Power grid business data management service terminal based on artificial intelligence

Abstract

The present invention relates to the field of data management technology, and in particular to a power grid business data management service terminal based on artificial intelligence, the power grid business data management service terminal comprising a first acquisition module for acquiring a time series sequence to be compressed; a second acquisition module for acquiring a reference target subsequence and a target subsequence set to be analyzed; a third acquisition module for acquiring a target data sequence to be compressed; and a compression management module for compressing the target data sequence to be compressed to obtain compressed data, and storing the compressed data in the power grid business management service terminal. The present invention can not only improve the compression effect, but also save storage space, reduce storage costs, and improve the management efficiency of power grid business data.

Description

Technical Field

[0001] The present invention relates to the technical field of data management, and in particular to an artificial intelligence-based power grid business data management service terminal. Background Art

[0002] The power grid refers to a huge network system composed of power generation, transmission, distribution and power consumption facilities. It is responsible for transmitting electricity from power stations to end users. Therefore, there will be a variety of power grid business data during the operation of the power grid, such as power grid load data, equipment operation status data, user power consumption data, etc.; and because these power grid business data can help power grid operators optimize power dispatching, predict load demand and prevent fault risks, thereby ensuring the stability, safety and efficient operation of the power grid, so at present, in order to ensure the stability, safety and efficient operation of the power grid, the collected power grid business data will be transmitted to the power grid business management service terminal for storage, and then the stored power grid business data will be analyzed later, and the purpose of optimizing power dispatching, predicting load demand and preventing fault risks will be achieved based on the analysis results.

[0003] As the operation time of the power grid increases, the data that needs to be stored will gradually increase. In order to save storage space and reduce storage costs, the collected data is usually compressed before storage. However, the prior art generally uses lossless methods such as Huffman coding to directly compress the power grid business data that needs to be compressed. This compression method will result in poor compression effect, and thus cannot effectively achieve the purpose of saving storage space and reducing storage costs, that is, it cannot effectively manage data. Summary of the invention

[0004] In order to solve the above problems, the present invention provides a power grid business data management service terminal based on artificial intelligence, and the technical solution adopted is as follows:

[0005] An embodiment of the present invention provides an artificial intelligence-based power grid business data management service terminal, the power grid business data management service terminal comprising:

[0006] A first acquisition module is used to acquire a time series sequence to be compressed, where the data in the time series sequence to be compressed is power grid service data;

[0007] A second acquisition module is used to divide the time series to be compressed to obtain target subsequences, obtain an overall change characterization value of the target subsequence according to the difference between adjacent power grid service data in the target subsequences, and obtain a reference target subsequence and a set of target subsequences to be analyzed according to the overall change characterization value;

[0008] A third acquisition module is used to perform a difference on the reference target subsequence to obtain a target to-be-compressed subsequence of the reference target subsequence, obtain a target to-be-compressed subsequence of the target subsequence to be analyzed based on the target to-be-compressed subsequence of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed, and obtain a target to-be-compressed data sequence based on the target to-be-compressed subsequences of all the target subsequences;

[0009] The compression management module is used to compress the target data sequence to be compressed to obtain compressed data, and store the compressed data in the power grid business management service terminal.

[0010] Beneficial effects: The present invention includes a first acquisition module for acquiring a time series sequence to be compressed; a second acquisition module for dividing the time series sequence to be compressed to obtain a target subsequence, obtaining an overall change characterization value of the target subsequence according to the difference between adjacent power grid business data in the target subsequence, and obtaining a reference target subsequence and a set of target subsequences to be analyzed according to the overall change characterization value; a third acquisition module for performing a difference on the reference target subsequence to obtain a target subsequence to be compressed of the reference target subsequence, obtaining a target subsequence to be compressed of the target subsequence to be analyzed according to the target subsequence to be compressed of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed, and obtaining a target data sequence to be compressed according to the target subsequence to be compressed of all target subsequences; a compression management module for compressing the target data sequence to be compressed to obtain compressed data, and storing the compressed data in a power grid business management service terminal. In addition, the present invention can improve the compression effect in the subsequent compression process, save storage space, reduce storage costs, and improve the management efficiency of power grid business data, etc. according to the target data sequence to be compressed obtained by pre-processing each target subsequence obtained after the division. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0012] Figure 1 This is a structural block diagram of an artificial intelligence-based power grid business data management service terminal of the present invention. DETAILED DESCRIPTION

[0013] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the embodiments of the present invention.

[0014] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0015] This embodiment provides a power grid business data management service terminal based on artificial intelligence, which is described in detail as follows:

[0016] like Figure 1 As shown, this embodiment provides an artificial intelligence-based power grid business data management service terminal, including:

[0017] The first acquisition module 01 is used to acquire a time series sequence to be compressed, where the data in the time series sequence to be compressed is power grid service data.

[0018] This embodiment mainly achieves the purpose of saving storage space and reducing storage costs by improving the compression effect when compressing the time series to be compressed, that is, achieving the purpose of effectively managing power grid business data; and the difference between this embodiment and the traditional compression process is: the traditional compression process generally uses a compression algorithm to directly compress the collected data, while this embodiment mainly performs pre-processing on the collected data, and then uses a compression algorithm to encode and compress the pre-processed data. The purpose of pre-processing is to make the data repetition rate in the subsequent sequence used for compression higher, so as to achieve better effect in subsequent compression.

[0019] Therefore, this embodiment first needs to obtain the time series sequence to be compressed, and the specific acquisition process is:

[0020] First, the time series sequence composed of the same type of power grid business data collected during the power grid operation time period is recorded as the power grid business data time series sequence corresponding to the power grid operation time period, and is recorded as the time series sequence to be compressed; and the types of power grid business data in this embodiment include but are not limited to power grid load data, equipment operation status data, and user electricity consumption data; and for the sake of ease of understanding and analysis, this embodiment only describes the pre-processing process of any type of power grid business data, that is, the time series sequence to be compressed that appears later in this embodiment is a sequence formed by the same type of power grid business data. For example, this embodiment can set the power grid business data that appears later to refer to power grid load data, that is, the data in the time series sequence to be compressed that appears later all refer to power grid load data.

[0021] In addition, in specific applications, the implementer needs to set the power grid operation time period and the collection time interval between adjacent data in the compressed time series according to actual conditions. For example, in this embodiment, the power grid operation time period can be set to the time interval between adjacent transmission moments. If the current moment is the transmission moment, and the transmission time interval in this embodiment is 5 minutes, then the power grid operation time period in this embodiment refers to the 5 minutes before the current moment, and this embodiment can also set the collection time interval between adjacent data to 1 second or 0.5 seconds.

[0022] Therefore, this embodiment can obtain the time series to be compressed through the above process; in addition, this embodiment realizes data collection through the acquisition module, and the acquisition module will also transmit the collected power grid business data to the power grid business management service terminal through the wireless transmission module arranged inside it, and the subsequent pre-processing process and compression process are all carried out in the management service terminal.

[0023] The second acquisition module 02 is used to divide the time series sequence to be compressed to obtain target subsequences, obtain the overall change characterization value of the target subsequence according to the difference between adjacent power grid business data in the target subsequences, and obtain the reference target subsequence and the set of target subsequences to be analyzed according to the overall change characterization value.

[0024] Since the change of power grid business data is related to the amount of power-consuming equipment in the power grid, and the usage of power grid power-consuming equipment in different time periods is different, the power grid business data has different or different characteristics in the change mode and amplitude in different time periods, and this change characteristic will make the repetition rate of the data in the above-mentioned acquired time series to be compressed low, so that when the Huffman algorithm is used to directly compress the acquired time series to be compressed, it is necessary to build a larger Huffman tree, which will also lead to poor compression effect and fail to achieve the purpose of effectively saving storage space and reducing storage costs. In order to achieve better compression effect, this embodiment will then pre-process the above-mentioned acquired time series to be compressed, and the first step of the pre-processing is to divide the time series to be compressed according to the change of the power grid business data, that is, this embodiment first divides the time series to be compressed to obtain the target subsequence, and the specific process is:

[0025] First, a preset sliding window and a preset moving step are obtained, and each window subsequence on the time series sequence to be compressed is obtained according to the preset sliding window, the preset moving step and the time series sequence to be compressed. In order to realize the segmentation of the sequence, this embodiment requires that the preset moving step is smaller than the length of the preset sliding window, and the parity of the preset moving step and the length of the preset sliding window are different. Then, in all window subsequences on the time series sequence to be compressed, the sequence group formed by two adjacent window subsequences is recorded as a window subsequence group, and the difference characterization value corresponding to each window subsequence group is obtained. Then, according to the sequence in the window subsequence group, the sequence is sorted in the time series sequence to be compressed. The difference characterization values ​​corresponding to all window subsequence groups are sorted according to the order on the columns, and the sequence after sorting is recorded as the difference characterization value sequence. For example, if the sequence in a window subsequence group is composed of the 1st window subsequence and the 2nd window subsequence on the time series sequence to be compressed, then the first data in the difference characterization value sequence is the difference characterization value corresponding to the window subsequence group. That is, if the sequence in a window subsequence group is composed of the pth window subsequence and the p+1th window subsequence on the time series sequence to be compressed, then the pth data in the difference characterization value sequence is the difference characterization value corresponding to the corresponding window subsequence group.

[0026] When a relatively obvious peak appears in the difference characterization value sequence, it indicates that the middle position of the overlapping area in the window subsequence group corresponding to the peak is the segmentation position. Therefore, this embodiment uses the AMPD algorithm to perform peak detection on the difference characterization value sequence, and marks the difference characterization value detected as the peak as the peak. Then, the window subsequence group corresponding to the difference characterization value marked as the peak in the difference characterization value sequence is recorded as the target window subsequence group. The process of performing peak detection using the AMPD algorithm is a well-known technology, so it will not be described in detail in this embodiment.

[0027] Then, the overlapping subsequence corresponding to each target window subsequence group is obtained, and the data located in the middle position of each overlapping subsequence is marked as b. The overlapping subsequence corresponding to the target window subsequence group is composed of the data of the overlapping area of ​​the two window subsequences in the corresponding target window subsequence group. Since the preset moving step size is less than the length of the preset sliding window, and the window subsequence group is composed of adjacent window subsequences, there must be an overlapping part between the two window subsequences in the target window subsequence group, that is, there must be an overlapping area. For example, for any target window Subsequence group, if the target window subsequence group consists of sequence Z1 and sequence Z2, and sequence Z1 consists of the zth data to the z+4th data on the time series sequence to be compressed, and includes the zth data and the z+4th data, and sequence Z2 consists of the z+2th data and the z+6th data on the time series sequence to be compressed, and includes the z+2th data and the z+6th data, then the overlapping subsequence corresponding to the target window subsequence group is composed of the z+2th data to the z+4th data on the time series sequence to be compressed, and includes the z+2th data and the z+4th data.

[0028] Then, the position corresponding to the power grid business data marked as b on the time series sequence to be compressed is used as the split position on the time series sequence to be compressed, and the time series sequence to be compressed is divided according to the split position on the time series sequence to be compressed, and each subsequence obtained by the division is recorded as a target subsequence; for example, if the z0th data, z0+5th data, and z0+12th data on the time series sequence to be compressed are marked as b, then the positions of the z0th data, z0+5th data, and z0+12th data are all split positions, so that the time series sequence to be compressed can be divided into four target subsequences, the first target subsequence is composed of the 1st data to the z0th data on the time series sequence to be compressed, and the second target subsequence is composed of the z0th data to the z0th data on the time series sequence to be compressed. The target subsequence consists of z0+5 data, the third target subsequence consists of the z0+5th data to the z0+12th data on the time series sequence to be compressed, the fourth target subsequence consists of the z0+12th data to the last data on the time series sequence to be compressed, and the first target subsequence includes the first data and the z0th data, the second target subsequence does not include the z0th data but includes the z0+5th data, the third target subsequence does not include the z0+5th data but includes the z0+12th data, the fourth target subsequence does not include the z0+12th data but includes the last data; in addition, the z0th data is not the first data on the time series sequence to be compressed, and the z0+12th data is not the last data on the time series sequence to be compressed.

[0029] In this embodiment, according to the preset sliding window, the preset moving step and the time series sequence to be compressed, the specific process of obtaining each window subsequence on the time series sequence to be compressed is: slide the preset sliding window on the time series sequence to be compressed with the preset moving step until the last data in the time series sequence to be compressed appears in the sliding window for the first time, then stop sliding immediately, and obtain all sliding windows on the time series sequence to be compressed, and the first data in the first sliding window on the time series sequence to be compressed is the first data in the time series sequence to be compressed; then the sequence formed by the data in each sliding window on the time series sequence to be compressed is recorded as a window subsequence, that is, the kth time series sequence to be compressed on the time series sequence to be compressed refers to the sequence formed by the data in the kth sliding window on the time series sequence to be compressed.

[0030] In addition, in a specific application, the implementer needs to set the size of the preset sliding window and the size of the preset moving step according to the actual situation. For example, in this embodiment, the size of the preset sliding window can be , that is, the length of the preset sliding window is 5. Since the preset moving step is required to be smaller than the length of the preset sliding window and the parity of the preset moving step and the length of the preset sliding window are different, the preset moving step can be set to 2 in this embodiment. Then, there will be 3 overlapping data between the window subsequence groups obtained above.

[0031] In this embodiment, the process of obtaining the difference characterization value corresponding to the window subsequence group is as follows: for any window subsequence group A: the two window subsequences in the window subsequence group A are respectively recorded as sequence A1 and sequence A2; then the absolute value of the Pearson correlation coefficient between sequence A1 and sequence A2 is obtained, and recorded as the first correlation coefficient; then the absolute value of the difference between the first correlation coefficient and the preset first constant is recorded as the first characterization value; then the absolute value of the difference between the extreme difference value of sequence A1 and the extreme difference value of sequence A2 is obtained, and recorded as the second characterization value; finally, the product of the first characterization value and the second characterization value is recorded as the difference characterization value corresponding to the window subsequence group A; and in this embodiment, the calculation expression of the difference characterization value corresponding to the window subsequence group A is:

[0032]

[0033] in, is the difference characterization value corresponding to the window subsequence group A, PER(A1,A2) is the Pearson correlation coefficient between sequence A1 and sequence A2, c1 is the preset first constant, is the first characterization value, R(A1) is the range value of sequence A1, and R(A2) is the range value of sequence A2. is the second characterization value, and the extreme value of the sequence refers to the difference between the maximum value and the minimum value in the sequence; and in this embodiment, the value of c1 is set to 1.

[0034] In addition, when The smaller and The larger the value, the greater the difference between sequence A1 and sequence A2 in window subsequence group A. The bigger and The smaller the value, the smaller the difference between sequence A1 and sequence A2 in window subsequence group A.

[0035] Therefore, this embodiment can obtain all target subsequences through the above process, and after obtaining the target subsequence, the target subsequence will be analyzed to obtain the reference target subsequence and the target subsequence set to be analyzed, and the purpose of obtaining the reference target subsequence and the target subsequence set to be analyzed is to obtain the target to-be-compressed subsequence of each target subsequence, and the target to-be-compressed subsequence is the key to constituting the target to-be-compressed data sequence, so this embodiment will obtain the reference target subsequence and the target subsequence set to be analyzed, and before obtaining the reference target subsequence and the target subsequence set to be analyzed, it is necessary to first obtain the reference target subsequence according to the adjacent power grid business in the target subsequence. The difference between the data is used to obtain the overall change characterization value of the target subsequence. Therefore, the specific method for obtaining the overall change characterization value of the target subsequence is as follows: for any target subsequence, the first difference sequence of the target subsequence is obtained, and the mean of the first difference sequence is used as the overall change characterization value of the target subsequence. The larger the mean of the first difference sequence is, the greater the overall change rate of the data in the target subsequence is, that is, the less similar the data in the target subsequence is or the lower the data repetition rate in the target subsequence is, and the cth first difference in the first difference sequence of the target subsequence is the absolute value of the difference between the cth power grid business data and the c+1th power grid business data in the target subsequence.

[0036] Therefore, the overall change characterization value of each target subsequence can be obtained through the above process. After obtaining the overall change characterization value of the target subsequence, this embodiment then obtains the reference target subsequence and the target subsequence set to be analyzed according to the overall change characterization value. The specific process is as follows:

[0037] Among all the target subsequences, the target subsequence with the smallest overall change representation value is selected as the reference target subsequence, and all target subsequences except the reference target subsequence are recorded as target subsequences to be analyzed, and then the set constructed from all the target subsequences to be analyzed is recorded as the target subsequence set to be analyzed.

[0038] The third acquisition module 03 is used to differentiate the reference target subsequence to obtain the target subsequence to be compressed of the reference target subsequence, obtain the target subsequence to be compressed of the target subsequence to be analyzed based on the target subsequence to be compressed of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed, and obtain the target data sequence to be compressed based on the target subsequences to be compressed of all the target subsequences.

[0039] After obtaining the reference target subsequence and the set of target subsequences to be analyzed, the present embodiment will first perform differential processing on the reference target subsequence to obtain the target subsequence to be compressed of the reference target subsequence, and then determine the target subsequence to be compressed of the target subsequence to be analyzed based on the target subsequence to be compressed of the reference target subsequence. The differential processing can make the repetition rate in the sequence involved in compression larger, so the specific process of obtaining the target subsequence to be compressed of the reference target subsequence is as follows:

[0040] First, calculate the effect representation value corresponding to the reference target subsequence and the effect representation value corresponding to the first-order difference sequence of the reference target subsequence; then determine whether the effect representation value corresponding to the first-order difference sequence of the reference target subsequence is greater than the effect representation value corresponding to the reference target subsequence. If so, then obtain the effect representation value corresponding to the second-order difference sequence of the reference target subsequence, and continue to determine whether the effect representation value corresponding to the second-order difference sequence of the reference target subsequence is greater than the effect representation value corresponding to the first-order difference sequence of the reference target subsequence. If so, obtain the effect representation value corresponding to the third-order difference sequence of the reference target subsequence, and continue to determine the third-order difference sequence of the reference target subsequence. Whether the effect characterization value corresponding to the differential sequence is greater than the effect characterization value corresponding to the second-order differential sequence of the reference target subsequence; if it is determined that the effect characterization value corresponding to the third-order differential sequence of the reference target subsequence is not greater than the effect characterization value corresponding to the second-order differential sequence of the reference target subsequence, then it is determined that the effect characterization value corresponding to the second-order differential sequence of the reference target subsequence is better, that is, the repetition rate of data in the second-order differential sequence of the reference target subsequence is large, then the second-order differential sequence of the reference target subsequence can be used as the target subsequence to be compressed of the reference target subsequence; in this embodiment, the process of obtaining the differential sequence of the sequence is a well-known technology, so this embodiment will not be described in detail.

[0041] In addition, in order to facilitate subsequent data restoration or decompression, in the present embodiment, in the process of obtaining the target to-be-compressed subsequence of the reference target subsequence, it is also necessary to record the data used for subsequent data restoration, that is, it is also necessary to obtain the feature vector group corresponding to the reference target subsequence. The feature vector group exists for the purpose of being used for subsequent data restoration. The feature vector group corresponding to the reference target subsequence is composed of the initial data sequence corresponding to the reference target subsequence and the length of the target to-be-compressed subsequence of the reference target subsequence; and the data and length in the initial data sequence corresponding to the reference target subsequence are related to the order difference sequence to which the target to-be-compressed subsequence of the reference target subsequence belongs. If the target to-be-compressed subsequence of the reference target subsequence is the second-order difference sequence of the reference target subsequence, then the number of data in the initial data sequence corresponding to the reference target subsequence is 2, and the first data in the initial data sequence corresponding to the reference target subsequence is the first data in the reference target subsequence, and the second data in the initial data sequence corresponding to the reference target subsequence is the first data in the first-order difference sequence of the reference target subsequence.

[0042] In this embodiment, the specific process of obtaining the effect characterization value is as follows: for the reference target subsequence, first obtain the coefficient of variation and the range value of the reference target subsequence, then take the product of the coefficient of variation of the reference target subsequence and the range value of the reference target subsequence as the first indicator value, and add the first indicator value to the preset second constant to obtain the inverse of the result obtained as the effect characterization value corresponding to the reference target subsequence. Moreover, when the coefficient of variation of the reference target subsequence is smaller and the range value of the reference target subsequence is smaller, it indicates that the effect characterization value corresponding to the reference target subsequence is larger, and the larger the effect characterization value corresponding to the reference target subsequence is, the greater the repetition rate or the concentration of the data in the reference target subsequence is, and the greater the possibility that the reference target subsequence is the target subsequence to be compressed of the reference target subsequence is; the coefficient of variation of the sequence can characterize the degree of discreteness, and the larger the value of the coefficient of variation is, the more discrete it is, and the process of obtaining the coefficient of variation of the sequence is a well-known technology, so it will not be described in detail.

[0043] In addition, the effect characterization value corresponding to the above-mentioned differential sequence is obtained in the same method as the effect characterization value corresponding to the reference target subsequence, so this embodiment will not be described in detail; and in specific applications, the implementer also needs to set the value of the preset second constant according to actual conditions. For example, in this embodiment, the value of the preset second constant can be set to 0.01. The purpose of adding the preset second constant in this embodiment is to prevent the calculation result or the denominator from being 0.

[0044] After obtaining the target to-be-compressed subsequence of the reference target subsequence, the present embodiment will obtain the target to-be-compressed subsequence of the target subsequence to be analyzed according to the target to-be-compressed subsequence of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed. For ease of understanding, the present embodiment will take the specific process of obtaining the target to-be-compressed subsequence of any target subsequence S to be analyzed as an example for description, that is, the specific process of obtaining the target to-be-compressed subsequence of the target subsequence S to be analyzed is:

[0045] First, obtain the characteristic index value corresponding to the target subsequence S to be analyzed and the characteristic index value corresponding to the first-order difference sequence of the target subsequence S to be analyzed, and record them as the initial characteristic index value and the first-order characteristic index value respectively; then determine whether the first-order characteristic index value is greater than the initial characteristic index value, if so, obtain the second-order characteristic index value, the second-order characteristic index value is the characteristic index value corresponding to the second-order difference sequence of the target subsequence S to be analyzed, and continue to determine whether the second-order characteristic index value is greater than the first-order characteristic index value, if so, obtain the third-order characteristic index value, the third-order characteristic index value is the characteristic index value corresponding to the third-order difference sequence of the target subsequence S to be analyzed , and continue to judge whether the third-order characteristic index value is greater than the second-order characteristic index value. If it is judged that the third-order characteristic index value is not greater than the second-order characteristic index value, it indicates that among all the differential sequences of the target subsequence S to be analyzed, the second-order differential sequence of the target subsequence S to be analyzed is most similar to the target subsequence to be compressed of the reference target subsequence. Since the data repetition rate in the target subsequence to be compressed of the reference target subsequence is relatively high, the data repetition rate in the second-order differential sequence of the target subsequence S to be analyzed is also relatively high. Therefore, in this embodiment, the second-order differential sequence of the target subsequence S to be analyzed is directly used as the target subsequence to be compressed of the target subsequence S to be analyzed.

[0046] Similarly, in order to facilitate subsequent data restoration or decompression, in the process of obtaining the target to-be-compressed subsequence of the target subsequence S to be analyzed, it is also necessary to record the data used for subsequent data restoration, that is, it is also necessary to obtain the feature vector group corresponding to the target subsequence S to be analyzed, and the feature vector group corresponding to the target subsequence S to be analyzed is also composed of the initial data sequence corresponding to the target subsequence S to be analyzed and the length of the target to-be-compressed subsequence of the target subsequence S to be analyzed; if the target to-be-compressed subsequence of the target subsequence S to be analyzed is the second-order difference sequence of the target subsequence S to be analyzed, then the number of data in the initial data sequence corresponding to the target subsequence S to be analyzed is 2, and the first data in the initial data sequence corresponding to the target subsequence S to be analyzed is the first data in the target subsequence S to be analyzed, and the second data in the initial data sequence corresponding to the target subsequence S to be analyzed is the first data in the first-order difference sequence of the target subsequence S to be analyzed.

[0047] In this embodiment, the method for obtaining the characteristic index value is as follows: for the target subsequence S to be analyzed: first, the total number of data in the target subsequence S to be analyzed is obtained, and recorded as the first quantity characterization value; then, the total number of data marked as repeated in the target subsequence S to be analyzed is obtained by counting, and recorded as the second quantity characterization value; then, the ratio of the second quantity characterization value to the first quantity characterization value is obtained, and recorded as the first characteristic value; then, the range value of the target subsequence S to be analyzed is obtained and recorded as the first range value, and the range value of the target subsequence to be compressed of the reference target subsequence is obtained and recorded as the second characteristic value. Two extreme difference values ​​are obtained, and then the absolute value of the difference between the first extreme difference value and the second extreme difference value is recorded as the characteristic difference value, and then the characteristic difference value and the preset second constant are added, and the inverse of the result obtained by the addition is recorded as the second eigenvalue; finally, the product of the first eigenvalue and the second eigenvalue is obtained, and recorded as the characteristic index value corresponding to the target subsequence S to be analyzed; and in this embodiment, the method for obtaining the characteristic index value corresponding to the differential sequence is the same as the method for obtaining the characteristic index value corresponding to the target subsequence S to be analyzed, so this embodiment no longer describes the process of obtaining the characteristic index value corresponding to the differential sequence.

[0048] In this embodiment, the method for marking the data in the target subsequence S to be analyzed is: for the f-th data in the target subsequence S to be analyzed, if there is data identical to the f-th data in the target to-be-compressed subsequence of the reference target subsequence, then the f-th data is marked as repeated.

[0049] In this embodiment, the specific expression for obtaining the characteristic index value corresponding to the target subsequence S to be analyzed is:

[0050]

[0051] in, is the characteristic index value corresponding to the target subsequence S to be analyzed, N1 is the first quantity representation value, N2 is the second quantity representation value, c2 is the preset second constant, R(S) is the range value of the target subsequence S to be analyzed, that is, the first range value, and R0 is the range value of the target subsequence to be compressed of the reference target subsequence, that is, the second range value.

[0052] And when the value of N2 is larger and The smaller the value of is, the larger the characteristic index value corresponding to the target subsequence S to be analyzed is, and the larger the characteristic index value corresponding to the target subsequence S to be analyzed is, the more similar the target subsequence S to be analyzed is to the target subsequence to be compressed of the reference target subsequence, that is, the higher the repetition rate between the target subsequence S to be analyzed and the target subsequence to be compressed of the reference target subsequence is, the greater the possibility that the target subsequence S to be analyzed is the target subsequence to be compressed of the target subsequence S to be analyzed is, and vice versa. The larger the value of , the smaller the possibility that the target subsequence S to be analyzed is the target subsequence to be compressed of the target subsequence S to be analyzed.

[0053] In addition, the method for obtaining the target to-be-compressed subsequences and the corresponding feature vector groups of other target subsequences to be analyzed is the same as the method for obtaining the target to-be-compressed subsequences and the corresponding feature vector groups of the target subsequence S to be analyzed, so this embodiment will not be described in detail.

[0054] Therefore, the present embodiment can obtain the target subsequences to be compressed and the corresponding feature vector groups of all target subsequences; and the reason why the present embodiment determines the target subsequences to be compressed through the above process is that the differential data has a linear or near-linear trend, and after the data is differentiated with a suitable order, the amplitude of the data change is usually reduced, and the data distribution is relatively concentrated, but when the differential data is excessively differentiated, it may cause its amplitude of change to increase, and the volatility of the data points will also increase, so the present embodiment can obtain the target subsequences to be compressed with a higher repetition rate through the above process.

[0055] After obtaining the target to-be-compressed subsequences of all target subsequences, this embodiment will then concatenate the target to-be-compressed subsequences to obtain the target to-be-compressed data sequence, that is, the specific process of obtaining the target to-be-compressed data sequence is as follows:

[0056] First, the target subsequences to be compressed of all target subsequences are sorted according to their positions in the time series to be compressed, and the sequence obtained after sorting is recorded as the first sequence, and then each target subsequence to be compressed in the first sequence is spliced ​​in turn, and the sequence after splicing is recorded as the target data sequence to be compressed. Since all target subsequences to be compressed are relatively similar or close, the compression effect after subsequent compression will be better. For example, if the number of target subsequences is 3, and they are respectively recorded as target subsequence Y1, target subsequence Y2 and target subsequence Y3, and if in the time series to be compressed, the target subsequence Y2 is between the target subsequence Y1 and the target subsequence Y3, then when splicing the target subsequences to be compressed, the target subsequence Y2 is first spliced ​​to the tail end of the target subsequence Y1, and then the target subsequence Y3 is spliced ​​to the tail end of the target subsequence Y2.

[0057] Similarly, according to the order of the target subsequences in the time series to be compressed, the feature vector groups corresponding to all target subsequences are sorted, and the sequence obtained after sorting is recorded as a decoding vector sequence.

[0058] Therefore, this embodiment can obtain the target to-be-compressed data sequence and the decoding vector sequence through the above process.

[0059] The compression management module 04 is used to compress the target data sequence to be compressed, obtain compressed data, and store the compressed data in the power grid business management service terminal.

[0060] Next, this embodiment uses the Huffman algorithm to compress the target data sequence to be compressed, and stores the compressed data in the storage space of the power grid business management service terminal, and then stores the decoding vector sequence in the storage space of the power grid business management service terminal. Since the compression process is a well-known technology, it will not be described in detail. In addition, the decoding vector sequence is mainly used in the subsequent decoding and restoration process; and the specific decoding and restoration process is a well-known technology, so this embodiment will not be described in detail.

[0061] Thus, this embodiment improves the compression effect, thereby achieving the purpose of saving storage space and reducing storage costs, that is, achieving the purpose of effectively managing power grid service data.

[0062] In summary, the present embodiment includes a first acquisition module for acquiring a time series sequence to be compressed; a second acquisition module for dividing the time series sequence to be compressed to obtain a target subsequence, obtaining an overall change characterization value of the target subsequence according to the difference between adjacent power grid business data in the target subsequence, and obtaining a reference target subsequence and a set of target subsequences to be analyzed according to the overall change characterization value; a third acquisition module for performing a differential operation on the reference target subsequence to obtain a target subsequence to be compressed of the reference target subsequence, obtaining a target subsequence to be compressed of the target subsequence to be analyzed according to the target subsequence to be compressed of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed, and obtaining a target data sequence to be compressed according to the target subsequences to be compressed of all target subsequences; a compression management module for compressing the target data sequence to be compressed to obtain compressed data, and storing the compressed data in the power grid business management service terminal. In addition, the present embodiment can improve the compression effect in the subsequent compression process by pre-processing each target subsequence obtained after the division, and can also save storage space, reduce storage costs, and improve the management efficiency of power grid business data.

[0063] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. An artificial intelligence-based power grid business data management service terminal, characterized in that: The power grid business data management service terminal includes: A first acquisition module is used to acquire a time series sequence to be compressed, where the data in the time series sequence to be compressed is power grid service data; A second acquisition module is used to divide the time series to be compressed to obtain target subsequences, obtain an overall change characterization value of the target subsequence according to the difference between adjacent power grid service data in the target subsequences, and obtain a reference target subsequence and a set of target subsequences to be analyzed according to the overall change characterization value; A third acquisition module is used to perform a difference on the reference target subsequence to obtain a target to-be-compressed subsequence of the reference target subsequence, obtain a target to-be-compressed subsequence of the target subsequence to be analyzed based on the target to-be-compressed subsequence of the reference target subsequence and the target subsequence to be analyzed in the set of target subsequences to be analyzed, and obtain a target to-be-compressed data sequence based on the target to-be-compressed subsequences of all the target subsequences; The compression management module is used to compress the target data sequence to be compressed to obtain compressed data, and store the compressed data in the power grid business management service terminal.

2. The power grid business data management service terminal based on artificial intelligence according to claim 1, characterized in that: The method for acquiring the target subsequence comprises: Obtaining a preset sliding window and a preset moving step, and obtaining each window subsequence on the time series sequence to be compressed according to the preset sliding window, the preset moving step and the time series sequence to be compressed, wherein the preset moving step is smaller than the length of the preset sliding window; Among all window subsequences on the time series to be compressed, sequence groups formed by two adjacent window subsequences are recorded as window subsequence groups, and difference representation values ​​corresponding to the window subsequence groups are obtained; According to the order of the sequences in the window subsequence group on the time series to be compressed, the difference characterization values ​​corresponding to all window subsequence groups are sorted, and the sequence after sorting is recorded as a difference characterization value sequence; in the difference characterization value sequence, the window subsequence group corresponding to the difference characterization value belonging to the peak value is recorded as a target window subsequence group; Obtaining the overlapping subsequence corresponding to each of the target window subsequence groups, and marking the data located at the middle position of the overlapping subsequence as b, wherein the overlapping subsequence corresponding to the target window subsequence group is composed of the data of the overlapping area of ​​two window subsequences in the target window subsequence group; The position corresponding to the power grid business data marked as b on the time series sequence to be compressed is used as the segmentation position on the time series sequence to be compressed, and the time series sequence to be compressed is divided according to the segmentation position on the time series sequence to be compressed, and each subsequence obtained by the division is recorded as a target subsequence; The method for obtaining the difference characterization value corresponding to the window subsequence group includes: for any window subsequence group, recording the two window subsequences in the window subsequence group as sequence A1 and sequence A2 respectively, recording the absolute value of the Pearson correlation coefficient between the sequence A1 and the sequence A2 as a first correlation coefficient, recording the absolute value of the difference between the first correlation coefficient and a preset first constant as a first characterization value, recording the absolute value of the difference between the extreme value of the sequence A1 and the extreme value of the sequence A2 as a second characterization value, and recording the product of the first characterization value and the second characterization value as the difference characterization value corresponding to the window subsequence group.

3. The power grid business data management service terminal based on artificial intelligence according to claim 2, characterized in that: The method for obtaining the window subsequence includes: The preset sliding window is slid on the time series sequence to be compressed with the preset moving step length until the last data in the time series sequence to be compressed appears in the sliding window for the first time, and the sliding is stopped immediately to obtain each sliding window located on the time series sequence to be compressed; the sequences formed by the data in each sliding window located on the time series sequence to be compressed are all recorded as window subsequences.

4. The artificial intelligence-based power grid business data management service terminal according to claim 1, characterized in that: The method for obtaining the overall change representation value of the target subsequence includes: For any target subsequence, obtain the first difference sequence of the target subsequence, and take the mean of the first difference sequence as the overall change representation value of the target subsequence, the cth first difference value in the first difference sequence of the target subsequence is the absolute value of the difference between the cth power grid business data and the c+1th power grid business data in the target subsequence.

5. The power grid business data management service terminal based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the reference target subsequence and the target subsequence set to be analyzed includes: Among all the target subsequences, the target subsequence with the smallest overall change representation value is selected as the reference target subsequence, and all target subsequences except the reference target subsequence are recorded as target subsequences to be analyzed, and the set constructed from all the target subsequences to be analyzed is recorded as the target subsequence set to be analyzed.

6. The power grid business data management service terminal based on artificial intelligence according to claim 1, characterized in that: The method for obtaining the target to-be-compressed subsequence of the reference target subsequence includes: Calculate and obtain the effect characterization value corresponding to the reference target subsequence and the effect characterization value corresponding to the first-order difference sequence of the reference target subsequence; determine whether the effect characterization value corresponding to the first-order difference sequence is greater than the effect characterization value corresponding to the reference target subsequence; if so, obtain the effect characterization value corresponding to the second-order difference sequence of the reference target subsequence, and continue to determine whether the effect characterization value corresponding to the second-order difference sequence is greater than the effect characterization value corresponding to the first-order difference sequence; if so, obtain the effect characterization value corresponding to the third-order difference sequence of the reference target subsequence, and continue to determine whether the effect characterization value corresponding to the third-order difference sequence is greater than the effect characterization value corresponding to the second-order difference sequence; if it is determined that the effect characterization value corresponding to the third-order difference sequence is not greater than the effect characterization value corresponding to the second-order difference sequence, use the second-order difference sequence as the target subsequence to be compressed of the reference target subsequence.

7. The power grid business data management service terminal based on artificial intelligence according to claim 6, characterized in that: The method for obtaining the effect characterization value includes: For the reference target subsequence, the coefficient of variation and the range value of the reference target subsequence are obtained, and the result of multiplying the coefficient of variation of the reference target subsequence by the range value of the reference target subsequence is used as the first index value. The reciprocal of the result obtained by adding the first index value to a preset second constant is recorded as the effect characterization value corresponding to the reference target subsequence. The method for obtaining the effect characterization value corresponding to the differential sequence is the same as the method for obtaining the effect characterization value corresponding to the reference target subsequence.

8. The power grid business data management service terminal based on artificial intelligence according to claim 1, characterized in that: The method for acquiring the target to-be-compressed subsequence of the target subsequence to be analyzed comprises: For any target subsequence S to be analyzed in the set of target subsequences to be analyzed: Obtaining a characteristic index value corresponding to the target subsequence S to be analyzed and a characteristic index value corresponding to the first-order difference sequence of the target subsequence S to be analyzed, and recording them as an initial characteristic index value and a first-order characteristic index value respectively; Determine whether the first-order characteristic index value is greater than the initial characteristic index value. If so, obtain a second-order characteristic index value, which is a characteristic index value corresponding to the second-order difference sequence of the target subsequence S to be analyzed, and continue to determine whether the second-order characteristic index value is greater than the first-order characteristic index value. If so, obtain a third-order characteristic index value, which is a characteristic index value corresponding to the third-order difference sequence of the target subsequence S to be analyzed, and continue to determine whether the third-order characteristic index value is greater than the second-order characteristic index value. If it is determined that the third-order characteristic index value is not greater than the second-order characteristic index value, the second-order difference sequence of the target subsequence S to be analyzed is used as the target subsequence to be compressed of the target subsequence S to be analyzed.

9. The artificial intelligence-based power grid business data management service terminal according to claim 8, characterized in that: The method for obtaining the characteristic index value includes: For the target subsequence S to be analyzed: The total number of data in the target subsequence S to be analyzed is recorded as a first quantity characterization value; for the f-th data in the target subsequence S to be analyzed, if there is data identical to the f-th data in the target subsequence to be compressed of the reference target subsequence, the f-th data is marked as repeated; the total number of data marked as repeated in the target subsequence S to be analyzed is counted and recorded as a second quantity characterization value; the ratio of the second quantity characterization value to the first quantity characterization value is recorded as a first eigenvalue; the range value of the target subsequence S to be analyzed is recorded as a first range value, the range value of the target subsequence to be compressed of the reference target subsequence is recorded as a second range value, the absolute value of the difference between the first range value and the second range value is recorded as a characteristic difference value, and the reciprocal of the result obtained by adding the characteristic difference value to a preset second constant is recorded as a second eigenvalue; the product of the first eigenvalue and the second eigenvalue is recorded as a characteristic index value corresponding to the target subsequence S to be analyzed; The method for obtaining the characteristic index value corresponding to the differential sequence is the same as the method for obtaining the characteristic index value corresponding to the target subsequence S to be analyzed.

10. The power grid business data management service terminal based on artificial intelligence according to claim 1, characterized in that: The method for acquiring the target data sequence to be compressed comprises: According to the sequence positions of the target subsequences in the time series to be compressed, the target subsequences to be compressed of all target subsequences are sorted, and the sequence obtained after sorting is recorded as the first sequence; and each target subsequence to be compressed in the first sequence is spliced ​​in turn, and the sequence after splicing is completed is recorded as the target data sequence to be compressed.

Images