Power equipment asset management method and system

Abstract

The invention relates to the technical field of data processing, in particular to a power equipment asset management method and system. The method comprises the following steps: determining the selection probability of each dimension under child nodes according to the distribution of power data in the child nodes of an isolated tree of power equipment, determining the segmentation necessity of the child nodes according to the time sequence regularity and the selection probability of the power data in the child nodes and the number of the power data in the child nodes, and when the segmentation necessity is greater than a segmentation threshold value, determining the segmentation necessity of the child nodes; according to the selection probability, randomly selecting one dimension in an unequal probability manner to divide the sub-nodes, continuously updating the isolated forest according to the newly generated power data, identifying the current state of the power equipment according to the abnormal score of the current power data, and synchronizing the current state of the power equipment to the digital twin model of the power equipment in real time. And power equipment asset monitoring and management are realized. According to the invention, the state identification of the power equipment is accurate and efficient, and the real-time performance of power equipment asset monitoring and management is improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and system for managing power equipment assets. Background Art

[0002] Power equipment asset management is a key link to ensure the stable, safe and efficient operation of the power system. With the continuous growth of power demand and the complexity of the power system, the importance of power equipment asset management has become increasingly prominent. Traditional power equipment management methods mainly rely on manual records and simple maintenance plans, and often face problems such as inaccurate equipment failure prediction, unreasonable allocation of maintenance resources, and low asset utilization efficiency, which affect the reliability and economy of the power system.

[0003] With the development of information technology and data analysis technology, modern power equipment asset management is gradually transforming towards intelligence and data-driven, and real-time monitoring and predictive maintenance of equipment status are carried out through sensor technology. For example, the patent document with publication number CN117556347A discloses a method for power equipment fault prediction and health management based on industrial big data, including: S1: constructing a data set of operating parameters containing fault information; S2: processing outliers and missing values, and realizing normalization; S3: performing feature selection, selecting feature groups with strong correlation with equipment operating index parameters; S4: dividing the data of the feature group into training set, test set and validation set; S5: creating a normal operating parameter prediction model of MSA-BLSTM for prediction; S6: using the validation set to verify the prediction model, using the predicted operating parameters output by MSA-BLSTM as input, and using the isolation forest method model to perform fault judgment to realize equipment fault prediction; S7: evaluating equipment fault warning and health status based on the fault prediction results.

[0004] The above methods can realize fault prediction of power equipment to a certain extent. However, since the power data of power equipment is streaming data, in order to realize real-time monitoring of the power equipment status, it is necessary to continuously build new isolated forests based on the newly generated power data, which is inefficient and occupies a lot of computing resources. At the same time, it affects the real-time performance of power equipment status monitoring and is not conducive to power equipment asset monitoring and management. Summary of the invention

[0005] In order to solve the problem that new isolated forests are continuously constructed based on newly generated power data, which is inefficient and occupies a large amount of computing resources, affects the real-time performance of power equipment status monitoring, and is not conducive to power equipment asset monitoring and management, the present invention provides a power equipment asset management method and system.

[0006] In a first aspect, the present invention provides a method for managing power equipment assets, which adopts the following technical solution: A method for managing power equipment assets, comprising the steps of: An isolation forest is constructed based on the historical power data of the power equipment; in the process of constructing the isolation forest: for each child node in the isolation tree, the selection probability of each dimension under the child node is determined according to the distribution of the power data in the child node in each dimension, and the selection probability is positively correlated with the fluctuation range of the power data in the child node in each dimension, and negatively correlated with the information entropy of the power data in the child node in each dimension; the necessity of segmentation of the child node is determined, and the necessity of segmentation is negatively correlated with the temporal regularity of the power data in the child node, and positively correlated with the maximum value of the selection probability of each dimension under the child node and the amount of power data contained in the child node; in response to the segmentation The necessity is greater than the preset segmentation threshold, and a dimension is randomly selected with unequal probability according to the selection probability of each dimension under the sub-node as the segmentation dimension of the sub-node, and the power data in the sub-node is divided into two new sub-nodes using the segmentation dimension; the newly generated power data is added to the isolated forest and the isolated forest is updated; the anomaly score of the newly generated power data is determined according to the position of the newly generated power data in each isolated tree in the isolated forest, and the current state of the power equipment is identified according to the size of the anomaly score; the current state of the power equipment is synchronized to the digital twin model of the power equipment in real time to realize power equipment asset monitoring and management.

[0007] The effect is as follows: the present invention constructs an isolation forest based on historical power data, and based on the constructed isolation forest, updates the isolation forest according to the newly generated power data, thereby avoiding the continuous reconstruction of the isolation forest, saving computing resources, and improving the efficiency of power equipment anomaly identification, thereby improving the real-time monitoring and management of power equipment assets; the present invention calculates the selection probability of each dimension when splitting the node according to the distribution of each dimension of power data in the node, and randomly selects a dimension with unequal probability according to the selection probability to divide the node, thereby enhancing the probability of selecting a dimension that may contain anomalies, avoiding the possibility that the traditional isolation forest randomly selects dimensions with equal probability for division, which may cause the dimension containing anomalies to be selected later, and improving the iteration of the isolation forest. Efficiency is improved, so that abnormal power data is included in the nodes of the shallower layer of the isolated tree as much as possible, thereby enhancing the accuracy of identifying abnormal power data; power data is time series data, and abnormal power data usually appears continuously. The present invention determines the necessity of segmenting the sub-node according to the time series regularity of the power data in the sub-node, the selection probability of each dimension under the sub-node and the amount of power data contained in the sub-node, and segments the sub-node according to the necessity of segmentation, thereby avoiding that normal power data with periodic regularity or abnormal power data with continuous time series distribution is segmented into multiple sub-nodes, thereby improving the segmentation efficiency of the isolated tree and the identification efficiency of abnormal power data, improving the accuracy and real-time performance of power equipment status identification, and further improving the real-time performance of power equipment asset monitoring and management.

[0008] Preferably, the selection probability of each dimension under the sub-node satisfies the expression: ;in, For the child node The probability of selection of each dimension; For all power data in the child node The maximum value of the dimension; For all power data in the child node The minimum value of the dimension; is the number of dimensions of power data; For all power data in the child node Information entropy of dimensions; For all power data in the child node The number of different values ​​in each dimension; is the maximum value function.

[0009] The effect is that: therefore, the present invention calculates the selection probability of each dimension when splitting the node according to the distribution of each dimension of the power data in the node, so as to randomly select a dimension for division with unequal probability according to the selection probability, thereby enhancing the probability of selecting the dimension that may contain anomalies, so that the dimension that may contain anomalies is selected as early as possible, and the abnormal power data is included in the nodes at the shallower level of the isolation tree as much as possible, thereby improving the recognition accuracy of abnormal power data and the efficiency of building the isolation forest.

[0010] Preferably, the temporal regularity of the power data in the subnode satisfies the expression: ; in, is the temporal regularity of the power data in the child nodes; For the child node The length of the power data segment; is the number of power data segments in the child node; is the average length of all power data segments in the child node; Indicates the child node The power data segment and The time interval between power data segments.

[0011] The effect is that the present invention takes into account the periodic fluctuation law of power data, measures whether there may be anomalies in the power data by calculating the timing regularity of the power data in the sub-nodes, determines the necessity of segmenting the sub-nodes according to the timing regularity, improves the accuracy of segmenting abnormal power data, and further improves the accuracy of identifying anomalies.

[0012] Preferably, the necessity of segmentation of the sub-nodes satisfies the expression: ;in, The necessity of splitting the child nodes; is the temporal regularity of the power data in the child nodes; For the child node The probability of selection of each dimension; is the number of dimensions of power data; is the number of power data contained in the child node; is the number of all power data; is the natural exponential function; is the maximum value function.

[0013] Preferably, the adding of the currently newly generated power data to the isolated forest and updating the isolated forest includes: for the currently newly generated power data, determining the leaf node to which the power data belongs in each isolated tree, and adding the power data to the corresponding leaf node; obtaining the necessity of segmentation of the corresponding leaf node after the power data is added; in response to the necessity of segmentation of the leaf node being greater than a preset segmentation threshold, selecting a dimension with unequal probability according to the selection probabilities of each dimension under the leaf node as the segmentation dimension of the leaf node, and dividing the power data in the leaf node into two new child nodes by using the segmentation dimension; repeating the segmentation judgment on the new child nodes until the segmentation necessity of all new child nodes is no greater than the preset segmentation threshold, and stopping the iteration to achieve the update of the isolated forest.

[0014] The effect is that: based on the constructed isolation forest, the present invention updates the isolation forest according to the newly generated power data, avoiding the continuous reconstruction of the isolation forest, saving computing resources, improving the efficiency of abnormal identification of power equipment, and thus improving the real-time monitoring and management of power equipment assets.

[0015] Preferably, the abnormal score of the currently newly generated power data satisfies the expression: ;in, Indicates the abnormal score of the currently newly generated power data; Indicates the newly generated power data in the The depth of the leaf node in an isolated tree; Indicates The height of an isolated tree; Indicates the newly generated power data in the The number of power data contained in the leaf nodes of each isolated tree; Indicates The mean value of the number of power data contained in all leaf nodes in an isolated tree; Indicates the newly generated power data in the The number of power data segments in the leaf nodes of an isolated tree; Indicates The average number of power data segments in all leaf nodes in an isolated tree.

[0016] The effect is that: the present invention takes into account that the abnormal data of the power equipment is continuously changing, and judges whether the power data in the leaf node where the current newly generated power data is located in the isolated tree is continuous in time series according to the number of power data segments in the leaf node where the current newly generated power data is located in the isolated tree. If the power data contained in the leaf node where the current newly generated power data is located is more continuous in time series, it means that the difference between the power data in the most recent period and the previous power data is greater, and the power equipment is more likely to be abnormal at this time. The present invention combines the layer depth of the leaf node where the current newly generated power data is located in the isolated tree, as well as the number of power data and the number of power data segments in the leaf node to obtain the abnormality score, and the result is more accurate.

[0017] Preferably, the method for acquiring the power data segments is: dividing all power data in the subnode into multiple power data segments in time sequence, all power data in each power data segment are continuous in time sequence, and different power data segments are not continuous in time sequence.

[0018] Preferably, the identifying the current state of the power equipment according to the size of the abnormality score includes: in response to the abnormality score of the currently newly generated power data being greater than an abnormality threshold, the power equipment state is abnormal; otherwise, the power equipment state is normal.

[0019] In a second aspect, the present invention provides an electric power equipment asset management system, which adopts the following technical solution: An electric power equipment asset management system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned electric power equipment asset management method is implemented.

[0020] By adopting the above technical solution, the above-mentioned power equipment asset management method is generated into a computer program and stored in a memory so as to be loaded and executed by a processor, thereby making a terminal device based on the memory and the processor for easy use.

[0021] The present invention has the following technical effects: based on the constructed isolation forest, the present invention updates the isolation forest according to the newly generated power data, improves the efficiency of identifying abnormalities of power equipment, and thus improves the real-time performance of power equipment asset monitoring and management; the present invention randomly selects a dimension with unequal probability to divide nodes, which enhances the probability of selecting dimensions that may contain abnormalities, improves the iteration efficiency of the isolation forest, and enhances the accuracy of identifying abnormal power data; the present invention splits sub-nodes according to the necessity of splitting, improves the splitting efficiency of the isolation tree and the efficiency of identifying abnormal power data, improves the accuracy and real-time performance of power equipment status identification, and further improves the real-time performance of power equipment asset monitoring and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] By reading the detailed description below with reference to the accompanying drawings, the above and other purposes, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts.

[0023] Figure 1 is a flow chart of a method for managing power equipment assets according to an embodiment of the present invention; Figure 2 It is a schematic diagram of the periodic variation of current; Figure 3 is a schematic diagram of temperature anomaly distribution; Figure 4 It is a schematic diagram of the current distribution of all power data in the subnode when the current changes in a normal cycle; Figure 5 It is a schematic diagram of abnormal periodic changes in current data; Figure 6 It is a schematic diagram of current distribution of all power data in the subnode when the current changes abnormally periodically. DETAILED DESCRIPTION

[0024] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0025] It should be understood that when the terms "first", "second", etc. are used in the claims, descriptions, and drawings of the present invention, they are only used to distinguish different objects, rather than to describe a specific order. The terms "include" and "comprise" used in the description and claims of the present invention indicate the presence of the described features, wholes, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their collections.

[0026] The embodiment of the present invention discloses a method for managing power equipment assets, referring to Figure 1 , comprising steps S1 to S5: S1: Collect power data of power equipment.

[0027] For each power device, power data of the power device is collected, and the power data includes multiple dimensions, such as current, voltage, temperature, etc. In the embodiment of the present invention, the collection interval for power data is once per second, and the implementer can set the collection interval according to the actual implementation situation.

[0028] At this point, the collection of power data from power equipment has been achieved.

[0029] S2: Construct an isolation forest based on the historical power data of power equipment.

[0030] Step S2 includes steps S21 to S25, specifically: S21: The power data of the power equipment over a period of time in history is organized into a sequence as a root node.

[0031] It should be noted that the implementers can set the length of the historical period according to the actual implementation situation, for example, the power data of the power equipment in the past day.

[0032] S22: Determine the selection probability of each dimension under the root node according to the distribution of each dimension of the power data in the root node, and randomly select a dimension with unequal probability according to the selection probability of each dimension under the root node as the segmentation dimension of the root node. Use the segmentation dimension of the root node to divide all the power data in the root node to obtain two child nodes.

[0033] It should be noted that when the traditional isolation forest algorithm divides the node, it randomly selects a dimension from all dimensions with medium probability for division. The data difference of the power data may be very small in a certain dimension, and the data difference in a certain dimension is large. The anomaly of the power data is more likely to be reflected in the dimension with large data difference. The equal probability of randomly selecting dimensions for division may cause the dimension containing the anomaly to be selected later, affecting the iteration efficiency of the isolation forest, and may also cause the abnormal power data to be included in the leaf nodes of the deeper layer of the isolation tree. Because the isolation forest algorithm usually calculates the abnormal score of the data point according to the layer depth of the leaf node where the data point is located, the shallower the layer depth, the larger the abnormal score. When the abnormal power data is included in the leaf node of the deeper layer of the isolation tree, the corresponding abnormal score is small, which may cause the abnormal power data to be unable to be identified. Therefore, the present invention calculates the selection probability of each dimension when the node is split according to the distribution of each dimension of the power data in the node, so as to randomly select a dimension for division according to the selection probability with unequal probability, thereby enhancing the probability of the dimension that may contain the anomaly being selected.

[0034] Specifically, determining the selection probability of each dimension under the root node according to the distribution of each dimension of the power data in the root node includes: ; in, The root node The probability of selection of each dimension; All power data in the root node The maximum value of the dimension; All power data in the root node The minimum value of the dimension; is the number of dimensions of power data; All power data in the root node Information entropy of dimensions; All power data in the root node The number of different values ​​in each dimension; is the maximum value function, This is to prevent the denominator from being 0. This is to prevent the index from being 0.

[0035] when When the value is larger, all power data in the root node The fluctuation range of the data in each dimension is relatively large. Used for Normalize. The larger the fluctuation range of the data in the first dimension, the The dimension is more likely to contain anomalies, then The greater the probability of selecting a dimension.

[0036] Under normal circumstances, some dimensions of power data may change periodically. For example, when the power equipment is AC, the current and voltage change periodically. Figure 2 This is a schematic diagram of the periodic change of current. At this time, the fluctuation of the data of this dimension within its fluctuation range may be normal fluctuation. When the frequencies of different values ​​appearing in the fluctuation range of this dimension are basically the same, the information entropy of this dimension is large, and the data of this dimension is more likely to be periodic changes, and the necessity of selecting this dimension for division is small; when the values ​​appearing in the fluctuation range of this dimension are mostly concentrated in one place, the information entropy of this dimension is small, and the values ​​not concentrated in this dimension are more likely to be abnormal. Figure 3 This is a schematic diagram of temperature anomaly distribution. At this time, it is necessary to select this dimension for division. Therefore, the present invention divides all the power data in the root node into Information entropy of dimensions right When all power data in the root node When the frequencies of different values ​​in the dimensions are the same, the frequencies of each value are At this time, all the power data in the root node The information entropy of the dimension reaches the maximum, Therefore, the present invention utilizes Information Entropy Normalize. When the information entropy The more hours, The smaller it is, the The greater the increase, the higher the The selection probability of each dimension is The bigger it is, The larger it is, and the closer it is to 1, the The smaller the increase.

[0037] The method uses the segmentation dimension of the root node to divide all the power data in the root node to obtain two child nodes, including: The maximum and minimum values ​​of the segmentation dimensions of all power data in the root node are obtained, and a value between the maximum and minimum values ​​is randomly selected as the segmentation threshold of the root node. All power data in the root node whose segmentation dimension is greater than the segmentation threshold are divided into one category as a child node of the root node, and all power data in the root node whose segmentation dimension is not greater than the segmentation threshold are divided into one category as another child node of the root node.

[0038] S23: For any child node, the necessity of segmentation of the child node is calculated according to the temporal regularity of the power data in the child node and the selection probability of each dimension under the child node. In response to the necessity of segmentation of the child node being greater than a preset segmentation threshold, a dimension is selected with unequal probability according to the selection probability of each dimension under the child node as the segmentation dimension of the child node, and all power data in the child node are divided using the segmentation dimension of the child node to obtain two child nodes.

[0039] Specifically, the method for acquiring the temporal regularity of the power data in the subnode is: All power data in the child node are divided into multiple power data segments according to the time sequence. All power data in each power data segment are continuous in time sequence, and different power data segments are not continuous in time sequence. For any two power data segments, the time difference between the last power data in the previous power data segment and the first power data in the next power data segment is taken as the time interval between the two power data segments. For example, when the child node contains power data at the 1st, 2nd, 3rd, 8th, 9th, 10th, 15th, 16th, and 17th moments, the power data at the 1st, 2nd, and 3rd moments are continuous in time sequence, then the power data at the 1st, 2nd, and 3rd moments constitute one power data segment. Similarly, the power data at the 8th, 9th, and 10th moments constitute one power data segment, and the power data at the 15th, 16th, and 17th moments constitute one power data segment. The time interval between the power data segment composed of the power data at the 1st, 2nd, and 3rd moments and the power data segment composed of the power data at the 15th, 16th, and 17th moments is 12.

[0040] The temporal regularity of the power data in the subnode is determined based on the length of the power data segment contained in the subnode and the time interval between adjacent power data segments: ; in, is the temporal regularity of the power data in the child nodes; For the child node The length of the power data segment; is the number of power data segments in the child node; is the average length of all power data segments in the child node; Indicates the child node The power data segment and The time interval between power data segments.

[0041] is the average difference between the length of each power data segment in the child node and the mean length of all power data segments, is the average difference between the time interval between the power data segments in the child node and the average time interval. and The smaller the value is, the stronger the temporal regularity of the power data in the child node is. The power data of the segmentation dimension corresponding to the parent node of the child node is more likely to change periodically. The power data in the child node is regularly distributed on the segmentation dimension corresponding to the parent node of the child node. For example, the segmentation dimension corresponding to the parent node of the child node is current, and there is no abnormality in the current. The schematic diagram of its periodic change can be seen in Figure 2 When the current distribution diagram of all power data in the child node is Figure 4 .when or When the value is larger, the temporal regularity of the power data in the child node is weaker, and the power data in the child node is irregularly distributed on the segmentation dimension corresponding to the parent node of the child node. For example, the segmentation dimension corresponding to the parent node of the child node is current, and there are anomalies in the current data. For a schematic diagram of its periodic change, see Figure 5 When the current distribution diagram of all power data in the child node is Figure 6 .

[0042] The selection probability of each dimension under the child node is calculated in the same manner as the selection probability of each dimension under the root node in S22, and will not be repeated here.

[0043] It should be noted that power data is time series data, and abnormal power data usually appears continuously. Therefore, the present invention determines the necessity of splitting the sub-nodes based on the time series regularity of the power data in the sub-nodes, the selection probability of each dimension under the sub-nodes, and the number of power data contained in the sub-nodes, so as to split the sub-nodes according to the necessity of splitting, so as to avoid normal power data with periodic regularity or abnormal power data with continuous time series distribution from being split into multiple sub-nodes, which affects the accuracy of abnormal power data identification.

[0044] Specifically, the necessity of segmenting the subnode is calculated according to the temporal regularity of the power data in the subnode, the selection probability of each dimension under the subnode, and the amount of power data contained in the subnode, including: ; in, The necessity of splitting the child nodes; is the temporal regularity of the power data in the child nodes; For the child node The probability of selection of each dimension; is the number of dimensions of power data; is the number of power data contained in the child node; is the number of all power data; is the natural exponential function; is the maximum value function.

[0045] The stronger the temporal regularity of the power data in the subnode, the smaller the possibility that the power data in the subnode contains anomalies, and the smaller the necessity to split the subnode at this time; when the selection probability of each dimension under the subnode is small, it means that the fluctuation range of each dimension of the power data of the subnode is small, and the data of each dimension is regularly distributed, and the smaller the necessity to split the subnode at this time; when the subnode contains less power data, the smaller the necessity to split the subnode; conversely, when the temporal regularity of the power data in the subnode is weaker, or the selection probability of a dimension under the subnode is larger, or the subnode contains more power data, the greater the necessity to split the subnode.

[0046] The segmentation threshold is set by the implementer according to the actual implementation situation. Since the value range of segmentation necessity is [0, 1], the segmentation threshold to be set must also be between [0, 1], for example, the segmentation threshold is 0.5.

[0047] The method uses the segmentation dimension of the sub-node to divide all the power data in the sub-node to obtain two sub-nodes, including: Get the maximum and minimum values ​​of the segmentation dimension of all power data in the child node, and randomly select a value between the maximum and minimum values ​​as the segmentation threshold of the child node. Divide all power data in the child node whose segmentation dimension is greater than the segmentation threshold into one category as a child node of the child node, and divide all power data in the child node whose segmentation dimension is not greater than the segmentation threshold into one category as another child node of the child node.

[0048] S24: Repeat S23 until the segmentation necessity of all child nodes is no greater than a preset segmentation threshold, and then stop the iteration to achieve the construction of an isolated tree.

[0049] S25: Repeat S21-S24 to construct multiple isolated trees to form an isolation forest.

[0050] Specifically, set the number of repetitions , repeat steps S21-S24 times, achieve The construction of an isolated tree. Since each time an isolated tree is constructed, the segmentation dimension of each node is randomly selected according to the selection probability of each dimension under each node, and the segmentation threshold of each node is also randomly selected, the constructed Each isolated tree is different from the others. Isolated trees form an isolation forest. It is set by the implementers according to the actual implementation situation and is not limited to specific ones. For example .

[0051] At this point, the construction of the isolation forest has been achieved.

[0052] S3: Add the newly generated power data to the isolation forest and update the isolation forest.

[0053] It should be noted that the power data of the power equipment is streaming data, which is generated in real time. In the traditional isolation forest, to realize the anomaly detection of the newly generated power data, it is necessary to rebuild the isolation forest based on the newly generated power data and the historical power data. In order to realize the real-time anomaly detection of the power equipment, each time a new power data is generated, it is necessary to rebuild the isolation forest, which is inefficient and occupies a lot of computing resources. Therefore, the present invention updates the isolation forest based on the newly generated power data on the basis of the existing isolation forest whenever a new power data is generated, thereby saving computing resources and improving the efficiency of anomaly detection.

[0054] Step S3 includes steps S21 to S35, specifically: S31: For the newly generated power data, according to the segmentation dimension and segmentation threshold of each node in the isolated tree, the leaf node to which the power data belongs in the isolated tree is obtained, and the power data is added to the corresponding leaf node.

[0055] Specifically, for the currently newly generated power data, according to the segmentation dimension and segmentation threshold of each node in the isolated tree, obtaining the leaf node to which the power data belongs in the isolated tree includes: Determine whether the value of the newly generated power data under the segmentation dimension of the root node of the isolation tree is greater than the segmentation threshold of the root node. When it is greater than the segmentation threshold of the root node, use the child nodes corresponding to all power data in the root node whose segmentation dimension is greater than the segmentation threshold as the first target child nodes of the newly generated power data. Otherwise, use the child nodes corresponding to all power data in the root node whose segmentation dimension is not greater than the segmentation threshold as the first target child nodes of the newly generated power data. When the first target child node is not a leaf node, determine whether the value of the newly generated power data under the segmentation dimension of the first target child node is greater than the segmentation threshold of the first target child node. When it is greater than the segmentation threshold of the first target child node, use the child nodes corresponding to all power data in the first target child node whose segmentation dimension is greater than the segmentation threshold as the second target child nodes of the newly generated power data. Otherwise, use the first target child node. The subnodes corresponding to all the power data whose split dimension is not greater than the split threshold in the target subnode are taken as the second target subnodes of the newly generated power data; when the second target subnode is not a leaf node, it is determined whether the value of the newly generated power data under the split dimension of the second target subnode is greater than the split threshold of the second target subnode. When it is greater than the split threshold of the second target subnode, the subnodes corresponding to all the power data whose split dimension is greater than the split threshold in the second target subnode are taken as the third target subnodes of the newly generated power data; otherwise, the subnodes corresponding to all the power data whose split dimension is not greater than the split threshold in the second target subnode are taken as the third target subnodes of the newly generated power data; and so on, the iteration is stopped until the target subnode is a leaf node, and the target subnode is the leaf node to which the newly generated power data belongs in the isolated tree.

[0056] S32: Using the method in step S23, obtain the necessity of segmentation of the corresponding leaf node after adding the currently newly generated power data. In response to the necessity of segmentation of the leaf node being greater than the preset segmentation threshold, select a dimension with unequal probability according to the selection probabilities of each dimension under the leaf node as the segmentation dimension of the leaf node, and use the segmentation dimension of the leaf node to divide all power data in the leaf node to obtain two child nodes.

[0057] S33: For any newly obtained child node, the necessity of segmentation of the child node is calculated according to the temporal regularity of the power data in the child node and the selection probability of each dimension under the child node. In response to the necessity of segmentation of the child node being greater than a preset segmentation threshold, a dimension is selected with unequal probability according to the selection probability of each dimension under the child node as the segmentation dimension of the child node, and all power data in the child node are divided using the segmentation dimension of the child node to obtain two child nodes.

[0058] S34: Repeat S33 until the segmentation necessity of all newly obtained child nodes is no greater than the preset segmentation threshold and then stop the iteration to implement the update of the isolated tree.

[0059] S35: Execute steps S31 to S34 for each isolated tree in the isolation forest to update each isolated tree in the isolation forest, thereby implementing the update of the isolation forest.

[0060] At this point, the update of the isolated forest has been achieved.

[0061] S4: Determine the abnormality score of the currently newly generated power data, and identify the current state of the power equipment according to the abnormality score.

[0062] It should be noted that the newly generated power data can reflect the current state of the power equipment. Therefore, the present invention determines the abnormality score of the newly generated power data based on the leaf nodes in each isolated tree of the newly generated power data in the isolated forest, thereby identifying the current state of the power equipment.

[0063] Specifically, the abnormal score of the newly generated power data is determined according to the layer depth of the leaf node where the newly generated power data is located in each isolated tree of the isolated forest, the amount of power data contained in the leaf node, and the number of power data segments contained in the leaf node: ; in, Indicates the abnormal score of the currently newly generated power data; Indicates the newly generated power data in the The depth of the leaf node in an isolated tree; Indicates The height of an isolated tree; Indicates the newly generated power data in the The number of power data contained in the leaf nodes of each isolated tree; Indicates The mean value of the number of power data contained in all leaf nodes in an isolated tree; Indicates the newly generated power data in the The number of power data segments in the leaf nodes of an isolated tree; Indicates The average number of power data segments in all leaf nodes in an isolated tree.

[0064] when The smaller the time, the more the newly generated power data is. The earlier it is isolated in the isolated tree, the greater the difference between the power data contained in the leaf node where the current newly generated power data is located and the rest of the power data, the more abnormal the current newly generated power data is. The smaller the value, the greater the difference between the newly generated power data and the rest of the power data, and the more abnormal the newly generated power data is. The smaller it is, the more continuous the power data contained in the leaf node where the newly generated power data is located is in terms of time series, which means that the difference between the power data in the recent period and the previous power data is greater, and the power equipment is more likely to be abnormal. , , The smaller the time, the more newly generated power data is. The present invention integrates the anomaly scores of the currently newly generated power data in all isolated trees of the isolation forest to obtain the anomaly score of the currently newly generated power data.

[0065] In response to the abnormal score of the newly generated power data being greater than the abnormal threshold, the power equipment is considered to be abnormal, otherwise, the power equipment is considered to be in normal state. The abnormal threshold is set by the implementer according to the actual implementation situation. Since the value range of the abnormal score is [0,1], the set abnormal threshold must also be between [0,1]. The empirical value of the abnormal threshold is 0.9.

[0066] S5: Synchronize the current status of the power equipment to the digital twin model of the power equipment in real time to realize power equipment asset monitoring and management.

[0067] Build a digital twin model of power equipment to achieve 3D visualization of all power equipment.

[0068] The current status of each power equipment is synchronized to the digital twin model of the power equipment in real time. Abnormal power equipment is highlighted in the digital twin model to remind staff to repair the abnormal power equipment, thereby realizing power equipment asset monitoring and management.

[0069] An embodiment of the present invention further discloses an electric power equipment asset management system, including a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an electric power equipment asset management method according to the present invention is implemented.

[0070] The above system also includes other components well known to those skilled in the art, such as a communication bus and a communication interface, and their configuration and functions are known in the art, so they will not be described in detail here.

[0071] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium may be any suitable magnetic storage medium or magneto-optical storage medium, or any other medium that can be used to store the required information and can be accessed by an application, module, or both. Any such computer storage medium may be part of the device or accessible or connectable to the device.

[0072] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

[0073] The above are all preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Therefore, any equivalent changes made based on the structure, shape, and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for managing power equipment assets, characterized in that: include: Construct an isolation forest based on the historical power data of power equipment; In the process of constructing the isolation forest: for each child node in the isolation tree, the selection probability of each dimension under the child node is determined according to the distribution of the power data in the child node in each dimension, and the selection probability is positively correlated with the fluctuation range of the power data in the child node in each dimension, and negatively correlated with the information entropy of the power data in the child node in each dimension; the necessity of segmentation of the child node is determined, and the necessity of segmentation is negatively correlated with the temporal regularity of the power data in the child node, and positively correlated with the maximum value of the selection probability of each dimension under the child node and the amount of power data contained in the child node; in response to the necessity of segmentation being greater than the preset segmentation threshold, a dimension is randomly selected with unequal probability according to the selection probability of each dimension under the child node as the segmentation dimension of the child node, and the power data in the child node is divided into two new child nodes using the segmentation dimension; Add the newly generated power data to the isolated forest and update the isolated forest; determine the anomaly score of the newly generated power data according to the position of the newly generated power data in each isolated tree in the isolated forest, and identify the current state of the power equipment according to the size of the anomaly score; The current status of power equipment is synchronized to the digital twin model of power equipment in real time to realize power equipment asset monitoring and management.

2. A method for managing power equipment assets according to claim 1, characterized in that: The selection probability of each dimension under the sub-node satisfies the expression: ； in, For the child node The probability of selection of each dimension; For all power data in the child node The maximum value of the dimension; For all power data in the child node The minimum value of the dimension; is the number of dimensions of power data; For all power data in the child node Information entropy of dimensions; For all power data in the child node The number of different values ​​in each dimension; is the maximum value function.

3. A method for managing power equipment assets according to claim 1, characterized in that: The temporal regularity of the power data in the subnode satisfies the expression: ； in, is the temporal regularity of the power data in the child nodes; For the child node The length of the power data segment; is the number of power data segments in the child node; is the average length of all power data segments in the child node; Indicates the child node The power data segment and The time interval between power data segments.

4. A method for managing power equipment assets according to claim 1, characterized in that: The necessity of splitting the sub-nodes satisfies the expression: ； in, Necessity of splitting the child nodes; is the temporal regularity of the power data in the child nodes; For the child node The probability of selection of each dimension; is the number of dimensions of power data; is the number of power data contained in the child node; is the number of all power data; is the natural exponential function; is the maximum value function.

5. A method for managing power equipment assets according to claim 1, characterized in that: The method of adding the currently newly generated power data to the isolation forest and updating the isolation forest includes: For the newly generated power data, determine the leaf node to which the power data belongs in each isolated tree, and add the power data to the corresponding leaf node; obtain the necessity of segmentation of the corresponding leaf node after adding the power data; in response to the necessity of segmentation of the leaf node being greater than the preset segmentation threshold, select a dimension with unequal probability according to the selection probability of each dimension under the leaf node as the segmentation dimension of the leaf node, and divide the power data in the leaf node into two new child nodes by using the segmentation dimension; repeat the segmentation judgment on the new child nodes until the segmentation necessity of all new child nodes is not greater than the preset segmentation threshold, and stop the iteration to achieve the update of the isolated forest.

6. A method for managing power equipment assets according to claim 1, characterized in that: The abnormal score of the currently newly generated power data satisfies the expression: ； in, Indicates the abnormal score of the currently newly generated power data; Indicates that the newly generated power data is in The depth of the leaf node in an isolated tree; Indicates The height of an isolated tree; Indicates that the newly generated power data is in The number of power data contained in the leaf nodes of each isolated tree; Indicates The mean value of the number of power data contained in all leaf nodes in an isolated tree; Indicates that the newly generated power data is in The number of power data segments in the leaf nodes of an isolated tree; Indicates The average number of power data segments in all leaf nodes in an isolated tree.

7. A method for managing power equipment assets according to claim 3 or 6, characterized in that: The method for obtaining the power data segment is: All power data in the child node are divided into multiple power data segments according to time sequence, all power data in each power data segment are continuous in time sequence, and different power data segments are not continuous in time sequence.

8. A method for managing power equipment assets according to any one of claim 1, characterized in that: The identifying the current state of the electric equipment according to the abnormality score includes: In response to the abnormality score of the currently newly generated power data being greater than the abnormality threshold, the power device state is abnormal; otherwise, the power device state is normal.

9. An electric power equipment asset management system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, an electric power equipment asset management method according to any one of claims 1-8 is implemented.

Images