User voltage abnormity identification method and system oriented to group characteristic deviation detection

Abstract

The invention discloses a group feature deviation detection-oriented user voltage anomaly identification method and system, and relates to the technical field of big data analysis. Historical voltage data including voltage deviation degree, fluctuation amplitude and fluctuation frequency of each user of a voltage detection group are acquired, and a feature mean value is calculated according to a time interval; and on the basis of historical total voltage anomaly events, statistics of each feature anomaly frequency analysis is carried out, a deviation degree feature weight coefficient set is obtained, objective weight quantification is realized, and subjective deviation is avoided. The method comprises the following steps: calling historical data of a target user, dividing time periods to construct a standard voltage portrait after removing anomalies, updating real-time data into a real-time standard voltage portrait through weighted fusion, connecting a historical reference and a real-time state, and adapting to individual difference and law attenuation. Group users are synchronously processed to obtain a comprehensive deviation degree, a threshold value is set to screen anomalies and upload information, individual differences in a group are adapted, misjudgment and missed judgment of a unified constant value are avoided, support is provided for operation and maintenance of a power grid, and recognition accuracy and timeliness are comprehensively improved.

Description

Technical Field

[0001] This invention relates to the field of big data analytics, specifically to a method and system for identifying user voltage anomalies based on the detection of deviations in group characteristics. Background Technology

[0002] User voltage, as a core indicator of power supply quality at the end of the power system, is directly related to the operational stability and service life of various electrical equipment on the user side. With the improvement of electrification level, users' demands for power supply reliability and power quality continue to upgrade. Accurate and efficient user voltage detection has become a key prerequisite for ensuring normal power supply for users, identifying grid risks in advance, and optimizing grid operation and maintenance management. It is of irreplaceable importance for maintaining the stable operation of the power system and improving the user's power experience.

[0003] While individual users exhibit historical voltage patterns, these patterns are gradually weakening due to the integration of distributed energy resources, the widespread adoption of power electronic devices, and increased load fluctuations. Furthermore, the source voltage received by different users varies due to differences in distribution lines and wire diameters. These differences, coupled with individualized needs such as load type, time of day distribution, and equipment operating characteristics, further exacerbate the variations in individual voltage characteristics. Existing voltage anomaly identification methods often rely on standardized alarm values ​​or single statistical models, failing to adequately consider the dynamic changes resulting from the weakening of user patterns and the causes of individual voltage characteristic differences. This leads to the potential for missed detection of latent anomalies or misjudgments of normal fluctuations, making them unsuitable for accurate identification in complex power consumption scenarios. Therefore, a user voltage anomaly identification method and system oriented towards detecting deviations in group characteristics is urgently needed. Summary of the Invention

[0004] The purpose of this invention is to provide a user voltage anomaly identification method and system for detecting deviations in group characteristics, so as to solve the problems raised in the prior art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a user voltage anomaly identification method for detecting deviations in group characteristics, the user voltage anomaly identification method comprising the following steps: Step S1: Acquire and analyze the historical voltage data of the voltage detection group to obtain the feature weight coefficients of each user in the voltage detection group when calculating the deviation based on the voltage profile. This set is denoted as the deviation feature weight coefficient set. Step S1-1: Obtain historical voltage detection data of each user in the voltage detection group through a voltage sensor. The historical voltage detection data is represented as a continuous record of the voltage value of each user changing over time, including voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency. Step S1-2: Process the historical voltage detection data of each user in the voltage detection group, and filter out the voltage deviation, voltage fluctuation amplitude and voltage fluctuation frequency corresponding to each user. The voltage deviation is expressed as the absolute value of the difference between the user's actual voltage value and the rated voltage value of the voltage detection group. Step S1-3: Divide the historical voltage detection data of all users in the voltage detection group into the same time interval. For each time interval, calculate the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of all users in the voltage detection group within that time interval. Steps S1-4: Analyze the voltage anomaly identification terminal data to extract the total voltage anomaly events from the historical data of the voltage detection group; for each total voltage anomaly event, determine whether there is an anomaly in voltage deviation, voltage fluctuation amplitude, or voltage fluctuation frequency; when a certain feature in the total voltage anomaly event is abnormal, count the number of times that feature is abnormal separately; after completing the statistics of all total voltage anomaly events simultaneously, obtain the total number of times voltage deviation is abnormal, the total number of times voltage fluctuation amplitude is abnormal, and the total number of times voltage fluctuation frequency is abnormal. Steps S1-5: Sum the total number of voltage deviation anomalies, the total number of voltage fluctuation amplitude anomalies, and the total number of voltage fluctuation frequency anomalies to obtain the comprehensive number of anomalies for the total voltage anomaly events. Calculate the ratio of the total number of anomalies for each feature to the comprehensive number of anomalies. Use the ratio as the weight coefficient of the corresponding feature in the deviation calculation. Integrate the weight coefficients of the three features to obtain the deviation feature weight coefficient set. By acquiring historical voltage data for each user in the group, including voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency, and calculating the feature mean by time interval, and then counting the number of anomalies for each feature based on the total historical voltage anomaly events, the weight is determined by the ratio of the number of feature anomalies to the total number of anomalies, and a coefficient set is constructed. This achieves the quantification and objective setting of feature weights, which can accurately reflect the contribution of each voltage feature to the anomaly event, avoid the subjectivity of weights, provide a reliable basis for subsequent deviation calculation that fits the actual causes of anomalies, and improve the accuracy and relevance of the weight coefficients.

[0006] Step S2: Acquire and analyze real-time voltage data of the voltage detection group to obtain the real-time voltage variation of the voltage detection group; Step S2-1: Collect the real-time voltage values ​​of each user in the voltage detection group through the voltage sensor, and synchronously retrieve the preset rated voltage value of the voltage detection group; Step S2-2: For each user in the voltage detection group, calculate the ratio of the real-time voltage value of each user to the preset rated voltage value of the voltage detection group, and record it as the rated voltage ratio; sum the rated voltage ratios of all users to obtain the total real-time voltage ratio of the voltage detection group; synchronously read the number of users in the voltage detection group, divide the total real-time voltage ratio by the number of users to obtain the real-time fluctuating voltage of the voltage detection group. The formula for calculating the real-time voltage fluctuations of the voltage detection group is as follows: ; In the formula, Ugroup represents the real-time changing voltage of the voltage detection group; It represents the sum of the rated voltage ratios of all users in the voltage detection group; N represents the total number of users in the voltage detection group; By collecting real-time voltage values ​​of each user in the voltage detection group using voltage sensors and simultaneously retrieving rated voltage values, the ratio of rated voltages of each user is calculated, summed, and then divided by the number of users to obtain the real-time fluctuating voltage. This achieves real-time quantification of the overall voltage status of the group. The calculation process relies on real-time data from users, ensuring the timeliness and objectivity of the real-time fluctuating voltage. It provides a benchmark that fits the dynamics of the group for subsequent calculation of the real-time voltage deviation of target users, avoiding the problem of a fixed benchmark being out of sync with the actual voltage status of the group.

[0007] Step S3: Select any user in the voltage detection group as the research object, and denote it as the target user; by combining the real-time voltage variation with the real-time voltage analysis of the target user, obtain the real-time voltage deviation of the target user; The real-time voltage value of the target user is collected by a voltage sensor, the real-time voltage fluctuation of the voltage detection group is retrieved, and the absolute value of the difference between the real-time voltage value of the target user and the real-time voltage fluctuation is calculated. This absolute value of the difference is recorded as the real-time voltage deviation of the target user. Step S4: Obtain historical voltage detection data of the target user, construct a voltage analysis set for the target user after data cleaning, perform data analysis based on the voltage analysis set for the target user, and obtain a standard voltage profile of the target user. Step S4-1: Retrieve multiple sets of historical daily voltage detection data of the target user through the voltage anomaly identification terminal. The multiple sets of daily voltage detection data represent continuous records of the target user's voltage value changing with the time period of the day, with the day as the time unit. Combine the voltage deviation characteristics, voltage fluctuation amplitude characteristics, and voltage fluctuation frequency characteristics selected in step S1, identify the abnormal data of voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency contained in the target user's historical voltage detection data, remove the identified abnormal data from the target user's historical voltage detection data, and retain the remaining normal historical voltage detection data to construct the target user voltage analysis set. Step S4-2: Based on the target user voltage analysis set, the target user's daily electricity consumption time is divided into preset equal time periods. The average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of the target user under normal electricity consumption conditions are extracted in each time period. The average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of each time period are integrated to form a standard voltage profile of the target user. The standard voltage profile represents the comprehensive benchmark state of the voltage characteristics of the target user under normal electricity consumption conditions, which is composed of the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of each time period. By retrieving multiple sets of historical daily voltage detection data for the target user through a voltage anomaly identification terminal, and combining the characteristics of voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency to identify and eliminate abnormal data, a voltage analysis set is constructed. The average value of the features of each time period is extracted to form a standard voltage profile, realizing the construction of an individual voltage benchmark based on the user's normal historical data. The elimination of abnormal data ensures the authenticity of the profile, and the integration of time-based features fits the user's electricity consumption time pattern, providing an accurate reference for the individual normal state for subsequent real-time standard voltage profile updates and deviation calculations.

[0008] Step S5: Using the real-time voltage deviation and the real-time voltage detection data of the target user, the standard voltage profile is updated by weighted fusion calculation to obtain the real-time standard voltage profile. Based on the real-time standard voltage profile and the deviation feature weight coefficient set, the real-time voltage data is analyzed and calculated to obtain the comprehensive voltage deviation of the target user. Step S5-1: Collect real-time voltage detection data of the target user through a voltage sensor, and synchronously record the collection period corresponding to the real-time voltage detection data. Extract the real-time voltage fluctuation amplitude and real-time voltage fluctuation frequency of the target user from the real-time voltage detection data. Retrieve the real-time voltage deviation of the target user obtained in step S3, use the real-time voltage deviation to cover the voltage deviation in the real-time voltage detection data, and integrate the real-time voltage deviation, real-time voltage fluctuation amplitude and real-time voltage fluctuation frequency to form the first voltage profile of the target user. Step S5-2: Retrieve the preset weighted fusion weight coefficients, which include preset weights for voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency. Step S5-2-1: For the voltage deviation characteristics, multiply the real-time voltage deviation in the first voltage profile by the preset weight of voltage deviation, and then multiply the average voltage deviation of the time period in the standard voltage profile that is consistent with the collection time of the first voltage profile by the preset weight of voltage deviation. Add the two together to obtain the comprehensive voltage deviation. Step S5-2-2: Based on the voltage fluctuation amplitude characteristics, multiply the real-time voltage fluctuation amplitude in the first voltage profile by the preset weight of voltage fluctuation amplitude, and then multiply the average voltage fluctuation amplitude of the time period in the standard voltage profile that is consistent with the collection time of the first voltage profile by the preset weight of voltage fluctuation amplitude. Add the two together to obtain the comprehensive voltage fluctuation amplitude deviation. Step S5-2-3: Based on the voltage fluctuation frequency characteristics, multiply the real-time voltage fluctuation frequency in the first voltage profile by the preset weight of voltage fluctuation frequency, and then multiply the average voltage fluctuation frequency of the time period in the standard voltage profile that is consistent with the time period of the first voltage profile by the preset weight of voltage fluctuation frequency. Add the two together to obtain the comprehensive voltage fluctuation frequency deviation. Step S5-2-4: Integrate the comprehensive voltage deviation, comprehensive voltage fluctuation amplitude deviation, and comprehensive voltage fluctuation frequency deviation to form a real-time standard voltage profile for the target user; Step S5-3: Normalize the features of the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency in the real-time standard voltage profile of the target user. Read the deviation feature weight coefficient set in step S1, multiply the normalized value of each feature by the corresponding weight coefficient to obtain the deviation of each feature, and sum the deviations of each feature to obtain the comprehensive voltage deviation of the target user. By collecting real-time voltage detection data of the target user using a voltage sensor and recording the collection period, real-time fluctuation features are extracted and combined with real-time voltage deviation to form a first voltage profile. Preset weighted fusion coefficients are retrieved, and the features of the first voltage profile are weighted and summed with the average features of the same period in the standard profile to form a real-time standard voltage profile. Then, its features are normalized and combined with the deviation feature weight coefficient set to calculate the comprehensive voltage deviation, realizing the dynamic updating of the standard voltage profile. It not only connects with the user's historical normal benchmark but also incorporates the real-time status. Combined with the comprehensive calculation of objective feature weights, the timeliness, individual adaptability, and accuracy of voltage deviation assessment are improved.

[0009] Step S6: Simultaneously process other users in the voltage detection group according to steps S3 to S5 to obtain the comprehensive voltage deviation of each user in the voltage detection group. Construct a voltage detection group deviation set by summarizing and storing the comprehensive voltage deviation of each user. Process the voltage detection group deviation set in combination with the three sigma principle to filter out users with abnormal voltage. Step S6-1: Based on the voltage detection group deviation set, calculate the arithmetic mean of the comprehensive voltage deviation of all users in the voltage detection group deviation set, and denot it as the mean of the comprehensive deviation of the group; simultaneously calculate the square root of the arithmetic mean of the squares of the differences between the comprehensive voltage deviation of all users in the voltage detection group deviation set and the mean of the comprehensive deviation of the group, and denot it as the standard deviation of the comprehensive deviation of the group. Step S6-2: Sum the mean of the overall deviation of the group with three times the standard deviation of the overall deviation of the group to obtain the voltage overall deviation threshold; compare the voltage overall deviation of each user in the voltage detection group deviation set with the voltage overall deviation threshold; when a user's voltage overall deviation exceeds the voltage overall deviation threshold, mark the user as a voltage abnormal user; when a user's voltage overall deviation does not exceed the voltage overall deviation threshold, mark the user as a voltage normal user; integrate all users marked as voltage abnormal, upload the unique identifier of the voltage abnormal user and the corresponding voltage overall deviation to the voltage abnormality identification terminal to complete the user voltage abnormality identification; By synchronously processing the voltage detection group of all users to obtain the comprehensive voltage deviation of each user, a voltage detection group deviation set is constructed. Based on the three sigma principle, the mean and standard deviation of the comprehensive deviation of the group are calculated to determine the voltage deviation threshold. After comparison, abnormal users are marked and their unique identifiers and corresponding comprehensive voltage deviations are uploaded to the voltage anomaly identification terminal. This achieves voltage anomaly screening across the entire group. The threshold is set based on the statistical characteristics of the group to adapt to individual voltage differences, avoiding misjudgments and omissions caused by uniform values, ensuring the comprehensiveness and accuracy of identification, and providing clear information to provide direct support for power grid operation and maintenance.

[0010] Furthermore, a user voltage anomaly identification system for detecting deviations in group characteristics is provided. This user voltage anomaly identification system includes a feature weight calculation module, a group real-time voltage module, a target real-time deviation module, a standard voltage profile module, and a voltage anomaly identification module. The feature weight calculation module is used to acquire historical voltage data of the voltage detection group, extract voltage features, count the number of anomalies, and calculate feature weight coefficients to construct a deviation feature weight coefficient set. The group real-time voltage module is used to collect real-time voltage values ​​of the voltage detection group, retrieve rated voltage values, and calculate the group's real-time voltage fluctuations. The target real-time deviation module is used to select target users, collect their real-time voltage values, and calculate the real-time voltage deviation in conjunction with the group's real-time voltage fluctuations. The standard voltage profile module is used to retrieve historical voltage data of target users, clean the data, divide it into time periods, extract feature mean values, and construct a standard voltage profile. The voltage anomaly identification module is used to update the target user's real-time standard voltage profile, calculate the comprehensive deviation, and filter users with voltage anomalies in the group using the three Sigma principle. The output of the feature weight calculation module is electrically connected to the input of the group real-time voltage module; the output of the group real-time voltage module is electrically connected to the input of the target real-time deviation module; the output of the target real-time deviation module is electrically connected to the input of the standard voltage profile module; and the output of the standard voltage profile module is electrically connected to the input of the voltage anomaly identification module. The feature weight calculation module includes a historical data parsing unit and a weight coefficient generation unit. The historical data parsing unit is used to obtain historical voltage data of each user in the group, filter voltage features and calculate the average value according to the time interval, and count the number of times each feature is abnormal in the total voltage abnormal events. The weight coefficient generation unit is used to calculate the ratio of the number of times each feature is abnormal to the total number of abnormal events, and integrate the ratios to form a set of deviation feature weight coefficients. The group real-time voltage module includes a real-time data acquisition unit and a group voltage calculation unit; the real-time data acquisition unit is used to acquire the real-time voltage values ​​of each user in the group through voltage sensors and synchronously retrieve the preset rated voltage value of the group; the group voltage calculation unit is used to calculate the ratio of the rated voltage of each user and sum them, and calculate the real-time changing voltage of the group in combination with the number of users. The target real-time deviation module includes a target user selection unit and a real-time deviation calculation unit; the target user selection unit is used to select any user from the voltage detection group as the target user; the real-time deviation calculation unit is used to collect the real-time voltage value of the target user, retrieve the real-time changing voltage of the group, and calculate the absolute value of the difference between the two. The standard voltage profile module includes a historical data cleaning unit and a standard profile generation unit. The historical data cleaning unit is used to retrieve the historical daily voltage data of the target user, identify and remove abnormal data, and retain normal data to construct a voltage analysis set. The standard profile generation unit is used to divide the electricity consumption time period of the target user, extract the average value of the features of each time period, and integrate them to form a standard voltage profile. The voltage anomaly identification module includes a real-time profile update unit and an abnormal user screening unit. The real-time profile update unit is used to collect real-time data of target users to form a first voltage profile, weighted and fused to update the real-time standard profile and calculate the voltage comprehensive deviation. The abnormal user screening unit is used to calculate the mean and standard deviation of the group comprehensive deviation, determine the threshold, and mark and upload the voltage anomaly user information.

[0011] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention acquires historical voltage data from each user in a voltage monitoring group, including voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency. It calculates the feature mean over time intervals and then counts the number of anomalies for each feature based on the total historical voltage anomaly events. A deviation feature weight coefficient set is constructed using the ratio of the number of feature anomalies to the total number of anomalies. This method quantifies and objectively sets feature weights, accurately reflecting the contribution of each feature to anomaly events, avoiding subjective weighting, and providing a reliable basis for subsequent calculation of overall voltage deviation that aligns with the actual causes of anomalies.

[0012] 2. This invention retrieves historical daily voltage data of target users through a voltage anomaly identification terminal, identifies and removes abnormal data to construct a voltage analysis set, extracts the average value of features by time period to form a standard voltage profile, and then updates the real-time standard voltage profile by weighted fusion of real-time data. This approach combines the user's historical normal benchmark with the real-time status, adapts to the attenuation of electricity consumption patterns and individual characteristics, solves the problem of disconnect from fixed benchmarks, and improves the timeliness and individual adaptability of voltage deviation assessment.

[0013] 3. This invention obtains the comprehensive voltage deviation of all users in a voltage detection group through synchronous processing, constructs a group deviation set, calculates the mean and standard deviation of the comprehensive group deviation based on the three sigma principle to determine a threshold, compares and marks abnormal users, and uploads the information. This group deviation detection method adapts to individual voltage differences, avoids misjudgments and omissions caused by uniform values, achieves accurate screening across the entire domain, provides clear support for power grid operation and maintenance, and ensures the comprehensiveness and accuracy of identification. Attached Figure Description

[0014] Figure 1 This is a flowchart illustrating a user voltage anomaly identification method based on group characteristic deviation detection according to the present invention. Figure 2 This is a schematic diagram of the structure of a user voltage anomaly identification system for detecting deviations in group characteristics according to the present invention. Detailed Implementation

[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0016] Example 1: As Figure 1 As shown, the present invention provides a technical solution, a user voltage anomaly identification method for detecting deviations in group characteristics, the user voltage anomaly identification method comprising the following steps: Step S1: Acquire and analyze the historical voltage data of the voltage detection group to obtain the feature weight coefficients of each user in the voltage detection group when calculating the deviation based on the voltage profile. This set is denoted as the deviation feature weight coefficient set. Step S1-1: Obtain historical voltage detection data of each user in the voltage detection group through a voltage sensor. The historical voltage detection data is represented as a continuous record of the voltage value of each user changing over time, including voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency. Step S1-2: Process the historical voltage detection data of each user in the voltage detection group, and filter out the voltage deviation, voltage fluctuation amplitude and voltage fluctuation frequency corresponding to each user. The voltage deviation is expressed as the absolute value of the difference between the user's actual voltage value and the rated voltage value of the voltage detection group. Step S1-3: Divide the historical voltage detection data of all users in the voltage detection group into the same time interval. For each time interval, calculate the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of all users in the voltage detection group within that time interval. Steps S1-4: Analyze the voltage anomaly identification terminal data to extract the total voltage anomaly events from the historical data of the voltage detection group; for each total voltage anomaly event, determine whether there is an anomaly in voltage deviation, voltage fluctuation amplitude, or voltage fluctuation frequency; when a certain feature in the total voltage anomaly event is abnormal, count the number of times that feature is abnormal separately; after completing the statistics of all total voltage anomaly events simultaneously, obtain the total number of times voltage deviation is abnormal, the total number of times voltage fluctuation amplitude is abnormal, and the total number of times voltage fluctuation frequency is abnormal. Steps S1-5: Sum the total number of voltage deviation anomalies, the total number of voltage fluctuation amplitude anomalies, and the total number of voltage fluctuation frequency anomalies to obtain the comprehensive number of anomalies for the total voltage anomaly events. Calculate the ratio of the total number of anomalies for each feature to the comprehensive number of anomalies. Use the ratio as the weight coefficient of the corresponding feature in the deviation calculation. Integrate the weight coefficients of the three features to obtain the deviation feature weight coefficient set. In practical implementation, this approach focuses on a voltage monitoring group within a specific transformer area. The principle involves analyzing the correlation between three characteristics—voltage deviation, fluctuation amplitude, and fluctuation frequency—and total voltage anomaly events in historical voltage data. The weight of each characteristic is quantified by its proportion of anomalies, ensuring that the weights objectively reflect the contribution of each characteristic to the anomaly. It is crucial that historical data cover different electricity consumption seasons and time periods, and that the judgment of anomalies in total voltage events is strictly based on historical alarm records stored in the voltage anomaly identification terminal to avoid human bias.

[0017] Step S2: Acquire and analyze real-time voltage data of the voltage detection group to obtain the real-time voltage variation of the voltage detection group; Step S2-1: Collect the real-time voltage values ​​of each user in the voltage detection group through the voltage sensor, and synchronously retrieve the preset rated voltage value of the voltage detection group; Step S2-2: For each user in the voltage detection group, calculate the ratio of the real-time voltage value of each user to the preset rated voltage value of the voltage detection group, and record it as the rated voltage ratio; sum the rated voltage ratios of all users to obtain the total real-time voltage ratio of the voltage detection group; synchronously read the number of users in the voltage detection group, divide the total real-time voltage ratio by the number of users to obtain the real-time fluctuating voltage of the voltage detection group. In practice, the average ratio of all users' real-time voltage to their rated voltage is calculated to dynamically capture the overall voltage status of the group, providing a benchmark that closely matches real-time operating conditions for calculating individual deviations. It is crucial that the voltage sensor acquisition times be strictly synchronized to avoid data inaccuracies caused by time differences, and that the number of users be read from the latest ledger in real time to ensure that the calculation base matches the actual group size.

[0018] Step S3: Select any user in the voltage detection group as the research object, and denote it as the target user; by combining the real-time voltage variation with the real-time voltage analysis of the target user, obtain the real-time voltage deviation of the target user; The real-time voltage value of the target user is collected by a voltage sensor, the real-time voltage fluctuation of the voltage detection group is retrieved, and the absolute value of the difference between the real-time voltage value of the target user and the real-time voltage fluctuation is calculated. This absolute value of the difference is recorded as the real-time voltage deviation of the target user. In practice, the absolute value of the difference between the target user's real-time voltage and the group's real-time fluctuating voltage directly reflects the degree of deviation of an individual from the group's dynamic benchmark. It is important to note that the target user's real-time voltage acquisition and the group's real-time fluctuating voltage calculation must be performed at the same time point, and the difference calculation must strictly use the absolute value to ensure the non-negative quantification of the deviation.

[0019] Step S4: Obtain historical voltage detection data of the target user, construct a voltage analysis set for the target user after data cleaning, perform data analysis based on the voltage analysis set for the target user, and obtain a standard voltage profile of the target user. Step S4-1: Retrieve multiple sets of historical daily voltage detection data of the target user through the voltage anomaly identification terminal. The multiple sets of daily voltage detection data represent continuous records of the target user's voltage value changing with the time period of the day, with the day as the time unit. Combine the voltage deviation characteristics, voltage fluctuation amplitude characteristics, and voltage fluctuation frequency characteristics selected in step S1, identify the abnormal data of voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency contained in the target user's historical voltage detection data, remove the identified abnormal data from the target user's historical voltage detection data, and retain the remaining normal historical voltage detection data to construct the target user voltage analysis set. Step S4-2: Based on the target user voltage analysis set, the target user's daily electricity consumption time is divided into preset equal time periods. The average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of the target user under normal electricity consumption conditions are extracted in each time period. The average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of each time period are integrated to form a standard voltage profile of the target user. The standard voltage profile represents the comprehensive benchmark state of the voltage characteristics of the target user under normal electricity consumption conditions, which is composed of the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency of each time period. In practice, the daily voltage monitoring data of the target user over the past three months is retrieved. The principle is to remove abnormal historical data and extract the average value of features under normal conditions according to time periods to construct an individual baseline profile that fits the user's electricity consumption patterns during different time periods. It is important to note that the identification of abnormal data must accurately correspond to the three types of features and must not be confused. Furthermore, the division of time periods throughout the day must be combined with the user's peak and valley electricity consumption characteristics to ensure that the profile matches the actual electricity consumption rhythm.

[0020] Step S5: Using the real-time voltage deviation and the real-time voltage detection data of the target user, the standard voltage profile is updated by weighted fusion calculation to obtain the real-time standard voltage profile. Based on the real-time standard voltage profile and the deviation feature weight coefficient set, the real-time voltage data is analyzed and calculated to obtain the comprehensive voltage deviation of the target user. Step S5-1: Collect real-time voltage detection data of the target user through a voltage sensor, and synchronously record the collection period corresponding to the real-time voltage detection data. Extract the real-time voltage fluctuation amplitude and real-time voltage fluctuation frequency of the target user from the real-time voltage detection data. Retrieve the real-time voltage deviation of the target user obtained in step S3, use the real-time voltage deviation to cover the voltage deviation in the real-time voltage detection data, and integrate the real-time voltage deviation, real-time voltage fluctuation amplitude and real-time voltage fluctuation frequency to form the first voltage profile of the target user. Step S5-2: Retrieve the preset weighted fusion weight coefficients, which include preset weights for voltage deviation, voltage fluctuation amplitude, and voltage fluctuation frequency. Step S5-2-1: For the voltage deviation characteristics, multiply the real-time voltage deviation in the first voltage profile by the preset weight of voltage deviation, and then multiply the average voltage deviation of the time period in the standard voltage profile that is consistent with the collection time of the first voltage profile by the preset weight of voltage deviation. Add the two together to obtain the comprehensive voltage deviation. Step S5-2-2: Based on the voltage fluctuation amplitude characteristics, multiply the real-time voltage fluctuation amplitude in the first voltage profile by the preset weight of voltage fluctuation amplitude, and then multiply the average voltage fluctuation amplitude of the time period in the standard voltage profile that is consistent with the collection time of the first voltage profile by the preset weight of voltage fluctuation amplitude. Add the two together to obtain the comprehensive voltage fluctuation amplitude deviation. Step S5-2-3: Based on the voltage fluctuation frequency characteristics, multiply the real-time voltage fluctuation frequency in the first voltage profile by the preset weight of voltage fluctuation frequency, and then multiply the average voltage fluctuation frequency of the time period in the standard voltage profile that is consistent with the time period of the first voltage profile by the preset weight of voltage fluctuation frequency. Add the two together to obtain the comprehensive voltage fluctuation frequency deviation. Step S5-2-4: Integrate the comprehensive voltage deviation, comprehensive voltage fluctuation amplitude deviation, and comprehensive voltage fluctuation frequency deviation to form a real-time standard voltage profile for the target user; Step S5-3: Normalize the features of the average voltage deviation, average voltage fluctuation amplitude, and average voltage fluctuation frequency in the real-time standard voltage profile of the target user. Read the deviation feature weight coefficient set in step S1, multiply the normalized value of each feature by the corresponding weight coefficient to obtain the deviation of each feature, and sum the deviations of each feature to obtain the comprehensive voltage deviation of the target user. In practice, processing can be carried out for the off-peak electricity consumption periods of the target users. By weighted fusion of real-time features and standard profiles, the baseline for updating features during the same period is obtained. Then, after normalization and weighted summation of feature weights, a comprehensive deviation degree that takes into account both historical patterns and real-time status is obtained.

[0021] Step S6: Simultaneously process other users in the voltage detection group according to steps S3 to S5 to obtain the comprehensive voltage deviation of each user in the voltage detection group. Construct a voltage detection group deviation set by summarizing and storing the comprehensive voltage deviation of each user. Process the voltage detection group deviation set in combination with the three sigma principle to filter out users with abnormal voltage. Step S6-1: Based on the voltage detection group deviation set, calculate the arithmetic mean of the comprehensive voltage deviation of all users in the voltage detection group deviation set, and denot it as the mean of the comprehensive deviation of the group; simultaneously calculate the square root of the arithmetic mean of the squares of the differences between the comprehensive voltage deviation of all users in the voltage detection group deviation set and the mean of the comprehensive deviation of the group, and denot it as the standard deviation of the comprehensive deviation of the group. Step S6-2: Sum the mean of the overall deviation of the group with three times the standard deviation of the overall deviation of the group to obtain the voltage overall deviation threshold; compare the voltage overall deviation of each user in the voltage detection group deviation set with the voltage overall deviation threshold; when a user's voltage overall deviation exceeds the voltage overall deviation threshold, mark the user as a voltage abnormal user; when a user's voltage overall deviation does not exceed the voltage overall deviation threshold, mark the user as a voltage normal user; integrate all users marked as voltage abnormal, upload the unique identifier of the voltage abnormal user and the corresponding voltage overall deviation to the voltage abnormality identification terminal to complete the user voltage abnormality identification; In practical implementation, a comprehensive approach can be taken for voltage monitoring groups in residential communities. Using the three-sigma principle, thresholds can be set based on the statistical characteristics of the group's overall deviation, identifying users whose deviations exceed the normal range as abnormal. The group deviation set must include all user data without omissions, and the calculation of the mean and standard deviation must strictly follow arithmetic logic to ensure that the thresholds conform to the statistical distribution characteristics of the group.

[0022] Example 2, as Figure 2 As shown, the present invention provides a user voltage anomaly identification system for detecting deviations in group characteristics. The user voltage anomaly identification system includes a feature weight calculation module, a group real-time voltage module, a target real-time deviation module, a standard voltage profile module, and a voltage anomaly identification module. The feature weight calculation module is used to acquire historical voltage data of the voltage detection group, extract voltage features, count the number of anomalies, and calculate feature weight coefficients to construct a deviation feature weight coefficient set. The group real-time voltage module is used to collect real-time voltage values ​​of the voltage detection group, retrieve rated voltage values, and calculate the group's real-time voltage fluctuations. The target real-time deviation module is used to select target users, collect their real-time voltage values, and calculate the real-time voltage deviation in conjunction with the group's real-time voltage fluctuations. The standard voltage profile module is used to retrieve historical voltage data of target users, clean the data, divide it into time periods, extract feature mean values, and construct a standard voltage profile. The voltage anomaly identification module is used to update the target user's real-time standard voltage profile, calculate the comprehensive deviation, and filter users with voltage anomalies in the group using the three Sigma principle. The output of the feature weight calculation module is electrically connected to the input of the group real-time voltage module; the output of the group real-time voltage module is electrically connected to the input of the target real-time deviation module; the output of the target real-time deviation module is electrically connected to the input of the standard voltage profile module; and the output of the standard voltage profile module is electrically connected to the input of the voltage anomaly identification module. The feature weight calculation module includes a historical data parsing unit and a weight coefficient generation unit. The historical data parsing unit is used to obtain historical voltage data of each user in the group, filter voltage features and calculate the average value according to the time interval, and count the number of times each feature is abnormal in the total voltage abnormal events. The weight coefficient generation unit is used to calculate the ratio of the number of times each feature is abnormal to the total number of abnormal events, and integrate the ratios to form a set of deviation feature weight coefficients. The group real-time voltage module includes a real-time data acquisition unit and a group voltage calculation unit; the real-time data acquisition unit is used to acquire the real-time voltage values ​​of each user in the group through voltage sensors and synchronously retrieve the preset rated voltage value of the group; the group voltage calculation unit is used to calculate the ratio of the rated voltage of each user and sum them, and calculate the real-time changing voltage of the group in combination with the number of users. The target real-time deviation module includes a target user selection unit and a real-time deviation calculation unit; the target user selection unit is used to select any user from the voltage detection group as the target user; the real-time deviation calculation unit is used to collect the real-time voltage value of the target user, retrieve the real-time changing voltage of the group, and calculate the absolute value of the difference between the two. The standard voltage profile module includes a historical data cleaning unit and a standard profile generation unit. The historical data cleaning unit is used to retrieve the historical daily voltage data of the target user, identify and remove abnormal data, and retain normal data to construct a voltage analysis set. The standard profile generation unit is used to divide the electricity consumption time period of the target user, extract the average value of the features of each time period, and integrate them to form a standard voltage profile. The voltage anomaly identification module includes a real-time profile update unit and an abnormal user screening unit. The real-time profile update unit is used to collect real-time data of target users to form a first voltage profile, weighted and fused to update the real-time standard profile and calculate the voltage comprehensive deviation. The abnormal user screening unit is used to calculate the mean and standard deviation of the group comprehensive deviation, determine the threshold, and mark and upload the voltage anomaly user information.

[0023] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of equivalents of the claims be included within the present invention.

Claims

1. A user voltage anomaly identification method for group feature deviation detection, characterized in that: The user voltage anomaly identification method comprises the following steps: Step S1, obtaining and analyzing the historical voltage data of the voltage detection group, obtaining the characteristic weight coefficients of the voltage portrait when each user in the voltage detection group calculates the deviation degree according to the voltage portrait, denoted as the deviation degree characteristic weight coefficient set; Step S2, obtaining and analyzing the real-time voltage data of the voltage detection group, obtaining the real-time variable voltage of the voltage detection group; Step S3, selecting any one user in the voltage detection group as the research object, denoted as the target user; through real-time voltage analysis combined with the real-time voltage of the target user, the real-time voltage deviation of the target user is obtained; Step S4, obtaining the historical voltage detection data of the target user, constructing the target user voltage analysis set after data cleaning, and obtaining the standard voltage portrait of the target user according to the target user voltage analysis set; Step S5, using the real-time voltage deviation combined with the real-time voltage detection data of the target user, updating the characteristics of the standard voltage portrait through weighted fusion calculation, obtaining the real-time standard voltage portrait, and analyzing and calculating the real-time voltage data based on the real-time standard voltage portrait combined with the deviation degree characteristic weight coefficient set, obtaining the voltage comprehensive deviation degree of the target user; Step S6, according to steps S3 to S5, the other users in the voltage detection group are processed synchronously, the voltage comprehensive deviation degrees of each user in the voltage detection group are obtained, the voltage detection group deviation degree set is constructed by inductively storing the voltage comprehensive deviation degrees of each user, and the voltage detection group deviation degree set is processed combined with the three sigma principle, and the users with voltage anomaly are screened out. 2.The user voltage anomaly identification method for group feature deviation detection according to claim 1, characterized in that: The specific steps of step S1 are as follows: Step S1-1, obtaining the historical voltage detection data of each user in the voltage detection group through the voltage sensor, wherein the historical voltage detection data is represented as continuous recording data of the voltage value of each user changing with time, including voltage deviation, voltage fluctuation amplitude and voltage fluctuation frequency; Step S1-2, processing the historical voltage detection data of each user in the voltage detection group respectively, and screening out the voltage deviation, voltage fluctuation amplitude and voltage fluctuation frequency corresponding to each user, wherein the voltage deviation is represented as the absolute value of the difference between the actual voltage value of the user and the rated voltage value of the voltage detection group; Step S1-3, dividing the historical voltage detection data of all users in the voltage detection group according to the same time interval, and calculating the voltage deviation average value, voltage fluctuation amplitude average value and voltage fluctuation frequency average value of all users in the voltage detection group in each time interval; Step S1-4, analyzing according to the day system of the voltage anomaly identification terminal, and extracting the total voltage abnormal event in the historical data of the voltage detection group; For each total voltage abnormal event, it is judged whether voltage deviation anomaly, voltage fluctuation amplitude anomaly and voltage fluctuation frequency anomaly occur respectively; When a certain feature in the total voltage abnormal event is abnormal, the abnormal times of the feature are counted separately; When a certain feature in the total voltage abnormal event is abnormal, the abnormal times of the feature are counted separately; After synchronously completing the statistics of all total voltage abnormal events, the total number of voltage deviation abnormal events, the total number of voltage fluctuation amplitude abnormal events, and the total number of voltage fluctuation frequency abnormal events are obtained; In step S1-5, the total number of voltage deviation abnormal events, the total number of voltage fluctuation amplitude abnormal events, and the total number of voltage fluctuation frequency abnormal events are summed to obtain the comprehensive abnormal number of total voltage abnormal events, and the ratio of the total number of each feature to the comprehensive abnormal number is calculated to obtain the weight coefficient of the corresponding feature in the deviation calculation, and the weight coefficients of the three features are integrated to obtain the deviation feature weight coefficient set. 3.The user voltage anomaly identification method for group feature deviation detection according to claim 2, characterized in that: The specific steps of step S2 are as follows: In step S2-1, the real-time voltage values of each user in the voltage detection group are collected by the voltage sensor, and the preset rated voltage value of the voltage detection group is synchronously called. In step S2-2, for each user in the voltage detection group, the ratio of the real-time voltage value of each user to the preset rated voltage value of the voltage detection group is calculated, which is denoted as the rated voltage ratio, and the sum of the rated voltage ratios of all users is calculated to obtain the real-time voltage ratio sum of the voltage detection group. In step S2-3, the number of users in the voltage detection group is synchronously read, and the real-time voltage ratio sum is divided by the number of users to obtain the real-time fluctuating voltage of the voltage detection group.

4. The user voltage anomaly identification method for group feature deviation detection according to claim 3, characterized in that: In step S3, the real-time voltage value of the target user is collected by the voltage sensor, the real-time fluctuating voltage of the voltage detection group is called, and the absolute value of the difference between the real-time voltage value of the target user and the real-time fluctuating voltage is calculated, which is denoted as the real-time voltage deviation of the target user.

5. The user voltage anomaly identification method for group feature deviation detection according to claim 4, characterized in that: The specific steps of step S4 are as follows: In step S4-1, a plurality of sets of historical daily voltage detection data of the target user are called by the voltage abnormality recognition terminal, and the plurality of sets of daily voltage detection data are represented as continuous recording data of the target user voltage value changing with the time period of the day; combined with the voltage deviation feature, the voltage fluctuation amplitude feature, and the voltage fluctuation frequency feature screened in step S1, the voltage deviation abnormal data, the voltage fluctuation amplitude abnormal data, and the voltage fluctuation frequency abnormal data contained in the historical voltage detection data of the target user are identified, the identified abnormal data is excluded from the historical voltage detection data of the target user, and the remaining normal historical voltage detection data is retained to construct the target user voltage analysis set; In step S4-2, based on the target user voltage analysis set, the all-day power consumption time of the target user is divided into preset equal time periods, the voltage deviation mean value, the voltage fluctuation amplitude mean value, and the voltage fluctuation frequency mean value of the target user in the normal power consumption state in each time period are extracted, and the voltage deviation mean value, the voltage fluctuation amplitude mean value, and the voltage fluctuation frequency mean value of each time period are integrated to form the standard voltage portrait of the target user, which is represented as the corresponding time period voltage feature comprehensive reference state of the target user in the normal power consumption state composed of the voltage deviation mean value, the voltage fluctuation amplitude mean value, and the voltage fluctuation frequency mean value of each time period.

6. The user voltage anomaly identification method for group feature deviation detection according to claim 5, characterized in that: The specific steps of step S5 are as follows: Step S5-1, collecting real-time voltage detection data of the target user through the voltage sensor, synchronously recording the collection time period corresponding to the real-time voltage detection data, extracting the real-time voltage fluctuation amplitude and the real-time voltage fluctuation frequency of the target user from the real-time voltage detection data; calling the real-time voltage deviation degree of the target user obtained in step S3, covering the voltage deviation degree in the real-time voltage detection data with the real-time voltage deviation degree, integrating the real-time voltage deviation degree, the real-time voltage fluctuation amplitude and the real-time voltage fluctuation frequency to form a first voltage portrait of the target user; Step S5-2, calling the preset weighted fusion weight coefficient, the weighted fusion weight coefficient including a voltage deviation degree preset weight, a voltage fluctuation amplitude preset weight and a voltage fluctuation frequency preset weight; Step S5-2-1, for the voltage deviation degree feature, multiplying the real-time voltage deviation degree in the first voltage portrait by the voltage deviation degree preset weight, and then multiplying the voltage deviation degree mean value in the standard voltage portrait corresponding to the collection time period of the first voltage portrait by the voltage deviation degree preset weight, and adding the two to obtain a comprehensive voltage deviation degree; Step S5-2-2, for the voltage fluctuation amplitude feature, multiplying the real-time voltage fluctuation amplitude in the first voltage portrait by the voltage fluctuation amplitude preset weight, and then multiplying the voltage fluctuation amplitude mean value in the standard voltage portrait corresponding to the collection time period of the first voltage portrait by the voltage fluctuation amplitude preset weight, and adding the two to obtain a comprehensive voltage fluctuation amplitude deviation degree; Step S5-2-3, for the voltage fluctuation frequency feature, multiplying the real-time voltage fluctuation frequency in the first voltage portrait by the voltage fluctuation frequency preset weight, and then multiplying the voltage fluctuation frequency mean value in the standard voltage portrait corresponding to the collection time period of the first voltage portrait by the voltage fluctuation frequency preset weight, and adding the two to obtain a comprehensive voltage fluctuation frequency deviation degree; Step S5-2-4, integrating the comprehensive voltage deviation degree, the comprehensive voltage fluctuation amplitude deviation degree and the comprehensive voltage fluctuation frequency deviation degree to form a real-time standard voltage portrait of the target user; Step S5-3, performing normalization processing on the features of the voltage deviation degree mean value, the voltage fluctuation amplitude mean value and the voltage fluctuation frequency mean value in the real-time standard voltage portrait of the target user, reading the deviation degree feature weight coefficient set in step S1, multiplying the normalized value of each feature by the corresponding weight coefficient to obtain the deviation degree of each feature, and summing the deviation degrees of the features to obtain the voltage comprehensive deviation degree of the target user.

7. The user voltage anomaly identification method for group feature deviation detection according to claim 6, characterized in that: The specific steps of step S6 are as follows: Step S6-1, based on the voltage detection group deviation degree set, calculating the arithmetic mean value of the voltage comprehensive deviation degrees of all users in the voltage detection group deviation degree set, denoted as group comprehensive deviation degree mean value; synchronously calculating the square root of the arithmetic mean value of the square of the difference between the voltage comprehensive deviation degrees of all users in the voltage detection group deviation degree set and the group comprehensive deviation degree mean value, denoted as group comprehensive deviation degree standard deviation; Step S6-2, summing the mean of the population comprehensive deviation and three times of the standard deviation of the population comprehensive deviation to obtain a voltage comprehensive deviation threshold; comparing the voltage comprehensive deviation of each user in the voltage detection population deviation set with the voltage comprehensive deviation threshold one by one; when the voltage comprehensive deviation of a user exceeds the voltage comprehensive deviation threshold, marking the user as a voltage abnormal user; when the voltage comprehensive deviation of a user does not exceed the voltage comprehensive deviation threshold, marking the user as a voltage normal user; integrating all users marked as voltage abnormal, uploading the unique identification of the voltage abnormal user and the corresponding voltage comprehensive deviation to the voltage abnormality identification terminal to complete the user voltage abnormality identification.

8. A user voltage anomaly identification system for group feature deviation detection, applied to the user voltage anomaly identification method for group feature deviation detection in any one of claims 1-7, characterized in that: The user voltage abnormality identification system comprises a feature weight calculation module, a population real-time voltage module, a target real-time deviation module, a standard voltage portrait module and a voltage abnormality identification module; The feature weight calculation module is used to obtain historical voltage data of the voltage detection population, extract voltage features and count abnormal times, calculate feature weight coefficients to construct a deviation feature weight coefficient set; the population real-time voltage module is used to collect real-time voltage values of the voltage detection population, call rated voltage values and calculate population real-time variable voltage; the target real-time deviation module is used to select a target user, collect its real-time voltage value and calculate real-time voltage deviation in combination with the population real-time variable voltage; the standard voltage portrait module is used to call historical voltage data of the target user, divide time intervals after data cleaning to extract feature mean to construct a standard voltage portrait; The voltage abnormality identification module is used to update the real-time standard voltage portrait of the target user, calculate comprehensive deviation, and screen voltage abnormal users in the population in combination with the three-sigma principle.

9. The user voltage abnormality identification system for population feature deviation detection according to claim 8, characterized in that: The feature weight calculation module comprises a historical data analysis unit and a weight coefficient generation unit; the historical data analysis unit is used to obtain historical voltage data of each user in the population, screen voltage features and calculate mean values according to time intervals, and count abnormal times of each feature in total voltage abnormal events; the weight coefficient generation unit is used to calculate the ratio of abnormal times of each feature to comprehensive abnormal times, and integrate the ratio to form a deviation feature weight coefficient set; The population real-time voltage module comprises a real-time data collection unit and a population voltage calculation unit; the real-time data collection unit is used to collect real-time voltage values of each user in the population through a voltage sensor, and synchronously call preset rated voltage values of the population; The population voltage calculation unit is used to calculate the rated voltage ratio of each user and sum, and calculate the population real-time variable voltage in combination with the number of users; The target real-time deviation module comprises a target user selection unit and a real-time deviation calculation unit; the target user selection unit is used to select any one user in the voltage detection population as a target user; The real-time deviation calculation unit is used to collect real-time voltage values of the target user, call the population real-time variable voltage and calculate the absolute value of the difference between them.

10. The user voltage anomaly identification system for group feature deviation detection according to claim 8, characterized in that: the standard voltage portrait module comprises a historical data cleaning unit and a standard portrait generating unit; the historical data cleaning unit is used to call the target user historical daily voltage data, identify and remove abnormal data, retain normal data to construct a voltage analysis set; the standard portrait generating unit is used to divide the target user power consumption period, extract the feature mean value of each period and integrate to form a standard voltage portrait; the voltage anomaly identification module comprises a real-time portrait updating unit and an abnormal user screening unit; the real-time portrait updating unit is used to collect the target user real-time data to form a first voltage portrait, weighted fusion updates the real-time standard portrait and calculates the voltage comprehensive deviation; the abnormal user screening unit is used to calculate the mean value and standard deviation of the group comprehensive deviation, determine the threshold value and mark and upload the voltage abnormal user information.

Images