Power distribution terminal with power supply fault abnormity monitoring and early warning functions

Abstract

The invention relates to the technical field of data processing, in particular to a power distribution terminal with a power supply fault abnormity monitoring and early warning function, and the power distribution terminal comprises a data collection module which is used for obtaining a K value range according to the difference between a real-time data sequence of the power distribution terminal in a current period and a historical data sequence in a historical period; the data analysis module is used for acquiring mutation data in the real-time data sequence; the data processing module is used for obtaining the mutation distribution density according to the distribution condition of the difference between the mutation data and the data in the local range, and obtaining the time distribution density according to the data distribution in the local range of the mutation data; and the abnormity early warning module is used for obtaining a self-adaptive K value according to the abrupt change distribution density, the time distribution density and the K value range of the abrupt change data, and obtaining an LOF value of the abrupt change data by using an LOF algorithm, so that power supply fault abnormity monitoring and early warning are carried out on the power distribution terminal, and the accuracy of fault abnormity early warning is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and particularly to a distribution terminal with a power supply failure abnormal monitoring and early warning function. Background Art

[0002] In the power system, as an important part of the power grid, the stability and reliability of the distribution terminal are directly related to the operation safety of the entire power system. Therefore, it is necessary to monitor the abnormal power supply data of the distribution terminal to give early warning of potential faults and prevent large economic losses and safety hazards.

[0003] The traditional algorithm for abnormal detection of the power supply data of the distribution terminal is the LOF algorithm. The LOF algorithm is an abnormal detection algorithm based on the local density of data, which is used to identify abnormal points in the data set. In the process of early warning of power supply data faults and abnormalities of the distribution terminal, the LOF algorithm can be effectively applied to the abnormal monitoring of power supply data. The K value in the LOF algorithm is a key parameter, which defines the number of the nearest neighbor data points considered when calculating the local density of each data point. However, the traditional LOF algorithm uses a fixed K value. Since the power supply data is affected by various factors, such as the residential power consumption period, the temperature change of the distribution box, the operation of surrounding electrical equipment, etc., it may cause the power supply data to present different data characteristics in different periods. Therefore, the fixed K value cannot be adjusted according to the change of data, which may lead to errors in the identification of abnormal power supply data.

[0004] Therefore, how to adaptively obtain the K value in the LOF algorithm to improve the accuracy of early warning of power supply data faults and abnormalities of the distribution terminal has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, an embodiment of the present invention provides a distribution terminal with a power supply failure abnormal monitoring and early warning function to solve the problem of how to adaptively obtain the K value in the LOF algorithm to improve the accuracy of early warning of power supply data faults and abnormalities of the distribution terminal.

[0006] An embodiment of the present invention provides a distribution terminal with a power supply failure abnormal monitoring and early warning function, and the terminal includes the following steps: A data acquisition module, configured to obtain the real-time power supply data of the distribution terminal at each sampling moment in the current period, obtain a real-time data sequence, obtain the historical data sequence of the distribution terminal in each historical period, and obtain the K value range in the LOF algorithm according to the difference between the real-time data sequence and each historical data sequence; A data analysis module, configured to obtain mutant data in a real-time data sequence according to data fluctuations in the real-time data sequence, and obtain a local change difference value of each mutant data according to the difference between each mutant data and each data within its local range; A data processing module, configured to obtain a mutant distribution density of each mutant data according to the distribution of all local change difference values, obtain a window of each mutant data in the real-time data sequence according to the K value range, and obtain a time distribution density of each mutant data according to the number of mutant data within the window of each mutant data; An abnormal warning module, configured to obtain an adaptive K value of each mutant data according to the mutant distribution density and time distribution density of each mutant data, and the K value range, obtain an LOF value of each mutant data by using the LOF algorithm according to the adaptive K value of each mutant data, and perform power supply fault abnormal monitoring and warning on the distribution terminal according to all LOF values.

[0007] The beneficial effects of the embodiments of the present invention compared with the prior art are: The present invention includes a data acquisition module, which is used to obtain the real-time power supply data of the distribution terminal at each sampling moment in the current cycle, obtain a real-time data sequence, obtain the historical data sequences of the distribution terminal in each historical cycle, and obtain the range of K values in the LOF algorithm according to the differences between the real-time data sequence and each historical data sequence; a data analysis module, which is used to obtain the mutation data in the real-time data sequence according to the data fluctuation condition in the real-time data sequence, and obtain the local change difference value of each mutation data according to the differences between each mutation data and each data within its local range; a data processing module, which is used to obtain the mutation distribution density of each mutation data according to the distribution condition of all local change difference values, obtain the window of each mutation data in the real-time data sequence according to the range of K values, and obtain the time distribution density of each mutation data according to the number of mutation data within the window of each mutation data; an abnormal warning module, which is used to obtain the adaptive K value of each mutation data according to the mutation distribution density and time distribution density of each mutation data, and the range of K values, obtain the LOF value of each mutation data by using the LOF algorithm, and perform abnormal monitoring and warning on the power supply failure of the distribution terminal according to all LOF values. Among them, according to the differences between the real-time power supply data sequence and the historical data sequences, the real-time power supply data is converted into two-dimensional data, and then the range of K values in the LOF algorithm is obtained; according to the differences between each mutation data and each data within its local range, the local change difference value of each mutation data is obtained, so as to convert the mutation data into two-dimensional data, and then the mutation distribution density of each mutation data is obtained; the time distribution density of each mutation data is obtained according to the data distribution within the window of each mutation data, so as to make up for the time characteristics lost by the mutation distribution density of each mutation data. Combining the range of K values, the mutation distribution density and the time distribution density, the adaptive K value of each mutation data is obtained, so that the abnormal detection result obtained by using the LOF algorithm is more accurate, and the probability of missed detection or false detection of the distribution terminal is smaller. Description of the Drawings

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments or the description of the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0009] Figure 1 It is a structural block diagram of a distribution terminal with a function of abnormal monitoring and warning of power supply failure provided in Embodiment 1 of the present invention; Figure 2 It is a clustering scatter diagram after clustering the historical differences of all data in the real-time data sequence provided in the embodiment of the present invention; Figure 3 It is a two-dimensional scatter plot obtained by clustering the local change difference values of all mutation data provided by an embodiment of the present invention. Detailed implementation manners

[0010] The embodiments of the present disclosure will be described in detail below. The examples of the embodiments are shown in the accompanying drawings. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present disclosure, and should not be construed as a limitation to the present disclosure.

[0011] It should be noted that the terms "first", "second", etc. in the specification of the present disclosure and the above accompanying drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present disclosure described herein can be implemented in an order other than those illustrated or described herein. The implementation manners described in the following exemplary embodiments do not represent all implementation manners consistent with the present disclosure. On the contrary, they are merely examples of devices and methods consistent with some aspects of the present disclosure.

[0012] In order to illustrate the technical solution of the present invention, it will be described below through specific embodiments.

[0013] See Figure 1 , which is a structural block diagram of a power distribution terminal with a power supply failure abnormal monitoring and warning function provided by Embodiment 1 of the present invention. As Figure 1 shown, the terminal may include: A data acquisition module 11, configured to obtain real-time power supply data of the power distribution terminal at each sampling moment in the current period, obtain a real-time data sequence, obtain historical data sequences of the power distribution terminal in each historical period, and obtain the K value range in the LOF algorithm according to the difference between the real-time data sequence and each historical data sequence.

[0014] The power supply data of the power distribution terminal includes voltage data and current data. In the embodiments of the present invention, taking the current data as an example, real-time power supply data (i.e., current data) of the power distribution terminal at each sampling moment in the current period is obtained through a current sensor, a real-time data sequence is obtained, one day is set as a period, and the sampling frequency is 1 Hz. There is no limitation here, and the implementer can set it according to the specific scenario.

[0015] The traditional algorithm for anomaly detection of power supply data of distribution terminals is the LOF algorithm. The LOF algorithm is an anomaly detection algorithm based on the local density of data. It identifies outliers by calculating the local density deviation of each data point relative to its neighborhood. In the process of fault anomaly warning for the power supply data of distribution terminals, the LOF algorithm can be effectively applied to the anomaly monitoring of power supply data. The K value in the LOF algorithm is a key parameter, which defines the number of nearest neighbor data points considered when calculating the local density of each data point. However, the traditional LOF algorithm uses a fixed K value. Since the power supply data is affected by various factors, such as the electricity consumption period of residents, the temperature change of the distribution box, the operation of surrounding electrical equipment, etc., it may cause the power supply data to exhibit different data characteristics in different periods. Therefore, in the embodiments of the present invention, the K value in the LOF algorithm is adaptively obtained in combination with the data characteristics of the real-time data sequence of the distribution terminal in the current cycle, so as to improve the accuracy of fault anomaly warning for the power supply data of the distribution terminal.

[0016] Since the principle of the LOF algorithm is to measure the anomaly degree of data points according to the density of each data point and its neighboring points, therefore, first, the range of the K value needs to be obtained according to the data distribution density of the real-time power supply data to prevent unreasonable setting of the K value. Considering that the real-time power supply data sequence is one-dimensional time-series data, while the data distribution density requires two-dimensional data for analysis, so it is necessary to convert the one-dimensional time-series data into two-dimensional data. Also considering that the change of the power supply data of the distribution terminal is mainly affected by the electricity consumption habits of users, and the electricity consumption habits of residential users have a certain regularity, so, in the embodiments of the present invention, the historical data sequence of the distribution terminal in each historical cycle is obtained to convert the real-time power supply data into two-dimensional data according to the difference between the real-time data sequence and each historical data sequence. Specifically: For any data in the real-time data sequence, according to the sampling time of the any data, historical data with the same sampling time as the any data is obtained in each historical data sequence, denoted as target historical data, and the historical mean value of all target historical data is calculated. The absolute value of the difference between the any data and the historical mean value is denoted as the historical difference of the any data.

[0017] A scatter plot is constructed based on all historical differences to obtain a two-dimensional scatter plot of real-time power supply data, where the abscissa represents the real-time power supply data (current data in the embodiments of the present invention), and the ordinate represents the historical difference.

[0018] Furthermore, according to the distribution density of the historical differences of each real-time power supply data in the real-time data sequence, the range of the K value in the LOF algorithm is obtained. The specific steps are as follows: (1) Cluster the data points in the two-dimensional scatter plot of real-time power supply data, and obtain the density distribution characteristic value according to the clustering result.

[0019] Specifically, the density clustering algorithm, such as the DBSCAN algorithm, is used to cluster the data points in the two-dimensional scatter plot of real-time power supply data (i.e., the historical differences of all data in the real-time data sequence), and the clustering result is obtained. There is no limitation here. The implementer can select the clustering algorithm according to the specific scenario. The DBSCAN algorithm is an existing technology and will not be elaborated here. Refer to Figure 2 , which is a clustering scatter plot after clustering the historical differences of all data in the real-time data sequence. Under normal circumstances, the power supply data fluctuates within a certain range. If there is more data within a certain range, it indicates that the power supply data within this range is normal, such as Figure 2 the data points in the No. 1 circle in Figure 2 ; due to the uncontrollability of user behavior, there will be situations where the historical difference is large, such as Figure 2 the data in the No. 2 circle in

[0020] ; there are also some data that are significantly different from other data, such as the data in the No. 3 circle in . These data are very likely to be abnormal data, which are shown as isolated points in the clustering scatter plot. When calculating the K value range, the isolated points are not used as reference data. Therefore, the isolated points are removed according to the clustering result to obtain at least one clustering cluster, and the number of elements in each clustering cluster is recorded as the corresponding cluster density; the cluster density mean of all cluster densities is obtained, and 1 minus the reciprocal of the cluster density mean is obtained to get the first variable; the number of clustering clusters greater than or equal to the cluster density mean is counted, and 1 minus the reciprocal of the number of clustering clusters is obtained to get the second variable; the average value between the first variable and the second variable is calculated to obtain the density distribution characteristic value of all clustering clusters.

[0021] It should be noted that the larger the cluster density mean and the larger the number of clustering clusters greater than or equal to the cluster density mean, the more concentrated the data distribution and the larger the density distribution characteristic value.

[0022] (2) Obtain the K value range in the LOF algorithm according to the density distribution characteristic value.

[0023] The larger the density distribution eigenvalue is, the more elements there are in each clustering cluster. If the value of K in the LOF algorithm is small at this time, it may lead to insufficient points in the neighborhood to accurately estimate the local density, and then lead to inaccurate identification of outliers using the LOF algorithm. Therefore, it is necessary to appropriately increase the value of K at this time; conversely, it is necessary to appropriately decrease the value of K to prevent mixing data in clustering clusters with different densities, resulting in inaccurate estimation of local density, and then leading to inaccurate identification of outliers using the LOF algorithm. The specific method for obtaining the range of the value of K in the LOF algorithm is as follows: Take the minimum cluster density as the lower limit of the initial value of K in the LOF algorithm, denoted as Take the maximum cluster density as the upper limit of the initial value of K in the LOF algorithm, denoted as Calculate the difference between the upper limit of the initial value of K and the lower limit of the initial value of K to obtain the difference of the initial value range of K, that is Round up the product of the density distribution eigenvalue and the difference of the initial value range of K to obtain the adjustment coefficient of the K value range, that is ; In the embodiments of the present invention, according to the idea of rounding, the preset density distribution eigenvalue is set to 0.5, which is not limited here, and the implementer can set it according to the specific scenario. If the density distribution eigenvalue is less than 0.5, it is necessary to appropriately reduce the upper limit of the initial value of K, that is, subtract the adjustment coefficient of the K value range from the upper limit of the initial value of K to obtain the final upper limit of the K value, denoted as According to the lower limit of the initial value of K and the final upper limit of the K value, obtain the range of the value of K in the LOF algorithm Among them, the calculation formula for the final upper limit of the K value is: Among them, represents the final upper limit of the K value, represents the upper limit of the initial value of K, represents the lower limit of the initial value of K, A represents the density distribution eigenvalue, represents the rounding-up symbol; If the density distribution eigenvalue is less than 0.5, it is necessary to appropriately increase the lower limit of the initial value of K, that is, calculate the sum of the lower limit of the initial value of K and the adjustment coefficient of the K value range to obtain the final lower limit of the K value. According to the final lower limit of the K value and the upper limit of the initial value of K, obtain the range of the value of K in the LOF algorithm Among them, the calculation formula for the final lower limit of the K value is: Among them, represents the final lower limit of the K value, represents the upper limit of the initial value of K, represents the lower limit of the initial value of K, A represents the density distribution eigenvalue, represents the rounding-up symbol.

[0024] So far, the range of the K value in the LOF algorithm is obtained and denoted as .

[0025] The data analysis module 12 is configured to obtain the mutation data in the real-time data sequence according to the data fluctuation condition in the real-time data sequence, and obtain the local change difference value of each mutation data according to the difference between each mutation data and each data within its local range.

[0026] Since abnormal data generally manifests as deviating from the normal value or surrounding values, in the embodiments of the present invention, according to the data fluctuation condition in the real-time data sequence, the data deviating from the normal value or surrounding values are screened out and denoted as mutation data, so as to further analyze the mutation data and obtain the adaptive K value of each mutation data in the LOF algorithm. The steps for obtaining the mutation data in the real-time data sequence are as follows: (1) According to the change rate of each data in the real-time data sequence, obtain the mutation degree of each data, and screen out the suspected mutation data according to the mutation degree.

[0027] In the embodiments of the present invention, taking the i-th data in the real-time data sequence as an example, the specific method for obtaining the mutation degree of the i-th data is as follows: Construct a scatter plot according to the real-time data sequence, where the abscissa of the scatter plot represents the sampling time, and the ordinate represents the real-time power supply data; If the i-th data is not the first data in the scatter plot and is not the last data in the scatter plot, calculate the left slope between the i-th data and its left adjacent data, and the right slope between the i-th data and its right adjacent data, calculate the absolute value of the difference between the left slope and the right slope to obtain the change rate of the i-th data, and calculate 1 minus the reciprocal of the change rate to obtain the fluctuation degree of the i-th data. In the real-time data sequence, if the number of data before the i-th data is greater than or equal to the first preset number, obtain the first preset number of data before the i-th data in the real-time data sequence, denoted as reference data, and the positions of the reference data are continuous with the position of the i-th data. In the embodiments of the present invention, the first preset number is set to 3, which is not limited herein, and the implementer can set it according to the specific scenario. Calculate the mean value of all reference data to obtain the reference value of the i-th data, calculate the absolute value of the difference between the i-th data and the reference value to obtain the reference difference, calculate 1 minus the reciprocal of the reference difference to obtain the mutation probability of the i-th data, and calculate the average value between the mutation probability of the i-th data and the fluctuation degree to obtain the mutation degree of the i-th data.

[0028] In an embodiment, the calculation formula for the mutation degree of the i-th data is as follows: Among them, represents the mutation degree of the i-th data, represents the right slope, represents the left slope, represents the value of the i-th data, represents the first preset quantity, that is, the quantity of reference data, represents the value of the j-th reference data, and 1 represents a constant, represents the absolute value symbol.

[0029] It should be noted that is the change rate of the i-th data. Specifically, if the i-th data is the first data in the scatter plot, then calculate the right slope between the i-th data and its right adjacent data to obtain the change rate of the i-th data, that is ; if the i-th data is the last data in the scatter plot, then calculate the left slope between the i-th data and its left adjacent data to obtain the change rate of the i-th data, that is . is the mutation probability of the i-th data. Specifically, if the number of data before the i-th data is less than the first preset quantity, then the mutation probability of the i-th data is not calculated, and the fluctuation degree of the i-th data is used as the mutation degree of the i-th data. That is, the calculation formula for the mutation degree of the i-th data is: . The larger the change rate, the more turning points the i-th data has, and the more obvious the change of the i-th data. Furthermore the larger it is, the greater the mutation degree of the i-th data; the larger it is, the greater the difference between the i-th data and the previous first preset quantity of data, and the more likely the i-th data is a mutated data. Furthermore the larger it is, the greater the mutation degree of the i-th data.

[0030] Considering that the distribution terminal belongs to the power system and is inevitably affected by electromagnetic interference, there may be a situation where data superposition causes abnormal power supply data. Therefore, in the embodiments of the present invention, according to experimental statistics, the preset mutation degree threshold is set to 0.6. This is not limited here, and the implementer can set it according to the specific scenario. If the mutation degree of the i-th data is greater than 0.6, then the i-th data is marked as a suspected mutated data. Similarly, all suspected mutated data in the real-time data sequence are obtained to further analyze the suspected mutated data, screen out the interference data that may be affected by electromagnetic interference, and then obtain the truly abnormal mutated data, improving the accuracy of the determination of mutated data.

[0031] (2) In the real-time data sequence, according to the change trend of each suspected mutated data within its local range, the mutated data in the real-time data sequence is obtained.

[0032] Since the interference data affected by electromagnetic interference has instantaneous recoverability, the change trends between the data before and after the interference data are similar, and the data difference between the data before and after it is small. Based on the above characteristics, taking the f-th suspected mutation data as an example, the interference probability of the f-th suspected mutation data is obtained, and then the truly abnormal mutation data is screened out. Specifically: In the real-time data sequence, if the f-th suspected mutation data is not the first data and the number of data before the f-th suspected mutation data is less than the second preset number, then all the data before the f-th suspected mutation data and the second preset number of data after the f-th suspected mutation data are recorded as target data. The positions of the target data are continuous with the position of the f-th suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before the f-th suspected mutation data and the target data after the f-th suspected mutation data to obtain the first trend similarity of the f-th suspected mutation data; calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of the f-th suspected mutation data to obtain the second trend similarity of the f-th suspected mutation data; perform weighted summation on the first trend similarity and the first trend similarity to obtain the interference probability of the f-th suspected mutation data; If the f-th suspected mutation data is not the last data and the number of data after the f-th suspected mutation data is less than the second preset number, then all the data after the f-th suspected mutation data and the second preset number of data before the f-th suspected mutation data are recorded as target data. The positions of the target data are continuous with the position of the f-th suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before the f-th suspected mutation data and the target data after the f-th suspected mutation data to obtain the first trend similarity of the f-th suspected mutation data, calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of the f-th suspected mutation data to obtain the second trend similarity of the f-th suspected mutation data, and perform weighted summation on the first trend similarity and the first trend similarity to obtain the interference probability of the f-th suspected mutation data; If the number of data before the f-th suspected mutation data is greater than or equal to the second preset quantity, and the number of data after the f-th suspected mutation data is greater than or equal to the second preset quantity, then the second preset quantity of data after the f-th suspected mutation data and the second preset quantity of data before the f-th suspected mutation data are recorded as target data. The positions of the target data are consecutive with the position of the f-th suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before the f-th suspected mutation data and the target data after the f-th suspected mutation data to obtain the first trend similarity of the f-th suspected mutation data. Calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of the f-th suspected mutation data to obtain the second trend similarity of the f-th suspected mutation data. Perform weighted summation on the first trend similarity and the second trend similarity to obtain the interference probability of the f-th suspected mutation data. To reduce the influence of the fluctuations of other data on the f-th suspected mutation data, in the embodiments of the present invention, the second preset quantity is set to 3, which is not limited herein, and the implementer can set it according to the specific scenario.

[0033] In one embodiment, the calculation formula for the interference probability of the f-th suspected mutation data is: Wherein, represents the interference probability of the f-th suspected mutation data, represents the first weight, represents the first trend similarity, represents the second weight, represents the left adjacent data of the f-th suspected mutation data in the real-time data sequence, represents the right adjacent data of the f-th suspected mutation data in the real-time data sequence, represents the absolute value symbol.

[0034] It should be noted that, the larger the , the more similar the change trends of the data before and after the f-th suspected mutation data, and thus the larger the , the more likely the f-th suspected mutation data is interference data affected by electromagnetic interference; the smaller the , the smaller the data difference between the data before and after the f-th suspected mutation data, and thus the larger the , the more likely the f-th suspected mutation data is interference data affected by electromagnetic interference. Since the change trends of the data before and after the f-th suspected mutation data can better reflect the recoverability of electromagnetic interference, so is set, which is not limited herein, and the implementer can set it according to the specific scenario.

[0035] Specifically, if the f-th suspected mutation data is the first data in the real-time data sequence, then obtain the second preset number of data after the f-th suspected mutation data, denoted as target data. The positions of the target data are consecutive with the position of the f-th suspected mutation data. Calculate the absolute value of the difference between the f-th suspected mutation data and each target data to obtain a data difference. Calculate the average of all data differences to obtain an average data difference. Subtract the average data difference from the constant 1 to obtain the interference probability of the f-th suspected mutation data. That is, the calculation formula for the interference probability of the f-th suspected mutation data is: , where represents the interference probability of the f-th suspected mutation data, represents the second preset number, represents the f-th suspected mutation data in the real-time data sequence, represents the -th target data after the f-th suspected mutation data in the real-time data sequence; If the f-th suspected mutation data is the last data in the real-time data sequence, then obtain the second preset number of data before the f-th suspected mutation data, denoted as target data. The positions of the target data are consecutive with the position of the f-th suspected mutation data. Calculate the absolute value of the difference between the f-th suspected mutation data and each target data to obtain a data difference. Calculate the average of all data differences to obtain an average data difference. Subtract the average data difference from the constant 1 to obtain the interference probability of the f-th suspected mutation data. That is, the calculation formula for the interference probability of the f-th suspected mutation data is: , where represents the interference probability of the f-th suspected mutation data, represents the second preset number, represents the f-th suspected mutation data in the real-time data sequence, represents the h-th target data before the f-th suspected mutation data in the real-time data sequence.

[0036] According to experimental statistics, set the preset interference probability threshold to 0.7. There is no limit here, and the implementer can set it according to the specific scenario. If the interference probability of the f-th suspected mutation data is less than or equal to 0.7, then determine that the f-th suspected mutation data is a truly abnormal mutation data. At this time, record the f-th suspected mutation data as a mutation data; if the interference probability of the f-th suspected mutation data is greater than 0.7, then determine that the f-th suspected mutation data is an interference data affected by electromagnetic interference. At this time, record the f-th suspected mutation data as an interference data.

[0037] Similarly, all mutation data in the real-time data sequence is obtained. If there is no mutation data, it is determined that the distribution terminal is normal; if there is mutation data, further analysis of the mutation data is required to obtain the adaptive K value of each mutation data in the LOF algorithm.

[0038] Since the principle of the LOF algorithm measures the abnormality degree of a data point according to the density between each data point and its neighboring points, therefore, first, the adaptive K value needs to be obtained according to the data distribution density of the mutation data. Considering that the mutation data is one-dimensional time-series data, while the data distribution density requires two-dimensional data for analysis, so the one-dimensional time-series data needs to be transformed into two-dimensional data. Therefore, in the embodiment of the present invention, according to the difference between each mutation data and each data within its local range, the local change difference value of each mutation data is obtained, so as to transform the mutation data into two-dimensional data according to the local change difference value, which is convenient for the subsequent density distribution of the mutation data.

[0039] Considering that interference data will cause more or less influence, so before obtaining the local change difference value of each mutation data, it is necessary to determine whether there is interference data in the real-time data sequence. If there is interference data, the value of the interference data needs to be modified. Specifically: Obtain the target quantity of all target data corresponding to any interference data, obtain the mean value of the target quantity of neighboring non-interference data of the any interference data in the real-time data sequence, and replace the any interference data with the mean value of all neighboring non-interference data to obtain a new real-time data sequence.

[0040] If there is no interference data, the real-time data sequence is used as the new real-time data sequence.

[0041] Further, according to the new real-time data sequence, the local change difference value of each mutation data is obtained. Specifically: In the new real-time data sequence, for any mutation data, obtain the third preset quantity of neighboring data of the any mutation data, that is, according to the position of each data in the new real-time data sequence, calculate the distance between the any mutation data and other data in the new real-time data sequence respectively, sort all the distances in ascending order to obtain a distance sequence, and select the data in the new real-time data sequence corresponding to the first third preset quantity of distances in the distance sequence as the neighboring data of the any mutation data. In order to reduce the influence of the fluctuation of other data on the any mutation data, in the embodiment of the present invention, the third preset quantity is set to 4, which is not limited here, and the implementer can set it according to the specific scenario. Calculate the absolute value of the difference between the any mutation data and each neighboring data respectively, and calculate the average value of all absolute values of the differences to obtain the local change difference value of the any mutation data.

[0042] Similarly, the local change difference value of each mutation data is obtained.

[0043] The data processing module 13 is configured to obtain the mutation distribution density of each mutation data according to the distribution of all local change difference values, obtain the window of each mutation data in the real-time data sequence according to the K value range, and obtain the time distribution density of each mutation data according to the number of mutation data within the window of each mutation data.

[0044] Since the principle of the LOF algorithm is to measure the abnormality degree of a data point according to the density between each data point and its neighboring points, in the embodiment of the present invention, a density clustering algorithm, such as the DBSCAN algorithm, is used to cluster all local change difference values to obtain at least one clustering cluster. There is no limitation here, and the implementer can select a clustering algorithm according to the specific scenario. The DBSCAN algorithm is a prior art and will not be elaborated here. Refer to Figure 3 which is a two-dimensional scatter plot after clustering the local change difference values of all mutation data. Figure 3 In , the abscissa represents the mutation data, and the ordinate represents the local change difference value. For the local change difference value of the t-th mutation data, the number of members in the clustering cluster where the local change difference value of the t-th mutation data is located is used as the mutation distribution density of the t-th mutation data, denoted as Similarly, the mutation distribution densities of all mutation data are obtained to obtain its adaptive K value according to the mutation distribution density of each mutation data: the greater the mutation distribution density, the higher the possibility that the power supply data of the distribution terminal is abnormal. At this time, a smaller K value should be selected to capture subtle changes; on the contrary, a larger K value should be selected to avoid false detection.

[0045] Considering that the scatter plot after clustering the local change difference values of all mutation data loses the time characteristics, in the embodiment of the present invention, according to the K value range in the LOF algorithm obtained by the data acquisition module 11, a window for each mutation data is constructed to obtain the time distribution density of each mutation data according to the data distribution within the window.

[0046] Taking the t-th mutation data as an example, the specific method for obtaining the window of the t-th mutation data according to the K value range is as follows: Calculate the difference between the maximum value and the minimum value in the K value range to obtain the K value range difference. If the K value range difference is an even number, the sum of the constant 1 and the K value range difference is recorded as the window size. If the K value range difference is an odd number, the K value range difference is recorded as the window size.

[0047] In the new real-time data sequence, if the number of data before the t-th mutation data is greater than or equal to the fourth preset number, and the number of data after the t-th mutation data is greater than or equal to the fourth preset number, then in the new real-time data sequence, according to the window size, a window centered on the t-th mutation data is constructed, and the fourth preset number is half of the difference between the K value range and the constant 1; If the number of data before the t-th mutation data is less than the fourth preset number, then in the new real-time data sequence, according to the window size, all the data before the t-th mutation data, and some of the data after the t-th mutation data, a window centered on the t-th mutation data is constructed; If the number of data after the t-th mutation data is less than the fourth preset number, then in the new real-time data sequence, according to the window size, all the data after the t-th mutation data, and some of the data before the t-th mutation data, a window centered on the t-th mutation data is constructed.

[0048] Further, according to the data distribution within the window of the t-th mutation data, the time distribution density of the t-th mutation data is obtained. Specifically: Obtain the number of other mutation data within the window of the t-th mutation data except the t-th mutation data. If the number is 0, then set the time distribution density of the t-th mutation data to 0; If the number is not 0, then calculate the constant 1 minus the reciprocal of the number to obtain the first density; calculate the time interval between the sampling times corresponding to every two adjacent mutation data within the window of the t-th mutation data, and take the reciprocal of the mean of all the time intervals as the second density; calculate the reciprocal of the variance of all the time intervals to obtain the third density; calculate the average value among the first density, the second density, and the third density to obtain the time distribution density of the t-th mutation data.

[0049] In an embodiment, the calculation formula for the time distribution density of the t-th mutation data is: Wherein, represents the time distribution density of the t-th mutation data, N represents the number of other mutation data within the window of the t-th mutation data except the t-th mutation data, represents the time interval between the sampling times corresponding to the -th mutation data and the -th mutation data within the window of the t-th mutation data, and S represents the variance of all the time intervals.

[0050] It should be noted that the larger N is, the more mutation data there are within the window of the t-th mutation data, that is, within the time range corresponding to the window of the t-th mutation data, the more mutation data there are. Furthermore, the larger it is, the greater the time distribution density of the t-th mutation data; the smaller it is, the smaller the time interval between two adjacent mutation data. Furthermore, the larger it is, the greater the time distribution density of the t-th mutation data; the smaller S is, the more similar the time intervals between every two adjacent mutation data are, and the more balanced the distribution of the mutation data within the window of the t-th mutation data is. Furthermore, the larger it is, the greater the time distribution density of the t-th mutation data.

[0051] Thus, the mutation distribution density and time distribution density of the t-th mutation data are obtained.

[0052] The anomaly warning module 14 is used to obtain the adaptive K value of each mutation data according to the mutation distribution density and time distribution density of each mutation data, and the K value range, and use the LOF algorithm to obtain the LOF value of each mutation data according to the adaptive K value of each mutation data, and perform power supply fault anomaly monitoring and warning on the distribution terminal according to all LOF values.

[0053] After obtaining the mutation distribution density and time distribution density of the t-th mutation data through the data processing module 13, according to the mutation distribution density and time distribution density of the t-th mutation data, and the K value range in the LOF algorithm obtained by the data acquisition module 11, the adaptive K value of the t-th mutation data in the LOF algorithm is obtained: the greater the mutation distribution density and the greater the time distribution density, the higher the possibility that the power supply data of the distribution terminal is abnormal. At this time, a smaller K value should be selected to capture subtle changes; on the contrary, a larger K value should be selected to avoid false detection. Specifically: Subtract the mutation distribution density of the t-th mutation data from the constant 1 to obtain the first eigenvalue of the t-th mutation data. Subtract the time distribution density of the t-th mutation data from the constant 1 to obtain the second eigenvalue of the t-th mutation data, and perform weighted summation on the first eigenvalue and the second eigenvalue to obtain the weighted summation result; Calculate the difference between the maximum value and the minimum value in the K value range to obtain the K value range difference. Perform a floor operation on the product of the K value range difference and the weighted summation result to obtain the K value adjustment coefficient, and calculate the sum of the minimum value in the K value range and the K value adjustment coefficient to obtain the adaptive K value of the t-th mutation data.

[0054] In an embodiment, the calculation formula for the adaptive K value of the t-th mutation data is: Among them, represents the adaptive K value of the t-th mutation data, represents the minimum value in the K value range, represents the difference between the maximum value and the minimum value in the K value range, that is, the K value range difference, represents the third weight, represents the mutation distribution density of the t-th mutation data, represents the fourth weight, represents the time distribution density of the t-th mutation data, represents the floor symbol.

[0055] It should be noted that the greater the mutation distribution density of the t-th mutation data and the greater the time distribution density, the higher the possibility that the power supply data of the distribution terminal is abnormal. At this time, a smaller K value should be selected to capture subtle changes, and then the smaller. Since the mutation distribution density is obtained based on the differences between the mutation data and each data within its local range, it has more reference value. Therefore, in the embodiments of the present invention, , , which is not limited here, and the implementer can set it according to the specific scenario.

[0056] Similarly, the adaptive K value of each mutation data is obtained. Further, according to the adaptive K value of each mutation data, the LOF value of each mutation data is obtained by using the LOF algorithm.

[0057] Since the LOF value is usually a non-negative real number, if the LOF value is greater than 1, it means that the possibility of being an outlier is higher. Therefore, in the embodiments of the present invention, the preset abnormal threshold is set to 1, which is not limited here, and the implementer can set it according to the specific scenario. Among all the LOF values, if there is at least one LOF value greater than 1, an abnormal power supply fault warning is given to the distribution terminal to notify the relevant staff in time; If all LOF values are less than or equal to 1, according to the real-time data sequence, the ARIMA algorithm is used to predict the power supply data of the distribution terminal at a future moment to obtain predicted data. The predicted data is added to the real-time data sequence as the last data in the real-time data sequence. According to the data processing module 13, the mutation distribution density and time distribution density of the predicted data are obtained. Further, according to the K value range, the mutation distribution density and time distribution density of the predicted data, the adaptive K value of the predicted data is obtained. Further, according to the adaptive K value of the predicted data, the LOF value of the predicted data is obtained by using the LOF algorithm. If the LOF value of the predicted data is greater than 1, an abnormal power supply fault warning is given to the distribution terminal to notify the relevant staff in time. Among them, the LOF algorithm and the ARIMA algorithm are existing technologies and will not be elaborated here.

[0058] In summary, the present invention includes a data acquisition module for obtaining real-time power supply data of a distribution terminal at each sampling moment in the current cycle to obtain a real-time data sequence, obtaining historical data sequences of the distribution terminal in each historical cycle, and obtaining the range of the K value in the LOF algorithm according to the differences between the real-time data sequence and each historical data sequence; a data analysis module for obtaining mutation data in the real-time data sequence according to the data fluctuation condition in the real-time data sequence, and obtaining the local change difference value of each mutation data according to the differences between each mutation data and each data within its local range; a data processing module for obtaining the mutation distribution density of each mutation data according to the distribution condition of all local change difference values, obtaining the window of each mutation data in the real-time data sequence according to the range of the K value, and obtaining the time distribution density of each mutation data according to the number of mutation data within the window of each mutation data; an abnormal warning module for obtaining the adaptive K value of each mutation data according to the mutation distribution density and time distribution density of each mutation data, and the range of the K value, obtaining the LOF value of each mutation data by using the LOF algorithm according to the adaptive K value of each mutation data, and performing abnormal monitoring and warning on the power supply failure of the distribution terminal according to all LOF values. Among them, according to the differences between the real-time power supply data sequence and the historical data sequence, the real-time power supply data is converted into two-dimensional data, and then the range of the K value in the LOF algorithm is obtained; according to the differences between each mutation data and each data within its local range, the local change difference value of each mutation data is obtained to convert the mutation data into two-dimensional data, and then the mutation distribution density of each mutation data is obtained; according to the data distribution within the window of each mutation data, the time distribution density of each mutation data is obtained to make up for the loss of the time characteristics of the mutation distribution density of each mutation data. Combining the range of the K value, the mutation distribution density, and the time distribution density, the adaptive K value of each mutation data is obtained, so that the abnormal detection result obtained by using the LOF algorithm is more accurate, and the probability of missed detection or false detection of the distribution terminal is smaller.

[0059] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A distribution terminal with a power supply failure abnormal monitoring and early warning function, characterized in that, The power distribution terminal with power supply failure abnormality monitoring and early warning function includes: The data acquisition module is used to obtain the real-time power supply data of the distribution terminal at each sampling moment in the current cycle, obtain the real-time data sequence, obtain the historical data sequence of the distribution terminal in each historical cycle, and obtain the K value range in the LOF algorithm based on the difference between the real-time data sequence and each historical data sequence; The data analysis module is used to obtain mutation data in the real-time data sequence according to the data fluctuation in the real-time data sequence, and obtain the local change difference value of each mutation data according to the difference between each mutation data and each data in its local range; The data processing module is used to obtain the mutation distribution density of each mutation data according to the distribution of all local change difference values, obtain the window of each mutation data in the real-time data sequence according to the K value range, and obtain the time distribution density of each mutation data according to the number of mutation data in the window of each mutation data; The abnormal warning module is used to obtain the adaptive K value of each mutation data according to the mutation distribution density and time distribution density of each mutation data, as well as the K value range. According to the adaptive K value of each mutation data, the LOF value of each mutation data is obtained by using the LOF algorithm. According to all LOF values, the distribution terminal is monitored and warned of power supply failure anomalies.

2. The distribution terminal with the power supply fault abnormal monitoring and early warning function according to claim 1, characterized in that In the data acquisition module, the K value range in the LOF algorithm is obtained according to the difference between the real-time data sequence and each historical data sequence, including: For any data in the real-time data sequence, according to the sampling time of any data, obtain the historical data with the same sampling time as the any data in each historical data sequence, record them as target historical data, calculate the historical mean of all target historical data, and record the absolute value of the difference between any data and the historical mean as the historical difference of any data; Clustering the historical differences of all data in the real-time data sequence to obtain a clustering result, removing isolated points according to the clustering result to obtain at least one cluster, and recording the number of elements in each cluster as the corresponding cluster density; Obtaining a cluster density mean of all cluster densities, and subtracting the reciprocal of the cluster density mean from a constant 1 to obtain a first variable; Counting the number of clusters whose cluster density is greater than or equal to the mean value of the cluster density, and subtracting the reciprocal of the number of clusters from a constant 1 to obtain a second variable; Calculate the average value between the first variable and the second variable to obtain density distribution characteristic values of all clusters; The minimum cluster density is used as the lower limit of the initial K value in the LOF algorithm, and the maximum cluster density is used as the upper limit of the initial K value in the LOF algorithm. The difference between the upper limit of the initial K value and the lower limit of the initial K value is calculated to obtain the initial K value range difference. The product between the density distribution characteristic value and the initial K value range difference is rounded up to obtain the K value range adjustment coefficient; If the density distribution eigenvalue is less than the preset density distribution eigenvalue, subtract the K - value range adjustment coefficient from the upper limit of the initial K - value to obtain the upper limit of the final K - value. According to the lower limit of the initial K - value and the upper limit of the final K - value, obtain the K - value range in the LOF algorithm; If the density distribution eigenvalue is greater than or equal to the preset density distribution eigenvalue, calculate the sum of the lower limit of the initial K - value and the K - value range adjustment coefficient to obtain the lower limit of the final K - value. According to the lower limit of the final K - value and the upper limit of the initial K - value, obtain the K - value range in the LOF algorithm.

3. The power distribution terminal with the power supply failure abnormal monitoring and early warning function according to claim 1, characterized in that, In the data analysis module, according to the data fluctuation situation in the real - time data sequence, obtain the mutation data in the real - time data sequence, including: Construct a scatter plot according to the real - time data sequence, where the abscissa of the scatter plot represents the sampling time and the ordinate represents the real - time power supply data; For any data in the scatter plot, if the any data is the first data in the scatter plot, calculate the right slope between the any data and its right - adjacent data to obtain the change rate of the any data; If the any data is the last data in the scatter plot, calculate the left slope between the any data and its left - adjacent data to obtain the change rate of the any data; If the any data is not the first data in the scatter plot and not the last data in the scatter plot, calculate the left slope between the any data and its left - adjacent data, and the right slope between the any data and its right - adjacent data, and calculate the absolute value of the difference between the left slope and the right slope to obtain the change rate of the any data; Calculate the reciprocal of 1 minus the change rate to obtain the fluctuation degree of the any data; In the real - time data sequence, if the number of data before the any data is greater than or equal to the first preset number, obtain the first preset number of data before the any data in the real - time data sequence, denoted as reference data. The positions of the reference data are continuous with the position of the any data. Calculate the mean value of all reference data to obtain the reference value of the any data. Calculate the absolute value of the difference between the any data and the reference value to obtain the reference difference. Calculate the reciprocal of 1 minus the reference difference to obtain the mutation probability of the any data. Calculate the average value between the mutation probability of the any data and the fluctuation degree to obtain the mutation degree of the any data; If the number of data before the any data is less than the first preset number, use the fluctuation degree of the any data as the mutation degree of the any data; If the mutation degree of the any data is greater than the preset mutation degree threshold, mark the any data as suspected mutation data; In the real - time data sequence, according to the change trend of each suspected mutation data in its local range, obtain the mutation data in the real - time data sequence.

4. The power distribution terminal with a power supply failure abnormal monitoring and early warning function according to claim 3, characterized in that, In the data analysis module, according to the change trend of each suspected mutation data in its local range, obtain the mutation data in the real - time data sequence, including: For any suspected mutation data, in the real-time data sequence, if the any suspected mutation data is the first data, obtain the second preset number of data after the any suspected data, denoted as target data. The positions of the target data are consecutive with the position of the any suspected mutation data. Calculate the absolute value of the difference between the any suspected mutation data and each target data to obtain the data difference. Calculate the mean of all data differences to obtain the average data difference. Subtract the average data difference from the constant 1 to obtain the interference probability of the any suspected mutation data; If the any suspected mutation data is the last data, obtain the second preset number of data before the any suspected data, denoted as target data. The positions of the target data are consecutive with the position of the any suspected mutation data. Calculate the absolute value of the difference between the any suspected mutation data and each target data to obtain the data difference. Calculate the mean of all data differences to obtain the average data difference. Subtract the average data difference from the constant 1 to obtain the interference probability of the any suspected mutation data; If the any suspected mutation data is not the first data and the number of data before the any suspected mutation data is less than the second preset number, then denote all the data before the any suspected mutation data and the second preset number of data after the any suspected mutation data as target data. The positions of the target data are consecutive with the position of the any suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before the any suspected mutation data and the target data after the any suspected mutation data to obtain the first trend similarity of the any suspected mutation data; calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of the any suspected mutation data to obtain the second trend similarity of the any suspected mutation data; perform weighted summation on the first trend similarity and the first trend similarity to obtain the interference probability of the any suspected mutation data; If the any suspected mutation data is not the last data and the number of data after the any suspected mutation data is less than the second preset number, then denote all the data after the any suspected mutation data and the second preset number of data before the any suspected mutation data as target data. The positions of the target data are consecutive with the position of the any suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before the any suspected mutation data and the target data after the any suspected mutation data to obtain the first trend similarity of the any suspected mutation data, calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of the any suspected mutation data to obtain the second trend similarity of the any suspected mutation data, and perform weighted summation on the first trend similarity and the first trend similarity to obtain the interference probability of the any suspected mutation data; If the number of data before any of the suspected mutation data is greater than or equal to the second preset number, and the number of data after any of the suspected mutation data is greater than or equal to the second preset number, then the second preset number of data after any of the suspected mutation data and the second preset number of data before any of the suspected mutation data are recorded as target data. The positions of the target data are continuous with the position of any of the suspected mutation data. Calculate the reciprocal of the DTW distance between the target data before any of the suspected mutation data and the target data after any of the suspected mutation data to obtain the first trend similarity of any of the suspected mutation data. Calculate the reciprocal of the absolute value of the difference between the left adjacent data and the right adjacent data of any of the suspected mutation data to obtain the second trend similarity of any of the suspected mutation data. Perform weighted summation on the first trend similarity and the first trend similarity to obtain the interference probability of any of the suspected mutation data; If the interference probability of any of the suspected mutation data is less than or equal to the preset interference probability threshold, then any of the suspected mutation data is recorded as mutation data.

5. The power distribution terminal with a power supply failure abnormal monitoring and early warning function according to claim 4, characterized in that, In the data analysis module, according to the differences between each mutation data and each data within its local range, the local change difference value of each mutation data is obtained, including: In the real-time data sequence, if the interference probability of any of the suspected mutation data is greater than the preset interference probability threshold, then any of the suspected mutation data is recorded as interference data. If there is interference data in the real-time data sequence, then obtain the target number of all target data corresponding to any interference data. Obtain the mean value of the target number of near-neighbor non-interference data of any interference data in the real-time data sequence. Replace any interference data with the mean value of all near-neighbor non-interference data to obtain a new real-time data sequence. If there is no interference data in the real-time data sequence, then the real-time data sequence is used as the new real-time data sequence. In the new real-time data sequence, for any mutation data, obtain the third preset number of near-neighbor data of any mutation data, calculate the absolute value of the difference between any mutation data and each near-neighbor data respectively, and calculate the average value of all absolute values of differences to obtain the local change difference value of any mutation data.

6. The power distribution terminal with a power supply failure abnormal monitoring and early warning function according to claim 1, characterized in that, In the data processing module, according to the distribution of all local change difference values, the mutation distribution density of each mutation data is obtained, including: Cluster all local change difference values to obtain at least one clustering cluster. For any local change difference value, obtain the number of members of the clustering cluster where any local change difference value is located to obtain the mutation distribution density of the mutation data corresponding to any local change difference value.

7. The distribution terminal with the power supply failure abnormal monitoring and early warning function according to claim 5, characterized in that In the data processing module, according to the K value range, obtain the window of each mutation data in the real-time data sequence, including: Calculate the difference between the maximum value and the minimum value in the K value range to obtain the K value range difference. If the K value range difference is an even number, then record the sum between the constant 1 and the K value range difference as the window size. If the K value range difference is an odd number, then record the K value range difference as the window size. In the new real-time data sequence, for any mutation data, if the number of data before the any mutation data is greater than or equal to a fourth preset quantity, and the number of data after the any mutation data is greater than or equal to the fourth preset quantity, then in the new real-time data sequence, according to the window size, a window centered on the any mutation data is constructed, and the fourth preset quantity is one half of the difference obtained by subtracting 1 from the K value range difference; If the number of data before the any mutation data is less than the fourth preset quantity, then in the new real-time data sequence, according to the window size, all the data before the any mutation data, and some of the data after the any mutation data, a window centered on the any mutation data is constructed; If the number of data after the any mutation data is less than the fourth preset quantity, then in the new real-time data sequence, according to the window size, all the data after the any mutation data, and some of the data before the any mutation data, a window centered on the any mutation data is constructed.

8. The power distribution terminal with the power supply failure abnormal monitoring and early warning function according to claim 1, characterized in that, In the data processing module, according to the number of mutation data within the window of each mutation data, the time distribution density of each mutation data is obtained, including: For any mutation data, obtain the number of other mutation data within the window of the any mutation data except the any mutation data itself. If the number is 0, then set the time distribution density of the any mutation data to 0; If the number is not 0, then calculate 1 minus the reciprocal of the number to obtain a first density; calculate the time interval between the sampling times corresponding to every two adjacent mutation data within the window of the any mutation data, and take the reciprocal of the average value of all the time intervals as a second density; calculate the reciprocal of the variance of all the time intervals to obtain a third density; calculate the average value among the first density, the second density, and the third density to obtain the time distribution density of the any mutation data.

9. The power distribution terminal with the function of abnormal monitoring and early warning for power supply faults according to claim 1, characterized in that In the anomaly warning module, according to the mutation distribution density and the time distribution density of each mutation data, and the K value range, the adaptive K value of each mutation data is obtained, including: For any mutation data, subtract the mutation distribution density of the any mutation data from 1 to obtain a first eigenvalue of the any mutation data, subtract the time distribution density of the any mutation data from 1 to obtain a second eigenvalue of the any mutation data, and perform a weighted sum on the first eigenvalue and the second eigenvalue to obtain a weighted sum result; Calculate the difference between the maximum value and the minimum value in the K value range to obtain a K value range difference, round down the product of the K value range difference and the weighted sum result to obtain a K value adjustment coefficient, and calculate the sum of the minimum value in the K value range and the K value adjustment coefficient to obtain the adaptive K value of the any mutation data.

10. The power distribution terminal with a power supply failure abnormal monitoring and early warning function according to claim 1, characterized in that, The abnormal power supply fault monitoring and warning of the distribution terminal according to all the LOF values includes: Among all the LOF values, if there is at least one LOF value greater than a preset abnormal threshold, then an abnormal power supply fault warning is given to the distribution terminal; If all LOF values are less than or equal to the preset anomaly threshold, then based on the real-time data sequence, using the preset prediction algorithm, predict the power supply data of the distribution terminal at a future moment to obtain prediction data, obtain the adaptive K value of the prediction data, and based on the adaptive K value of the prediction data, obtain the LOF value of the prediction data. If the LOF value of the prediction data is greater than the preset anomaly threshold, then issue an anomaly warning for the abnormal power supply fault of the distribution terminal.

Images