Transformer area load intelligent operation and maintenance management system based on machine learning

Abstract

The invention relates to the technical field of electric power operation and maintenance, in particular to an intelligent operation and maintenance management system for court loads based on machine learning, which comprises a processor and a memory, the processor executes a computer program stored in the memory to realize the following steps: according to a time sequence decomposition result of a historical voltage data sequence of a to-be-analyzed node in a to-be-analyzed node set, a voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment, and a target influenced credible value of the to-be-analyzed node, calculating a target influenced credible value of the to-be-analyzed node; obtaining a target influence degree characterization value of the node to be adjusted on each node to be analyzed in a node set to be analyzed corresponding to the node to be adjusted, and obtaining a target voltage adjustment value of the node to be adjusted according to the target influence degree characterization value and the minimum voltage adjustment value, and performing voltage adjustment on the to-be-adjusted node by using the target voltage adjustment value of the to-be-adjusted node. Moreover, the method can improve the operation and maintenance management effect of the transformer area load, and guarantees the stability and safety of power supply.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power operation and maintenance, and particularly relates to a transformer area load intelligent operation and maintenance management system based on machine learning. BACKGROUND

[0002] The transformer area load operation and maintenance management system refers to a comprehensive management system for monitoring, analyzing, scheduling and maintaining the load in the transformer area of the power system, aiming to ensure that the power loads in the transformer area are reasonably distributed, the equipment is normally operated, the load changes and system faults are timely responded, and the stability and safety of power supply are ensured. Since node voltage monitoring and adjustment is a core component of transformer area load operation and maintenance management, monitoring and adjusting the node voltage in the transformer area is crucial for transformer area load intelligent operation and maintenance management.

[0003] Currently, it is usually determined whether to adjust the node based on the threshold or the required voltage range. After determining that the voltage needs to be adjusted, the voltage adjustment value is determined based on the difference between the threshold or the required voltage range and the current node voltage, and then the voltage is adjusted based on the determined voltage adjustment value. However, the nodes in the transformer area may have mutual influence. This mutual influence between nodes may result in poor voltage adjustment effect of the existing voltage adjustment value determination method, thereby resulting in poor operation and maintenance management effect of the transformer area load, and further affecting the stability and safety of power supply. For example, if it is determined that the voltage of node h1 needs to be adjusted, adjusting the voltage of node h1 may cause the voltage of node h2 to change. Therefore, if only the voltage value of node h1 itself is considered when determining the voltage adjustment value of node h1, without considering the influence of the adjustment of node h1 on node h2, it may result in the voltage of node h2 exceeding the specified voltage range when the voltage of node h1 is adjusted, thereby resulting in poor adjustment effect, i.e. poor operation and maintenance management effect of the transformer area load. Therefore, how to improve the operation and maintenance management effect of the transformer area load becomes a problem to be solved. SUMMARY

[0004] To solve the above problems, the present application provides a transformer area load intelligent operation and maintenance management system based on machine learning, which adopts the following technical solutions: One embodiment of the present application provides a transformer area load intelligent operation and maintenance management system based on machine learning, comprising a processor and a memory. The processor executes a computer program stored in the memory to realize the following steps: obtain current voltage data and historical voltage data sequence of all nodes in the target transformer area at the current monitoring time; obtaining the target affected credibility value of each node according to the difference between adjacent historical voltage data in the historical voltage data sequence and the time series decomposition result of the historical voltage data sequence; obtaining the target voltage adjustment value of the to-be-adjusted node according to the time series decomposition result of the historical voltage data sequence of the to-be-analyzed node in the to-be-analyzed node set of the to-be-adjusted node, the voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment, and the target affected credibility value of the to-be-analyzed node, and obtaining the target voltage adjustment value of the to-be-adjusted node according to the target influence degree representation value and the minimum voltage adjustment value; adjusting the voltage of the to-be-adjusted node by using the target voltage adjustment value of the to-be-adjusted node.

[0005] Beneficial effects: the present application firstly obtains the current voltage data and the historical voltage data sequence of all nodes in the target area at the current monitoring time; then obtains the target affected credibility value of each node according to the difference between adjacent historical voltage data in the historical voltage data sequence and the time series decomposition result of the historical voltage data sequence; then obtains the to-be-adjusted node, the to-be-analyzed node set of the to-be-adjusted node, and the minimum voltage adjustment value of the to-be-adjusted node at the current monitoring time according to the preset voltage interval and the current voltage data; obtains the target influence degree representation value of each to-be-analyzed node in the to-be-analyzed node set of the corresponding to-be-adjusted node according to the time series decomposition result of the historical voltage data sequence of the to-be-analyzed node in the to-be-analyzed node set, the voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment, and the target affected credibility value of the to-be-analyzed node; obtains the target voltage adjustment value of the to-be-adjusted node according to the target influence degree representation value and the minimum voltage adjustment value; finally adjusts the voltage of the to-be-adjusted node by using the target voltage adjustment value of the to-be-adjusted node. The voltage adjustment value of the to-be-adjusted node determined based on the influence of the voltage adjustment of the to-be-adjusted node on the voltage of other nodes, that is, the voltage adjustment value of the to-be-adjusted node determined according to the target influence degree representation value of the to-be-adjusted node on other nodes, can not only reduce the risk of exceeding the preset voltage interval of the node that does not need to be adjusted due to the influence of the voltage adjustment of the to-be-adjusted node to the maximum extent, but also can make the adjusted voltage of the to-be-adjusted node tend to the rated voltage of the to-be-adjusted node to the maximum extent, thereby improving the effect of the load operation and maintenance management of the area, and ensuring the stability and safety of power supply. BRIEF DESCRIPTION OF DRAWINGS

[0006] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, and the advantages thereof, the drawings required by the description of the embodiments or the prior art will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained based on these drawings without creative labor.

[0007] Figure 1 A flowchart of a kind of based on machine learning's district load intelligent operation and maintenance management method of the present application. DETAILED DESCRIPTION

[0008] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the scope of protection of the embodiments of the present application.

[0009] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0010] The present embodiment provides a kind of based on machine learning's district load intelligent operation and maintenance management system, including processor and memory, the processor executes the computer program stored in the memory to realize a kind of based on machine learning's district load intelligent operation and maintenance management, as Figure 1 As shown in the figure, the based on machine learning's district load intelligent operation and maintenance management method includes the following steps: Step S001, obtain the current voltage data and historical voltage data sequence of all nodes in the target district at the current monitoring time.

[0011] The embodiment mainly obtains the voltage adjustment value of the node needing voltage adjustment by analyzing the influence of the node needing voltage adjustment on other nodes when the node is adjusted, and then performs voltage adjustment based on the determined voltage adjustment value. The voltage adjustment value of the node needing voltage adjustment obtained based on the influence of the node needing voltage adjustment on other nodes when the node is adjusted can not only minimize the risk of exceeding the voltage specification range of other nodes caused by the voltage adjustment of the node needing voltage adjustment, but also maximize the voltage of the node to be adjusted after adjustment to approach the rated voltage of the node to be adjusted, thereby improving the effect of the substation load operation and maintenance management, ensuring the stability and safety of power supply, and the node voltage monitoring and adjustment is the core component of the substation load operation and maintenance management. In addition, the embodiment will take the node voltage monitoring and adjustment process in any substation covered as an example for analysis, and be recorded as a target substation. The embodiment will take the process of obtaining the voltage adjustment value of the node needing voltage adjustment in the target substation as an example for description. The embodiment only optimizes the voltage adjustment value of the node needing voltage adjustment, and does not change the process of adjusting the voltage of the node after determining the voltage adjustment value. In addition, the nodes and voltages appearing in the subsequent embodiment belong to the nodes and node voltages in the target substation.

[0012] In the embodiment, the current voltage data and historical voltage data sequence of all nodes in the target substation at the current monitoring time need to be obtained first. The current voltage data and historical voltage data sequence are mainly used to determine whether the node needs voltage adjustment and the influence of the node needing voltage adjustment on other nodes when the node is adjusted. The specific obtaining process of the current voltage data and historical voltage data sequence of all nodes in the target substation at the current monitoring time is as follows: Firstly, all nodes located in the target area are obtained and recorded as nodes in the target area, then the voltage data of each node in the target area at the current monitoring moment is obtained and recorded as the current voltage data of the corresponding node; then the voltage data of each node in the target area at each historical monitoring moment in the preset historical monitoring period before the current monitoring moment is obtained and recorded as the historical voltage data of the corresponding node; then in the preset historical monitoring period before the current monitoring moment, the time sequence sequence composed of all historical voltage data of the same node is recorded as the historical voltage data sequence of the corresponding node at the current monitoring moment, that is, for any node, the historical voltage data sequence of the node refers to the sequence composed of all historical voltage data of the node obtained in the preset historical monitoring period before the current monitoring moment; and in this embodiment, the implementer needs to set the preset historical monitoring period before the current monitoring moment and the time interval between adjacent monitoring moments according to the actual situation, for example, the preset historical monitoring period before the current monitoring moment can be set as one day or one week before the current monitoring moment, and the time interval between adjacent monitoring moments can be set as 1 second, that is, the voltage data is uploaded once every 1 second by the device for monitoring the voltage of the node, and the voltage of all nodes in the target area in this embodiment is synchronously collected.

[0013] In addition, it should be noted that the node in this embodiment refers to the electrical equipment node in the power supply range of the area, including physical quantity measuring points such as distribution transformers, branch switches and electric meters, mainly used for power distribution, voltage regulation and line loss management; and the acquisition method of the node voltage in the area mainly relies on intelligent monitoring equipment and digital technology system, and since the acquisition method is known, it will not be described in detail.

[0014] Therefore, the current voltage data and the historical voltage data sequence of all nodes in the target area at the current monitoring moment are obtained through the above process.

[0015] Step S002, according to the difference between adjacent historical voltage data in the historical voltage data sequence and the time sequence decomposition result of the historical voltage data sequence, the target affected confidence value of each node is obtained.

[0016] Since the voltage change of the node with unstable power consumption is more likely to be caused by the frequent switching of its own internal power supply, and the voltage change of the node with stable power consumption is more likely to be changed when it is affected by the adjustment of the voltage of some nodes, the voltage of the node with more stable historical voltage data change is more likely to be affected by the voltage adjustment of other nodes, so the embodiment will next reflect the possibility or credibility of the voltage change of the node being affected by the voltage adjustment of other nodes based on the stability characteristics of the historical voltage data of the node in different dimensions, that is, the embodiment will next reflect the credibility or possibility of the voltage change of the corresponding node being affected by the voltage adjustment of the node by analyzing the change of the voltage data in the historical voltage data sequence of the node, that is, the target affected credibility value of the node, and the specific acquisition process of the target affected credibility value of the node is as follows: For any node a in the target transformer area: First, the history voltage data sequence of node a at the current monitoring moment is recorded as sequence A, then the fluctuation determination value of each history voltage data in sequence A is obtained according to the difference between adjacent history voltage data in sequence A, and the specific obtaining process of the fluctuation determination value of each history voltage data in sequence A in this embodiment is that the fluctuation characteristic value of each history voltage data in sequence A is obtained according to the difference between adjacent history voltage data in sequence A, and the fluctuation characteristic value of the bth history voltage data in sequence A is the absolute value of the difference between the bth history voltage data in sequence A and the (b-1)th history voltage data in sequence A, b is greater than 1, and the fluctuation determination value of each history voltage data in sequence A is obtained according to the difference between the fluctuation characteristic values of the history voltage data in sequence A, and the fluctuation characteristic value and the fluctuation determination value of the first history voltage data in sequence A are not calculated, and the fluctuation determination value is mainly to filter out the data with large fluctuation degree for analysis; Since small amplitude voltage fluctuation will also occur when the node device corresponding to the node is in normal operation, but the voltage fluctuation at this time is relatively high in aggregation degree compared with the fluctuation caused by frequent switching of the power supply and the voltage adjustment of the node, that is, the fluctuation determination value is small, so the smaller the fluctuation determination value is, the smaller the possibility that node a is affected by the voltage adjustment of the node or the probability that the voltage change of node a is caused by the voltage adjustment of the node is greater, and the voltage adjustment of the node at this time may be the voltage adjustment of itself or the voltage adjustment of other nodes; Therefore, after obtaining the fluctuation determination value of the history voltage data, it is necessary to judge whether the fluctuation determination values of all history voltage data in sequence A are less than the preset fluctuation threshold, if yes, it is judged that node a is not affected by the voltage adjustment of other nodes or the voltage adjustment of other nodes almost does not cause the voltage change of node a, so at this time the first affected credibility value of node a is directly recorded as 0, that is, the voltage stability degree of node a in the voltage fluctuation frequency is assigned as 0, otherwise, it indicates that node a has the possibility of being affected by the voltage adjustment of the node, and further judgment is needed, that is, when the fluctuation determination values of all history voltage data in sequence A are not less than the preset fluctuation threshold, all history voltage data with the fluctuation determination value greater than the preset fluctuation threshold in sequence A are obtained and recorded as the history voltage data to be analyzed, then the first affected credibility value of node a is obtained according to the monitoring time interval of all adjacent history voltage data to be analyzed in sequence A, and the first affected credibility value of node a is the voltage stability degree of node a in the voltage fluctuation frequency, that is, the greater the first affected credibility value of node a is, the higher the voltage stability degree of node a in the voltage fluctuation frequency is.And since the frequency of voltage adjustment is lower than the frequency of frequent switching, that is, the voltage fluctuation caused by frequent switching is more frequent, and the voltage fluctuation caused by voltage adjustment is less frequent, the influence factors of the historical voltage change of node a can be further determined by analyzing the monitoring time interval of the historical voltage data to be analyzed, and then in this embodiment, according to the monitoring time interval between all adjacent historical voltage data to be analyzed in sequence A, the specific process of obtaining the first affected credibility value of node a is: first, the time sequence composed of all historical voltage data to be analyzed in sequence A is recorded as a feature sequence, then the monitoring time interval between adjacent historical voltage data to be analyzed in the feature sequence is obtained, and the sequence composed of all monitoring time intervals between adjacent historical voltage data to be analyzed in the feature sequence is recorded as a time interval sequence, and the cth monitoring time interval in the time interval sequence is the time interval between the monitoring time corresponding to the cth historical voltage data to be analyzed in the feature sequence and the monitoring time corresponding to the c+1th historical voltage data to be analyzed in the feature sequence, then the mean normalized value of the time interval sequence is recorded as the first affected credibility value of node a, the mean normalized value of the time interval sequence is Norm(T0), Norm() is a normalization function, T0 is the mean of the time interval sequence, and when T0 is smaller, that is, Norm(T0) is smaller, it indicates that the voltage fluctuation of node a is more affected by the frequent switching power supply or the voltage fluctuation stability of node a is lower, on the contrary, when T0 is larger, that is, Norm(T0) is larger, it indicates that the voltage fluctuation of node a is more likely to be affected by node voltage adjustment or the credibility is higher, or the voltage fluctuation stability of node a is higher, that is, the first affected credibility value of node a is larger, it indicates that the voltage fluctuation of node a is more likely to be affected by node voltage adjustment or the credibility is higher or the voltage fluctuation stability of node a is lower, the first affected credibility value of node a is one of the key parameters constituting the target affected credibility value corresponding to node a. In addition, in specific applications, the implementer needs to set a preset fluctuation threshold according to the actual situation, experimental statistics, and the value interval of the fluctuation determination value, such as setting the preset fluctuation threshold to 0.5 in this embodiment.

[0017] In the embodiment, according to the difference between the fluctuation characteristic values of the historical voltage data in sequence A, the specific process of obtaining the fluctuation judgment value of each historical voltage data in sequence A is as follows: for the bth historical voltage data in sequence A, b is greater than 1, a set composed of all historical voltage data in sequence A except the bth historical voltage data is denoted as the characteristic set of the bth historical voltage data, the first historical voltage data in sequence A is not included in the characteristic set, the normalized value of the absolute value accumulation result of the difference between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of each historical voltage data in the characteristic set is obtained and taken as the fluctuation judgment value of the bth historical voltage data, and the specific calculation formula of the fluctuation judgment value of the bth historical voltage data is as follows:

[0018] wherein, is the fluctuation judgment value of the bth historical voltage data in sequence A, is the total number of data in the characteristic set of the bth historical voltage data, is the fluctuation characteristic value of the jth historical voltage data in the characteristic set of the bth historical voltage data, is the fluctuation characteristic value of the bth historical voltage data, is the absolute value accumulation result of the difference between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of each historical voltage data in the characteristic set.

[0019] and the smaller is, that is, the smaller is, the smaller the difference between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of the historical voltage data in the characteristic set of the bth historical voltage data is, and it also indicates that the distribution between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of the historical voltage data in the characteristic set of the bth historical voltage data is more concentrated, and then it indicates that the fluctuation characteristic value of the bth historical voltage data is caused by the normal fluctuation under the normal operation of the node, and the larger is, that is, the larger is, the larger the difference between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of the historical voltage data in the characteristic set of the bth historical voltage data is, and it also indicates that the distribution between the fluctuation characteristic value of the bth historical voltage data and the fluctuation characteristic value of the historical voltage data in the characteristic set of the bth historical voltage data is less concentrated, and then it indicates that the fluctuation characteristic value of the bth historical voltage data is caused by the frequent switching of the power supply or the voltage adjustment.

[0020] In order to further improve the accuracy of the analysis on the reliability or possibility of the node voltage change affected by the node voltage adjustment, the embodiment will analyze the stability degree of the corresponding node voltage on the periodic change rule by combining the decomposition result obtained by the on-time decomposition of the historical voltage data sequence, that is, the second affected credibility value of node a. The specific obtaining process of the second affected credibility value of node a is as follows: the residual components of each historical voltage data in sequence A are obtained by performing ASTL time decomposition on sequence A, and the cumulative sum of the absolute values of the residual components of all historical voltage data in sequence A is recorded as a feature cumulative value. The inverse function mapping value of the result obtained by adding a preset constant to the feature cumulative value is taken as the second affected credibility value of node a. The inverse function mapping value of the result obtained by adding the preset constant to the feature cumulative value is the normalized value of the reciprocal of the result obtained by adding the preset constant to the feature cumulative value, and the second affected credibility value of node a is wherein, is the total number of data in sequence A, is the residual component of the bottom m historical voltage data in sequence A, c0 is a preset constant, and c0 is a constant greater than 0, which is set according to the actual situation, but the value of c0 is a constant greater than 0, for example, c0 can be set to 0.1; in addition, when is greater, it indicates that the voltage change of node a is less consistent with the periodic rule or the voltage stability degree is lower, and when the stability degree is higher, it indicates that the reliability or possibility of the voltage change of node a affected by the node voltage adjustment is higher. Therefore, when is greater, that is, the second affected credibility value of node a is smaller, the reliability or possibility of the voltage change of node a affected by the node voltage adjustment is lower, and when is smaller, that is, the second affected credibility value of node a is greater, the reliability or possibility of the voltage change of node a affected by the node voltage adjustment is higher; and the time decomposition process of the time sequence is known, so it will not be described.

[0021] After obtaining the first affected credibility value of node a and the second affected credibility value of node a, the first affected credibility value and the second affected credibility value of node a are fused to obtain the target affected credibility value. Specifically, the product of the first affected credibility value of node a and the second affected credibility value of node a is obtained and taken as the target affected credibility value of node a.

[0022] Therefore, the target affected credibility value of all nodes in the target transformer area can be obtained by the above process of obtaining the target affected credibility value of node a.

[0023] In step S003, according to the preset voltage interval and the current voltage data, a node to be adjusted, a set of nodes to be analyzed of the node to be adjusted, and a minimum voltage adjustment value of the node to be adjusted at a current monitoring time are obtained; according to a time sequence decomposition result of a historical voltage data sequence of a node to be analyzed in the set of nodes to be analyzed, a voltage change of the node to be analyzed when the node to be adjusted performs historical voltage adjustment, and a target affected trust value of the node to be analyzed, a target influence degree representation value of the node to be adjusted on each node to be analyzed in the set of nodes to be analyzed corresponding to the node to be adjusted is obtained, and according to the target influence degree representation value and the minimum voltage adjustment value, a target voltage adjustment value of the node to be adjusted is obtained.

[0024] After obtaining the target affected trust value of the node, the rated voltage of each node is combined to obtain the node to be adjusted, and the minimum voltage adjustment value of the node to be adjusted at the current monitoring time, that is, the specific process of obtaining the node to be adjusted and the minimum voltage adjustment value of the node to be adjusted at the current monitoring time is as follows: For any node in the area, the preset voltage interval of the node is obtained, and then it is judged whether the current voltage data of the node belongs to the preset voltage interval of the node. If yes, it indicates that the node is in normal operation or the voltage of the node is normal, and no adjustment is needed. If the current voltage data of the node is not in the preset voltage interval of the node, it is determined that the node has a risk of abnormal operation or is in abnormal operation, or it is determined that the node has voltage abnormality or has a high risk of voltage abnormality, and the node is recorded as the node to be adjusted.

[0025] In specific application, the preset voltage interval of different nodes generally refers to a floating range allowed near the rated voltage of the node, and the current allowed floating range is generally plus or minus 7% to 10% of the rated voltage of the node, that is, for any node, if the rated voltage of the node is V, the preset voltage interval of the node is [V-Vx10%, V+Vx7%].

[0026] When a node is determined to be an adjustment node, the minimum voltage adjustment value of the adjustment node at the current monitoring time is obtained according to the current voltage data of the adjustment node at the current monitoring time and the preset voltage interval of the adjustment node. Specifically, for any adjustment node, first, the maximum and minimum values in the preset voltage interval of the adjustment node are recorded as the limit voltage data of the adjustment node, that is, one adjustment node corresponds to two limit voltage data. Then, among the two limit voltage data of the adjustment node, the limit target adjustment voltage value of the adjustment node is selected as the limit target adjustment voltage value of the adjustment node, and the absolute value of the difference between the current voltage data of the adjustment node and the limit target adjustment voltage value of the adjustment node is taken as the minimum voltage adjustment value of the adjustment node at the current monitoring time. The minimum voltage adjustment value of the adjustment node in this embodiment is the minimum value of the voltage adjustment of the corresponding node. In the future, the minimum voltage adjustment value will be increased to different degrees. That is, the minimum voltage adjustment value of the adjustment node is the lower limit value of the adjustment of the adjustment node at the current monitoring time. However, the adjustment node in this embodiment is not adjusted by the minimum voltage adjustment value of the corresponding adjustment node because frequent voltage adjustment may occur, which may cause the service life of the device to be shortened, the efficiency of the power grid to be reduced, the economic cost to be increased, and other hazards. In addition, the voltage adjustment value of any node refers to the voltage value that needs to be adjusted, such as adjusting the voltage of a node from 30 volts to 36 volts. At this time, the voltage adjustment value of the node is 6 volts.

[0027] Then, this embodiment analyzes the influence of the voltage adjustment of the adjustment node on other nodes. Before the analysis, this embodiment needs to determine the node set that may be affected by the voltage adjustment of the adjustment node, that is, the analysis node set of the adjustment node. For any adjustment node, the set of all nodes in the target area except the adjustment node is recorded as the analysis node set of the adjustment node.

[0028] After the set of to-be-analyzed nodes of each to-be-adjusted node is obtained, the target influence degree representation value of each to-be-adjusted node on each to-be-analyzed node in the set of to-be-analyzed nodes of the corresponding to-be-adjusted node is obtained according to the time sequence decomposition result of the historical voltage data sequence of each to-be-analyzed node in the set of to-be-analyzed nodes of each to-be-adjusted node, the voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment, and the target affected trust value of the to-be-analyzed node. In order to facilitate analysis and understanding, the acquisition process of the target influence degree representation value of any to-be-adjusted node f on the dth to-be-analyzed node in the set of to-be-analyzed nodes of the to-be-adjusted node f is taken as an example for description, that is, the acquisition process of the target influence degree representation value of the to-be-adjusted node f on the dth to-be-analyzed node in the set of to-be-analyzed nodes of the to-be-adjusted node f is as follows: First, in a preset historical monitoring time period before the current monitoring moment, each historical voltage adjustment experienced by the to-be-adjusted node f is obtained and is recorded as a to-be-analyzed historical voltage adjustment experienced by the to-be-adjusted node f. For example, if the to-be-adjusted node f experiences five voltage adjustments in the preset historical monitoring time period before the current monitoring moment, then the five voltage adjustments experienced by the to-be-adjusted node f in the preset historical monitoring time period before the current monitoring moment are all to-be-analyzed historical voltage adjustments experienced by the to-be-adjusted node f. Then, the residual component curve obtained by performing ASTL time sequence decomposition on the historical voltage data sequence of the dth to-be-analyzed node is obtained and is recorded as the residual component curve of the historical voltage data sequence of the dth to-be-analyzed node. The abscissa of the xth point on the residual component curve of the historical voltage data sequence of the dth to-be-analyzed node is the monitoring time of the xth historical voltage data in the historical voltage data sequence of the dth to-be-analyzed node, and the ordinate is the residual component of the xth historical voltage data. Then, the mutation point on the residual component curve of the historical voltage data sequence of the dth to-be-analyzed node is obtained by using the 3σ criterion, and the abscissa value of each mutation point is recorded as the historical period mutation moment of the dth to-be-analyzed node. In addition, the process of obtaining the mutation point by using the 3σ criterion is known, and thus will not be described in detail.

[0029] If the adjustment start time of the to-be-adjusted node f experienced s-th to-be-analyzed historical voltage adjustment coincides with the historical period mutation time of the d-th to-be-analyzed node, it indicates that the voltage change of the d-th to-be-analyzed node is not completely caused by the voltage adjustment, and thus the to-be-analyzed historical voltage adjustment in which the adjustment start time coincides with the historical period mutation time of the d-th to-be-analyzed node is excluded when analyzing the target influence degree representation value of the to-be-adjusted node f on the d-th to-be-analyzed node. Therefore, it is determined whether the adjustment start time of the to-be-adjusted node f experienced s-th to-be-analyzed historical voltage adjustment coincides with all historical period mutation times of the d-th to-be-analyzed node. If not, the to-be-adjusted node f experienced s-th to-be-analyzed historical voltage adjustment is recorded as a target historical voltage adjustment of the to-be-adjusted node f. For example, if the adjustment start time of the to-be-adjusted node f experienced first voltage adjustment does not coincide with all historical period mutation times of the d-th to-be-analyzed node in a preset historical monitoring time period before the current monitoring time, the to-be-adjusted node f experienced first voltage adjustment is a target historical voltage adjustment of the to-be-adjusted node f. If the voltage of the to-be-adjusted node f is out of the preset voltage interval of the to-be-adjusted node f at a historical monitoring time, the system triggers an initial time point of the voltage adjustment control strategy, and the historical monitoring time is the adjustment start time of the voltage adjustment of the to-be-adjusted node f. If the voltage of the to-be-adjusted node f is out of the preset voltage interval of the to-be-adjusted node f at the historical monitoring time w1, the system triggers the voltage adjustment control strategy to adjust the voltage of the to-be-adjusted node f. However, the voltage of the to-be-adjusted node f returns to the normal range at the historical monitoring time w2, that is, the voltage of the to-be-adjusted node f is in the preset voltage interval of the to-be-adjusted node f at the historical monitoring time w2. Therefore, the historical monitoring time w2 is the adjustment end time of the voltage adjustment of the to-be-adjusted node f.

[0030] After obtaining all target historical voltage adjustments of the to-be-adjusted node f, the voltage change ratio between the to-be-adjusted node f and the d-th to-be-analyzed node is obtained under each target historical voltage adjustment of the to-be-adjusted node f. The voltage change ratio can reflect the influence of the voltage adjustment of the to-be-adjusted node f on the voltage of the d-th to-be-analyzed node. The specific obtaining process of the voltage change ratio between the to-be-adjusted node f and the d-th to-be-analyzed node under each target historical voltage adjustment of the to-be-adjusted node f is as follows: for the g-th target historical voltage adjustment of the to-be-adjusted node f, First, the voltage adjustment value of the to-be-adjusted node f when performing the gth target historical voltage adjustment is obtained, and is recorded as a first voltage change value, which is the voltage change value of the to-be-adjusted node f under the gth target historical voltage adjustment; then, the historical voltage data of the dth to-be-analyzed node at the adjustment start time of the gth target historical voltage adjustment is obtained, and is recorded as first voltage data, the historical voltage data of the dth to-be-analyzed node at the adjustment end time of the gth target historical voltage adjustment is obtained, and is recorded as second voltage data, the absolute value of the difference between the second voltage data and the first voltage data is calculated, and is recorded as a second voltage change value, which is the voltage change value of the dth to-be-analyzed node under the gth target historical voltage adjustment; then, the ratio of the second voltage change value to the first voltage change value is obtained, and is recorded as the voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node under the gth to-be-analyzed voltage adjustment experienced by the to-be-adjusted node f, and the greater the voltage change ratio, the greater the influence of the voltage adjustment of the to-be-adjusted node f on the voltage of the dth to-be-analyzed node, and vice versa.

[0031] After the voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node under each target historical voltage adjustment experienced by the to-be-adjusted node f is obtained, the average of all voltage change ratios between the to-be-adjusted node f and the dth to-be-analyzed node is calculated, and is recorded as the initial influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node. Under each target historical voltage adjustment experienced by the to-be-adjusted node f, a voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node can be obtained, that is, the number of voltage change ratios between the to-be-adjusted node f and the dth to-be-analyzed node is consistent with the number of target historical voltage adjustments experienced by the to-be-adjusted node f. After the initial influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node is obtained, the initial influence degree representation value of the dth to-be-analyzed node is adjusted by using the target affected credibility value of the dth to-be-analyzed node, so as to obtain a target influence degree representation value. Specifically, the product of the initial influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node and the target affected credibility value of the dth to-be-analyzed node is obtained, and is recorded as the target influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node. The greater the target influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node, the greater the influence of the voltage adjustment of the to-be-adjusted node f on the voltage of the dth to-be-analyzed node. Conversely, the smaller the target influence degree representation value of the to-be-adjusted node f to the dth to-be-analyzed node, the smaller the influence of the voltage adjustment of the to-be-adjusted node f on the voltage of the dth to-be-analyzed node.

[0032] In order to make the node needing voltage adjustment be within the voltage specification requirement after adjustment, and also to reduce the risk of the voltage of other nodes exceeding the voltage specification requirement caused by the adjustment of the node needing voltage adjustment as much as possible, that is, to reduce the influence on other nodes as much as possible when the node needing voltage adjustment is adjusted, the subsequent requirement of the embodiment is that the smaller the voltage adjustment degree of the to-be-adjusted node f, the greater the influence on other nodes. However, the minimum voltage adjustment value of the to-be-adjusted node f is not less than the minimum voltage adjustment value. Therefore, the target voltage adjustment value of each to-be-adjusted node is obtained according to the target influence degree representation value and the minimum voltage adjustment value in the embodiment. Specifically, for any to-be-adjusted node f: The inverse normalization processing result of the result obtained by accumulating the target influence degree characteristic value of the to-be-adjusted node f to each to-be-analyzed node in the to-be-analyzed node set of the to-be-adjusted node f is taken as the target adjustment coefficient of the to-be-adjusted node f, the absolute value of the difference between the standard voltage of the to-be-adjusted node f and the limit target adjustment voltage value of the to-be-adjusted node f is taken as a characteristic difference value, and the result of multiplying the characteristic difference value by the target adjustment coefficient and then adding the minimum voltage adjustment value of the to-be-adjusted node f at the current monitoring moment is taken as the target voltage adjustment value of the to-be-adjusted node f at the current monitoring moment, that is, the acquisition formula of the target voltage adjustment value of the to-be-adjusted node f at the current monitoring moment is:

[0033] wherein, is the target voltage adjustment value of the to-be-adjusted node f at the current monitoring moment, is the minimum voltage adjustment value of the to-be-adjusted node f at the current monitoring moment, is the limit target adjustment voltage value of the to-be-adjusted node f, is the standard voltage of the to-be-adjusted node f, and K is the number of to-be-analyzed nodes in the to-be-analyzed node set of the to-be-adjusted node f, is the target influence degree characteristic value of the to-be-adjusted node f to the kth to-be-analyzed node in the to-be-analyzed node set of the to-be-adjusted node f; and is smaller, the target voltage adjustment value of the to-be-adjusted node f at the current monitoring moment is larger, and the voltage value of the adjusted to-be-adjusted node f is closer to the standard voltage of the to-be-adjusted node f, the standard voltage of the to-be-adjusted node f is the rated voltage of the to-be-adjusted node f, and Norm() is a normalization function.

[0034] In step S004, the target voltage adjustment value of the to-be-adjusted node is used to adjust the voltage of the to-be-adjusted node.

[0035] Finally, the target voltage adjustment value of the to-be-adjusted node is used to adjust the voltage of the corresponding to-be-adjusted node at the current monitoring moment. For example, if the current voltage data of a to-be-adjusted node at this time is G0, and G0 is less than the minimum value of the preset voltage interval of the to-be-adjusted node, and the target voltage adjustment value of the to-be-adjusted node obtained at the current monitoring moment is G1, then the voltage of the to-be-adjusted node at the current monitoring moment needs to be adjusted to be larger, and the adjusted voltage is G0+G1.

[0036] Thus, the embodiment completes the monitoring and adjustment of the nodes in the transformer area, and the voltage adjustment value of the to-be-adjusted node determined based on the influence of the voltage adjustment of the to-be-adjusted node on the voltages of other nodes can not only maximize the reduction of the risk of exceeding the voltage specification range of other nodes due to the voltage adjustment of the to-be-adjusted node, but also maximize the voltage of the to-be-adjusted node after adjustment to approach the rated voltage of the to-be-adjusted node, that is, the embodiment can maximize the reduction of the risk of exceeding the preset voltage interval of the node that does not need to be adjusted due to the voltage adjustment of the to-be-adjusted node, and also maximize the voltage of the to-be-adjusted node after adjustment to approach the rated voltage of the to-be-adjusted node, thereby improving the effect of transformer area load operation and maintenance management and ensuring the stability and safety of power supply.

[0037] In summary, the embodiment first acquires the current voltage data and the historical voltage data sequence of all nodes in the target transformer area at the current monitoring time; then obtains the target influence credibility value of each node according to the difference between adjacent historical voltage data in the historical voltage data sequence and the time sequence decomposition result of the historical voltage data sequence; then obtains the to-be-adjusted node, the to-be-analyzed node set of the to-be-adjusted node, and the minimum voltage adjustment value of the to-be-adjusted node at the current monitoring time according to the preset voltage interval and the current voltage data; obtains the target influence degree representation value of the to-be-adjusted node on each to-be-analyzed node in the to-be-analyzed node set of the to-be-adjusted node according to the time sequence decomposition result of the historical voltage data sequence of the to-be-analyzed node in the to-be-analyzed node set, the voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment, and the target influence credibility value of the to-be-analyzed node; obtains the target voltage adjustment value of the to-be-adjusted node according to the target influence degree representation value and the minimum voltage adjustment value; and finally performs voltage adjustment on the to-be-adjusted node by using the target voltage adjustment value of the to-be-adjusted node. The voltage adjustment value of the to-be-adjusted node determined based on the influence of the voltage adjustment of the to-be-adjusted node on the voltages of other nodes, that is, the voltage adjustment value of the to-be-adjusted node determined based on the target influence degree representation value of the to-be-adjusted node on other nodes, can not only maximize the reduction of the risk of exceeding the preset voltage interval of the node that does not need to be adjusted due to the voltage adjustment of the to-be-adjusted node, but also maximize the voltage of the to-be-adjusted node after adjustment to approach the rated voltage of the to-be-adjusted node, thereby improving the effect of transformer area load operation and maintenance management and ensuring the stability and safety of power supply.

[0038] The above-described embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to part 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 application, and should be included in the protection scope of the present application.

Claims

1. A machine learning-based intelligent management system for transformer area load operation and maintenance, comprising a processor and a memory, characterized in that, The processor executes the computer program stored in the memory to implement the following steps: Obtaining current voltage data and historical voltage data sequence of all nodes in the target area at the current monitoring moment; According to the difference between adjacent historical voltage data in the historical voltage data sequence and the time series decomposition result of the historical voltage data sequence, obtaining the target affected credibility value of each node; According to the preset voltage interval and the current voltage data, obtaining the to-be-adjusted node, the to-be-analyzed node set of the to-be-adjusted node and the minimum voltage adjustment value of the to-be-adjusted node at the current monitoring moment; according to the time series decomposition result of the historical voltage data sequence of the to-be-analyzed node in the to-be-analyzed node set, the voltage change of the to-be-analyzed node when the to-be-adjusted node performs historical voltage adjustment and the target affected credibility value of the to-be-analyzed node, obtaining the target influence degree representation value of each to-be-analyzed node in the to-be-analyzed node set corresponding to the to-be-adjusted node by the to-be-adjusted node, and obtaining the target voltage adjustment value of the to-be-adjusted node according to the target influence degree representation value and the minimum voltage adjustment value; Adjusting the voltage of the to-be-adjusted node by using the target voltage adjustment value of the to-be-adjusted node.

2. The machine learning-based intelligent management system for transformer area load operation and maintenance according to claim 1, wherein, The method for obtaining the target affected credibility value of each node comprises: For any node a: record the historical voltage data sequence of the node a as sequence A, obtain the fluctuation judgment value of each historical voltage data in the sequence A according to the difference between adjacent historical voltage data in the sequence A, and record the historical voltage data with the fluctuation judgment value greater than a preset fluctuation threshold as to-be-analyzed historical voltage data, obtain the first affected credibility value of the node a according to the monitoring time interval between all adjacent to-be-analyzed historical voltage data in the sequence A; perform ASTL time series decomposition on the sequence A to obtain the residual component of each historical voltage data in the sequence A, and take the inverse function mapping value of the cumulative sum of the absolute values of all residual components of the historical voltage data in the sequence A as the second affected credibility value of the node a; take the product of the first affected credibility value and the second affected credibility value as the target affected credibility value of the node a. 3.The machine learning-based intelligent management system for transformer area load operation and maintenance, according to claim 2, wherein, The method for obtaining the fluctuation judgment value of each historical voltage data in the sequence A comprises: According to the difference between adjacent historical voltage data in the sequence A, obtaining the fluctuation characteristic value of each historical voltage data in the sequence A, and the fluctuation characteristic value of the bth historical voltage data in the sequence A is the absolute value of the difference between the bth historical voltage data and the b-1th historical voltage data in the sequence A, and b is greater than 1; according to the difference between the fluctuation characteristic values of the historical voltage data, obtaining the fluctuation judgment value of each historical voltage data in the sequence A.

4. The machine learning-based intelligent management system for transformer area load operation and maintenance according to claim 3, characterized in that, The method for obtaining the fluctuation judgment value of each historical voltage data in the sequence A according to the difference between the fluctuation characteristic values of the historical voltage data comprises: For the bth historical voltage data in the sequence A, a set composed of all historical voltage data in the sequence A except the bth historical voltage data is denoted as a feature set, and a normalized value of an accumulated sum of absolute values of differences between a fluctuation feature value of the bth historical voltage data and fluctuation feature values of each historical voltage data in the feature set is taken as a fluctuation determination value of the bth historical voltage data. 5.The machine learning based intelligent management system for transformer area load operation and maintenance, according to claim 2, wherein, The method for obtaining the first affected trust value of the node a comprises: a time sequence composed of all to-be-analyzed historical voltage data in the sequence A is denoted as a feature sequence, and a normalized value of a mean value of monitoring time intervals between all adjacent to-be-analyzed historical voltage data in the feature sequence is taken as the first affected trust value of the node a. 6.The machine learning based intelligent management system for transformer area load operation and maintenance, according to claim 1, wherein, The method for obtaining the to-be-adjusted node and the to-be-analyzed node set of the to-be-adjusted node comprises: For any node, if current voltage data of the node is not within a preset voltage interval of the node, the node is taken as a to-be-adjusted node; For any to-be-adjusted node, a set composed of all nodes in the target area except the to-be-adjusted node is taken as a to-be-analyzed node set of the to-be-adjusted node.

7. The machine learning-based intelligent management system for transformer area load operation and maintenance according to claim 1, wherein, The method for obtaining the minimum voltage adjustment value of the to-be-adjusted node at a current monitoring moment comprises: For any to-be-adjusted node, a maximum value and a minimum value in a preset voltage interval of the to-be-adjusted node are both taken as limit voltage data of the to-be-adjusted node, among the two limit voltage data of the to-be-adjusted node, a limit target adjustment voltage value closest to current voltage data of the to-be-adjusted node is selected as a limit target adjustment voltage value of the to-be-adjusted node, and an absolute value of a difference between the current voltage data of the to-be-adjusted node and the limit target adjustment voltage value of the to-be-adjusted node is taken as a minimum voltage adjustment value of the to-be-adjusted node at a current monitoring moment. 8.The machine learning based intelligent management system for transformer area load operation and maintenance, according to claim 7, wherein, The method for obtaining the target influence degree representation value of each to-be-analyzed node in the to-be-analyzed node set of the corresponding to-be-adjusted node by the to-be-adjusted node comprises: For any to-be-adjusted node f and a dth to-be-analyzed node in the to-be-analyzed node set of the to-be-adjusted node f: in a preset historical monitoring time period before a current monitoring moment, each historical voltage adjustment experienced by the to-be-adjusted node f is obtained and is taken as a to-be-analyzed historical voltage adjustment experienced by the to-be-adjusted node f, a residual component curve corresponding to a historical voltage data sequence of the dth to-be-analyzed node is obtained, a mutation point on the residual component curve is obtained by using a 3σ criterion, and an abscissa value of the mutation point is taken as a historical period mutation moment of the dth to-be-analyzed node, if an adjustment start moment of an s th to-be-analyzed historical voltage adjustment experienced by the to-be-adjusted node f does not repeat all historical period mutation moments of the dth to-be-analyzed node, the s th to-be-analyzed historical voltage adjustment experienced by the to-be-adjusted node f is taken as a target historical voltage adjustment experienced by the to-be-adjusted node f. The voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node is obtained under each target historical voltage adjustment experienced by the to-be-adjusted node f, and the average of all voltage change ratios between the to-be-adjusted node f and the dth to-be-analyzed node is taken as an initial influence degree representation value of the to-be-adjusted node f on the dth to-be-analyzed node, and the product of the initial influence degree representation value and a target affected credibility value of the dth to-be-analyzed node is taken as a target influence degree representation value of the to-be-adjusted node f on the dth to-be-analyzed node. 9.The machine learning based intelligent management system for transformer area load operation and maintenance, according to claim 8, wherein, The method for obtaining the voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node comprises the following steps of: For the gth target historical voltage adjustment experienced by the to-be-adjusted node f, a voltage adjustment value of the to-be-adjusted node f when the gth target historical voltage adjustment is performed is obtained and recorded as a first voltage change value, historical voltage data of the dth to-be-analyzed node at an adjustment start time of the gth target historical voltage adjustment is recorded as first voltage data, historical voltage data of the dth to-be-analyzed node at an adjustment end time of the gth target historical voltage adjustment is recorded as second voltage data, an absolute value of a difference between the second voltage data and the first voltage data is recorded as a second voltage change value, and a ratio of the second voltage change value to the first voltage change value is recorded as the voltage change ratio between the to-be-adjusted node f and the dth to-be-analyzed node under the gth target historical voltage adjustment experienced by the to-be-adjusted node f.

10. The machine learning-based intelligent management system for transformer area load operation and maintenance according to claim 7, wherein, The method for obtaining the target voltage adjustment value of the to-be-adjusted node comprises the following steps of: For any to-be-adjusted node f, a reverse normalization processing result of a result obtained by accumulating the target influence degree representation values between the to-be-adjusted node f and each to-be-analyzed node in a to-be-analyzed node set of the to-be-adjusted node f is taken as a target adjustment coefficient of the to-be-adjusted node f, an absolute value of a difference between a standard voltage of the to-be-adjusted node f and a limit target adjustment voltage of the to-be-adjusted node f is recorded as a feature difference value, and a result obtained by multiplying the feature difference value and the target adjustment coefficient and then adding a minimum voltage adjustment value of the to-be-adjusted node f at a current monitoring time is taken as a target voltage adjustment value of the to-be-adjusted node f at the current monitoring time.

Images