Intelligent processing method for data of electric energy meter

Abstract

The invention relates to the technical field of data processing, in particular to an intelligent electric energy meter data processing method, which comprises the following steps of: acquiring a monitoring value and a predicted value of electric energy data of any user in a current preset monitoring period and historical electric energy data in a preset historical time period; according to the difference between the monitoring value and the predicted value of each electric energy data in the current preset monitoring period, judging whether the electric energy data in the current preset monitoring period has overall deviation or not; if the electric energy data in the current preset monitoring period has overall deviation, acquiring the real abnormal degree of the electric energy data in the current preset monitoring period according to the change similarity degree of the electric energy data of any user and the reference user; according to the difference between the monitoring value and the predicted value of each electric energy data of any user in the current preset monitoring period and the real abnormal degree, the electric energy data abnormal degree of any user in the current preset monitoring period is obtained, and the accuracy of abnormal electricity consumption identification of any user is improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to an intelligent data processing method for electricity meters. Background Technology

[0002] Electricity meters collect data primarily for calculating user electricity consumption and managing user electricity usage behavior, playing a crucial role in ensuring user experience and safety, improving electricity efficiency, and reducing energy consumption. With technological advancements, electricity meters are no longer limited to simply recording data; they are increasingly focusing on intelligently processing the collected data to identify user electricity usage behavior and abnormal data, enabling intelligent regulation, management, and optimization of the power system.

[0003] Existing intelligent data processing methods for electricity meters typically predict users' electricity consumption by analyzing historical usage patterns. When the actual monitored values ​​differ significantly from the predicted values, an anomaly warning is issued. However, when users experience changes in electricity demand due to external factors, such as a significant increase in electricity consumption during extreme heat and extended duration of such consumption, existing intelligent data processing methods often identify this as abnormal electricity usage, resulting in false alarms and impacting the efficiency of power system regulation and management.

[0004] Therefore, how to reduce the interference of changes in electricity demand under the influence of external factors on the identification of abnormal electricity use and improve the accuracy of abnormal electricity use identification has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide an intelligent data processing method for electricity meters to solve the problem of how to reduce the interference of changes in electricity demand under the influence of external factors on the identification of abnormal electricity use and improve the accuracy of abnormal electricity use identification.

[0006] This invention provides an intelligent data processing method for electricity meters, which includes the following steps: For any user, the electricity data of the user at each moment is monitored using an electricity meter to obtain the monitored value of the electricity data within the current preset monitoring period, and the historical electricity data within a preset historical period before the current preset monitoring period is obtained. The electricity data includes current data and voltage data. Using historical energy data within a preset historical period, the predicted value of each energy data point within the current preset monitoring period is obtained. Based on the difference between the monitored value and the predicted value of each energy data point within the current preset monitoring period, it is determined whether there is an overall deviation in the energy data within the current preset monitoring period. If the power data within the current preset monitoring period shows an overall deviation, then at least two reference users are obtained for any user, and the true degree of anomaly in the power data within the current preset monitoring period is obtained based on the similarity of the power data changes between any user and each reference user. Based on the difference between the monitored and predicted values ​​of each power data point for any user within the current preset monitoring period, and the actual degree of anomaly, the degree of power data anomaly for any user within the current preset monitoring period is obtained. Based on the degree of power data anomaly, an abnormal warning is issued for the power consumption of any user.

[0007] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows: This invention targets any user and uses an electricity meter to monitor the user's electricity data at each moment to obtain the monitored values ​​of the electricity data within the current preset monitoring period. It also acquires historical electricity data from a preset historical period prior to the current preset monitoring period, including current and voltage data. Using the historical electricity data from the preset historical period, it obtains the predicted values ​​of each electricity data point within the current preset monitoring period. Based on the difference between the monitored and predicted values ​​of each electricity data point within the current preset monitoring period, it determines whether there is an overall deviation in the electricity data within the current preset monitoring period. If there is an overall deviation, it acquires at least two reference users for the user and, based on the similarity of the changes in the electricity data of the user and each reference user, obtains the true degree of anomaly in the electricity data within the current preset monitoring period. Based on the difference between the monitored and predicted values ​​of each electricity data point for the user within the current preset monitoring period, and the true degree of anomaly, it obtains the degree of anomaly in the electricity data for the user within the current preset monitoring period. Based on the degree of anomaly in the electricity data, it provides an abnormal warning for the user's electricity consumption. Specifically, based on the difference between the monitored and predicted values ​​of each power data point within the current preset monitoring period, it is determined whether there is an overall deviation in the power data within the current preset monitoring period, i.e., whether the user's power consumption within the current monitoring period deviates from historical power consumption patterns. Then, by combining the similarity between the user's power data changes and those of a reference user, the degree of abnormality in the user's power data within the current preset monitoring period is calculated. This reduces the interference of changes in power demand under the influence of external factors on the identification of abnormal power consumption, improves the accuracy of monitoring abnormal power consumption behavior of users, and enhances the efficiency of power system management. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a flowchart of a smart data processing method for electricity meters provided in Embodiment 1 of the present invention. Detailed Implementation

[0010] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0011] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0012] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0013] See Figure 1 This is a flowchart of a smart data processing method for electricity meters provided in Embodiment 1 of the present invention, as follows: Figure 1 As shown, the method may include: Step S101: For any user, use an electricity meter to monitor the electricity data of the user at each moment to obtain the monitored value of the electricity data within the current preset monitoring period, and obtain the historical electricity data within a preset historical period before the current preset monitoring period. The electricity data includes current data and voltage data.

[0014] Electricity meters collect data primarily for calculating user electricity consumption and managing user electricity usage behavior, playing a crucial role in ensuring user experience and safety, improving electricity efficiency, and reducing energy consumption. With technological advancements, electricity meters are no longer limited to simply recording data; they are increasingly focusing on intelligently processing the collected data to identify user electricity usage behavior and abnormal data, enabling intelligent regulation, management, and optimization of the power system.

[0015] For any given user, designated as the target user, this embodiment uses the target user as an example. The electricity data of the target user is monitored in real time using the resistive voltage divider and current transformer in the electricity meter. The electricity data includes voltage and current data. Since abnormal electricity consumption needs to be observed in conjunction with changes in the data, this embodiment obtains the monitored values ​​of the electricity data within the current preset monitoring period to analyze abnormal situations of the target user during this period. In this embodiment, the real-time monitoring frequency of the electricity data is once per second, and the current preset monitoring period is 5 minutes including the current time, meaning that abnormal electricity consumption of the target user is analyzed every 5 minutes. This is not a limitation and can be set according to the specific implementation scenario.

[0016] Since it is necessary to analyze the target user's electricity consumption patterns based on historical electricity data to determine any abnormalities in the target user's current preset monitoring period, this embodiment obtains the target user's initial electricity data within a preset historical period through an electricity meter. However, electricity meters may experience malfunctions such as poor contact, which could lead to abnormal data in the target user's initial electricity data monitored by the electricity meter. This would interfere with the analysis of the target user's electricity consumption patterns. Therefore, it is necessary to correct the abnormal data caused by the electricity meter malfunction in the initial electricity data to obtain historical electricity data within a preset historical period before the current preset monitoring period (i.e., the corrected electricity data). The preset historical period is the week before the current preset monitoring period, but this is not limited and can be set according to the specific implementation scenario.

[0017] The method for correcting abnormal data caused by electricity meter malfunctions in the initial electricity data to obtain historical electricity data for a preset historical period prior to the current preset monitoring cycle is as follows: (1) According to the type of each initial power data, all initial power data are divided into two initial power data sequences, namely the initial current data sequence and the initial voltage data sequence. For any initial power data sequence, the mutation index of each initial power data in any initial power data sequence is obtained according to the data fluctuation characteristics of any initial power data sequence.

[0018] Specifically, for any initial power data in any initial power data sequence, the initial power data and the first preset number of initial power data before it are combined to form a subsequence. The difference between each initial power data in the subsequence and the initial power data at the previous moment is obtained to obtain the change amount of each initial power data. In this embodiment, the first preset number is set to 10, but it is not limited here and can be set according to the specific implementation scenario. The absolute value of the difference between the change in each initial energy data (excluding the initial energy data) in the subsequence and the change in the initial energy data is accumulated to obtain the cumulative value of the change difference. The cumulative value of the change difference is then normalized to obtain the mutation index of the initial energy data.

[0019] In one embodiment, taking the i-th initial current data in the initial current data sequence as an example, the formula for calculating the mutation index of the i-th initial current data is:

[0020] in, This is the abrupt change index for the i-th initial current data; This represents the change in the i-th initial current data. It represents the change in the j-th initial current data in the subsequence, excluding the i-th initial current data. The number of initial current data in the subsequence excluding the i-th initial current data; It is the absolute value symbol; This is the normalization function.

[0021] It should be noted that, The larger the value, the greater the difference between the i-th initial current data and the j-th initial current data (excluding the i-th initial current data) in the subsequence, indicating that the i-th initial current data is more likely to have undergone a sudden change. The larger it is.

[0022] If the number of initial energy data before the i-th initial current data in the initial current data sequence is less than 10, then the initial energy data before the i-th initial current data is insufficient to analyze the mutation index of the i-th initial current data. Moreover, this situation will only occur in the first 9 initial current data in the initial current data sequence, and the amount of data is small. Therefore, the mutation index of the i-th initial current data in the initial current energy data sequence is marked as 0.

[0023] (2) According to the method of obtaining the mutation index of the i-th initial current data in the initial current data sequence, obtain the mutation index of each initial energy data in each initial energy data sequence, and obtain the fault energy data based on the mutation index of each initial energy data in each initial energy data sequence.

[0024] Since the trend of energy data monitored by an energy meter under normal conditions does not change significantly within a certain period, the consistency of the abrupt change indices of the initial current and initial voltage data under normal conditions should be high. However, the degree of data abrupt change caused by an energy meter fault may not be completely consistent and usually only affects one type of data (current or voltage). If clustering is performed based solely on the degree of abrupt change in one type of data, it may not be able to identify the fault. Therefore, in this embodiment, a abrupt change index graph is established, with the abrupt change index of the initial current data in the initial current data sequence as the x-axis and the abrupt change index of the initial voltage data in the initial voltage data sequence as the y-axis. The abrupt change indexes of the initial current data and the initial voltage data at each historical moment constitute a graph. For a given data point, DBSCAN is used to cluster the data points in the mutation index graph, resulting in at least one cluster to exclude discrete and outlier points. Data points within a cluster are those within the normal fluctuation range (the voltage and current data corresponding to the data points in the cluster are within the normal fluctuation range). Since electrical energy data includes both voltage and current data, the maximum mutation index of the initial current data is obtained from all clusters and denoted as the current mutation index threshold. The maximum mutation index of the initial voltage data is obtained and denoted as the voltage mutation index threshold. All initial current data with mutation indices greater than the current mutation index threshold, and all initial voltage data with mutation indices greater than the voltage mutation index threshold, are denoted as fault electrical energy data. DBSCAN is existing technology and will not be elaborated upon here.

[0025] (3) Correct the faulty power data to obtain historical power data within the preset historical period.

[0026] For any initial power data sequence, if there is at least one faulty power data in the initial power data sequence, then the faulty power data in the initial power data sequence is corrected to obtain a corrected power data sequence.

[0027] Specifically, for any faulty power data, the 10 non-faulty power data closest to the time of the faulty power data are obtained from any initial power data sequence, forming a target sequence. The least squares method is used to obtain the fitting curve of the target sequence, and the fitting value of the time of the faulty power data is obtained from the fitting curve as the correction value of the faulty power data. The least squares method is an existing technology and will not be elaborated here. Obtain the correction value of each faulty power data in any initial power data sequence, and combine it with all non-faulty power data in any initial power data sequence to form a corrected power data sequence.

[0028] If no faulty power data is found in any of the initial power data sequences, then the initial power data sequence is used as a corrected power data sequence, and the corrected power data sequence of each initial power data sequence is obtained to form historical power data within a preset historical period.

[0029] Existing intelligent data processing methods for electricity meters typically predict users' electricity consumption by analyzing historical usage patterns. When the actual monitored values ​​differ significantly from the predicted values, an anomaly warning is issued. However, when users experience changes in electricity demand due to external factors, such as a significant increase in electricity consumption during extreme heat and extended duration of such consumption, existing intelligent data processing methods often identify this as abnormal electricity usage, resulting in false alarms and impacting the efficiency of power system regulation and management.

[0030] Therefore, in this embodiment, based on the difference between the monitored value and the predicted value of each power data point within the current preset monitoring period, it is determined whether there is an overall deviation in the power data within the current preset monitoring period, that is, whether the user's power consumption within the current monitoring period deviates from historical power consumption patterns. Then, by combining the similarity between the user's power data changes and those of a reference user, the degree of abnormality in the user's power data within the current preset monitoring period is calculated. This reduces the interference of changes in power demand under the influence of external factors on the identification of abnormal power consumption, improves the accuracy of monitoring abnormal power consumption behavior of users, and enhances the efficiency of power system management.

[0031] Step S102: Using historical energy data within a preset historical time period, obtain the predicted value of each energy data point within the current preset monitoring period. Based on the difference between the monitored value and the predicted value of each energy data point within the current preset monitoring period, determine whether there is an overall deviation in the energy data within the current preset monitoring period.

[0032] After obtaining the historical energy data of the target user within a preset historical period, the predicted value of each energy data point within the current preset monitoring period is obtained using the LSTM algorithm based on the historical energy data. The LSTM algorithm is an existing technology and will not be elaborated here. Then, based on the difference between the monitored value and the predicted value of each energy data point within the current preset monitoring period, it is determined whether there is an overall deviation in the energy data within the current preset monitoring period.

[0033] The method for determining whether there is an overall deviation in the power data within the current preset monitoring period, based on the difference between the monitored and predicted values ​​of each power data point within the current preset monitoring period, is as follows: (1) Take any type of electrical energy data within the current preset monitoring period to form an electrical energy data sequence, obtain the difference between the monitored value and the predicted value of each electrical energy data in the electrical energy data sequence, obtain the corresponding difference accumulation value, normalize the difference accumulation value, and obtain the deviation degree of the electrical energy data sequence.

[0034] In one embodiment, taking the current data within the current preset monitoring period as an example, a current data sequence is formed, and the formula for calculating the deviation of the current data sequence is:

[0035] in, The degree of deviation of the current data sequence; This represents the monitored value of the r-th current data point in the current data sequence. Let r be the predicted value of the r-th current data in the current data sequence; This represents the number of current data points in the current data sequence. It is the absolute value symbol; This is the normalization function.

[0036] It should be noted that, The larger the value, the greater the difference between the monitored value and the predicted value of the r-th current data point in the current data sequence; that is, the more the monitored value of the r-th current data point deviates from the predicted value. The larger it is.

[0037] (2) Obtain the deviation degree of each power data sequence. If the deviation degree of at least one power data sequence is greater than the preset deviation degree threshold, it is confirmed that the power data in the current preset monitoring period has an overall deviation. The preset deviation degree threshold in this embodiment is set to 0.5 based on historical experience. It is not limited here and can be set according to the specific implementation scenario.

[0038] Thus, the overall deviation results of the power data within the current preset monitoring period are obtained.

[0039] Step S103: If there is an overall deviation in the power data within the current preset monitoring period, then at least two reference users are obtained for any user, and the true degree of abnormality of the power data within the current preset monitoring period is obtained based on the similarity of the power data changes between any user and each reference user.

[0040] When users are affected by external factors, their electricity demand may change. For example, the use of air conditioners increases during hot weather, resulting in users consuming significantly more electricity than usual. On the night of holidays such as New Year's Eve, most residents consume more electricity than usual. Although the electricity data monitored by the electricity meter deviates from the original electricity consumption pattern, it does not mean that the user has abnormal electricity consumption behavior.

[0041] External influencing factors typically affect multiple users' electricity consumption rather than just a single user. Therefore, if the target user's overall electricity consumption data deviates within the current preset monitoring period due to external factors, this will also be reflected in the electricity consumption data of other users similar to the target user. However, since different users may have different electricity consumption patterns, it's not possible to directly compare the similarity of user electricity consumption data. Furthermore, different users may exhibit certain similarities in their electricity consumption patterns; for example, users with similar electricity needs may both increase their consumption during midday and decrease it in the evening. Therefore, if there is an overall deviation in the electricity consumption data within the current preset monitoring period, all users located under the same transformer as the target user are identified, resulting in at least two reference users. The degree of similarity between the target user's electricity consumption data and that of each reference user is then used to determine the true degree of anomaly in the target user's electricity consumption data within the current preset monitoring period.

[0042] Taking the current data of the target user within the current preset monitoring period as an example, the method for obtaining the true degree of anomaly in the current data of the target user within the current preset monitoring period based on the similarity of the current data changes between the target user and each reference user is as follows: (1) Record any reference user as the target reference user, obtain the historical reference current data sequence of the target reference user and the historical current data sequence of the target user within a preset historical period, and obtain the historical change similarity of the current data between the target reference user and the target user based on the data change difference between the historical current data sequence and the historical reference current data sequence.

[0043] Because current data fluctuates to some extent, it is difficult to identify the changing trend of the data collected by the electricity meter from continuous data over a short period of time. Therefore, in the historical reference current data sequence, a second preset number of historical reference subsequences are selected according to a preset interval time. Each historical reference subsequence includes a third preset number of historical reference current data. In this embodiment, the preset interval time is 20 seconds, the second preset number is 5, and the third preset number is 1000. The uniform and dispersed selection can cover the historical reference current data as much as possible and show the historical reference current data that reflects the pattern of the data collected by the electricity meter. There are no restrictions here, and it can be set according to the specific implementation scenario. For any historical reference subsequence, historical energy data that is at the same time as each historical reference energy data in any historical reference subsequence is obtained from the historical energy data sequence to form a historical subsequence; First-order difference processing is performed on any historical reference subsequence and the historical subsequence respectively to obtain a historical reference difference sequence and a historical difference sequence. Two elements with the same position number in the historical reference difference sequence and the historical difference sequence are grouped into an element group. First-order difference is an existing technology and will not be described in detail here. The absolute values ​​of the differences between the two elements in each element group are summed to obtain the difference accumulation value. The reciprocal of the sum of the difference accumulation value and the preset constant is normalized to obtain the initial similarity between any historical reference subsequence and its corresponding historical subsequence.

[0044] In one embodiment, taking the c-th historical reference subsequence as an example, the formula for calculating the initial similarity between the c-th historical reference subsequence and its corresponding historical subsequence is as follows:

[0045] in, Let be the initial similarity between the c-th historical reference subsequence and its corresponding historical subsequence; For the historical reference differential data of the target reference user in the s-th element group; represents the historical differential data of the target user in the s-th element group; N is the number of element groups. It is the absolute value symbol; This is the normalization function; As a preset constant, this embodiment sets This is used to ensure that the fraction is meaningful. There are no restrictions here, and it can be set according to the specific implementation scenario.

[0046] It should be noted that, The smaller the value, the smaller the difference between the historical reference differential data of the target reference user and the historical reference data of the target user in the s-th element group, the more similar the changing trends of the current data of the target reference user and the target user at the corresponding time in the s-th element group, and the greater the initial similarity between the c-th historical reference subsequence and its corresponding historical subsequence.

[0047] Following the method for obtaining the initial similarity between the c-th historical reference subsequence and its corresponding historical subsequence, the initial similarity between each historical reference subsequence and its corresponding historical subsequence is obtained, and the mean of the initial similarity is obtained as the historical change similarity of the current data between the target reference user and the target user.

[0048] (2) Obtain the current data sequence of the target user and the reference current data sequence of the target reference user within the current preset monitoring period. Based on the difference in data change between the current data sequence and the reference current data sequence, obtain the deviation similarity of the current data between the target reference user and the target user.

[0049] According to the above method for obtaining the deviation of the current data sequence, the deviation of the reference current data sequence is obtained, the absolute value of the difference between the deviation of the current data sequence and the deviation of the reference current data sequence is calculated to obtain the deviation difference, and the reciprocal of the sum of the deviation difference and a preset constant is obtained to obtain the similarity of the deviation of the current data between the target reference user and the target user. The difference between the monitored value and the predicted value of each current data in the current data sequence is obtained to obtain a difference sequence. The difference sequence is fitted to obtain a difference fitting curve. The slope of the difference fitting curve is obtained as the deviation trend value of the current data sequence. The difference between the monitored value and the predicted value of each reference current data in the reference current data sequence is obtained to obtain a reference difference sequence. The reference difference sequence is fitted to obtain a reference difference fitting curve. The slope of the reference difference fitting curve is obtained as the deviation trend value of the reference current data sequence. The absolute value of the difference between the deviation trend value of the current data sequence and the deviation trend value of the reference current data sequence is obtained and denoted as the deviation trend difference. The reciprocal of the sum of the deviation trend difference and a preset constant is obtained to obtain the deviation trend similarity of the current data between the target reference user and the target user. The deviation similarity between the target reference user and the target user is obtained by multiplying the deviation degree similarity with the deviation trend similarity.

[0050] In one embodiment, the formula for calculating the deviation similarity of current data between the target reference user and the target user is as follows:

[0051] in, The deviation similarity of current data between the target reference user and the target user; The degree of deviation of the current data sequence; The degree of deviation from the reference current data sequence; This represents the deviation trend value of the current data sequence; This refers to the deviation trend value of the reference current data sequence; It is the absolute value symbol; As a preset constant, this embodiment sets This is used to ensure that the fraction is meaningful. There are no restrictions here, and it can be set according to the specific implementation scenario.

[0052] It should be noted that, The similarity of the deviation between the current data of the target reference user and the target user. The larger the value, the greater the difference in deviation between the current data sequence and the reference current data sequence. The smaller, The smaller it is; The similarity of the deviation trend of current data between the target reference user and the target user is used as a reference. The larger the value, the greater the difference in trend between the current data sequence and the reference current data sequence. The smaller, The smaller it is.

[0053] (3) Obtain the true degree of anomaly in the current data of the target user within the current preset monitoring period.

[0054] Obtain the historical change similarity of current data between each reference user and the target user, and obtain the corresponding historical change similarity cumulative value. Obtain the ratio of the historical change similarity of current data between the target reference user and the target user to the historical change similarity cumulative value, and obtain the true anomaly weight of current data compared with the target reference user. Normalize the inverse of the deviation similarity to obtain the true outlier value of the current data compared with the target reference user, and obtain the product of the true outlier weight and the true outlier value to obtain the reference outlier degree of the current data compared with the target reference user. The reference anomaly level of the current data compared with each target reference user is obtained, and the cumulative value of the reference anomaly level is obtained as the true anomaly level of the current data within the current preset monitoring period.

[0055] In one embodiment, the formula for calculating the true degree of anomaly in the current data of the target user within the current preset monitoring period is as follows:

[0056] in, To determine the true degree of anomaly in the current data of the target user within the current preset monitoring period; The similarity of historical changes in current data between the Ath target reference user and the target user; The deviation similarity of current data between the Ath target reference user and the target user; The number of target reference users; This is the normalization function.

[0057] It should be noted that, The larger the value, the greater the weight of the reference anomaly degree of the current data compared with the Ath target reference user. The larger the value, the more similar the degree and trend of deviation between the current data of the A-th target reference user and the target user, and the more the changes in the target user's current data conform to the characteristics of changes in electricity demand caused by external factors. The smaller it is.

[0058] The actual degree of anomaly in the voltage data of the target user within the current preset monitoring period is obtained using the same method as the method used to obtain the actual degree of anomaly in the current data of the target user within the current preset monitoring period.

[0059] Thus, the true degree of anomaly in the target user's current data and voltage data within the current preset monitoring period are obtained.

[0060] Step S104: Based on the difference between the monitored value and the predicted value of each power data of any user in the current preset monitoring period, and the actual degree of anomaly, obtain the degree of power data anomaly of any user in the current preset monitoring period, and issue an anomaly warning for the power consumption of any user based on the degree of power data anomaly.

[0061] The higher the degree of real anomaly of each type of electrical energy data within the current preset monitoring period, the more likely the target user is to have abnormal electricity consumption behavior. Abnormal electricity consumption usually affects both current and voltage data simultaneously. However, if the data collected by the electricity meter is incorrect, it usually will not have the same impact on the current and voltage data. Therefore, the more similar the abnormal trends of current and voltage data within the current preset monitoring period, the higher the reliability of the current abnormal data.

[0062] Therefore, based on the difference between the monitored and predicted values ​​of each power data point for the target user within the current preset monitoring period, and the actual degree of abnormality of each type of power data within the current preset monitoring period, the degree of abnormality of the power data for the target user within the current preset monitoring period is obtained. Then, based on the degree of abnormality of the power data, an abnormal warning is issued for the target user's electricity consumption.

[0063] The method for obtaining the degree of abnormality of the target user's electricity data in the current preset monitoring period is as follows, based on the difference between the monitored and predicted values ​​of each type of electricity data for the target user within the current preset monitoring period, and the actual degree of abnormality of each type of electricity data within the current preset monitoring period: For any type of electrical energy data within the current preset monitoring period, the deviation degree of the electrical energy data sequence of the given electrical energy data is obtained by multiplying the actual abnormality degree of the given electrical energy data by the product of the deviation degree of the electrical energy data and the actual abnormality degree of the given electrical energy data. The abnormality degree of each type of electrical energy data is obtained, and the average abnormality degree is obtained accordingly. Obtain the absolute value of the difference between the deviation trend values ​​of the two electrical energy data within the current preset monitoring period to obtain the deviation difference value, and obtain the reciprocal of the sum of the deviation difference value and the constant 1 to obtain the data anomaly similarity. The product of the average anomaly level and the data anomaly similarity is obtained to determine the anomaly level of the target user's power data in the current preset monitoring period.

[0064] In one embodiment, the formula for calculating the degree of abnormality in the target user's electricity data during the current preset monitoring period is as follows:

[0065] Where Y represents the degree of abnormality in the target user's electricity data during the current preset monitoring period; The degree of deviation of the current data sequence; The degree of deviation of the voltage data sequence; To determine the true degree of anomaly in the current data of the target user within the current preset monitoring period; To determine the true degree of anomaly in the voltage data of the target user within the current preset monitoring period; This represents the deviation trend value of the current data sequence; This represents the deviation trend value of the voltage data sequence; It is the absolute value symbol.

[0066] It should be noted that, This represents the average degree of anomaly in current and voltage data within the current preset monitoring period, reflecting the abnormality of electrical energy data during the current preset monitoring period. The larger the value, the greater the possibility of abnormalities in the current preset monitoring period's power data, and the greater Y becomes; This indicates the degree of data anomaly similarity between current and voltage data within the current preset monitoring period. The larger the value, the more likely the abnormal situation in the power data during the previous preset monitoring period is a real anomaly, and the higher the credibility of Y.

[0067] Furthermore, if the abnormality level of the target user's electricity data in the current preset monitoring period is greater than 0.5, it is confirmed that the target user has abnormal electricity consumption behavior. An abnormality warning is issued for the target user's electricity consumption. It is necessary to continuously monitor the electricity data monitored by the electricity meter with abnormal electricity consumption behavior and take corresponding control measures. 0.5 is the preset threshold for the abnormality level of electricity data in this embodiment. It is not limited here and can be set according to the specific implementation scenario.

[0068] In summary, this embodiment of the invention monitors the electricity data of any user at any given time using an electricity meter to obtain the monitored value of the electricity data within the current preset monitoring period. It also acquires historical electricity data from a preset historical period prior to the current preset monitoring period, including current and voltage data. Using the historical electricity data from the preset historical period, it obtains the predicted value of each electricity data point within the current preset monitoring period. Based on the difference between the monitored value and the predicted value of each electricity data point within the current preset monitoring period, it determines whether there is an overall deviation in the electricity data within the current preset monitoring period. If there is an overall deviation in the electricity data within the current preset monitoring period, it acquires at least two reference users for the user. Based on the similarity of the electricity data changes between the user and each reference user, it obtains the true degree of anomaly in the electricity data within the current preset monitoring period. Based on the difference between the monitored value and the predicted value of each electricity data point for the user within the current preset monitoring period, and the true degree of anomaly, it obtains the degree of anomaly in the electricity data for the user within the current preset monitoring period. Based on the degree of anomaly in the electricity data, it provides an abnormal warning for the user's electricity consumption. Specifically, based on the difference between the monitored and predicted values ​​of each power data point within the current preset monitoring period, it is determined whether there is an overall deviation in the power data within the current preset monitoring period, i.e., whether the user's power consumption within the current monitoring period deviates from historical power consumption patterns. Then, by combining the similarity between the user's power data changes and those of a reference user, the degree of abnormality in the user's power data within the current preset monitoring period is calculated. This reduces the interference of changes in power demand under the influence of external factors on the identification of abnormal power consumption, improves the accuracy of monitoring abnormal power consumption behavior of users, and enhances the efficiency of power system management.

[0069] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for intelligent processing of electricity meter data, characterized in that, The method includes: For any user, the electricity data of the user at each moment is monitored using an electricity meter to obtain the monitored value of the electricity data within the current preset monitoring period, and the historical electricity data within a preset historical period before the current preset monitoring period is obtained. The electricity data includes current data and voltage data. Using historical energy data within a preset historical period, the predicted value of each energy data point within the current preset monitoring period is obtained. Based on the difference between the monitored value and the predicted value of each energy data point within the current preset monitoring period, it is determined whether there is an overall deviation in the energy data within the current preset monitoring period. If the power data within the current preset monitoring period shows an overall deviation, then at least two reference users are obtained for any user, and the true degree of anomaly in the power data within the current preset monitoring period is obtained based on the similarity of the power data changes between any user and each reference user. Based on the difference between the monitored and predicted values ​​of each power data point for any user within the current preset monitoring period, and the actual degree of anomaly, the degree of power data anomaly for any user within the current preset monitoring period is obtained. Based on the degree of power data anomaly, an abnormal warning is issued for the power consumption of any user.

2. The intelligent data processing method for electricity meters according to claim 1, characterized in that, The acquisition of historical energy data within a preset historical period prior to the current preset monitoring cycle includes: The initial energy data of any user within a preset historical period is obtained using an energy meter. According to the type of each initial energy data, all initial energy data are divided into two initial energy data sequences. For any initial energy data sequence, the mutation index of each initial energy data in the sequence is obtained based on the data fluctuation characteristics of the sequence. Obtain the mutation index of each initial power data in each initial power data sequence. Based on the mutation index of each initial power data in each initial power data sequence, use DBSCAN to cluster the initial power data within a preset historical period to obtain at least one cluster. Among all clusters, obtain the maximum mutation index of the initial current data, which is denoted as the current mutation index threshold. Obtain the maximum mutation index of the initial voltage data, which is denoted as the voltage mutation index threshold. Record the initial current data with mutation indices greater than the current mutation index threshold and the initial voltage data with mutation indices greater than the voltage mutation index threshold as fault power data. For any initial power data sequence, if there is at least one faulty power data in the initial power data sequence, the faulty power data in the initial power data sequence is corrected to obtain a corrected power data sequence. If there is no faulty power data in the initial power data sequence, the initial power data sequence is used as the corrected power data sequence. The corrected power data sequence of each initial power data sequence is obtained to form historical power data within a preset historical period.

3. The intelligent data processing method for electricity meters according to claim 2, characterized in that, The step of correcting faulty power data in any initial power data sequence to obtain a corrected power data sequence includes: For any faulty power data, a first preset number of non-faulty power data that are closest to the time when the faulty power data is located are obtained from any initial power data sequence, forming a target sequence, and a fitting curve of the target sequence is obtained. The fitting value of the time when the faulty power data is located is obtained from the fitting curve and used as the correction value of the faulty power data. Obtain the correction value of each faulty power data in any initial power data sequence, and combine it with all non-faulty power data in any initial power data sequence to form a corrected power data sequence.

4. The intelligent data processing method for electricity meters according to claim 2, characterized in that, The step of obtaining the mutation index of each initial power data in any initial power data sequence based on the data fluctuation characteristics of any initial power data sequence includes: For any initial energy data in any initial energy data sequence, the initial energy data is combined with the first preset number of initial energy data to form a subsequence, and the difference between each initial energy data in the subsequence and the initial energy data at the previous moment is obtained to obtain the change amount of each initial energy data. The absolute value of the difference between the change in each initial energy data (excluding the initial energy data) in the subsequence and the change in the initial energy data is accumulated to obtain the cumulative value of the change difference. The cumulative value of the change difference is then normalized to obtain the mutation index of the initial energy data. In any initial power data sequence, if the number of initial power data preceding any initial power data is less than a first preset number, then the mutation index of any initial power data is marked as 0.

5. The intelligent data processing method for electricity meters according to claim 1, characterized in that, The step of determining whether there is an overall deviation in the power data within the current preset monitoring period based on the difference between the monitored value and the predicted value of each power data point within the current preset monitoring period includes: The electrical energy data of any type within the current preset monitoring period are combined into an electrical energy data sequence. The difference between the monitored value and the predicted value of each electrical energy data in the electrical energy data sequence is obtained, and the cumulative difference value is obtained. The cumulative difference value is normalized to obtain the degree of deviation of the electrical energy data sequence. The deviation of each power data sequence is obtained. If the deviation of at least one power data sequence is greater than the preset deviation threshold, it is confirmed that there is an overall deviation in the power data within the current preset monitoring period.

6. The intelligent data processing method for an electricity meter according to claim 5, characterized in that, The step of obtaining the true degree of anomaly in the power data within the current preset monitoring period based on the similarity of power data changes between any user and each reference user includes: For any type of electrical energy data within the current preset monitoring period, any reference user is recorded as the target reference user. Within a preset historical time period, the historical reference electrical energy data sequence of the target reference user and the historical electrical energy data sequence of the user are obtained under the given electrical energy data. Based on the differences in data changes between the historical electrical energy data sequence and the historical reference electrical energy data sequence, the historical change similarity of the given electrical energy data between the target reference user and the user is obtained. Within the current preset monitoring period, acquire the power data sequence of any user and the reference power data sequence of the target reference user under any type of power data. Based on the data change difference between the power data sequence and the reference power data sequence, obtain the deviation similarity of the power data between the target reference user and any user. Obtain the historical change similarity of the electrical energy data between each reference user and any user, and obtain the corresponding historical change similarity cumulative value. Obtain the ratio of the historical change similarity of the electrical energy data between the target reference user and any user to the historical change similarity cumulative value, and obtain the true anomaly weight of the electrical energy data compared with the target reference user. Normalize the inverse of the deviation similarity to obtain the true outlier value of any power data compared with the target reference user, and obtain the product of the true outlier weight and the true outlier value to obtain the reference outlier degree of any power data compared with the target reference user. Obtain the reference anomaly level of any type of electrical energy data compared with each target reference user, and obtain the cumulative value of the reference anomaly level as the true anomaly level of any type of electrical energy data within the current preset monitoring period.

7. The intelligent data processing method for an electricity meter according to claim 6, characterized in that, The step of obtaining the historical change similarity of any type of electricity data between the target reference user and any user based on the data change differences between the historical electricity data sequence and the historical reference electricity data sequence includes: In the historical reference energy data sequence, a second preset number of historical reference subsequences are selected according to a preset interval time, and each historical reference subsequence includes a third preset number of historical reference energy data. For any historical reference subsequence, historical energy data that is at the same time as each historical reference energy data in any historical reference subsequence is obtained from the historical energy data sequence to form a historical subsequence; Perform first-order difference processing on any historical reference subsequence and the historical subsequence respectively to obtain a historical reference difference sequence and a historical difference sequence. Then, form an element group by combining two elements with the same position number in the historical reference difference sequence and the historical difference sequence. The absolute values ​​of the differences between the two elements in each element group are summed to obtain the difference accumulation value. The reciprocal of the sum of the difference accumulation value and the preset constant is normalized to obtain the initial similarity between any historical reference subsequence and its corresponding historical subsequence. The initial similarity between each historical reference subsequence and its corresponding historical subsequence is obtained, and the mean of the initial similarity is obtained as the historical change similarity between the target reference user and any user for any type of electrical energy data.

8. The intelligent data processing method for an electricity meter according to claim 6, characterized in that, The step of obtaining the deviation similarity of any type of electrical data between the target reference user and any user based on the data change difference between the electrical data sequence and the reference electrical data sequence includes: Obtain the deviation degree of the reference power data sequence, calculate the absolute value of the difference between the deviation degree of the power data sequence and the deviation degree of the reference power data sequence, obtain the deviation degree difference value, obtain the reciprocal of the sum of the deviation degree difference value and a preset constant, and obtain the deviation degree similarity between the target reference user and any user for any type of power data. The difference between the monitored value and the predicted value of each power data in the power data sequence is obtained to obtain a difference sequence. The difference sequence is fitted to obtain a difference fitting curve. The slope of the difference fitting curve is obtained as the deviation trend value of the power data sequence. The difference between the monitored value and the predicted value of each reference energy data in the reference energy data sequence is obtained to obtain a reference difference sequence. The reference difference sequence is fitted to obtain a reference difference fitting curve. The slope of the reference difference fitting curve is obtained as the deviation trend value of the reference energy data sequence. The absolute value of the difference between the deviation trend value of the power data sequence and the deviation trend value of the reference power data sequence is obtained and denoted as the deviation trend difference. The reciprocal of the sum of the deviation trend difference and a preset constant is obtained to obtain the deviation trend similarity between the target reference user and any user for any type of power data. The product of the deviation degree similarity and the deviation trend similarity is obtained to obtain the deviation similarity between the target reference user and any user for any type of electrical energy data.

9. The intelligent data processing method for an electricity meter according to claim 8, characterized in that, The step of obtaining the degree of abnormality of the power data of any user in the current preset monitoring period based on the difference between the monitored value and the predicted value of each power data point of any user in the current preset monitoring period, and the actual degree of abnormality, includes: For any type of electrical energy data within the current preset monitoring period, the deviation degree of the electrical energy data sequence of the given electrical energy data is obtained by multiplying the actual abnormality degree of the given electrical energy data by the product of the deviation degree of the electrical energy data and the actual abnormality degree of the given electrical energy data. The abnormality degree of each type of electrical energy data is obtained, and the average abnormality degree is obtained accordingly. Obtain the absolute value of the difference between the deviation trend values ​​of the two electrical energy data within the current preset monitoring period to obtain the deviation difference value, and obtain the reciprocal of the sum of the deviation difference value and the constant 1 to obtain the data anomaly similarity. The product of the average anomaly level and the data anomaly similarity is obtained to determine the degree of anomaly in the power data of any user in the current preset monitoring period.

10. The intelligent data processing method for an electricity meter according to claim 1, characterized in that, The step of issuing an abnormal warning for the electricity consumption of any user based on the degree of abnormality in the electricity data includes: If the abnormality level of any user's power consumption data in the current preset monitoring period is greater than the preset threshold for abnormality level of power consumption data, then an abnormality warning will be issued for the power consumption of any user.

Images