A security resilience measurement method for power management systems

Abstract

The present invention discloses a method for measuring the safety resilience of an electric power management system, and specifically relates to the technical field of safety resilience measurement. By acquiring node pressure information, block broadcast information, node fault information, and blockchain update information during the operation of a consensus mechanism, a consensus mechanism operation hidden danger perception model is constructed, the operation status of the consensus mechanism is perceived and evaluated, and a consensus mechanism operation hidden danger perception index is generated. Before an obvious failure of the consensus mechanism occurs, potential operation hidden dangers are intelligently perceived, and the validity of the safety resilience measurement data within the consensus mechanism operation hidden danger time interval is verified, thereby effectively reducing the risk of reduced validity of the safety resilience measurement data due to consensus mechanism failures, and improving the accuracy of the safety resilience assessment of the electric power management system.

Description

Technical Field

[0001] The present invention relates to the technical field of security resilience measurement, and more specifically, to a security resilience measurement method for a power management system. Background Art

[0002] In modern power systems, the security and resilience of power management systems are crucial. As the scale and complexity of power systems continue to increase, the threats facing power grids are becoming increasingly diverse, including natural disasters, equipment failures, cyber attacks, etc.

[0003] In the security resilience measurement of power management systems, blockchain technology can provide a reliable mechanism to record and analyze the behavior of power management systems under abnormal conditions or attacks. For example, recording the status information, operation logs, and fault events of each node on the blockchain can provide detailed data support for the security resilience analysis of the system. In large-scale power management systems, multiple blockchain platforms are usually used to manage different functional modules (such as energy trading, security monitoring, emergency response, etc.), and different blockchain platforms usually use different consensus mechanisms to optimize the performance of each module due to different functional requirements of management modules. The stable operation of different consensus mechanisms ensures the consistency and integrity of measurement data in security resilience measurement, and the consensus mechanism may have different degrees of operational risks before the occurrence of obvious operational failures. Due to the lack of intelligent perception of operational risks, the effectiveness of some measurement data in the security resilience measurement of the power management system is reduced, which in turn affects the accuracy of the security resilience measurement of the power management system. Summary of the invention

[0004] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a security resilience measurement method for a power management system to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for measuring the security resilience of a power management system comprises the following steps:

[0007] Step S1, obtaining node pressure information, block broadcast information, node failure information, and blockchain update information during the operation of the consensus mechanism;

[0008] Step S2, constructing a consensus mechanism operation hidden danger perception model, performing a perception assessment on the operation status of the consensus mechanism, and generating a consensus mechanism operation hidden danger perception index;

[0009] Step S3, comparing the consensus mechanism operation hidden danger perception index with a preset consensus mechanism operation hidden danger perception index threshold, classifying the operation status of the consensus mechanism, and determining the consensus mechanism operation hidden danger;

[0010] Step S4, verifying the validity of the resilience measurement data within the consensus mechanism operation risk time interval.

[0011] In a preferred embodiment, the node pressure information includes an attack pressure coefficient, the block broadcast information includes a broadcast abnormality delay coefficient, the node failure information includes a node disconnection abnormality coefficient, and the blockchain update information includes a block addition speed bump coefficient.

[0012] In a preferred embodiment, the acquisition logic of the attack pressure coefficient is as follows:

[0013] Within the preset time period TZ, record the time S1 when each node is attacked and the time S2 when it takes defensive measures, and calculate the node response coefficient xy, which is expressed as follows: Where S3 i represents the response time of the node to the i-th attack, S3 = |S2-S1|, min(S3 i ) is a minimum value function used to obtain the minimum value of the response time, and I is a positive integer;

[0014] Calculate the average value xyp of the node response coefficient, the expression is as follows Where xy r represents the node response coefficient of the rth node;

[0015] Calculate the standard deviation xyb of the node response coefficient, the expression is as follows

[0016] Calculate the attack pressure coefficient GJY, the expression is as follows

[0017] In a preferred embodiment, the logic for obtaining the broadcast abnormal delay coefficient is as follows:

[0018] During the preset time period TZ, when monitoring the broadcast process of the new block, obtain the broadcast time G1 of the new block from the source node to the target node;

[0019] Set the benchmark value GZ for the normal broadcast time of a new block from the source node to the target node;

[0020] Compare the broadcast time of the new block from the source node to the target node with the benchmark value. If the broadcast time of the new block from the source node to the target node is greater than the benchmark value, the broadcast time greater than the benchmark value is marked as delayed broadcast time YG;

[0021] Calculate the broadcast abnormal delay coefficient GBY, the expression is as follows Where YG n Indicates the nth delayed broadcast time, where N is a positive integer.

[0022] In a preferred embodiment, the logic for obtaining the node disconnection anomaly coefficient is as follows:

[0023] Acquire disconnection characteristic data of each node within a preset time period TZ, establish a disconnection characteristic data set, wherein the disconnection characteristic data includes disconnection frequency and disconnection duration, and perform standardization processing on the acquired disconnection characteristic data;

[0024] The K-Means algorithm is used to calculate the node disconnection anomaly coefficient based on the obtained disconnection feature data as follows:

[0025] Step A1, determine the number of clusters K using the elbow rule;

[0026] Step A2: randomly select K initial cluster centers C1, C2, C3, ..., C K , each initial cluster center has two coordinates, C k =(C k，F , C k，T ), where C k，F represents the normalized value of the disconnection frequency of the initial cluster center k, C k，T represents the normalized value of the disconnection duration of the initial cluster center k, k = {1, 2, 3, ..., K}, K is a positive integer;

[0027] Step A3, calculate each node X m =(F m , T m ) to each initial cluster center C k =(C k，F , C k，T ) is expressed as follows: In the formula, F m Represents node X m The normalized value of the disconnection frequency, T m Represents node X m The disconnection duration is standardized, and each node is assigned to the cluster with the closest Euclidean distance to establish a standard cluster of nodes;

[0028] Step A4, calculating the average disconnection frequency and the average disconnection duration in the node standard cluster, and using the average disconnection frequency and the average disconnection duration as two new coordinates of the node standard cluster respectively;

[0029] Step A5, repeating steps A3 and A4 to iteratively update the node standard cluster until the node standard cluster no longer changes;

[0030] Calculate the node disconnection abnormality coefficient JDY, the expression is as follows Where d m Represents node X m The Euclidean distance to the standard cluster of the node to which it belongs, dmin represents the node X m The minimum value of the Euclidean distance in the standard cluster of the node to which it belongs, dmax represents the minimum value of the Euclidean distance in the standard cluster of the node to which it belongs, and ... m The maximum value of the Euclidean distance among the standard clusters of the nodes.

[0031] In a preferred embodiment, the logic for obtaining the block adding speed bump coefficient is as follows:

[0032] Get the addition timestamp of each new block within the preset time period TZ, calculate the addition time interval between every two adjacent new blocks, and obtain the block addition speed time series y[t];

[0033] The block adding speed time series y[t] is decomposed by discrete wavelet, and Daubechies wavelet is selected as the wavelet basis function. The decomposition formula is as follows Where Y j，b represents the wavelet coefficient at the decomposed scale j and displacement b, ψ j,b [t] represents the wavelet basis function at scale j and displacement b, t = {1, 2, 3, ..., G}, G is a positive integer;

[0034] Calculate the average value xbp of the wavelet coefficients at scale j j , the expression is as follows Wherein b={1,2,3,...,B}, B is a positive integer;

[0035] Calculate the standard deviation xbb of the wavelet coefficients at scale j j , the expression is as follows

[0036] Calculate the block adding speed bump coefficient QKD, the expression is as follows

[0037] In a preferred implementation, in step S2, a consensus mechanism operation hidden danger perception model is constructed, the operation status of the consensus mechanism is perceived and evaluated, and a consensus mechanism operation hidden danger perception index is generated, as follows:

[0038] The attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient obtained within the time period TZ are normalized. According to the normalized attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient, a consensus mechanism operation hidden danger perception model is constructed to generate the consensus mechanism operation hidden danger perception index RPI. The model is based on the following formula: Wherein a1, a2, a3, and a4 represent the preset proportional coefficients of the attack pressure coefficient, the broadcast abnormal delay coefficient, the node disconnection abnormal coefficient, and the block addition speed bump coefficient, respectively, and a1, a2, a3, and a4 are all greater than 0.

[0039] In a preferred implementation, in step S3, the consensus mechanism operation hidden danger perception index is compared with a preset consensus mechanism operation hidden danger perception index threshold, the operation status of the consensus mechanism is classified, and the operation hidden danger of the consensus mechanism is determined, as follows:

[0040] If the consensus mechanism operation hidden danger perception index is greater than the consensus mechanism operation hidden danger perception index threshold, an operation hidden danger signal is generated; if the consensus mechanism operation hidden danger perception index is less than or equal to the consensus mechanism operation hidden danger perception index threshold, an operation stability signal is generated.

[0041] In a preferred implementation, in step S4, the validity of the security resilience measurement data within the consensus mechanism operation potential risk time interval is verified as follows:

[0042] Obtain historical security resilience measurement data of the power management system, including but not limited to fault frequency, fault repair time, number of intrusion detection events, emergency response time, emergency plan execution success rate, backup system switching time, and false alarm rate;

[0043] Preprocess the acquired historical security resilience measurement data, including data quality inspection and cleaning, data standardization, and dividing the data set into training set, validation set, and test set according to a certain ratio; build a neural network model and train the neural network model;

[0044] Obtain the safety resilience measurement data of the power management system within the time period TZ corresponding to the generation of the operation hidden danger signal, input the safety resilience measurement data of the power management system within the time period TZ into the trained neural network model, screen out the minimum output lower limit value from all output nodes, and compare the output lower limit value with the preset validity threshold. If the output lower limit value is greater than the validity threshold, a valid signal is generated; if the output lower limit value is less than or equal to the validity threshold, a data invalid risk signal is generated.

[0045] Technical effects and advantages of the present invention:

[0046] 1. The present invention obtains node pressure information, block broadcast information, node fault information, and blockchain update information during the operation of the consensus mechanism, constructs a consensus mechanism operation hidden danger perception model, perceives and evaluates the operation status of the consensus mechanism, and generates a consensus mechanism operation hidden danger perception index. Before the consensus mechanism has obvious failures, it can intelligently perceive potential operation hidden dangers and verify the validity of the safety resilience measurement data within the consensus mechanism operation hidden danger time interval, effectively reducing the risk of reduced validity of safety resilience measurement data due to consensus mechanism failures and improving the accuracy of safety resilience assessment of the power management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to facilitate understanding by those skilled in the art, the present invention is further described below in conjunction with the accompanying drawings;

[0048] Figure 1 Schematic diagram of the structure of the method of the embodiment of the present invention. DETAILED DESCRIPTION

[0049] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Example: Figure 1 A method for measuring the security resilience of a power management system is provided in the present invention, comprising the following steps:

[0051] Step S1, obtaining node pressure information, block broadcast information, node failure information, and blockchain update information during the operation of the consensus mechanism;

[0052] Step S2, constructing a consensus mechanism operation hidden danger perception model, performing a perception assessment on the operation status of the consensus mechanism, and generating a consensus mechanism operation hidden danger perception index;

[0053] Step S3, comparing the consensus mechanism operation hidden danger perception index with a preset consensus mechanism operation hidden danger perception index threshold, classifying the operation status of the consensus mechanism, and determining the consensus mechanism operation hidden danger;

[0054] Step S4, verifying the validity of the resilience measurement data within the consensus mechanism operation risk time interval;

[0055] In step S1, node pressure information, block broadcast information, node failure information, and blockchain update information during the operation of the consensus mechanism are obtained;

[0056] The node pressure information includes the attack pressure coefficient, the block broadcast information includes the broadcast abnormal delay coefficient, the node failure information includes the node disconnection abnormal coefficient, and the blockchain update information includes the block addition speed bump coefficient;

[0057] The attack pressure coefficient is used to measure the frequency of attacks on each node in the blockchain and the degree of response to attack behaviors when the consensus mechanism is running. The larger the attack pressure coefficient, the deeper the frequency of attacks on each node in the blockchain and the degree of response to attack behaviors, and the greater the probability that each node will become a malicious node. Malicious nodes may participate in the consensus by creating multiple false identities, thereby manipulating the consensus process and preventing other nodes from participating in the consensus process. A larger attack pressure coefficient reflects the greater probability that the consensus mechanism has potential operating risks. Conversely, a smaller attack pressure coefficient indicates that the operation of the consensus mechanism is more stable.

[0058] The logic for obtaining the attack pressure coefficient is as follows:

[0059] Within the preset time period TZ, record the time S1 when each node is attacked and the time S2 when it takes defensive measures, and calculate the node response coefficient xy, which is expressed as follows: Where S3 i represents the response time of the node to the i-th attack, S3 = |S2-S1|, min(S3 i ) is a minimum value function used to obtain the minimum value of the response time, and I is a positive integer;

[0060] Calculate the average value xyp of the node response coefficient, the expression is as follows Where xy r represents the node response coefficient of the rth node;

[0061] Calculate the standard deviation xyb of the node response coefficient, the expression is as follows

[0062]

[0063] Calculate the attack pressure coefficient GJY, the expression is as follows

[0064] The consensus mechanism is a core component used in blockchain and other distributed systems to ensure that all nodes in the network reach a consensus. In the consensus mechanism, the system will select one or more nodes from all nodes to propose new blocks. The proposed new blocks will be broadcast to all nodes in the blockchain through the peer-to-peer network. The broadcast abnormality delay coefficient is used to measure the broadcast delay of the new block to all nodes in the network through the peer-to-peer network. The larger the broadcast abnormality delay coefficient, the more serious the broadcast delay of the new block, which is obviously beyond normal expectations, and may cause network congestion and node performance degradation, resulting in inconsistent information in the consensus process. Nodes cannot receive the latest block or transaction information in time, which in turn leads to a decrease in the effectiveness of some measurement data in the safety resilience measurement of the power management system. On the contrary, a smaller broadcast abnormality delay coefficient indicates that the broadcast delay of the new block is within a controllable range, indicating that the operation of the consensus mechanism is more stable;

[0065] The logic for obtaining the broadcast abnormal delay coefficient is as follows:

[0066] During the preset time period TZ, when monitoring the broadcast process of the new block, obtain the broadcast time G1 of the new block from the source node to the target node;

[0067] Set the benchmark value GZ for the normal broadcast time of a new block from the source node to the target node;

[0068] It should be noted that the benchmark value of the normal broadcast time of a new block from a source node to a target node is determined by a person skilled in the art by comprehensively considering the network scale, the number of nodes and the network topology, and will not be elaborated here;

[0069] Compare the broadcast time of the new block from the source node to the target node with the benchmark value. If the broadcast time of the new block from the source node to the target node is greater than the benchmark value, it indicates that there is a delay in the broadcast of the new block. The broadcast time greater than the benchmark value is marked as delayed broadcast time YG.

[0070] Calculate the broadcast abnormal delay coefficient GBY, the expression is as follows

[0071] Where YG n Indicates the nth delayed broadcast time, where N is a positive integer;

[0072] The node disconnection anomaly coefficient is used to measure the frequency and severity of disconnection or disconnection anomalies of each node in the blockchain network, thereby quantifying the frequency of disconnection or disconnection of the node within a certain period of time. Frequent disconnections may mean that the node has a hardware failure. A higher node disconnection anomaly coefficient indicates that there are potential abnormal risks in the node or network, which may lead to a decrease in the efficiency and accuracy of the consensus mechanism, which may cause hidden dangers in the operation of the consensus mechanism, and then lead to a decrease in the effectiveness of some measurement data in the safety resilience measurement of the power management system. On the contrary, the smaller the node disconnection anomaly coefficient, the smaller the probability of causing hidden dangers in the operation of the consensus mechanism.

[0073] The logic for obtaining the node disconnection anomaly coefficient is as follows:

[0074] Acquire disconnection characteristic data of each node within a preset time period TZ, establish a disconnection characteristic data set, wherein the disconnection characteristic data includes disconnection frequency and disconnection duration, and perform standardization processing on the acquired disconnection characteristic data;

[0075] It should be noted that in order to ensure that the disconnection frequency and disconnection duration are on the same scale, the disconnection feature data needs to be standardized. Common standardization methods include Z-Score standardization, Min-Max standardization, Robust standardization, etc.

[0076] The K-Means algorithm is used to calculate the node disconnection anomaly coefficient based on the obtained disconnection feature data as follows:

[0077] Step A1, determine the number of clusters K using the elbow rule;

[0078] It should be noted that the elbow rule is used to determine the number of clusters K, starting from K=1, and gradually increasing the K value. The minimum Euclidean distance from each node to the nearest cluster center under different K values ​​is calculated, and the minimum Euclidean distances of all nodes are squared and accumulated to obtain the total square error and SSE. The K value is used as the horizontal axis, and the total square error and SSE are used as the vertical axis to draw a curve graph. In the curve graph, find the position where the total square error and SSE decrease significantly slower with the increase of K value. This point usually presents the shape of an "elbow". Calculate the slope of the curve at different K values, compare the slope of the curve at different K values ​​with the curve elbow slope threshold range, and take the K value corresponding to the curve slope that first appears within the curve elbow slope threshold range as the final determined value;

[0079] Step A2: randomly select K initial cluster centers C1, C2, C3, ..., C K , each initial cluster center has two coordinates, C k =(C k，F , C k，T ), where Ck，F represents the normalized value of the disconnection frequency of the initial cluster center k, C k，T represents the normalized value of the disconnection duration of the initial cluster center k, k = {1, 2, 3, ..., K}, K is a positive integer;

[0080] Step A3, calculate each node X m =(F m , T m ) to each initial cluster center C k =(C k，F , C k，T ) is expressed as follows

[0081] In the formula, F m Represents node X m The normalized value of the disconnection frequency, T m Represents node X m The disconnection duration is standardized, and each node is assigned to the cluster with the closest Euclidean distance to establish a standard cluster of nodes;

[0082] Step A4, calculating the average disconnection frequency and the average disconnection duration in the node standard cluster, and using the average disconnection frequency and the average disconnection duration as two new coordinates of the node standard cluster respectively;

[0083] Step A5, repeating steps A3 and A4 to iteratively update the node standard cluster until the node standard cluster no longer changes;

[0084] Calculate the node disconnection abnormality coefficient JDY, the expression is as follows Where d m Represents node X m Euclidean distance to the standard cluster of the node, d min Represents node X m The minimum value of the Euclidean distance in the standard cluster of the node, d max Represents node X m The maximum value of the Euclidean distance in the standard cluster of the node to which it belongs;

[0085] The block addition speed fluctuation coefficient is used to measure the volatility of the speed at which new blocks are added to the blockchain in the blockchain platform. This coefficient reflects the stability or instability of the block addition speed over a period of time. If the block addition speed fluctuates greatly, that is, the speed is fast and slow, it means that the block addition speed fluctuates greatly, and the consensus mechanism may have potential operating risks, which means that the efficiency of the consensus mechanism is affected. On the contrary, a smaller block addition speed fluctuation coefficient indicates that the block addition speed is relatively stable, and the probability of the consensus mechanism having operating risks is lower, and its operating status is more stable.

[0086] The logic for obtaining the block adding speed bump coefficient is as follows:

[0087] Get the addition timestamp of each new block within the preset time period TZ, calculate the addition time interval between every two adjacent new blocks, and obtain the block addition speed time series y[t];

[0088] The block adding speed time series y[t] is decomposed by discrete wavelet, and Daubechies wavelet is selected as the wavelet basis function. The decomposition formula is as follows Where Y j，b represents the wavelet coefficient at the decomposed scale j and displacement b, ψ j,b [t] represents the wavelet basis function at scale j and displacement b, t = {1, 2, 3, ..., G}, G is a positive integer;

[0089] Calculate the average value xbp of the wavelet coefficients at scale j j , the expression is as follows

[0090] Wherein b={1,2,3,...,B}, B is a positive integer;

[0091] Calculate the standard deviation xbb of the wavelet coefficients at scale j j , the expression is as follows

[0092]

[0093] Calculate the block adding speed bump coefficient QKD, the expression is as follows

[0094] In step S2, a consensus mechanism operation hidden danger perception model is constructed to perform a perception assessment on the operation status of the consensus mechanism and generate a consensus mechanism operation hidden danger perception index, as follows:

[0095] The attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient obtained within the time period TZ are normalized. According to the normalized attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient, a consensus mechanism operation hidden danger perception model is constructed to generate the consensus mechanism operation hidden danger perception index RPI. The model is based on the following formula: Where a1, a2, a3, and a4 represent the preset proportional coefficients of the attack pressure coefficient, the broadcast abnormal delay coefficient, the node disconnection abnormal coefficient, and the block addition speed bump coefficient, respectively, and a1, a2, a3, and a4 are all greater than 0;

[0096] It should be noted that a1, a2, a3, a4, a5, and a6 are set according to actual conditions. For example, the expert empowerment method is adopted, that is, experts in related fields are invited to determine the preset proportion coefficients of various indicators through professional opinion surveys and comprehensive evaluations. In addition, various methods such as hierarchical analysis method and fuzzy comprehensive evaluation method can also be considered to determine the preset proportion coefficients to ensure the objectivity and scientificity of the preset proportion coefficients, which will not be elaborated here;

[0097] From the above calculation expression, it can be seen that the greater the attack pressure coefficient, the greater the broadcast abnormal delay coefficient, the greater the node disconnection abnormal coefficient, and the greater the block addition speed bump coefficient, the greater the consensus mechanism operation hidden danger perception index, indicating that the probability of potential operation hidden dangers in the consensus mechanism is greater. On the contrary, the smaller the attack pressure coefficient, the smaller the broadcast abnormal delay coefficient, the smaller the node disconnection abnormal coefficient, and the smaller the block addition speed bump coefficient, the smaller the consensus mechanism operation hidden danger perception index, indicating that the probability of operation hidden dangers in the consensus mechanism is lower and its operation status is more stable.

[0098] In step S3, the consensus mechanism operation hidden danger perception index is compared with the preset consensus mechanism operation hidden danger perception index threshold, the operation status of the consensus mechanism is classified, and the operation hidden danger of the consensus mechanism is determined, as follows:

[0099] If the consensus mechanism operation hidden danger perception index is greater than the consensus mechanism operation hidden danger perception index threshold, it indicates that the consensus mechanism has potential operation hidden dangers, and an operation hidden danger signal is generated; if the consensus mechanism operation hidden danger perception index is less than or equal to the consensus mechanism operation hidden danger perception index threshold, it indicates that the consensus mechanism does not have potential operation hidden dangers, and its operation status is relatively stable, and an operation stability signal is generated;

[0100] In step S4, the validity of the security resilience measurement data within the consensus mechanism operation risk time interval is verified as follows:

[0101] Obtain historical security resilience measurement data of the power management system, including but not limited to fault frequency, fault repair time, number of intrusion detection events, emergency response time, emergency plan execution success rate, backup system switching time, and false alarm rate;

[0102] Fault frequency refers to the number of faults that occur in the system within a specified period of time. Frequent faults may indicate potential safety hazards in the system.

[0103] Fault repair time refers to the average time from the occurrence of a fault to the system returning to normal operation. A shorter fault repair time indicates a high system recovery efficiency.

[0104] The number of intrusion detection events refers to the number of network attacks or abnormal access events detected by the recording system;

[0105] Emergency response time refers to the time from the occurrence of an incident to the initiation of emergency measures. A shorter response time means a higher efficiency of the emergency mechanism.

[0106] The success rate of emergency plan execution refers to the measurement of the effectiveness of the emergency plan in actual application, including the rate of successful system restoration or prevention of disaster spread;

[0107] The backup system switching time refers to the time required for the backup system to take over from the failure of the main system, which is used to evaluate the effectiveness of the system redundancy design;

[0108] The false alarm rate refers to the frequency with which the power management system mistakenly identifies normal activities as security threats. A high false alarm rate may lead to waste of system resources and user distrust.

[0109] Preprocess the acquired historical security resilience measurement data, including data quality inspection and cleaning, data standardization, and dividing the data set into training set, validation set, and test set according to a certain ratio;

[0110] Construct a neural network model and select a multilayer perceptron as the neural network architecture. The multilayer perceptron includes an input layer, multiple hidden layers, and an output layer. The number of hidden layers and the number of neurons in each layer are adjusted according to the actual situation. Select an activation function for the nonlinear transformation of neurons and select the mean square error as the loss function to measure the error between the predicted value and the actual value.

[0111] It should be noted that common activation functions, such as ReLU (Rect if ied Li near Unit), Sigmoid or Tanh, are used to introduce nonlinearity and help neural networks capture complex data patterns;

[0112] The neural network model is trained, and the gradient of the loss function is calculated using the stochastic gradient descent algorithm. The weights of the neural network are updated according to the gradient of the loss function. The weights of the neural network are optimized through multiple iterations until the performance of the neural network model on the validation set is stable. For each neuron in the trained neural network, a linear approximation is performed on the input historical security resilience metric data based on the activation area corresponding to the neuron. According to the structure of the neural network, the linear approximation value is boundedly propagated in the neural network, and the output lower limit value of each output node is calculated layer by layer to obtain the output lower limit value. Boundary propagation helps to identify the output lower limit value in the worst case.

[0113] Obtain the safety resilience measurement data of the power management system within the time period TZ corresponding to the generation of the operation hidden danger signal, input the safety resilience measurement data of the power management system within the time period TZ into the trained neural network model, screen out the minimum output lower limit value from all output nodes, and compare the output lower limit value with the preset validity threshold. If the output lower limit value is greater than the validity threshold, it means that the safety resilience measurement data is valid and a valid signal is generated; if the output lower limit value is less than or equal to the validity threshold, it means that the safety resilience measurement data may have the risk of invalid data or system abnormality, and a data invalid risk signal is generated;

[0114] The present invention acquires node pressure information, block broadcast information, node fault information, and blockchain update information during the operation of the consensus mechanism, constructs a consensus mechanism operation hidden danger perception model, perceives and evaluates the operation status of the consensus mechanism, and generates a consensus mechanism operation hidden danger perception index. Before the consensus mechanism has obvious faults, it can intelligently perceive potential operation hidden dangers and verify the validity of the safety resilience measurement data within the consensus mechanism operation hidden danger time interval, effectively reducing the risk of reduced validity of safety resilience measurement data due to consensus mechanism failures and improving the accuracy of safety resilience assessment of the power management system.

[0115] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0116] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0117] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for measuring the security resilience of a power management system, characterized by: The steps include: Step S1, obtaining node pressure information, block broadcast information, node failure information, and blockchain update information during the operation of the consensus mechanism; Step S2, constructing a consensus mechanism operation hidden danger perception model, performing a perception assessment on the operation status of the consensus mechanism, and generating a consensus mechanism operation hidden danger perception index; Step S3, comparing the consensus mechanism operation hidden danger perception index with a preset consensus mechanism operation hidden danger perception index threshold, classifying the operation status of the consensus mechanism, and determining the consensus mechanism operation hidden danger; Step S4, verifying the validity of the resilience measurement data within the consensus mechanism operation risk time interval; The node pressure information includes the attack pressure coefficient, the block broadcast information includes the broadcast abnormal delay coefficient, the node failure information includes the node disconnection abnormal coefficient, and the blockchain update information includes the block addition speed bump coefficient; The logic for obtaining the attack pressure coefficient is as follows: Within the preset time period TZ, record the time S1 when each node is attacked and the time S2 when it takes defensive measures, and calculate the node response coefficient xy, which is expressed as follows: Where S3 i represents the response time of the node to the i-th attack, S3 = |S2-S1|, min(S3 i ) is a minimum value function used to obtain the minimum value of the response time, and I is a positive integer; Calculate the average value xyp of the node response coefficient, the expression is as follows Where xy r represents the node response coefficient of the rth node; Calculate the standard deviation xyb of the node response coefficient, the expression is as follows Calculate the attack pressure coefficient GJY, the expression is as follows The logic for obtaining the broadcast abnormal delay coefficient is as follows: During the preset time period TZ, when monitoring the broadcast process of the new block, obtain the broadcast time G1 of the new block from the source node to the target node; Set the benchmark value GZ for the normal broadcast time of a new block from the source node to the target node; Compare the broadcast time of the new block from the source node to the target node with the benchmark value. If the broadcast time of the new block from the source node to the target node is greater than the benchmark value, the broadcast time greater than the benchmark value is marked as delayed broadcast time YG; Calculate the broadcast abnormal delay coefficient GBY, the expression is as follows Where YG n Indicates the nth delayed broadcast time, where N is a positive integer; The logic for obtaining the node disconnection anomaly coefficient is as follows: Acquire disconnection characteristic data of each node within a preset time period TZ, establish a disconnection characteristic data set, wherein the disconnection characteristic data includes disconnection frequency and disconnection duration, and perform standardization processing on the acquired disconnection characteristic data; The node disconnection anomaly coefficient JDY is calculated using the K-Means algorithm based on the obtained disconnection feature data. The expression is as follows: Where d m Represents node X m The Euclidean distance to the standard cluster of the node to which it belongs, dmin represents the node X m The minimum value of the Euclidean distance in the standard cluster of the node to which it belongs, dmax represents the minimum value of the Euclidean distance in the standard cluster of the node to which it belongs, and ... m The maximum value of the Euclidean distance in the standard cluster of the node to which it belongs; The logic for obtaining the block adding speed bump coefficient is as follows: Get the addition timestamp of each new block within the preset time period TZ, calculate the addition time interval between every two adjacent new blocks, and obtain the block addition speed time series y[t]; The block adding speed time series y[t] is decomposed by discrete wavelet, and Daubechies wavelet is selected as the wavelet basis function. The decomposition formula is as follows Where Y j，b represents the wavelet coefficient at the decomposed scale j and displacement b, ψ j，b [t] represents the wavelet basis function at scale j and displacement b, t = {1, 2, 3, ..., G}, G is a positive integer; Calculate the average value xbp of the wavelet coefficients at scale j j , the expression is as follows Wherein b={1,2,3,...,B}, B is a positive integer; Calculate the standard deviation xbb of the wavelet coefficients at scale j j , the expression is as follows Calculate the block adding speed bump coefficient QKD, the expression is as follows 2. A method for measuring the security resilience of a power management system according to claim 1, characterized in that: Calculating the node disconnection anomaly coefficient using the K-Means algorithm based on the obtained disconnection feature data also includes the following steps: Step A1, determine the number of clusters K using the elbow rule; Step A2: randomly select K initial cluster centers C1, C2, C3, ..., C K , each initial cluster center has two coordinates, C k =(C k，F , C k，T ), where C k，F represents the normalized value of the disconnection frequency of the initial cluster center k, C k，T represents the normalized value of the disconnection duration of the initial cluster center k, k = {1, 2, 3, ..., K}, K is a positive integer; Step A3, calculate each node X m =(F m , T m ) to each initial cluster center C k =(C k，F , C k，T ) is expressed as follows: In the formula, F m Represents node X m The normalized value of the disconnection frequency, T m Represents node X m The disconnection duration is standardized, and each node is assigned to the cluster with the closest Euclidean distance to establish a standard cluster of nodes; Step A4, calculating the average disconnection frequency and the average disconnection duration in the node standard cluster, and using the average disconnection frequency and the average disconnection duration as two new coordinates of the node standard cluster respectively; Step A5, repeating steps A3 and A4 to iteratively update the node standard cluster until the node standard cluster no longer changes.

3. The method for measuring the security resilience of a power management system according to claim 1, characterized in that: In step S2, a consensus mechanism operation hidden danger perception model is constructed to perform a perception assessment on the operation status of the consensus mechanism and generate a consensus mechanism operation hidden danger perception index, as follows: The attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient obtained within the time period TZ are normalized. According to the normalized attack pressure coefficient, broadcast abnormal delay coefficient, node disconnection abnormal coefficient, and block addition speed bump coefficient, a consensus mechanism operation hidden danger perception model is constructed to generate the consensus mechanism operation hidden danger perception index RPI. The model is based on the following formula: Wherein a1, a2, a3, and a4 represent the preset proportional coefficients of the attack pressure coefficient, the broadcast abnormal delay coefficient, the node disconnection abnormal coefficient, and the block addition speed bump coefficient, respectively, and a1, a2, a3, and a4 are all greater than 0.

4. A method for measuring the security resilience of a power management system according to claim 3, characterized in that: In step S3, the consensus mechanism operation hidden danger perception index is compared with the preset consensus mechanism operation hidden danger perception index threshold, the operation status of the consensus mechanism is classified, and the operation hidden danger of the consensus mechanism is determined, as follows: If the consensus mechanism operation hidden danger perception index is greater than the consensus mechanism operation hidden danger perception index threshold, an operation hidden danger signal is generated; if the consensus mechanism operation hidden danger perception index is less than or equal to the consensus mechanism operation hidden danger perception index threshold, an operation stability signal is generated.

5. A method for measuring the security resilience of a power management system according to claim 4, characterized in that: In step S4, the validity of the security resilience measurement data within the consensus mechanism operation risk time interval is verified as follows: Obtain historical security resilience measurement data of the power management system, including but not limited to fault frequency, fault repair time, number of intrusion detection events, emergency response time, emergency plan execution success rate, backup system switching time, and false alarm rate; Preprocess the acquired historical security resilience measurement data, including data quality inspection and cleaning, data standardization, and dividing the data set into training set, validation set, and test set according to a certain ratio; build a neural network model and train the neural network model; Obtain the safety resilience measurement data of the power management system within the time period TZ corresponding to the generation of the operation hidden danger signal, input the safety resilience measurement data of the power management system within the time period TZ into the trained neural network model, screen out the minimum output lower limit value from all output nodes, and compare the output lower limit value with the preset validity threshold. If the output lower limit value is greater than the validity threshold, a valid signal is generated; if the output lower limit value is less than or equal to the validity threshold, a data invalid risk signal is generated.

Images