A method, system, and storage medium for power system fault data analysis

Abstract

This application relates to the field of fault tree analysis technology in power CPS (Continuous Power System Configuration) technology, specifically to a method, system, and storage medium for power system fault data analysis. The method includes: statistically analyzing all faulty nodes in the power CPS system as suspected faulty nodes to construct a fuzzy fault tree; obtaining the fault deviation value of each suspected faulty node; obtaining the forged evaluation value of each suspected faulty node; obtaining the real faulty nodes and interfering nodes; obtaining the average consistency of each real faulty node; obtaining the adaptive fuzzification coefficient of each node; constructing a fault tree based on the adaptive fuzzification coefficient, obtaining the importance of each node, and performing fault tree analysis on the power CPS according to the order of node importance. This application improves the accuracy and rationality of fuzzy fault trees.

Description

Technical Field

[0001] This application relates to the field of fault tree analysis technology in power CPS, specifically to a power system fault data analysis method, system, and storage medium. Background Technology

[0002] A power CPS (Cyber-Physical System for Power) is a power system that integrates power and information systems, and is a concept within the current smart grid. Power CPS deeply integrates computing, communication, and control systems. It collects operational data from various power nodes through sensors and uses the data transmitted from these nodes to control them. Typically, a power CPS system encompasses the operational status of all power and information nodes, making it a relatively complex power cyber-physical system, thus requiring reliability analysis.

[0003] Fault trees are a common method for analyzing the reliability of complex systems and can be applied to reliability analysis in power CPS (Cybersecurity System). However, traditional fault trees heavily rely on a large amount of precise operational data from each node, which is difficult to obtain. Therefore, fuzzy mathematics is introduced to construct fuzzy fault trees. Fuzzy fault trees use fuzzy subsets to quantify uncertainties in nodes. In power CPS systems, the levels of each device node are different, resulting in varying tolerances to anomalies. They are also susceptible to interference from data misidentified as faulty nodes due to network attacks. Traditional fuzzy fault trees use fixed-range fuzzy subsets for quantification, failing to reflect the correlation between device nodes and their level differences in fault tolerance. Consequently, fuzzy fault trees cannot accurately handle uncertainties at nodes, leading to biases in the stability evaluation of power CPS systems. Summary of the Invention

[0004] To address the aforementioned technical problems, the purpose of this application is to provide a method, system, and storage medium for power system fault data analysis. The specific technical solution adopted is as follows:

[0005] In a first aspect, embodiments of this application provide a method for analyzing power system fault data, the method comprising the following steps:

[0006] All faulty nodes in the power CPS system are used as initial fault nodes. Actual operating data and rated operating data of all nodes and the levels of each node are collected to construct a fuzzy fault tree. The operating data includes voltage, current and impedance, and the levels are transmission voltage levels.

[0007] The fault deviation value of each node is obtained based on the difference between the actual operating data and the rated operating data;

[0008] The falsified evaluation value of each initial fault node is obtained based on the discrete amplitude of the fault deviation value;

[0009] The real fault nodes are obtained by combining the falsified evaluation values ​​with a threshold segmentation algorithm;

[0010] The average consistency of each real fault node is obtained based on the degree of similarity between the actual operating data and the rated operating data of any two real fault nodes.

[0011] Based on the level of the real fault nodes and the dispersion of the average consistency, the adaptive fuzzification coefficients of each real fault node and interference node are obtained.

[0012] A fault tree is constructed based on the adaptive fuzzification coefficients, the importance of each node is obtained, and fault tree analysis is performed on the power CPS according to the order of node importance.

[0013] Furthermore, the method for obtaining the fault deviation value is as follows:

[0014] For each node, calculate the absolute value of the difference between the actual operating data and the rated operating data of the same category of the node, and use the sum of all such absolute values ​​of difference as the fault deviation value of each node.

[0015] Furthermore, the method for obtaining the forged evaluation value is as follows:

[0016] For each initial fault node, calculate the range of the fault deviation values ​​of the initial fault node and all nodes directly connected to the initial fault node, and calculate the normalized value of the ratio of the fault deviation value of the initial fault node to the range as the fake evaluation value of each initial fault node.

[0017] Furthermore, the method for obtaining the actual fault node is as follows:

[0018] The algorithm takes the fake evaluation values ​​of all initial fault nodes as input to the threshold segmentation algorithm, outputs the optimal segmentation threshold, and takes the initial fault nodes whose fake evaluation values ​​are greater than the optimal segmentation threshold as real fault nodes, and takes the other nodes as interference nodes.

[0019] Furthermore, the method for obtaining the average consistency is as follows:

[0020] The absolute value of the difference between the voltage in the actual operating data and the voltage in the rated operating data of each real fault node is calculated as the voltage operating difference of each real fault node. Using the same method as the voltage operating difference, the current operating difference and impedance operating difference of each real fault node are obtained. The vector composed of the voltage operating difference, current operating difference and impedance operating difference is used as the fault vector of each real fault node.

[0021] The formula for calculating the average consistency is: In the formula, This represents the average consistency of the i-th real fault node; This represents the number of nodes at the same level as the current real faulty node, and cos() represents the cosine similarity. and Let represent the fault vectors of the i-th and j-th real fault nodes, respectively.

[0022] Furthermore, the method for obtaining the adaptive fuzzification coefficient is as follows:

[0023] Sort all real fault nodes in ascending order of their severity level, and use the sorting number of each severity level as the sorting number of that severity level.

[0024] The formula for calculating the adaptive fuzzification coefficient of each real fault node is as follows: In the formula, K represents the adaptive fuzzification coefficient of each real fault node; round() represents the rounding function. This indicates the preset scaling factor. This indicates the sorting number representing the severity level of each actual faulty node. Represents the normalization function. This represents the standard deviation of the average consistency among all real fault nodes that have the same level as each real fault node.

[0025] For interfering nodes, the value 1 is used as the adaptive fuzzification coefficient for each interfering node.

[0026] Furthermore, the step of constructing a fault tree based on adaptive fuzzification coefficients and obtaining the importance of each node includes:

[0027] For each node, the product of the adaptive fuzzification coefficient of each node and the range of the initial fuzzification subset is calculated as the range of the adjusted fuzzification subset.

[0028] The mean values ​​of voltage operating difference, current operating difference, and impedance operating difference of the calculated nodes are used to obtain the mapping relationship between the operating data of each node and the fuzzy subset by using chaotic mapping and the range of the adjusted fuzzy subset. The element values ​​corresponding to the actual operating data of each node in the range of the adjusted fuzzy subset are output. The element values ​​corresponding to each node are input into the membership function to obtain multiple fuzzy probability values ​​of each node. The multiple fuzzy probability values ​​of each node are used as input for fuzzy probability analysis and the failure probability of each node is output.

[0029] Based on the probability of failure in the running data of all nodes, a fault tree is constructed, and the importance of each node in the fault tree is calculated using a conditional probability model.

[0030] Secondly, this application also provides a power system fault data analysis storage medium, wherein the power CPS fault tree analysis storage medium has a computer program, and the computer program, when executed by a processor, implements any of the power system fault data analysis methods described above.

[0031] Thirdly, embodiments of this application also provide a power system fault data analysis system, including:

[0032] The data acquisition module is used to statistically analyze all faulty nodes in the power CPS system as initial fault nodes, collect the actual operating data and rated operating data of each node, and the level of each node to construct a fuzzy fault tree; the operating data includes voltage, current, and impedance, and the level is the transmission voltage level;

[0033] The adaptive fuzzy coefficient acquisition module is used to obtain the fault deviation value of each initial fault node based on the difference between the actual operating data and the rated operating data; obtain the fake evaluation value of each initial fault node based on the fault deviation value; obtain the real fault node and interference node based on the fake evaluation value and the threshold segmentation algorithm; obtain the average consistency of each real fault node based on the similarity between the difference between the actual operating data and the rated operating data of any two real fault nodes; and obtain the adaptive fuzzification coefficient of each node based on the level and average consistency dispersion of the real fault node.

[0034] The fault tree analysis module is used to construct a fault tree based on adaptive fuzzification coefficients, obtain the importance of each node, and perform fault tree analysis on the power CPS according to the order of node importance.

[0035] This application has at least the following beneficial effects:

[0036] This application obtains historical fault and fault operation data for each node in a power CPS system. Based on the differences between the fault operation data and the rated operation data of faulty nodes, and the deviations of directly connected nodes, a falsified assessment value is obtained. This reflects the possibility that some nodes are misjudged as faulty due to data falsification, eliminating data interference from misjudged nodes. An average consistency is obtained based on the similarity of fault operation data of nodes at the same level, reflecting the differences between individual nodes and other faulty nodes, indicating the degree of influence of uncertain factors on faulty nodes. Finally, an adaptive fuzzification coefficient is obtained based on the node level. The probability of occurrence at each time point is obtained based on the fuzzy mapping to construct a fault tree. Furthermore, the probability of occurrence of the top event and the importance of basic events are obtained based on qualitative analysis. Compared with traditional fuzzy fault trees, this scheme analyzes the possibility of network attacks in the power CPS system causing some nodes to be misjudged as faulty, eliminating the influence of misjudged node data on the fault tree. Simultaneously, an adaptive fuzzy subset is obtained based on the influence of uncertain factors on each device node in a single basic event, improving the accuracy of the fuzzification probability affected by uncertain factors during fuzzy mapping, ultimately improving the accuracy and rationality of the fuzzy fault tree. Attached Figure Description

[0037] To more clearly illustrate the technical solutions and advantages in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 A flowchart illustrating the steps of a power system fault data analysis method provided in one embodiment of this application;

[0039] Figure 2 This is a schematic diagram of a power CPS system provided in one embodiment of this application. Detailed Implementation

[0040] To further illustrate the technical means and effects adopted by this application to achieve the intended purpose of the invention, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a power system fault data analysis method, system, and storage medium proposed in this application. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0041] 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 this application pertains.

[0042] The following description, in conjunction with the accompanying drawings, details the specific scheme of the power system fault data analysis method, system, and storage medium provided in this application.

[0043] Please see Figure 1 The diagram illustrates a flowchart of a power system fault data analysis method according to an embodiment of this application. The method includes the following steps:

[0044] Step S1: Collect all faulty nodes in the power CPS system as initial faulty nodes, collect the actual operating data and rated operating data of all nodes and the levels of each node, and construct a fuzzy fault tree.

[0045] In this application, the schematic diagram of the power CPS system is as follows: Figure 2 As shown, it mainly consists of three layers. The bottom layer is the equipment physical layer, which mainly targets the various physical nodes of the power system, namely the nodes that realize the production, transmission and consumption of electricity. The middle layer corresponds to the communication layer. The information nodes of the communication layer correspond to the power nodes, realize the collection of the operating data of each node and upload it to the decision layer. The top layer is the decision layer, which mainly realizes the rational allocation and scheduling of electricity based on the operating data of the power system nodes, and sends the final decision to the equipment physical nodes, thereby realizing the regulation and control of the entire power system.

[0046] This scheme takes a city's power CPS system as the research object. Based on the power CPS system log data, it identifies faulty nodes in the power CPS system as initial fault nodes, and obtains the actual operating data and rated operating data of all nodes. Specifically, the actual operating data and rated operating data are voltage, current, and impedance. Simultaneously, it obtains the current node's voltage level, where the level refers to the voltage level transmitted through that node, such as 110kV, 380V, 220V, etc. Each node corresponds to a fuzzy subset, which measures the deviation of the node's real-time operating data from its normal operating data and its membership in the fault occurrence. The initial fuzzy subset for each node ranges from [0,1].

[0047] Step S2: Obtain the fault deviation value of each node based on the difference between the actual operating data and the rated operating data; obtain the fake evaluation value of each initial fault node based on the discrete amplitude of the fault deviation value; and obtain the real fault node based on the fake evaluation value combined with the threshold segmentation algorithm.

[0048] In a power CPS system, physical device nodes and information nodes are interdependent and can be considered as a whole. Therefore, when a node fails, it may be due to a genuine equipment malfunction causing node anomalies and data fluctuations, or it may be due to a network attack that generates false and fabricated abnormal operating data, causing the power CPS system to mistakenly identify the node as a faulty node during dynamic adjustments.

[0049] The node that has failed is taken as the initial fault node. There is a possibility that the initial fault node may be misjudged as a fault node. For each initial fault node, whether it is an operational failure or a misjudged failure, the main judgment basis is that the current node's operating data deviates significantly from the data during normal rated operation. The larger the deviation, the higher the probability of the node's failure, and the smaller the deviation, the lower the probability of failure.

[0050] Based on the above analysis, and considering the deviation between the fault operation data and the rated operation data at the initial fault node, the fault deviation value is obtained as follows: In the formula, A represents the fault deviation value of each node, and N represents the number of types of collected operational data. In this embodiment, N is set to 3. This represents the absolute value of the difference between the operating data of the i-th node and the corresponding rated operating data, where the rated operating data is obtained through the power CPS system, which will not be described in detail in this embodiment.

[0051] For an initially faulty node, if the deviation between the node's faulty operating data and its rated operating data is greater when it is determined to be faulty, then... The larger the value, the larger the fault deviation value, indicating that the current node is more likely to actually fail.

[0052] However, when a node is attacked by the network, the fault data collected by the node is fake data forged by the attacker. Therefore, evaluation based solely on fault operation data will result in significant deviations, requiring further judgment.

[0053] For a genuine faulty node, when a node fails, it affects the operation of connected nodes, potentially causing the load of the current node to be randomly and passively distributed to a neighboring node—this is a passive adjustment. However, for false faults caused by network attacks, the current equipment is operating normally but is mistakenly identified as faulty. In this case, the power CPS system actively adjusts, ensuring that the load of the current node is evenly distributed to its neighboring nodes to guarantee the normal operation of the power grid—this is an active adjustment.

[0054] Based on the above analysis, forged evaluation values ​​are obtained based on the operating status of nodes adjacent to each initial fault node: In the formula, B represents the spoofed evaluation value of each initial fault node, norm() represents the normalization function, and A represents the fault deviation value of each initial fault node. This represents the range of fault deviation values ​​for each initial fault node and all nodes directly connected to it.

[0055] For spoofed nodes, because the system actively adjusts, the load of each initially faulty node is evenly distributed among all adjacent nodes, resulting in a smaller range of fault deviation values ​​between adjacent nodes and a larger spoofed evaluation value. Conversely, if the initially faulty node is a real faulty node, the load of the current node will be randomly distributed among connected nodes due to the node's failure, resulting in a larger range of fault deviation values ​​between adjacent nodes and a smaller spoofed evaluation value.

[0056] The operational data of a node that is mistakenly identified as a real fault due to a cyberattack is forged and false data. Therefore, the data of this node cannot be used as a reference when generating the fault tree, which can easily affect the stability assessment of the fault tree.

[0057] Based on the above analysis, the forged evaluation values ​​of all initial fault nodes are used as input to the Otsu thresholding method, which outputs the optimal segmentation threshold. Initial fault nodes with forged evaluation values ​​less than the segmentation threshold are marked as real fault nodes, while other nodes are designated as interference nodes. The Otsu thresholding method is a well-known technique and will not be elaborated upon in this embodiment.

[0058] Step S3: Based on the similarity of the differences between the actual operating data and the rated operating data of any two real fault nodes, obtain the average consistency of each real fault node; based on the level of the real fault node and the dispersion of the average consistency, obtain the adaptive fuzzification coefficient of each real fault node and the interference node.

[0059] Step S2 utilizes forged evaluation values ​​to exclude some nodes that are mistakenly identified as faulty due to network attacks. Based on the data distribution of truly faulty nodes, a fuzzy fault tree reflecting the actual state of the nodes is constructed. However, during fault tree construction, due to the different levels and fault tolerance of each device node, it is impossible to determine the probability of node failure using accurate fault thresholds. Therefore, it is necessary to combine fuzzy logic theory to further analyze anomalies and construct the fault tree.

[0060] For nodes of the same level, theoretically, when a device malfunctions, the deviation between its faulty operating data and the rated data should be relatively close. In this case, the fault boundary should be relatively clear, and a small fuzzy subset can be used for measurement. However, during device operation, inherent differences exist, causing the original deviation to become disordered, making it impossible to determine the fault using an accurate threshold. In this case, a larger fuzzy subset is needed for measurement.

[0061] Based on the above analysis, the absolute value of the difference between the voltage in the actual operating data and the voltage in the rated operating data of each real fault node is calculated as the voltage operating difference of each real fault node. Using the same method as the voltage operating difference, the current operating difference and impedance operating difference of each real fault node are obtained. The vector composed of the voltage operating difference, current operating difference and impedance operating difference is used as the fault vector of each real fault node.

[0062] Based on the above analysis, and considering the similarity of fault vectors with nodes of the same level as the faulty node, the average consistency of each real faulty node is obtained: In the formula, This represents the average consistency of the i-th real fault node; This represents the number of nodes at the same level as the current real faulty node, and cos() represents the cosine similarity. and Let represent the fault vectors of the i-th and j-th real fault nodes, respectively.

[0063] If the actual fault node is largely consistent with other device nodes of the same level, the probability of the current node failing is higher. Conversely, if the actual fault node deviates significantly from other device nodes of the same level, resulting in low similarity of fault vectors, the current node may be located at a faulty boundary, and the probability of failure is more uncertain.

[0064] Only nodes of the same level are comparable, and higher levels generally have higher tolerance for faults. Therefore, based on the average consistency distribution of all faulty nodes at the current level, the adaptive fuzzification coefficient K of the faulty nodes at the current level is obtained: In the formula, K represents the adaptive fuzzification coefficient of each real fault node, and round() represents the rounding function. This represents a preset scaling factor used to scale the fuzziness coefficients. In this embodiment... The value of is 2, and the implementer may choose other positive integers according to the actual situation; This represents the sorting number indicating the level of each actual fault node. The sorting number is determined by ranking the levels from smallest to largest. Represents the normalization function. It represents the standard deviation of the average consistency among all real fault nodes at the same level as each real fault node.

[0065] If the fault conditions of all real fault nodes at the same level are relatively consistent, it indicates that the possibility of the equipment node being affected by uncertain factors when it fails at the current level is relatively small. In this case, the standard deviation of the average consistency is small, and the value of the adaptive fuzzification coefficient is small, so a smaller fuzzification subset range can be used. Conversely, if the conditions of all real fault nodes at the current level are relatively disordered, it indicates that the equipment is greatly affected by its own uncertain factors when it fails. The standard deviation of the average consistency is large, and the value of the adaptive fuzzification coefficient is large, indicating that a larger fuzzification subset range is needed to improve the fault identification accuracy.

[0066] Step S4: Construct a fault tree based on the adaptive fuzzification coefficient, obtain the importance of each node, and perform fault tree analysis on the power CPS according to the order of node importance.

[0067] Based on step S3, the adaptive fuzzification coefficients of each real fault node are obtained. For interference nodes, the value 1 is used as the adaptive fuzzification coefficient of each interference node, and the fuzzification subset is scaled according to the adaptive fuzzification coefficients of each node.

[0068] By using fuzzy subsets, we can measure the offset of a node's real-time running data from its normal running data and its membership in the event of a fault. The initial range of the fuzzy subset is [0,1].

[0069] Furthermore, for each node, the product of the adaptive fuzzification coefficient of each node and the range of the initial fuzzification subset is calculated as the range of the adjusted fuzzification subset.

[0070] For each node, the mean values ​​of the voltage operating difference, current operating difference, and impedance operating difference are calculated. These mean values ​​are then mapped to the range of the adjusted fuzzy subset using chaotic mapping to obtain the mapping relationship between the operating data of each node and the fuzzy subset. The element values ​​corresponding to the actual operating data of each node within the range of the adjusted fuzzy subset are output. These element values ​​are then input into a membership function to obtain multiple fuzzy probability values ​​for each node. These multiple fuzzy probability values ​​are used as input to a fuzzy probability score (FPS) to output the failure probability of each node. The chaotic mapping, membership function, and fuzzy probability score are well-known techniques and will not be elaborated upon here.

[0071] A fault tree is constructed based on the fault values ​​of all nodes' operational data. A conditional probability model is used to calculate the importance of each node in the fault tree, and the nodes are then sorted according to their importance. Therefore, when a node fault occurs in the power CPS system, the faulty node can be quickly located based on its importance ranking and dealt with promptly, improving system reliability. The process of constructing the fault tree and the method of calculating importance based on the conditional probability model are well-known technologies and will not be elaborated upon here.

[0072] Based on the same inventive concept as the above method, this application embodiment also provides a power system fault data analysis storage medium, wherein the power CPS fault tree analysis storage medium has a computer program, and the computer program implements any of the power system fault data analysis methods when executed by a processor.

[0073] Furthermore, embodiments of this application also provide a power system fault data analysis system, including:

[0074] The data acquisition module is used to statistically analyze all faulty nodes in the power CPS system as initial fault nodes, collect the actual operating data and rated operating data of each node, and the level of each node to construct a fuzzy fault tree; the operating data includes voltage, current, and impedance, and the level is the transmission voltage level;

[0075] The adaptive fuzzy coefficient acquisition module is used to obtain the fault deviation value of each initial fault node based on the difference between the actual operating data and the rated operating data; obtain the fake evaluation value of each initial fault node based on the fault deviation value; obtain the real fault node and interference node based on the fake evaluation value and the threshold segmentation algorithm; obtain the average consistency of each real fault node based on the similarity between the difference between the actual operating data and the rated operating data of any two real fault nodes; and obtain the adaptive fuzzification coefficient of each node based on the level and average consistency dispersion of the real fault node.

[0076] The fault tree analysis module is used to construct a fault tree based on adaptive fuzzification coefficients, obtain the importance of each node, and perform fault tree analysis on the power CPS according to the order of node importance.

[0077] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, specific embodiments of this specification have been described above. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0078] The various embodiments in this application are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0079] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the principles of this application should be included within the protection scope of this application.

Claims

1. A method of power system fault data analysis, characterized by, The method comprises the following steps: All fault nodes in a statistical power CPS system are taken as initial fault nodes, actual operation data, rated operation data and levels of all nodes are collected to construct a fuzzy fault tree; the operation data comprises voltage, current and impedance, and the level is a transmission voltage level; A fault deviation value of each node is obtained based on a difference between the actual operation data and the rated operation data; A fake evaluation value of each initial fault node is obtained based on a discrete amplitude of the fault deviation value; Real fault nodes and interference nodes are obtained according to the fake evaluation value and a threshold segmentation algorithm; An average consistency of each real fault node is obtained based on a similarity degree of a difference between the actual operation data and the rated operation data of any two real fault nodes; An adaptive fuzzification coefficient of each real fault node is obtained based on a level of the real fault node and a discrete degree of the average consistency, and an adaptive fuzzification coefficient of the interference node is preset; A fault tree is constructed according to the adaptive fuzzification coefficient, an importance degree of each node is obtained, and a fault tree analysis of the power CPS is performed according to an order of the importance degree of the node; The fake evaluation value is obtained by: For each initial fault node, a range of the fault deviation value of the initial fault node and all nodes directly connected to the initial fault node is calculated, and a normalized value of a ratio of the fault deviation value of the initial fault node to the range is taken as the fake evaluation value of each initial fault node.

2. A power system fault data analysis method as claimed in claim 1, characterized by, The fault deviation value is obtained by: For each node, a difference absolute value between the actual operation data and the rated operation data of the same category of the node is calculated, and an average value of all the difference absolute values is taken as the fault deviation value of each node.

3. A power system fault data analysis method as claimed in claim 1, wherein, The real fault node is obtained by: The fake evaluation values of all initial fault nodes are taken as inputs of the threshold segmentation algorithm, an optimal segmentation threshold is output, and the initial fault nodes with the fake evaluation values greater than the optimal segmentation threshold are taken as the real fault nodes.

4. A power system fault data analysis method as claimed in claim 3, characterised in that, The interference node is obtained by: The initial fault nodes with the fake evaluation values less than or equal to the optimal segmentation threshold are taken as the interference nodes.

5. A power system fault data analysis method as recited in claim 1, wherein, The average consistency is obtained by: A difference absolute value between voltage in the actual operation data and voltage in the rated operation data of each real fault node is calculated as a voltage operation difference value of each real fault node, and the same method is used to obtain a current operation difference value and an impedance operation difference value of each real fault node, and a vector composed of the voltage operation difference value, the current operation difference value and the impedance operation difference value is taken as a fault vector of each real fault node. The calculation formula of the average consistency is: ; in the formula, represents the average consistency of the i th real faulty node; represents the number of nodes of the same level as the current real faulty node, and cos() represents the cosine similarity, and respectively represent the fault vectors of the i th real faulty node and the j th real faulty node.

6. A power system fault data analysis method as claimed in claim 3, wherein, The adaptive fuzzification coefficient is obtained by: All levels of the real fault nodes are sorted in ascending order, and a sorting serial number of each level is taken as a sorting serial number of each level. The calculation formula of the adaptive blurring coefficient of each real fault node is: ; in the formula, K represents the adaptive blurring coefficient of each real fault node; round() represents a rounding function, represents a preset scaling factor, represents the ranking serial number of the grade of each real fault node, represents a normalization function, represents the standard deviation of the average consistency of all real fault nodes with the same grade as each real fault node. For the interference nodes, a numerical value 1 is taken as the adaptive fuzzification coefficient of each interference node.

7. A power system fault data analysis method as claimed in claim 5, wherein, The fault tree is constructed according to the adaptive fuzzification coefficient, the importance degree of each node is obtained, and the fault tree analysis of the power CPS is performed according to the order of the importance degree of the node, comprising: For each node, a product of the adaptive fuzzification coefficient of each node and a range of an initial fuzzification subset is calculated as a range of an adjusted fuzzification subset. The mean values of the voltage operation difference, the current operation difference and the impedance operation difference of the computing nodes are calculated, the mean values are compared with the range of the adjusted fuzzy subset, the mapping relationship between each node operation data and the fuzzy subset is obtained through chaotic mapping, the actual operation data of each node is output in the range of the adjusted fuzzy subset, the corresponding element value of each node is input into the membership function to obtain a plurality of fuzzy probability values of each node, the plurality of fuzzy probability values of each node are taken as the input of the fuzzy possibility analysis, and the failure probability of each node is output. According to the failure probability of all node operation data, a fault tree is constructed, and the importance of each node in the fault tree is calculated using a conditional probability model.

8. An electric power system fault data analysis storage medium characterized by, The storage medium stores a computer program, and the computer program is run by the processor to realize the power system failure data analysis method of any one of claims 1-7.

9. A power system fault data analysis system implementing a power system fault data analysis method as claimed in claim 1, characterized by, The system comprises: A data acquisition module is configured to count all failure nodes in a power CPS system as initial failure nodes, acquire actual operation data and rated operation data of each node, acquire the grade of each node, and construct a fuzzy fault tree; the operation data includes voltage, current and impedance, and the grade is a transmission voltage grade; An adaptive fuzzy coefficient acquisition module is configured to acquire a failure deviation value of each initial failure node based on the difference between the actual operation data and the rated operation data, acquire a forgery evaluation value of each initial failure node based on the failure deviation value, acquire a real failure node and an interference node according to the forgery evaluation value and a threshold segmentation algorithm, acquire the average consistency of each real failure node based on the similarity degree of the difference between the actual operation data and the rated operation data of any two real failure nodes, and acquire the adaptive fuzzy coefficient of the real failure node based on the grade and the average consistency discrete degree of the real failure node, and preset the adaptive fuzzy coefficient of the interference node; A fault tree analysis module is configured to construct a fault tree according to the adaptive fuzzy coefficient, acquire the importance of each node, and perform fault tree analysis on the power CPS according to the order of the importance of the nodes.

Images