Robustness-based outlier abnormal electricity utilization detection result verification method and system

Abstract

The invention discloses an outlier abnormal electricity utilization detection result verification method and system based on robustness, and the method comprises the steps: obtaining the electricity utilization data of a transformer area user based on a metering automation system, and carrying out the data cleaning and preprocessing; a screening method and an embedding method are adopted, characteristic indexes related to the daily electric quantity of the user and the line loss of the transformer area are analyzed, heuristic search and factor analysis between the characteristics are carried out on the characteristic indexes, and key characteristics are selected out; building an outlier algorithm model based on the key features, carrying out model training to find a suitable parameter threshold, and outputting suspected users with abnormal power consumption behaviors; and performing data verification on the case set. According to the method, the stability and the reliability can still be kept in the face of data uncertainty, noise and external interference.

Description

Background Art

[0002] With the continuous deepening of power market reform and the increasingly diversified and complex trend of users' electricity consumption behavior, using the electricity consumption data of low-voltage users to analyze electricity consumption behavior in order to achieve more refined and effective governance of low-voltage areas has become an important issue that needs to be solved in current power management and operation. This demand stems from the continuous changes in the power market environment and the personalized and diversified needs of users for electricity use. However, in actual operation, using the electricity consumption data of low-voltage users to analyze electricity consumption behavior faces a series of challenges.

[0003] First, there is a relative lack of electricity consumption information in low-voltage areas, which is mainly due to the limitations of data collection technology and the lack of data sharing mechanisms. Due to the insufficient amount of data, it is difficult for analysts to obtain comprehensive and accurate user electricity consumption data, making it difficult to build an effective electricity consumption behavior analysis model.

[0004] Secondly, the large number of low-voltage areas increases the complexity and workload of data analysis. Each area has a certain number of users, and each user's electricity consumption behavior is unique, which means that the amount of data that analysts need to process is extremely large. How to efficiently process and analyze this data within limited resources and time has become an urgent problem to be solved.

[0005] Furthermore, the factors that affect user power consumption and line loss in the substation are complex. These factors include but are not limited to weather conditions, equipment conditions, user power usage habits, etc. The interweaving and interaction of these factors make the analysis of power consumption behavior more complex and difficult. At the same time, the changes in these factors are also highly uncertain, which further increases the difficulty of analysis.

[0006] In addition, the poor robustness of abnormal electricity usage detection is also a prominent problem in the current electricity usage behavior analysis. Due to the diversity and concealment of abnormal electricity usage behaviors, as well as the existence of data noise and interference, the existing detection methods have a high false alarm rate and missed alarm rate when identifying abnormal electricity usage behaviors, which not only reduces the accuracy of the analysis, but also increases the difficulty and cost of governance. Summary of the invention

[0007] In order to overcome the problem that the existing power consumption behavior analysis method cannot effectively analyze the abnormal power consumption of low-voltage users, the present invention provides a robust outlier abnormal power consumption detection result verification method and system.

[0008] The technical solution of the present invention is as follows:

[0009] In a first aspect, the present invention provides a robust outlier abnormal power consumption detection result verification method, comprising the following steps:

[0010] Step S1: obtaining the electricity consumption data of users in the substation area based on the metering automation system, and performing data cleaning and preprocessing;

[0011] Step S2: using the screening method and embedding method to analyze the characteristic indicators related to the daily power consumption of users and the line loss in the substation area, and performing heuristic search and factor analysis between the characteristics on the characteristic indicators to select key features;

[0012] Step S3: constructing an outlier algorithm model based on the key features, performing model training to find a suitable parameter threshold, and outputting suspected users with abnormal electricity usage behavior;

[0013] Step S4: perform data verification on the case set.

[0014] As a preferred solution of the present invention, step S1 specifically includes the following sub-steps:

[0015] Step S101, exclude the daily electricity data with the power supply of the substation area ≤ 0, and exclude the daily electricity data with the line loss rate <-2% or the line loss rate > 40%;

[0016] Step S102: Combining the horizontal and vertical comparison methods of the substations, interpolation processing is performed on the missing daily electricity consumption data of the user;

[0017] Step S103: using the confidence interval to identify and correct the abnormal power consumption data of the user.

[0018] As a preferred solution of the present invention, in step S102, when the number of consecutive missing days of the user's daily electricity consumption is ≤ 7 days, the cubic spline interpolation method is used to interpolate the missing daily electricity consumption data of the user, and the daily power supply of the substation after interpolation meets the line loss rate condition in step S101.

[0019] As a preferred solution of the present invention, in step S103, the formula of the confidence interval is as follows:

[0020]

[0021] Among them, the mean is the average value of the user's electricity consumption data, the standard deviation is a measure of the degree of dispersion of the user's electricity consumption data, z is the z-score of the standard normal distribution, and n is the number of data points.

[0022] As a preferred solution of the present invention, the user power consumption data outside the confidence interval is identified as abnormal power data and corrected to the boundary value of the confidence interval, and the power supply to the substation is corrected accordingly.

[0023] As a preferred solution of the present invention, the characteristic indicators include daily power, daily power change, line loss power, line loss power change rate, K value, and restored power, wherein the K value is the indicator influence degree, and the restored power is the actual power of the user;

[0024] The K value is calculated based on the actual power consumption of the user on the current day, the actual power consumption of the user on the previous day, the line loss in the substation on the current day, and the line loss in the substation on the previous day.

[0025] As a preferred solution of the present invention, the formula for calculating the K value based on the actual power consumption of the user on the current day, the actual power consumption of the user on the previous day, the line loss of the station area on the current day, and the line loss of the station area on the previous day is as follows:

[0026]

[0027] As a preferred solution of the present invention, the restored electric quantity is restored by constructing an electric quantity conservation equation and solving the user error rate using the least square method.

[0028] As a preferred solution of the present invention, step S3 specifically includes the following sub-steps:

[0029] Step S301, selecting power consumption jump sample data;

[0030] Step S302, selecting abnormal power consumption sample data;

[0031] Step S303, eliminating sample data with abnormal line loss power;

[0032] Step S304, judging whether the user is an abnormal suspected user according to the K value sample data obtained by screening in steps S301-S303 and the parameter threshold;

[0033] Step S305: Output a list of abnormal users.

[0034] In a second aspect, the present invention provides a robust outlier abnormal power consumption detection result verification system, comprising:

[0035] The data acquisition and preprocessing module is used to obtain the power consumption data of users in the substation area based on the metering automation system, and perform data cleaning and preprocessing;

[0036] The feature selection module is used to analyze the characteristic indicators related to the daily power consumption of users and the line loss of the substation area by using the screening method and the embedding method, and to perform heuristic search and factor analysis between the characteristics on the characteristic indicators to select the key features;

[0037] A model building and output module, used to build an outlier algorithm model based on the key features, perform model training to find suitable parameter thresholds, and output suspected users with abnormal electricity usage behavior;

[0038] The data verification module is used to perform data verification on the case set.

[0039] The present invention according to the above scheme has the following beneficial effects:

[0040] 1. The present invention provides a robust outlier abnormal electricity consumption detection result verification method, which mines outliers of user electricity consumption behavior in low-voltage areas, deeply extracts user electricity consumption behavior characteristics, and mines abnormal electricity consumption behavior within users. At the same time, the algorithm is optimized to improve robustness. In the face of various complex situations such as data uncertainty, noise, and external interference, the stability and reliability of the detection results can still be ensured;

[0041] 2. By introducing the impact of actual power consumption changes of users on line loss changes in the substation, the indicator influence degree K is constructed as the criterion for abnormal users, the model calculation results are objectively verified, the user status is comprehensively analyzed, the recognition efficiency, accuracy and robustness are improved, and an effective method is provided for the analysis of abnormal power consumption behavior of users. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a flow chart of a robust outlier abnormal electricity consumption detection result verification method in the present invention;

[0043] Figure 2 for Figure 1 Specific flow diagram of step S1 in FIG.

[0044] Figure 3 for Figure 1 Specific flow diagram of step S3;

[0045] Figure 4 It is a structural schematic diagram of the robustness-based outlier abnormal electricity consumption detection result verification system in the present invention. DETAILED DESCRIPTION

[0046] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. 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.

[0047] It should be noted that the terms "including" and "having" and any variations thereof in the specification and claims of the present invention are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0048] Please refer to Figure 1 This embodiment provides a robust outlier abnormal power consumption detection result verification method, comprising the following steps:

[0049] Step S1: Obtain the electricity consumption data of users in the substation area based on the metering automation system, and perform data cleaning and preprocessing.

[0050] In one embodiment, please refer to Figure 2 , step S1 specifically includes the following sub-steps:

[0051] Step S101: Remove the data of abnormal power supply and line loss caused by factors such as collection. After removing the abnormal data, the remaining data is closer to the actual situation, providing a more accurate basis for subsequent data analysis and processing, ensuring the data quality of subsequent analysis, and avoiding the interference of abnormal data on the overall results. Specifically, remove the daily power data with power supply ≤ 0, and remove the daily power data with line loss rate <-2% or line loss rate > 40%.

[0052] Step S102: Combine the methods of horizontal and vertical comparison of the substations to interpolate the missing daily electricity data of the user to fill the gaps in the missing data and make the data more complete. Specifically, when the number of consecutive missing days of the user's daily electricity is ≤7 days, the cubic spline interpolation method is used to interpolate the missing daily electricity data of the user, and the daily power supply of the substation after interpolation meets the line loss rate conditions in step S101, further ensuring the accuracy and rationality of the data. The use of cubic spline interpolation method for interpolation can make the interpolated data consistent with the original data in terms of trend, thereby maintaining the continuity of the data.

[0053] Step S103: Use confidence intervals to identify and correct abnormal power consumption of user power consumption data to improve the accuracy and reliability of the data. A confidence interval is an interval estimate that gives a range of possible values ​​for a parameter. This range has a certain probability (i.e., confidence level) that it contains the true value of the parameter. In the present invention, confidence intervals are used to evaluate the normal fluctuation range of user power consumption data. Specifically, the formula for the confidence interval is as follows:

[0054]

[0055] Among them, the mean is the average value of the user's electricity consumption data; the standard deviation is a measure of the degree of dispersion of the user's electricity consumption data; z is the z-score of the standard normal distribution, which is 1.96 here, indicating that the selected confidence level is 95%, that is, there is a 95% confidence that the true parameter value will fall within the calculated confidence interval; n is the number of data points, that is, the number of user electricity consumption data used to calculate the mean and standard deviation.

[0056] When using confidence intervals to identify and correct abnormal electricity consumption data of users, firstly, the electricity consumption data of users outside the confidence interval is identified as abnormal electricity consumption data; then, the identified abnormal electricity consumption data is corrected to the boundary value of the confidence interval; if the abnormal electricity consumption data is less than the lower limit of the confidence interval, it is corrected to the lower limit value; if the abnormal electricity consumption data is greater than the upper limit of the confidence interval, it is corrected to the upper limit value; finally, since the electricity consumption data of users are interrelated, after correcting the electricity consumption data of users, it is also necessary to correct the electricity supply of the substation accordingly to maintain the consistency of the data.

[0057] Step S2: Use the screening method and embedding method to analyze the characteristic indicators related to the user's daily electricity consumption and the line loss in the substation area, and perform heuristic search and factor analysis on the characteristic indicators to select key features.

[0058] The benefit of using the screening method and embedding method to analyze feature indicators is that it can systematically evaluate the correlation between each feature and the user's daily power consumption and the line loss in the substation area. The screening method can quickly narrow the range of features and remove those features that are irrelevant or weakly correlated with the target variable (such as substation line loss), while the embedding method can automatically select features that contribute significantly to model performance during the model construction process. Combining the screening method and the embedding method can help reduce noise and redundancy and improve the efficiency and accuracy of subsequent analysis.

[0059] Heuristic search is an effective feature selection strategy that can find a near-optimal feature subset in a reasonable time. Through heuristic search, features that are critical to line loss prediction or analysis can be further screened out, thereby improving the model's interpretability and prediction performance.

[0060] Factor analysis between features helps to reveal the potential relationship and structure between features. Through factor analysis, multiple related features can be summarized into a few independent factors, which can express the information of the original features more concisely, help to understand the causes and influencing factors of line loss in the substation area, and provide strong support for subsequent analysis and decision-making.

[0061] In one embodiment, the characteristic indicators include daily electricity, daily electricity change, line loss electricity, line loss electricity change rate, K value, restored electricity, etc.

[0062] The K value is the indicator influence. The K value is calculated based on the actual power consumption of the user on the day, the actual power consumption of the user on the previous day, the line loss of the station on the day, and the line loss of the station on the previous day. Specifically, the formula for calculating the K value based on the actual power consumption of the user on the day, the actual power consumption of the user on the previous day, the line loss of the station on the day, and the line loss of the station on the previous day is as follows:

[0063]

[0064] By comparing the changes in actual user electricity consumption with the changes in line losses in the substation, the K value can reflect the impact of changes in user electricity consumption on changes in line losses in the substation, which helps to more accurately evaluate the contribution of user behavior to line losses in the substation, thereby providing a basis for formulating targeted loss reduction measures.

[0065] The restored power is the actual power of the user. The restored power is obtained by constructing the power conservation equation and solving the user error rate using the least squares method. Specifically, the formula of the power conservation equation is as follows:

[0066]

[0067] Among them, y is the total power supply of the area, e j is the error of the jth user’s electric energy meter, is the power consumption of the jth user's electric energy meter, e y It refers to line loss rate, and e0 refers to fixed loss.

[0068] Specifically, the formula for the restored power is as follows:

[0069]

[0070] The restored electricity volume takes into account the impact of factors such as electricity meter errors and line loss rates on power supply, thereby improving the accuracy and reliability of electricity volume data; at the same time, the restored electricity volume data can more realistically reflect the user's electricity consumption behavior, providing strong support for subsequent electricity volume analysis and prediction.

[0071] Step S3: construct an outlier algorithm model based on key features, perform model training to find suitable parameter thresholds, and output suspected users with abnormal electricity usage behavior.

[0072] In one embodiment, please refer to Figure 3 , step S3 specifically includes the following sub-steps:

[0073] Step S301, select sample data of power consumption jump. Specifically, select sample data whose restored power consumption change rate is outside the observed value; select sample data whose daily power consumption change amount is outside the observed value. Observation value = upper / lower quartile ± 1.5 × upper quartile - lower quartile). By selecting sample data whose restored power consumption change rate and daily power consumption change amount exceed the observed value, it is possible to effectively capture abnormal situations of sudden increase or decrease in power consumption, provide important clues for the identification of abnormal power consumption behavior, and quickly locate user groups with abnormal power consumption behavior.

[0074] Step S302, select abnormal power consumption sample data. Specifically, select sample data whose line loss power change rate is greater than a threshold (10% or upper and lower quartiles), and whose user power consumption is proportional to the line loss power consumption. By setting a threshold for the line loss power change rate and combining the proportional relationship between user power consumption and line loss power consumption, it is possible to more accurately identify users whose power consumption does not change significantly but whose line loss power consumption increases abnormally, thereby improving the accuracy of identifying abnormal power consumption behavior.

[0075] Step S303, remove sample data with abnormal line loss electricity. Specifically, remove sample data whose line loss electricity change rate is within the observed value and whose line loss electricity of the previous day is not abnormal. By removing sample data whose line loss electricity change rate is within the observed value and whose line loss electricity of the previous day is not abnormal, misjudgment caused by accidental factors or data errors can be reduced, which helps to improve the model's recognition accuracy of abnormal electricity consumption behavior and reduce unnecessary interference.

[0076] Step S304: Determine whether the user is an abnormal suspected user based on the K value sample data and parameter thresholds obtained by screening in steps S301-S303. Specifically, when the K value sample data satisfies the standard deviation < 0.4 and the median > 0.2; or the standard deviation is within 0.4-0.6, the minimum observed value > 0, and the interquartile range < 0.6, the user is determined to be an abnormal suspected user. By introducing the indicator influence K value as an abnormal user judgment criterion, the model calculation results can be verified more objectively, and the accuracy and robustness of recognition can be improved.

[0077] It should be noted that the above thresholds can be adjusted according to the actual power consumption of users. By setting reasonable parameter thresholds, such as standard deviation, median, minimum observed value and interquartile range, the screening criteria for abnormal suspected users can be further refined to ensure the accuracy and reliability of identification. For users whose daily power consumption is less than 1 for a long time, the user needs to be included in continuous observation.

[0078] Step S305: Output the abnormal user list. The abnormal user list provides direct decision support for the power company. The power company can quickly locate users who may have abnormal electricity usage behavior based on this list and then take corresponding measures.

[0079] Step S4: Perform data verification on the case set. By performing data verification on the case set, the performance of the model in actual applications can be verified, which helps to discover possible problems and deficiencies in the model and then perform targeted optimization and improvement.

[0080] Specifically, by comparing the model recognition results with the actual situation, the accuracy of the recognition can be evaluated, and the parameters and thresholds of the model can be adjusted accordingly to improve the accuracy of the recognition. By testing in different scenarios and conditions, the model's ability to recognize various abnormal power consumption behaviors can be evaluated, and the model can be optimized accordingly to make it more suitable for different power consumption environments and scenarios.

[0081] In summary, the robust outlier abnormal electricity consumption detection result verification method provided in this embodiment mines outliers of user electricity consumption behavior in low-voltage substations, deeply extracts the user's electricity consumption behavior characteristics, mines abnormal electricity consumption behavior within the user, and optimizes the algorithm to improve robustness. It can still ensure the stability and reliability of the detection results in the face of various complex situations such as data uncertainty, noise, and external interference. In addition, by introducing the impact of the actual power consumption changes of users on the changes in the line loss of the substation, constructing the index influence degree K as an abnormal user judgment criterion, objectively verifying the model calculation results, and comprehensively analyzing the user status to improve the recognition efficiency, accuracy and robustness.

[0082] Please refer to Figure 4 In one embodiment, the present invention provides a robust outlier abnormal power consumption detection result verification system, comprising:

[0083] The data acquisition and preprocessing module 100 is used to acquire the power consumption data of users in the substation area based on the metering automation system, and perform data cleaning and preprocessing;

[0084] The feature selection module 200 is used to analyze the characteristic indicators related to the daily power consumption of users and the line loss of the substation area by using the screening method and the embedding method, and to perform heuristic search on the characteristic indicators and factor analysis between the characteristics to select the key features;

[0085] The model building and output module 300 is used to build an outlier algorithm model based on key features, perform model training to find suitable parameter thresholds, and output suspected users with abnormal electricity usage behavior;

[0086] The data verification module 400 is used to perform data verification on the case set.

[0087] The robust outlier abnormal electricity consumption detection result verification system provided in the embodiment of the present application can execute the above method embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0088] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as a system, method or program product. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as a "circuit", "module" or "system".

[0089] According to the modules, submodules, units, and subunits of the present embodiment, any multiple or at least part of the functions of any multiple thereof can be implemented in one module. According to the modules, submodules, units, and subunits of the present embodiment, any one or more thereof can be split into multiple modules for implementation. According to the modules, submodules, units, and subunits of the present embodiment, any one or more thereof can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware of any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation modes of software, hardware, and firmware or in any appropriate combination of any of them. Alternatively, according to the modules, submodules, units, and subunits of the present embodiment, one or more thereof can be at least partially implemented as a computer program module, and when the computer program module is run, the corresponding function can be executed.

[0090] In one embodiment, the present invention provides a robust outlier abnormal power consumption detection result verification device, which may include at least one processing unit and at least one storage unit. The storage unit stores program code, and when the program code is executed by the processing unit, the processing unit executes the steps of the robust outlier abnormal power consumption detection result verification according to various exemplary embodiments of the present invention described above in this specification. For example, the processing unit may execute the following steps: Figure 1 The process of the robust outlier abnormal electricity consumption detection result verification method is shown.

[0091] In one embodiment, the present invention provides a computer-readable medium storing computer-executable instructions for executing the steps of the robustness-based outlier abnormal electricity consumption detection result verification method according to various exemplary embodiments of the present invention described above in this specification.

[0092] The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, - but not limited to - a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination thereof. The readable signal medium can include a data signal propagated in a baseband or as a part of a carrier wave, wherein a readable program code is carried. The readable signal medium can also be any readable medium other than a readable storage medium, which can send, propagate or transmit a program used by an instruction execution system, device or device or used in combination with it. The program code contained on the readable medium can be transmitted with any appropriate medium, including - but not limited to - wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0093] In addition, although the operations of the method of the present invention are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

[0094] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all these improvements and changes should fall within the scope of protection of the appended claims of the present invention.

[0095] The above is an exemplary description of the present invention in conjunction with the accompanying drawings. It is obvious that the implementation of the present invention is not limited to the above-mentioned method. As long as various improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the present invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A robust outlier abnormal electricity consumption detection result verification method, characterized in that: The following steps are involved: Step S1: obtaining the electricity consumption data of users in the substation area based on the metering automation system, and performing data cleaning and preprocessing; Step S2: using the screening method and embedding method to analyze the characteristic indicators related to the daily power consumption of users and the line loss in the substation area, and performing heuristic search and factor analysis between the characteristics on the characteristic indicators to select key features; Step S3: constructing an outlier algorithm model based on the key features, performing model training to find a suitable parameter threshold, and outputting suspected users with abnormal electricity usage behavior; Step S4: perform data verification on the case set.

2. The robust outlier abnormal electricity consumption detection result verification method according to claim 1 is characterized in that: Step S1 specifically includes the following sub-steps: Step S101, exclude the daily electricity data with the power supply of the substation area ≤ 0, and exclude the daily electricity data with the line loss rate <-2% or the line loss rate > 40%; Step S102: Combining the horizontal and vertical comparison methods of the substations, interpolation processing is performed on the missing daily electricity consumption data of the user; Step S103: using the confidence interval to identify and correct the abnormal power consumption data of the user.

3. The robust outlier abnormal electricity consumption detection result verification method according to claim 2 is characterized in that: In step S102, when the number of consecutive missing days of the user's daily electricity consumption is ≤ 7 days, the cubic spline interpolation method is used to interpolate the missing daily electricity consumption data of the user, and the daily power supply of the substation after interpolation meets the line loss rate condition in step S101.

4. The robust outlier abnormal electricity consumption detection result verification method according to claim 3 is characterized in that: In step S103, the formula of the confidence interval is as follows: Among them, the mean is the average value of the user's electricity consumption data, the standard deviation is a measure of the degree of dispersion of the user's electricity consumption data, z is the z-score of the standard normal distribution, and n is the number of data points.

5. The robust outlier abnormal electricity consumption detection result verification method according to claim 3 is characterized in that: The user power consumption data outside the confidence interval is identified as abnormal power consumption data, and corrected to the boundary value of the confidence interval, and the power supply to the substation is corrected accordingly.

6. The robust outlier abnormal electricity consumption detection result verification method according to claim 1, characterized in that: The characteristic indicators include daily power consumption, daily power consumption change, line loss power consumption, line loss power consumption change rate, K value, and restored power consumption. The K value is the indicator influence degree, and the restored power consumption is the actual power consumption of the user. The K value is calculated based on the actual power consumption of the user on the current day, the actual power consumption of the user on the previous day, the line loss in the substation on the current day, and the line loss in the substation on the previous day.

7. The robust outlier abnormal electricity consumption detection result verification method according to claim 6, characterized in that: The formula for calculating the K value based on the actual power consumption of the user on the current day, the actual power consumption of the user on the previous day, the line loss of the station area on the current day, and the line loss of the station area on the previous day is as follows:

8. The robust outlier abnormal electricity consumption detection result verification method according to claim 6, characterized in that: The restored electric quantity is restored by constructing an electric quantity conservation equation and solving the user error rate using the least square method.

9. The robust outlier abnormal electricity consumption detection result verification method according to claim 6, characterized in that: Step S3 specifically includes the following sub-steps: Step S301, selecting power consumption jump sample data; Step S302, selecting abnormal power consumption sample data; Step S303, eliminating sample data with abnormal line loss power; Step S304, judging whether the user is an abnormal suspected user according to the K value sample data obtained by screening in steps S301-S303 and the parameter threshold; Step S305: Output a list of abnormal users.

10. A robust outlier abnormal electricity consumption detection result verification system, characterized in that: include: The data acquisition and preprocessing module is used to obtain the power consumption data of users in the substation area based on the metering automation system, and perform data cleaning and preprocessing; The feature selection module is used to analyze the characteristic indicators related to the daily power consumption of users and the line loss of the substation area by using the screening method and the embedding method, and to perform heuristic search and factor analysis between the characteristics on the characteristic indicators to select the key features; A model building and output module, used to build an outlier algorithm model based on the key features, perform model training to find suitable parameter thresholds, and output suspected users with abnormal electricity usage behavior; The data verification module is used to perform data verification on the case set.

Images