Electricity stealing detection method and device based on neural network model

Abstract

The application provides a kind of electric larceny detection method and device based on neural network model, it is suitable for electric power system technical field, including: the division is carried out to the region to be detected and obtains multiple detection regions;Obtain the historical power consumption data of each detection region, train neural network model, obtain power consumption prediction model;For each detection region, obtain the type of each power consumer and the proportion of different types of power consumers, determine the maximum allowable power fluctuation value of the detection region based on the proportion, predict power consumption for the detection region, compare the predicted power consumption data with the actual power consumption data to obtain the actual fluctuation value;Determine the abnormal area and the similar area of each abnormal area, screen the abnormal area based on the similar area, and determine the electric larceny area.The application predicts power consumption by power consumption prediction model, determines abnormal area, and determines electric larceny area through similar area, which ensures the accuracy of electric larceny area and effectively reduces the investigation range of electric larceny.

Description

Technical Field

[0001] This invention belongs to the field of power system technology, and in particular relates to a method and device for detecting electricity theft based on a neural network model. Background Technology

[0002] Electricity theft is extremely harmful to society, affecting the stable supply of electricity. More serious cases of electricity theft may also damage power equipment during the theft process, which in turn affects the safety of electricity supply and people's lives.

[0003] Existing methods for detecting electricity theft mostly analyze the power grid's operational status and then conduct manual on-site verification when theft is detected to confirm the user committing the theft. The inventors found in practice that this method cannot accurately account for the influence of various factors and relies heavily on human experience to determine whether electricity theft has occurred, resulting in low accuracy. Furthermore, existing methods have a large detection range, making it difficult to pinpoint the exact location of the theft when it occurs. Summary of the Invention

[0004] This invention provides a method and apparatus for detecting electricity theft based on a neural network model, in order to solve the problem that existing electricity theft detection methods have a large detection range and are not easy to pinpoint the specific location of electricity theft.

[0005] This invention is achieved through the following technical solution:

[0006] In a first aspect, embodiments of the present invention provide a method for detecting electricity theft based on a neural network model, comprising:

[0007] The area to be tested is divided into multiple testing zones;

[0008] Historical electricity consumption data for each detection area is acquired, and a neural network model is trained based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area.

[0009] For each detection area, obtain the type of each electricity user in the detection area and the proportion of different types of electricity users in the detection area, and determine the maximum allowable electricity fluctuation value for the detection area based on the proportion;

[0010] For each detection area, electricity consumption is predicted based on the electricity consumption prediction model. The predicted electricity consumption data of the detection area is compared with the actual electricity consumption data to obtain the actual fluctuation value.

[0011] Based on all actual fluctuation values, abnormal areas are identified, similar areas for each abnormal area are identified, and the corresponding abnormal areas are screened based on the actual electricity consumption data of the similar areas to identify areas where electricity is stolen.

[0012] In one possible implementation, the area to be detected is divided into multiple detection regions, including:

[0013] The number of transformer substations and their coverage area within the area to be detected are obtained, and the area to be detected is divided based on the number of transformer substations and their coverage area to obtain multiple detection areas.

[0014] Each detection area contains one or more testing zones.

[0015] In one possible implementation, acquiring historical electricity consumption data for each detection area and training a neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area includes:

[0016] A neural network model is constructed, historical electricity consumption data for each detection area is obtained, a training set and a validation set are obtained based on the historical electricity consumption data, and the neural network model is trained based on the training set to obtain an initial electricity consumption prediction model.

[0017] The initial electricity consumption prediction model is validated using time series cross-validation based on the validation set, and the electricity consumption prediction model for the corresponding detection area is determined based on the validation results.

[0018] In one possible implementation, the step of obtaining the type of each electricity user in each detection area and the proportion of different types of electricity users in the detection area, and determining the maximum allowable electricity fluctuation value for the detection area based on the proportion, includes:

[0019] For each detection area, perform the following steps:

[0020] Obtain the type of each electricity user within the detection area;

[0021] Obtain the proportion of each type of electricity user, and the historical electricity consumption data of each type of electricity user within a preset historical period;

[0022] Based on the historical electricity consumption data, determine the electricity consumption fluctuation value and total electricity consumption for each type of electricity user;

[0023] Substitute the electricity consumption fluctuation value and total electricity consumption of each type of electricity user in the detection area into the electricity consumption fluctuation prediction formula for the detection area to calculate the maximum allowable electricity consumption fluctuation value for the detection area.

[0024] The formula for predicting electricity consumption fluctuations is:

[0025]

[0026] Where, ki α represents the maximum allowable power fluctuation value for the i-th detection area. w α represents the temperature weight of the i-th detection region. m ΔW represents the economic weight of the i-th detection region. i Δm represents the temperature fluctuation coefficient of the i-th detection area. i E represents the economic volatility coefficient of the i-th detection region. i p represents the total electricity consumption in the historical electricity consumption data of the i-th detection area. j β represents the proportion of electricity users of the j-th type in the i-th detection area. j γ represents the influence factor of the first influencing factor of the j-th type in the i-th detection region. j Let represent the influence factor of the second influencing factor of the j-th type in the i-th detection region, and let represent the first influence value of the j-th type in the i-th detection region, ΔE(2). j This represents the second influence value of the j-th type in the i-th detection region.

[0027] In one possible implementation, the step of predicting electricity consumption for each detection area based on an electricity consumption prediction model, and comparing the predicted electricity consumption data of the detection area with the actual electricity consumption data to obtain the actual fluctuation value, includes:

[0028] For each detection area, perform the following steps:

[0029] Based on the historical electricity consumption data of the detection area and the electricity consumption prediction model, the electricity consumption of the detection area is predicted to obtain the predicted electricity consumption data within the preset time period.

[0030] Obtain actual electricity consumption data within a preset time period, compare the predicted electricity consumption data with the actual electricity consumption data, and obtain the actual fluctuation value based on the comparison result;

[0031] The historical electricity consumption data and the predicted electricity consumption data both include the electricity consumption of the detection area in different time periods and the line loss data of the detection area.

[0032] In one possible implementation, comparing the preset electricity consumption data with the actual electricity consumption data, and obtaining the actual fluctuation value based on the comparison result, includes:

[0033] For each detection area, the preset power consumption data, the power consumption in different time segments of the actual power consumption data, and the line loss data of the detection area are input into the actual fluctuation value calculation formula to obtain the actual fluctuation value.

[0034] The formula for calculating the actual fluctuation value is:

[0035]

[0036] Where, k i ′ represents the actual fluctuation value of the i-th detection region, t represents the t-th time segment, s(1) it Let s(2) represent the actual electricity consumption of the i-th detection area in the t-th time period. it Let represent the predicted electricity consumption of the i-th detection area in the t-th time period, and δ1 represent the line loss correction coefficient of the i-th detection area in the t-th time period, x(1) it Let x(2) represent the actual line loss of the i-th detection area in the t-th time period. it This represents the preset line loss of the i-th detection area in the t-th time period.

[0037] In one possible implementation, determining the abnormal region based on all actual fluctuation values ​​includes:

[0038] The actual fluctuation value of each detection area is compared with the corresponding maximum allowable power fluctuation value. If the actual fluctuation value exceeds the maximum allowable power fluctuation value, the detection area is identified as an abnormal area.

[0039] Otherwise, the detection area will be classified as a normal area.

[0040] In one possible implementation, the step of acquiring similar regions for each abnormal region and filtering the corresponding abnormal regions based on the actual electricity consumption data of the similar regions to determine the electricity theft areas includes:

[0041] Acquire feature data for each detection region, calculate the similarity between each abnormal region and the remaining detection regions based on the feature data, and determine the similar regions for each abnormal region based on the similarity.

[0042] For each abnormal region, perform the following steps:

[0043] Obtain the actual electricity consumption data of similar areas in the abnormal area, and compare the actual electricity consumption data of the abnormal area with the actual electricity consumption data of the corresponding similar areas;

[0044] Calculate the difference between the actual electricity consumption data of the abnormal area and the actual electricity consumption data of the corresponding similar area for each data point, and compare each difference with the corresponding threshold. Based on the comparison results, determine whether the abnormal area is an area of ​​electricity theft.

[0045] In one possible implementation, the method further includes:

[0046] Based on the electricity theft detection model, each electricity user in each electricity theft area is monitored to identify suspicious electricity users who may be involved in electricity theft.

[0047] The user ID of each suspicious electricity user is sent to the electricity theft alarm system, and staff are notified to conduct on-site verification.

[0048] Secondly, embodiments of the present invention provide an electricity theft detection device based on a neural network model, comprising:

[0049] The segmentation module is used to divide the area to be detected into multiple detection zones;

[0050] The training module is used to acquire historical electricity consumption data for each detection area and train the neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area.

[0051] The determination module is used to obtain the type of each electricity user in each detection area and the proportion of different types of electricity users in the detection area for each detection area, and determine the maximum allowable electricity fluctuation value for the detection area based on the proportion;

[0052] The comparison module is used to predict the electricity consumption of each detection area based on the electricity consumption prediction model, and compare the predicted electricity consumption data of the detection area with the actual electricity consumption data to obtain the actual fluctuation value.

[0053] The filtering module is used to determine abnormal areas based on all actual fluctuation values, determine similar areas for each abnormal area, and filter the corresponding abnormal areas based on the actual electricity consumption data of the similar areas to determine the electricity theft areas.

[0054] This invention provides a method and apparatus for detecting electricity theft based on a neural network model. The method involves dividing the area to be detected into multiple detection zones, training the neural network model using historical electricity consumption data from each zone to obtain an electricity consumption prediction model, predicting electricity consumption data, comparing the predicted data with actual electricity consumption data to obtain fluctuation values, determining the proportion of different types of electricity users in each detection zone, and determining the maximum allowable fluctuation value. Anomaly zones are identified by comparing these fluctuation values ​​with actual fluctuation values, and the theft zones are further verified based on similar zones. Determining the maximum allowable fluctuation value through proportions and verifying with similar zones ensures the accuracy of identifying theft zones, effectively narrowing the scope of electricity theft detection. Attached Figure Description

[0055] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0056] Figure 1 This is a flowchart illustrating a method for detecting electricity theft based on a neural network model, provided in an embodiment of the present invention.

[0057] Figure 2 This is an initial region partitioning map of an electricity theft detection method based on a neural network model provided in an embodiment of the present invention;

[0058] Figure 3 This is a region partitioning map of an electricity theft detection method based on a neural network model provided in an embodiment of the present invention;

[0059] Figure 4 This is a schematic diagram of the structure of an electricity theft detection device based on a neural network model provided in an embodiment of the present invention. Detailed Implementation

[0060] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of the invention. However, those skilled in the art will understand that the invention can be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods are omitted so as not to obscure the description of the invention with unnecessary detail.

[0061] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0062] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0063] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [described condition or event] is detected" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once [described condition or event] is detected," or "in response to detection of [described condition or event]."

[0064] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0065] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0066] Figure 1 This is a flowchart illustrating a method for detecting electricity theft based on a neural network model according to an embodiment of the present invention. (Refer to...) Figure 1 The following is a detailed description of this electricity theft detection method based on a neural network model:

[0067] S110 divides the area to be tested into multiple testing zones.

[0068] In an optional embodiment, the area to be detected is divided into multiple detection regions, including:

[0069] Obtain the number of transformer substations and their coverage area within the area to be tested, and divide the area to be tested based on the number of transformer substations and their coverage area to obtain multiple testing areas;

[0070] Each detection area contains one or more testing zones.

[0071] The specific division method is predetermined, such as... Figure 2 As shown, if there are 20 transformer substations in area A to be monitored, and each substation covers 50-80 households, then for ease of monitoring, area A can be divided into 5 initial areas based on its area. These initial areas can then be adjusted based on the coverage range of the transformer substations within them. For example, initial area A1 contains 3 complete transformer substations (substations 11, 12, and 13) and a portion of the users in 4 incomplete substations (substations 14, 15, 16, and 17). During the division, the number of households in the 4 incomplete substations within initial area A1 can be determined. If the number of households in an incomplete substation within initial area A1 exceeds 60% of the total number of households covered by that substation, then that substation is assigned to initial area A1; otherwise, it is assigned to another monitoring area.

[0072] For example, one of these four incomplete transformer substations (15) covers a total of 80 electricity users, with 67 of them within the initial area 1 and only 13 outside of it. However, the number of electricity users within the initial area 1 in the other substations (substations 14, 16, and 17) does not exceed 60% of the total number of electricity users covered by the corresponding substation. Therefore, it can be as follows: Figure 3 The initial area 1 is adjusted so that it does not contain the other three incomplete transformer areas, and the remaining 13 electricity users of this transformer area (15) are assigned to the area of ​​the initial area 1 to obtain the detection area 1. The area to be detected can also be divided in other ways. The specific division method is not limited here, but for the convenience of monitoring, each detection area after the division contains complete transformer areas.

[0073] By dividing the area to be tested into multiple testing zones based on the number of transformer substations and their coverage area, the potential scope of electricity theft is effectively reduced.

[0074] S120: Obtain historical electricity consumption data for each detection area, and train the neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area.

[0075] Optionally, historical electricity consumption data may include the total electricity consumption of each user within the detection range, electricity consumption at different time periods, harmonic distortion rate, records of solar power generation, abnormal electricity consumption status, and the magnitude and power consumption data of current, voltage, and power consumption at each historical moment.

[0076] In training the neural network model, in addition to historical electricity consumption data, it is also necessary to detect the weather data, economic data, and activity data corresponding to each historical electricity consumption data point in the region.

[0077] In this embodiment, the electricity consumption prediction model is used to predict future electricity consumption data for the corresponding detection area.

[0078] In an optional embodiment, historical electricity consumption data for each detection area is acquired, and a neural network model is trained based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area, including:

[0079] A neural network model is constructed, historical electricity consumption data for each detection area is obtained, a training set and a validation set are obtained based on the historical electricity consumption data, and the neural network model is trained based on the training set to obtain an initial electricity consumption prediction model.

[0080] The initial electricity consumption prediction model was validated using time series cross-validation based on the validation set, and the electricity consumption prediction model for the corresponding detection area was determined based on the validation results.

[0081] The training set and validation set are divided according to the chronological order. For example, if the historical electricity consumption data includes electricity consumption data from January 1, 2022 to December 31, 2022, then the electricity consumption data from January 1, 2022 to November 1, 2022 can be used as the training set, and the remaining electricity consumption data can be used as the validation set.

[0082] By using time series cross-validation to divide historical electricity consumption data into training and validation sets, and then training the neural network model on the training set to obtain the initial electricity consumption prediction model, the initial electricity consumption prediction model is validated through the validation set, which can ensure the accuracy of the electricity consumption prediction model.

[0083] S130: For each detection area, obtain the type of each electricity user in the detection area and the proportion of different types of electricity users in the detection area, and determine the maximum allowable electricity fluctuation value of the detection area based on the proportion.

[0084] In this embodiment, the types of electricity users can be urban residential electricity use, industrial electricity use, agricultural electricity use, commercial electricity use, rural residential electricity use, school electricity use, etc., or they can be divided into more detailed categories. The more detailed the classification, the more accurate the maximum allowable electricity fluctuation value will be.

[0085] For example, in testing area 1, there are a total of 360 electricity users, of which 200 are urban residents, 30 are schools, 50 are businesses, and 80 are industries. Therefore, the proportion of urban residents using electricity is 55.6%, the proportion of schools using electricity is 8.3%, the proportion of businesses using electricity is 13.9%, and the proportion of industries using electricity is 22.2%.

[0086] In an optional embodiment, for each detection area, the type of each electricity user in that detection area and the proportion of different types of electricity users in that detection area are obtained, and the maximum allowable electricity fluctuation value for that detection area is determined based on the proportion, including:

[0087] For each detection area, perform the following steps:

[0088] Obtain the type of each electricity user within the detection area;

[0089] Obtain the proportion of each type of electricity user, as well as the historical electricity consumption data of each type of electricity user within a preset historical period;

[0090] Based on historical electricity consumption data, determine the total electricity consumption for each type of electricity user;

[0091] Substitute the total electricity consumption of each type of electricity user in the detection area into the electricity fluctuation prediction formula for the detection area to calculate the maximum allowable electricity fluctuation value for the detection area.

[0092] The formula for predicting electricity consumption fluctuations is:

[0093]

[0094] Where, k i α represents the maximum allowable power fluctuation value for the i-th detection area. w α represents the temperature weight of the i-th detection region. m △W represents the economic weight of the i-th detection region. i Δm represents the temperature fluctuation coefficient of the i-th detection area. i E represents the economic volatility coefficient of the i-th detection region. i p represents the total electricity consumption in the historical electricity consumption data of the i-th detection area. j β represents the proportion of electricity users of the type in the i-th detection area. j γ represents the influence factor of the first influencing factor of the j-th type in the i-th detection region. j Let represent the influence factor of the second influencing factor of the i-th type in the i-th detection region, and let represent the first influence value of the j-th type in the i-th detection region, ΔE(2). j This represents the second influence value of the j-th type in the i-th detection region.

[0095] In this embodiment, the preset historical time period refers to the time period during which electricity theft detection needs to be carried out. For example, if the electricity consumption of detection area 1 is to be detected from August 1, 2023 to August 7, 2023, the preset historical time period can be from August 1, 2022 to August 7, 2022 and from August 1, 2021 to August 7, 2021. It is necessary to obtain the historical electricity consumption data of each type of electricity user from August 1, 2022 to August 7, 2022 and from August 1, 2021 to August 7, 2021.

[0096] The first and second influencing factors represent events that will affect electricity consumption data. These can include natural disasters, holidays, large-scale events, etc. Each type of electricity user can have 0-2 influencing factors. For example, for schools in monitoring area 1, the influencing factors could be holidays and large-scale events. Alternatively, if no events affecting electricity consumption occurred for urban residents in monitoring area 1 during a certain period, then there are no influencing factors.

[0097] The first and second impact values ​​represent the potential impact of influencing factors on electricity consumption. The potential impact value is a prediction obtained by forecasting the impact on electricity consumption when such an impact occurs in the past. For example, if there are three schools in monitoring area 1, and school A is going to hold activity D, while schools B and C are not holding any activities, and school A has held activity D many times before, then the impact value of monitoring area 1 can be determined based on the historical data of school A holding activity D.

[0098] The impact factor refers to the proportion of the average electricity consumption of the affected part to the total electricity consumption of that type of electricity user. For example, if the electricity consumption of school A is X% on a normal basis, then the impact factor of school A is X%.

[0099] By classifying each electricity user into different types and determining the proportion of different types of electricity users within the detection area, and further determining the maximum allowable electricity fluctuation value based on the proportion, the accuracy of the maximum allowable electricity fluctuation value is effectively improved by taking into account the characteristics of different electricity users.

[0100] S140, For each detection area, electricity consumption is predicted based on the electricity consumption prediction model, and the predicted electricity consumption data of the detection area is compared with the actual electricity consumption data to obtain the actual fluctuation value.

[0101] Specifically, for each detection area, the data of that detection area is input into the electricity consumption prediction model, which can realize the electricity consumption prediction for that detection area and obtain the predicted electricity consumption data for that detection area.

[0102] The actual electricity consumption data refers to the electricity consumption data recorded in the detection area within a certain time period.

[0103] For example, to obtain the actual fluctuation value of region A from Monday to Sunday of the previous week, we can first obtain the predicted electricity consumption data of region A from the electricity consumption forecasting model, and then collect the actual electricity consumption data of region A during that period. Comparing the predicted electricity consumption data with the actual electricity consumption data will give us the actual fluctuation value.

[0104] In an optional embodiment, for each detection area, electricity consumption is predicted based on an electricity consumption prediction model. The predicted electricity consumption data for the detection area is compared with the actual electricity consumption data to obtain the actual fluctuation value, including:

[0105] For each detection area, perform the following steps:

[0106] Based on the historical electricity consumption data of the detection area and the electricity consumption prediction model, the electricity consumption of the detection area is predicted to obtain the predicted electricity consumption data within the preset time period.

[0107] Obtain actual electricity consumption data within a preset time period, compare the predicted electricity consumption data with the actual electricity consumption data, and obtain the actual fluctuation value based on the comparison results;

[0108] The historical electricity consumption data and the predicted electricity consumption data both include the electricity consumption of the detection area in different time periods and the line loss data of the detection area.

[0109] In an optional embodiment, preset electricity consumption data and actual electricity consumption data are compared, and the actual fluctuation value is obtained based on the comparison result, including:

[0110] For each detection area, the preset electricity consumption data and the electricity consumption data in different time periods of the actual electricity consumption data, as well as the line loss data of the detection area, are input into the actual fluctuation value calculation formula to obtain the actual fluctuation value;

[0111] The formula for calculating the actual fluctuation value is:

[0112]

[0113] Where, k i ′ represents the actual fluctuation value of the i-th detection region, t represents the t-th time segment, s(1) it Let s(2) represent the actual electricity consumption of the i-th detection area in the t-th time period. it Let represent the predicted electricity consumption of the i-th detection area in the t-th time period, and δ1 represent the line loss correction coefficient of the i-th detection area in the t-th time period, x(1) it Let x(2) represent the actual line loss of the i-th detection area in the t-th time period. it This represents the preset line loss of the i-th detection area in the t-th time period.

[0114] In this embodiment, the preset time period is a past time period. For example, if the current date is August 8, 2023, then the preset time period can only be a period before August 8, 2023, and this period does not include August 1, 2023. It can be from August 1, 2023 to August 7, 2023.

[0115] Optionally, the actual fluctuation value is the fluctuation value of electricity consumption, which is obtained by combining the fluctuation value of electricity consumption with line loss data.

[0116] The division of time segments is illustrated with an example. If the preset time period is from August 1, 2023 to August 7, 2023, then the first time segment is 0:00-1:00 AM from August 1, 2023 to August 7, 2023; the second time segment is 1:00-2:00 AM from August 1, 2023 to August 7, 2023; the third time segment is 2:00-3:00 AM from August 1, 2023 to August 7, 2023, and so on, resulting in 24 time segments. The right endpoint of each time segment is included within that time segment.

[0117] By dividing the data into time segments, the electricity consumption and line loss data for each time segment are obtained, and the fluctuation value of each time segment is calculated. The fluctuation values ​​of each time segment are then superimposed to obtain the actual fluctuation value. This fully takes into account the different electricity consumption patterns in each time period, and the difference in line loss is used to correct the fluctuation value of electricity consumption, thus ensuring the accuracy of the actual fluctuation value.

[0118] S150, based on all actual fluctuation values, determines abnormal areas, determines similar areas for each abnormal area, and filters the corresponding abnormal areas based on the actual electricity consumption data of similar areas to determine the electricity theft areas.

[0119] The abnormal area refers to the detection area where the actual fluctuation value is greater than the maximum allowable fluctuation value of electricity consumption.

[0120] In this embodiment, the similarity region between the abnormal area and the abnormal area is calculated based on data such as the number of transformer substations, area, proportion of different types of electricity users, geographical relationship with the abnormal area, and average daily electricity consumption in historical electricity consumption data over a certain period of time. The similarity is then ranked from highest to lowest, and the region with the highest similarity is selected. For example, if the abnormal area is detection area 1, and the remaining areas are detection areas 2 and 3, the similarity between detection area 1 and detection area 2, and between detection area 1 and detection area 3, is calculated. If the similarity between detection area 1 and detection area 2 is 78%, and the similarity between detection area 1 and detection area 3 is 80%, then detection area 3 is selected as the similarity region of detection area 1. A threshold can also be set when selecting the similarity region; only areas exceeding the threshold can be considered similar regions to the abnormal area. Other methods can also be used to select the similarity region; this solution does not limit this approach.

[0121] In an optional embodiment, anomaly regions are determined based on all actual fluctuation values, including:

[0122] The actual fluctuation value of each detection area is compared with the corresponding maximum allowable power fluctuation value. If the actual fluctuation value exceeds the maximum allowable power fluctuation value, the detection area is identified as an abnormal area.

[0123] Otherwise, the detection area will be classified as a normal area.

[0124] In an optional embodiment, similar regions are obtained for each abnormal region, and the corresponding abnormal regions are filtered based on the actual electricity consumption data of the similar regions to determine the electricity theft areas, including:

[0125] Acquire feature data for each detection region, calculate the similarity between each abnormal region and the remaining detection regions based on the feature data, and determine the similar regions for each abnormal region based on the similarity.

[0126] For each abnormal region, perform the following steps:

[0127] Obtain the actual electricity consumption data of similar areas in the abnormal area, and compare the actual electricity consumption data of the abnormal area with the actual electricity consumption data of the corresponding similar areas;

[0128] Calculate the difference between the actual electricity consumption data of the abnormal area and the actual electricity consumption data of the corresponding similar area for each data point, and compare each difference with the corresponding threshold. Based on the comparison results, determine whether the abnormal area is an area of ​​electricity theft.

[0129] The characteristic data of the detection area includes the number of transformer substations, area, proportion of different types of electricity users, geographical relationship with abnormal areas, and average daily electricity consumption in historical electricity consumption data over a certain period of time. Other data may also be included, without any restrictions.

[0130] In this embodiment, the actual power consumption data of the abnormal area and the actual power consumption data of the similar area can include: total power consumption, power consumption at different time periods, harmonic distortion rate, solar power generation records, abnormal power consumption status, and data such as the magnitude of current and voltage and power consumption at each historical moment. Other data can also be included, and this solution does not limit them.

[0131] Optionally, an example illustrates the specific solution of this embodiment: For an abnormal region N, if data such as electricity consumption, peak-valley electricity price, electricity efficiency, and voltage stability are used to determine the abnormal region, and if two or three differences exceed a threshold, the abnormal region is designated as an electricity theft region. This requires calculating the difference h1 between the actual electricity consumption data of the abnormal region N and the actual electricity consumption data of the corresponding similar region; calculating the difference h2 between the peak-valley electricity price data of the abnormal region N and the peak-valley electricity price data of the corresponding similar region; and calculating the difference h3 between the actual electricity consumption data of the abnormal region N and the actual electricity consumption data of the corresponding similar region, resulting in three differences h1, h2, and h3. If h1 and h2 exceed the threshold, the abnormal region N is designated as an electricity theft region. If only h1 exceeds the threshold, the abnormal region N is corrected to a normal region; the specific correction method is not limited here.

[0132] Anomalies are identified by comparing the actual fluctuation values ​​of the detected areas with the corresponding maximum permissible fluctuation values. Similar areas are then calculated for each anomaly using the characteristic data of each detected area. The actual electricity consumption data of the anomaly areas is compared with the actual electricity consumption data of the corresponding similar areas to determine the areas where electricity theft occurs. Since it is extremely rare in reality for both anomalies and their similar areas to have electricity thieves, and the difference in the amount of electricity stolen by these thieves is very small and almost negligible, correcting for anomalies by using similar areas ensures the accuracy of the detection results in situations where unforeseen circumstances are not considered in the electricity consumption fluctuation calculations.

[0133] In an optional embodiment, the method further includes:

[0134] Based on the electricity theft detection model, each electricity user in each electricity theft area is monitored to identify suspicious electricity users who may be involved in electricity theft.

[0135] The user ID of each suspicious electricity user is sent to the electricity theft alarm system, and staff are notified to conduct on-site verification.

[0136] This method effectively identifies areas prone to electricity theft by dividing the region into multiple detection zones and acquiring historical electricity consumption data for each zone. Based on this data, a neural network model is trained to predict electricity consumption for each zone. The proportion of different types of electricity users in each zone is then determined, and the maximum allowable fluctuation value is calculated accordingly. The predicted and actual electricity consumption data are compared to obtain the actual fluctuation value. Finally, based on the actual electricity consumption data of similar areas within each anomaly zone, anomaly zones are excluded to identify areas prone to electricity theft. Electricity theft detection is then performed on each user within these zones to identify potential offenders, and staff are notified to conduct on-site verification. This method effectively identifies areas where electricity theft may occur without requiring the entire region to be monitored. Furthermore, the method considers the fluctuation impact of different types of electricity users during the identification process, ensuring the accuracy of anomaly zone identification. The method of excluding anomaly zones based on similar areas eliminates the influence of other unconsidered factors.

[0137] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0138] Corresponding to the above embodiment, a method for detecting electricity theft based on a neural network model, Figure 4 The diagram shows a schematic of a power theft detection device based on a neural network model provided by an embodiment of the present invention. For ease of explanation, only the parts related to the embodiment of the present invention are shown.

[0139] See Figure 4 An embodiment of the present invention provides a method for detecting electricity theft based on a neural network model, which may include:

[0140] The segmentation module 41 is used to segment the area to be detected, resulting in multiple detection areas;

[0141] Training module 42 is used to acquire historical electricity consumption data for each detection area and train the neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area.

[0142] The determination module 43 is used to obtain the type of each electricity user in each detection area and the proportion of different types of electricity users in the detection area for each detection area, and determine the maximum allowable electricity fluctuation value of the detection area based on the proportion;

[0143] Comparison module 44 is used to predict the electricity consumption of each detection area based on the electricity consumption prediction model, and compare the predicted electricity consumption data of the detection area with the actual electricity consumption data to obtain the actual fluctuation value.

[0144] The filtering module 45 is used to determine abnormal areas based on all actual fluctuation values, determine similar areas for each abnormal area, and filter the corresponding abnormal areas based on the actual electricity consumption data of the similar areas to determine the electricity theft areas.

[0145] In some embodiments, the partitioning module 41 is specifically used for:

[0146] Obtain the number of transformer substations and their coverage area within the area to be tested, and divide the area to be tested based on the number of transformer substations and their coverage area to obtain multiple testing areas;

[0147] Each detection area contains one or more testing zones.

[0148] In some embodiments, the training module 42 is specifically used for:

[0149] A neural network model is constructed, historical electricity consumption data for each detection area is obtained, a training set and a validation set are obtained based on the historical electricity consumption data, and the neural network model is trained based on the training set to obtain an initial electricity consumption prediction model.

[0150] The initial electricity consumption prediction model was validated using time series cross-validation based on the validation set, and the electricity consumption prediction model for the corresponding detection area was determined based on the validation results.

[0151] In some embodiments, the determining module 43 is specifically used for:

[0152] For each detection area, perform the following steps:

[0153] Obtain the type of each electricity user within the detection area;

[0154] Obtain the proportion of each type of electricity user, as well as the historical electricity consumption data of each type of electricity user within a preset historical period;

[0155] Based on historical electricity consumption data, determine the total electricity consumption for each type of electricity user;

[0156] Substitute the total electricity consumption of each type of electricity user in the detection area into the electricity fluctuation prediction formula for the detection area to calculate the maximum allowable electricity fluctuation value for the detection area.

[0157] The formula for predicting electricity consumption fluctuations is:

[0158]

[0159] Where, k i α represents the maximum allowable power fluctuation value for the i-th detection area. w α represents the temperature weight of the i-th detection region. m ΔW represents the economic weight of the i-th detection region.i Δm represents the temperature fluctuation coefficient of the i-th detection area. i E represents the economic volatility coefficient of the i-th detection region. i p represents the total electricity consumption in the historical electricity consumption data of the i-th detection area. j β represents the proportion of electricity users of the j-th type in the i-th detection area. j γ represents the influence factor of the first influencing factor of the j-th type in the i-th detection region. j Let represent the influence factor of the second influencing factor of the j-th type in the i-th detection region, and let represent the first influence value of the j-th type in the i-th detection region, ΔE(2). j This represents the second influence value of the j-th type in the i-th detection region.

[0160] In some embodiments, the comparison module 44 is specifically used for:

[0161] For each detection area, perform the following steps:

[0162] Based on the historical electricity consumption data of the detection area and the electricity consumption prediction model, the electricity consumption of the detection area is predicted to obtain the predicted electricity consumption data within the preset time period.

[0163] Obtain actual electricity consumption data within a preset time period, compare the predicted electricity consumption data with the actual electricity consumption data, and obtain the actual fluctuation value based on the comparison results;

[0164] The historical electricity consumption data and the predicted electricity consumption data both include the electricity consumption of the detection area in different time periods and the line loss data of the detection area.

[0165] In some embodiments, the comparison module 44 is further configured to:

[0166] For each detection area, the preset electricity consumption data and the electricity consumption data in different time periods of the actual electricity consumption data, as well as the line loss data of the detection area, are input into the actual fluctuation value calculation formula to obtain the actual fluctuation value;

[0167] The formula for calculating the actual fluctuation value is:

[0168]

[0169] Where, k i ′ represents the actual fluctuation value of the i-th detection region, t represents the t-th time segment, s(1) it Let s(2) represent the actual electricity consumption of the i-th detection area in the t-th time period. it Let represent the predicted electricity consumption of the i-th detection area in the t-th time period, and δ1 represent the line loss correction coefficient of the i-th detection area in the t-th time period, x(1) itLet x(2) represent the actual line loss of the i-th detection area in the t-th time period. it This represents the preset line loss of the i-th detection area in the t-th time period.

[0170] In some embodiments, the filtering module 45 is specifically used for:

[0171] The actual fluctuation value of each detection area is compared with the corresponding maximum allowable power fluctuation value. If the actual fluctuation value exceeds the maximum allowable power fluctuation value, the detection area is identified as an abnormal area.

[0172] Otherwise, the detection area will be classified as a normal area.

[0173] In some embodiments, the filtering module 45 is further configured to:

[0174] Acquire feature data for each detection region, calculate the similarity between each abnormal region and the remaining detection regions based on the feature data, and determine the similar regions for each abnormal region based on the similarity.

[0175] For each abnormal region, perform the following steps:

[0176] Obtain the actual electricity consumption data of similar areas in the abnormal area, and compare the actual electricity consumption data of the abnormal area with the actual electricity consumption data of the corresponding similar areas;

[0177] Calculate the difference between the actual electricity consumption data of the abnormal area and the actual electricity consumption data of the corresponding similar area for each data point, and compare each difference with the corresponding threshold. Based on the comparison results, determine whether the abnormal area is an area of ​​electricity theft.

[0178] In some embodiments, a neural network-based electricity theft detection device 4 can also be used for:

[0179] Based on the electricity theft detection model, each electricity user in each electricity theft area is monitored to identify suspicious electricity users who may be involved in electricity theft.

[0180] The user ID of each suspicious electricity user is sent to the electricity theft alarm system, and staff are notified to conduct on-site verification.

[0181] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0182] Those skilled in the art will recognize that the templates, units, and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0183] If the module / unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of each of the above embodiments of the electricity theft detection method based on a neural network model. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory, a random access memory, an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0184] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for detecting electricity theft based on a neural network model, characterized in that, include: The area to be tested is divided into multiple testing zones; Historical electricity consumption data for each detection area is acquired, and a neural network model is trained based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area. For each detection area, obtain the type of each electricity user in the detection area and the proportion of different types of electricity users in the detection area, and determine the maximum allowable electricity fluctuation value for the detection area based on the proportion; For each detection area, electricity consumption is predicted based on the electricity consumption prediction model. The predicted electricity consumption data of the detection area is compared with the actual electricity consumption data to obtain the actual fluctuation value. Based on all actual fluctuation values, abnormal areas are identified, similar areas for each abnormal area are identified, and the corresponding abnormal areas are screened based on the actual electricity consumption data of the similar areas to identify areas where electricity is stolen.

2. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, The area to be tested is divided into multiple testing zones, including: The number of transformer substations and their coverage area within the area to be detected are obtained, and the area to be detected is divided based on the number of transformer substations and their coverage area to obtain multiple detection areas. Each detection area contains one or more testing zones.

3. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, The process of acquiring historical electricity consumption data for each detection area and training a neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area includes: A neural network model is constructed, historical electricity consumption data for each detection area is obtained, a training set and a validation set are obtained based on the historical electricity consumption data, and the neural network model is trained based on the training set to obtain an initial electricity consumption prediction model. The initial electricity consumption prediction model is validated using time series cross-validation based on the validation set, and the electricity consumption prediction model for the corresponding detection area is determined based on the validation results.

4. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, For each detection area, the method involves obtaining the type of each electricity user in that area and the proportion of different types of electricity users within that area, and determining the maximum allowable electricity fluctuation value for that detection area based on the proportions, including: For each detection area, perform the following steps: Obtain the type of each electricity user within the detection area; Obtain the proportion of each type of electricity user, and the historical electricity consumption data of each type of electricity user within a preset historical period; Based on the historical electricity consumption data, the total electricity consumption of each type of electricity user is determined; Substitute the total electricity consumption of each type of electricity user in the detection area into the electricity fluctuation prediction formula for the detection area to calculate the maximum allowable electricity fluctuation value for the detection area. The formula for predicting electricity consumption fluctuations is: Where, k i α represents the maximum allowable power fluctuation value for the i-th detection area. w α represents the temperature weight of the i-th detection region. m ΔW represents the economic weight of the i-th detection region. i Δm represents the temperature fluctuation coefficient of the i-th detection area. i E represents the economic volatility coefficient of the i-th detection region. i p represents the total electricity consumption in the historical electricity consumption data of the i-th detection area. j β represents the proportion of electricity users of the j-th type in the i-th detection area. j γ represents the influence factor of the first influencing factor of the j-th type in the i-th detection region. j Let ΔE(1) represent the influence factor of the second influencing factor of the j-th type in the i-th detection region. j Let ΔE(2) represent the first influence value of the j-th type in the i-th detection region. j This represents the second influence value of the j-th type in the i-th detection region.

5. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, For each detection area, electricity consumption is predicted based on an electricity consumption prediction model. The predicted electricity consumption data for the detection area is compared with the actual electricity consumption data to obtain the actual fluctuation value, including: For each detection area, perform the following steps: Based on the historical electricity consumption data of the detection area and the electricity consumption prediction model, the electricity consumption of the detection area is predicted to obtain the predicted electricity consumption data within the preset time period. Obtain actual electricity consumption data within a preset time period, compare the predicted electricity consumption data with the actual electricity consumption data, and obtain the actual fluctuation value based on the comparison result; The historical electricity consumption data and the predicted electricity consumption data both include the electricity consumption of the detection area in different time periods and the line loss data of the detection area.

6. The electricity theft detection method based on a neural network model as described in claim 5, characterized in that, The step of comparing the preset electricity consumption data with the actual electricity consumption data, and obtaining the actual fluctuation value based on the comparison result, includes: For each detection area, the preset power consumption data, the power consumption in different time segments of the actual power consumption data, and the line loss data of the detection area are input into the actual fluctuation value calculation formula to obtain the actual fluctuation value. The formula for calculating the actual fluctuation value is: Where, k i ′ represents the actual fluctuation value of the i-th detection region, t represents the t-th time segment, s(1) it Let s(2) represent the actual electricity consumption of the i-th detection area in the t-th time period. it Let represent the predicted electricity consumption of the i-th detection area in the t-th time period, and δ1 represent the line loss correction coefficient of the i-th detection area in the t-th time period, x(1) it Let x(2) represent the actual line loss of the i-th detection area in the t-th time period. it This represents the preset line loss of the i-th detection area in the t-th time period.

7. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, The process of determining abnormal regions based on all actual fluctuation values ​​includes: The actual fluctuation value of each detection area is compared with the corresponding maximum allowable power consumption fluctuation value. If the actual fluctuation value exceeds the maximum allowable power consumption fluctuation value, the detection area is identified as an abnormal area. Otherwise, the detection area will be classified as a normal area.

8. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, The process of acquiring similar regions for each abnormal region and filtering the corresponding abnormal regions based on the actual electricity consumption data of the similar regions to determine the electricity theft areas includes: Acquire feature data for each detection region, calculate the similarity between each abnormal region and the remaining detection regions based on the feature data, and determine the similar regions for each abnormal region based on the similarity. For each abnormal region, perform the following steps: Obtain the actual electricity consumption data of similar areas in the abnormal area, and compare the actual electricity consumption data of the abnormal area with the actual electricity consumption data of the corresponding similar areas; Calculate the difference between the actual electricity consumption data of the abnormal area and the actual electricity consumption data of the corresponding similar area for each data point, and compare each difference with the corresponding threshold. Based on the comparison results, determine whether the abnormal area is an area of ​​electricity theft.

9. The electricity theft detection method based on a neural network model as described in claim 1, characterized in that, The method further includes: Based on the electricity theft detection model, each electricity user in each electricity theft area is monitored to identify suspicious electricity users who may be involved in electricity theft. The user ID of each suspicious electricity user is sent to the electricity theft alarm system, and staff are notified to conduct on-site verification.

10. A device for detecting electricity theft based on a neural network model, characterized in that, include: The segmentation module is used to divide the area to be detected into multiple detection zones; The training module is used to acquire historical electricity consumption data for each detection area and train the neural network model based on the historical electricity consumption data to obtain an electricity consumption prediction model for each detection area. The determination module is used to obtain the type of each electricity user in each detection area and the proportion of different types of electricity users in the detection area for each detection area, and determine the maximum allowable electricity fluctuation value for the detection area based on the proportion; The comparison module is used to predict the electricity consumption of each detection area based on the electricity consumption prediction model, and compare the predicted electricity consumption data of the detection area with the actual electricity consumption data to obtain the actual fluctuation value. The filtering module is used to determine abnormal areas based on all actual fluctuation values, determine similar areas for each abnormal area, and filter the corresponding abnormal areas based on the actual electricity consumption data of the similar areas to determine the electricity theft areas.

Images