needs to fit the user's load curve at each time point within 3-5 years.

Images