A smart grid security situation assessment method based on improved probabilistic neural network

Abstract

The present invention relates to a smart grid security situation assessment method based on an improved probabilistic neural network, and belongs to the field of Internet technology. The method specifically includes the following steps: S1, collecting smart grid security data information and security asset information, and establishing a smart grid security database; S2, performing data cleaning and normalization preprocessing on the security data in the database to obtain a smart grid security data set; S3, inputting the data set into XGBOOST for feature screening and extracting its key features; S4, constructing a sample set and a test set using a holdout method; S5, building an XGBOOST-ISSA-PNN smart grid security situation assessment model, and using the training set sample set to classify and train the model and retain its optimal parameters; S6, using the trained XGBOOST-ISSA-PNN attack detection model to perform attack detection on the sample test set, calculating the network security situation value according to the quantitative indicators of different attacks, and evaluating the security status of the smart grid.

Description

Technical Field

[0001] The present invention belongs to the field of Internet technology and relates to a smart grid security situation assessment method based on an improved probabilistic neural network. Background Art

[0002] With the continuous development of ultra-high voltage (UHV) power grids and next-generation industrial information technology, the scale and complexity of smart grids are increasing. While the diversity, complexity, and openness of power grids bring convenience to the power industry, they also pose significant challenges to the secure operation of smart grid systems. my country's power industry has achieved its UHV and large-scale power grid development goals, but the overall security structure of the grid remains relatively weak. Failures in power operations are increasing, causing widespread power outages and other incidents, which have serious impacts on the country, the economy, and society. Traditional security measures, such as access control and firewalls, are no longer sufficient to address the diverse and complex cybersecurity risks facing the current power grid. Cyber situational awareness technology, however, can assess cybersecurity risks in real time and further quantify the overall security posture of smart grid systems and the types of attacks they are subject to. By integrating indicators that influence cybersecurity, it can transform post-event response into pre-event response, mitigating the negative impacts of cyberattacks.

[0003] Although machine learning methods have been applied to network security situation assessment, these methods have primarily been applied to the general internet sector. There is a lack of security situation assessment techniques specifically tailored to smart grids, making traditional network security situation assessment techniques difficult to directly apply to smart grids. Therefore, researching more accurate smart grid security situation assessment methods within the industrial internet environment is crucial for ensuring the security of smart grids. Summary of the Invention

[0004] In view of this, the purpose of the present invention is to provide a smart grid security situation assessment method based on an improved probabilistic neural network, to solve the problem that traditional power grid security situation assessment methods cannot be directly applied to smart grids, and to propose a smart grid security situation assessment method with high assessment accuracy and fast convergence speed.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] A smart grid security situation assessment method based on an improved probabilistic neural network comprises the following steps:

[0007] S1: Collect smart grid security data and security asset information and establish a smart grid security database;

[0008] S2: Perform data cleaning and normalization preprocessing on the security data in the database to obtain the smart grid security dataset;

[0009] S3: Input the dataset into XGBOOST for feature screening and extract its key features;

[0010] S4: Use the holdout method to construct the training sample data set and the test sample data set;

[0011] S5: Build the XGBOOST-ISSA-PNN smart grid security situation assessment model, and use the training set sample set to classify and train the model and retain its optimal parameters;

[0012] S6: Use the trained XGBOOST-ISSA-PNN attack detection model to perform attack detection on the sample test set, calculate the network security situation value based on the quantitative indicators of different attacks, and evaluate the security status of the network.

[0013] Optionally, the step S2 specifically includes the following steps:

[0014] S21: Eliminate invalid data from smart grid security data through data cleaning operations;

[0015] S22: Normalization: Normalize the data to the range [0,1] according to the following formula:

[0016]

[0017] Where, X min and X max Respectively represent the minimum and maximum data in the experimental data set; X i is the i-th number in the experimental data, Y i For X i Normalized value.

[0018] Optionally, in S3, the number of features of the samples in the training sample data set is used as the original input feature, XGBOOST is used to obtain the importance of each feature, and then a recursive method is used to screen and retain its key features.

[0019] Optionally, in S4, 80% of the data in the smart grid security data set is determined as a training sample data set and 20% is determined as a test sample data set by using a holdout method, and a training sample set D=(X i ,Y i ) and the test sample set T=(X i ,Y i ).

[0020] Optionally, in S5, the chaos optimization algorithm and the improved Gaussian mutation algorithm are introduced into the sparrow algorithm to obtain an improved sparrow optimization algorithm. First, the chaos optimization algorithm is used to initialize the population size, maximum number of iterations, and space dimension in the sparrow algorithm. Then, the training sample data set constructed in step S4 is used to train the situation assessment model. The residual between the evaluated value obtained after training and the true value is used as the fitness function to find the optimal sparrow individual. Then, the improved Gaussian mutation algorithm is introduced to perform perturbation optimization on the optimal individual. Finally, the obtained optimal individual is used to optimize the probabilistic neural network PNN parameters to construct a smart grid security situation assessment model based on XGBOOST-ISSA-PNN.

[0021] Optionally, the S5 specifically includes the following steps:

[0022] S51: Model parameter initialization: set the initial parameters and smoothing factor of the PNN model σ Parameters, set n to the number of sparrow populations, iter max Indicates the maximum number of iterations, and uses the chaos optimization algorithm to initialize the sparrow population so that the population is evenly distributed and the positions are initialized. The formula of the Tent chaos optimization algorithm is as follows:

[0023]

[0024] Where: X t+1 is the chaotic value at the t+1th iteration; X t is the chaotic value at the t-th iteration; α is a constant in the interval [0,1], and α is set to α=0.75; T is the maximum number of iterations;

[0025] S52: PNN neural network training, using the residual between the trained evaluation value and the true value as the fitness function of the sparrow search algorithm to calculate the fitness value of each sparrow individual;

[0026] S53: Update the location of the discoverer, follower, and sentinel:

[0027] During each iteration, the update formula of the discoverer's position in the population is as follows:

[0028]

[0029] Where: t is the current iteration number; is the position of the i-th sparrow in the j-th dimension of the t-th generation, j = 1, 2, 3, m; iter maxis the maximum number of iterations; random value α∈(0,1); random value R2∈[0,1], is the warning value issued by the vigilant; safety threshold ST∈[0.5,1.0]; Q is a random value that satisfies the normal distribution of [0,1]; L is a one-dimensional matrix of 1×m, whose element values are all 1;

[0030] During each iteration, the follower's position update formula is as follows:

[0031]

[0032] Where: is the location of the best discoverer in the t+1 generation sparrow population; is the worst position of the t-th generation sparrow in the world; when i>n / 2, it means that the i-th follower with a poor position has no good food source and needs to fly to other places to forage, n is the total number of sparrows; A is a 1×m matrix, each element in the matrix is randomly 1 or -1, and the matrix A satisfies A * =A T (AA T ) -1 ;

[0033] During each iteration, the position update formula of the sentinel is as follows:

[0034]

[0035] Where: is the optimal position of the t-th generation sparrow population in the world; β and k are coefficients used to improve the step size, where β satisfies the standard normal distribution with mean 0 and variance 1; k is a random value in [-1, 1], indicating the direction of movement of the sparrow; f i is the fitness value of the current sparrow individual, and f w and f g are the fitness values of the global optimal position and the worst position of the current sparrow population respectively; ε is a constant;

[0036] S54: Perform an improved Gaussian mutation operation on the optimal sparrow individual and perform perturbation optimization on the optimal individual. The improved perturbation formula is as follows:

[0037] X nbest =X best [1+r]

[0038] r=0.05 / e (1-t / T)

[0039] Where: X nbest is the position after Gaussian perturbation; X bestis the current global optimal position; during the algorithm iteration process, r will decrease nonlinearly; the algorithm has a larger r value at the beginning, which can improve its global optimization ability; in the later stage of the algorithm iteration, the r value gradually decreases, and the algorithm's local optimization ability is improved;

[0040] S55: Determine whether the maximum number of iterations has been reached. If so, output the current optimal position to obtain the optimal smoothing factor σ parameter, thereby establishing the optimal smart grid security situation assessment model.

[0041] Optionally, the S6 specifically includes the following steps:

[0042] S61: Input the test sample data set obtained in step S4 into the XGBOOST-ISSA-PNN attack detection model with the optimal parameters to obtain its specific attack classification;

[0043] S62: Security Situation Value Calculation: Comprehensively consider the vulnerability types, vulnerability hazard levels, related equipment asset information, number of attacks, number of security products, and other information of the smart grid system to extract security factors. The security situation value of the smart grid is evaluated by considering the three security factors of vulnerability index V, threat index T, and asset security risk index A. The security situation value formula is then introduced to quantitatively evaluate the security situation of the smart grid, as shown below:

[0044] SA=g(ω1V+ω2T+ω3A)

[0045] Where SA represents the security situation value of a cycle; ω1, ω2, and ω3 are weights; V represents the quantitative value of the vulnerability index of the system after the attack; T represents the quantitative value of the threat index of the system after the attack; A represents the quantitative value of the asset security risk index of the system after the attack; and g(·) is the normalization function.

[0046] S63: Smart grid security situation assessment: Based on the attack detection situation and the situation value calculation formula, the security situation value of the smart grid is obtained. Combined with the security situation assessment level, the security situation assessment result of the smart grid is output.

[0047] The beneficial effects of the present invention are as follows: compared with traditional network security situation assessment methods, the XGBOOST feature selection method is first used to screen key features, solving the problem of high dimensionality of power grid security data. Then, the chaos optimization algorithm and the improved Gaussian mutation algorithm are used to improve the sparrow optimization algorithm, solving the problems of low solution accuracy and easy trapping in local optimal solutions of the sparrow optimization algorithm. The improved sparrow optimization algorithm is then used to optimize the smoothing factor parameters of the probabilistic neural network. Finally, the advantages of the probabilistic neural network model, such as fast training speed, high classification accuracy, and resistance to local optimal solutions, are utilized to evaluate the security situation value. Therefore, the present invention solves the problems of low accuracy in smart grid security situation assessment and unsatisfactory data feature selection.

[0048] Other advantages, objects, and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art upon examination of the following description or may be learned from practice of the present invention. The objects and other advantages of the present invention may be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below with reference to the accompanying drawings, in which:

[0050] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0051] The following describes the embodiments of the present invention by means of specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0052] Among them, the accompanying drawings are only for illustrative purposes and represent only schematic diagrams rather than actual pictures, and should not be understood as limiting the present invention. In order to better illustrate the embodiments of the present invention, some parts of the accompanying drawings may be omitted, enlarged or reduced, and do not represent the dimensions of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions may be omitted in the accompanying drawings.

[0053] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", "front", "back", etc. indicating directions or positional relationships, they are based on the directions or positional relationships shown in the drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0054] See also Figure 1 ,A flow chart of the smart grid security situation assessment algorithm based on XGBOOST-ISSA-PNN, which is suitable for smart grid security situation assessment,,the specific steps are as follows:

[0055] Step 1: Data preparation: Collect smart grid security data and security asset information to establish a smart grid security database;

[0056] Step 2: Data preprocessing: Perform data cleaning and normalization preprocessing on the security data in the database to obtain the smart grid security dataset;

[0057] Step 3: Take the number of features in the training sample dataset as the original input features, use XGBOOST to get the importance of each feature, and then use a recursive method to screen and retain its key features;

[0058] Step 4: Use the holdout method to determine 80% of the data in the smart grid security dataset as the training sample dataset and 20% as the test sample dataset, and construct the training sample set D = (X i ,Y i ) and the test sample set T=(X i ,Y i );

[0059] Step 5: Introduce the chaos optimization algorithm and the improved Gaussian mutation algorithm into the sparrow algorithm to obtain the improved sparrow optimization algorithm. First, use the chaos optimization algorithm to initialize the population size, maximum number of iterations, and space dimension in the sparrow algorithm. Then use the training sample data set constructed in step S4 to train the situation assessment model. Use the residual between the trained evaluation value and the true value as the fitness function to find the optimal sparrow individual. Then introduce the improved Gaussian mutation algorithm to perturb the optimal individual and finally use the obtained optimal individual to optimize the probabilistic neural network (PNN) parameters to construct a smart grid security situation assessment model based on XGBOOST-ISSA-PNN.

[0060] Step 6: Use the trained XGBOOST-ISSA-PNN attack detection model to perform attack detection on the sample test set, calculate the network security situation value based on the quantitative indicators of different attacks, and evaluate the security status of the network.

[0061] Optionally, in step 2, the data processing process includes the following steps:

[0062] Step 21: Through data cleaning operations, invalid data in the smart grid security data is eliminated, important information is retained, and the feature extraction effect is improved;

[0063] Step 22: Normalization: Normalize the data to the range [0,1] according to the following formula:

[0064]

[0065] Where, X min and X max Respectively represent the minimum and maximum data in the experimental data set. i is the i-th number in the experimental data, Y i For X i Normalized value.

[0066] In step 5, refer to Figure 1 As shown, according to the training sample D obtained in step 4, the optimal XGBOOST-ISSA-PNN model is trained, which specifically includes the following steps:

[0067] Step 51: Model parameter initialization: set the initial parameters of the PNN model, the smoothing factor σ parameter, set n to the number of sparrow populations, iter max Indicates the maximum number of iterations, and uses the chaos optimization algorithm to initialize the sparrow population so that the population is evenly distributed and the positions are initialized. The formula of the Tent chaos optimization algorithm is as follows:

[0068]

[0069] Where: X t+1 is the chaotic value at the t+1th iteration; X t is the chaotic value at the t-th iteration; α is a constant in the interval [0,1], and in this paper, α is taken as 0.75; T is the maximum number of iterations;

[0070] Step 52: PNN neural network training, using the residual between the trained evaluation value and the true value as the fitness function of the sparrow search algorithm to calculate the fitness value of each sparrow individual;

[0071] Step 53: Update the positions of the discoverer, follower, and sentinel: During each iteration, the position formula of the discoverer in the population is updated as follows:

[0072]

[0073] Where: t is the current iteration number; is the position of the i-th sparrow in the j-th dimension of the t-th generation, j = 1, 2, 3, m; iter max is the maximum number of iterations; random value α∈(0,1); random value R2∈[0,1], is the warning value issued by the vigilant; safety threshold ST∈[0.5,1.0]; Q is a random value that satisfies the normal distribution of [0,1]; L is a one-dimensional matrix of 1×m, whose element values are all 1;

[0074] During each iteration, the follower's position update formula is as follows:

[0075]

[0076] Where: is the location of the best discoverer in the t+1 generation sparrow population; is the worst position of the t-th generation sparrow in the world; when i>n / 2, it means that the i-th follower with a poor position has no good food source and needs to fly to other places to forage, n is the total number of sparrows; A is a 1×m matrix, each element in the matrix is randomly 1 or -1, and the matrix A satisfies A * =A T (AA T ) -1 ;

[0077] During each iteration, the position update formula of the sentinel is as follows:

[0078]

[0079] Where: is the optimal position of the t-th generation sparrow population in the world; β and k are coefficients used to improve the step size, where β satisfies the standard normal distribution with mean 0 and variance 1; k is a random value in [-1, 1], indicating the direction of movement of the sparrow; f i is the fitness value of the current sparrow individual, and f w and f g are the fitness values of the global optimal position and the worst position of the current sparrow population respectively; ε is a constant to avoid the situation where the denominator is equal to 0;

[0080] Step 54: Perform the improved Gaussian mutation operation on the optimal sparrow individual, and perform perturbation optimization on the optimal individual. The improved perturbation formula is as follows:

[0081] X nbest =X best [1+r]

[0082] r=0.05 / e (1-t / T)

[0083] Where: X nbest is the position after Gaussian perturbation; X best is the current global optimal position; during the algorithm iteration, r will decrease nonlinearly. The algorithm has a large r value at the beginning, which can improve its global optimization ability; in the later stages of the algorithm iteration, the r value gradually decreases, and the algorithm's local optimization ability is improved;

[0084] Step 55: Determine whether the maximum number of iterations has been reached. If so, output the current optimal position to obtain the optimal smoothing factor σ parameter, thereby establishing the optimal smart grid security situation assessment model.

[0085] In step 6, the test sample set T obtained in step 4 is input into the optimal attack detection model with retained parameters. Based on the detection results, the smart grid security situation value is quantitatively calculated to effectively evaluate the security status of the smart grid. Specifically, the following steps are included:

[0086] Step 61: Input the test sample dataset obtained in step S4 into the XGBOOST-ISSA-PNN attack detection model with the optimal parameters to obtain its specific attack classification;

[0087] Step 62: Calculate the security situation value: Comprehensively consider the vulnerability types, vulnerability hazard levels, related equipment asset information, number of attacks, number of security products, and other information of the smart grid system to extract security factors. The security situation value of the smart grid is evaluated by considering the three security factors of vulnerability index V, threat index T, and asset security risk index A. Then, the security situation value formula is introduced to quantitatively evaluate the security situation of the smart grid, as shown below:

[0088] SA=g(ω1V+ω2T+ω3A)

[0089] Where SA represents the security situation value of a cycle; ω1, ω2, and ω3 are weights; V represents the quantitative value of the vulnerability index of the system after the attack; T represents the quantitative value of the threat index of the system after the attack; A represents the quantitative value of the asset security risk index of the system after the attack; and g(·) is the normalization function.

[0090] Step 63: Smart grid security situation assessment: Based on the attack detection situation and the situation value calculation formula, the security situation value of the smart grid is obtained, and combined with the security situation assessment level, the security situation assessment result of the smart grid is output.

[0091] The present invention is applicable to the security situation assessment of smart grids. The smart grid security situation assessment method disclosed by the present invention is used. Since it integrates multiple algorithms and is performed in a sequential order, the data set is first cleaned, outlier processing is performed, and normalization operations are performed, which is beneficial to improving the assessment accuracy. Since the probabilistic neural network model tends to ignore the relevant features of a part of the category when processing high-dimensional data, resulting in poor generalization effect of the model, the XGBOOST feature screening method is used to effectively extract the features and extract the key features, so that the assessment accuracy is higher and the stability is better. The chaos optimization algorithm and the improved Gaussian mutation algorithm are used to avoid the sparrow algorithm from easily falling into the local optimal solution. The improved sparrow optimization algorithm is then used to select the optimal probabilistic neural network related parameters. Finally, the improved probabilistic neural network model is used to evaluate the security situation of the smart grid, which can achieve the effect of high assessment accuracy and strong convergence ability.

[0092] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions, which should all be included in the scope of the claims of the present invention.

Claims

1. A smart grid security situation assessment method based on an improved probabilistic neural network, characterized by: The method comprises the following steps: S1: Collect smart grid security data and security asset information and establish a smart grid security database; S2: Perform data cleaning and normalization preprocessing on the security data in the database to obtain the smart grid security dataset; S3: Input the dataset into XGBOOST for feature screening and extract its key features; S4: Use the holdout method to construct the training sample data set and the test sample data set; S5: Build an XGBOOST-ISSA-PNN smart grid security situation assessment model, and use the training set sample set to classify and train the model and retain its optimal parameters; introduce the chaos optimization algorithm and the improved Gaussian mutation algorithm into the sparrow algorithm to obtain an improved sparrow optimization algorithm. First, use the chaos optimization algorithm to initialize the population size, maximum number of iterations, and space dimension in the sparrow algorithm. Then use the training sample data set constructed in step S4 to train the situation assessment model. Use the residual between the trained evaluation value and the true value as the fitness function to find the optimal sparrow individual. Then introduce the improved Gaussian mutation algorithm to perturb the optimal individual and finally use the obtained optimal individual to optimize the probabilistic neural network parameters to construct a smart grid security situation assessment model based on XGBOOST-ISSA-PNN. The S5 specifically includes the following steps: S51: Model parameter initialization: set the initial parameters of the probabilistic neural network model, the smoothing factor σ parameter, set n to the number of sparrow populations, iter max Indicates the maximum number of iterations, and uses the chaos optimization algorithm to initialize the sparrow population so that the population is evenly distributed and the positions are initialized. The formula of the Tent chaos optimization algorithm is as follows: Where: X t+1 is the chaotic value at the t+1th iteration; X t is the chaotic value at the t-th iteration; α is a constant in the interval [0,1], and α is set to α=0.75; T is the maximum number of iterations; S52: PNN neural network training, using the residual between the trained evaluation value and the true value as the fitness function of the sparrow search algorithm to calculate the fitness value of each sparrow individual; S53: Update the location of the discoverer, follower, and sentinel: During each iteration, the update formula of the discoverer's position in the population is as follows: Where: t is the current iteration number; is the position of the i-th sparrow in the j-th dimension of the t-th generation, j = 1, 2, 3, m; iter max is the maximum number of iterations; random value α∈(0,1); random value R2∈[0,1], is the warning value issued by the vigilant; safety threshold ST∈[0.5,1.0]; Q is a random value that satisfies the normal distribution of [0,1]; L is a one-dimensional matrix of 1×m, whose element values are all 1; During each iteration, the follower's position update formula is as follows: Where: is the location of the best discoverer in the t+1 generation sparrow population; is the worst position of the t-th generation sparrow in the world; when i ＞ When n / 2, it means that the i-th follower with a poor position has no good food source and needs to fly to other places to find food. n represents the total number of sparrows. A is a 1×m matrix. Each element in the matrix is randomly 1 or -1, and the matrix A satisfies A * =A T (AA T ) -1 ; During each iteration, the position update formula of the sentinel is as follows: Where: is the optimal position of the t-th generation sparrow population in the world; β and k are coefficients used to improve the step size, where β satisfies the standard normal distribution with mean 0 and variance 1; k is a random value in [-1, 1], indicating the direction of movement of the sparrow; f i is the fitness value of the current sparrow individual, and f w and f g are the fitness values of the global optimal position and the worst position of the current sparrow population respectively; ε is a constant; S54: Perform an improved Gaussian mutation operation on the optimal sparrow individual and perform perturbation optimization on the optimal individual. The improved perturbation formula is as follows: X nbest =X best [1+r] r＝0.05 / e (1-t / T) Where: X nbest is the position after Gaussian perturbation; X best is the current global optimal position; during the algorithm iteration, r will decrease nonlinearly; the algorithm has a larger r value at the beginning, which can improve its global optimization ability; in the later iterations of the algorithm, the r value gradually decreases, and the local optimization ability of the algorithm is improved; S55: determining whether the maximum number of iterations has been reached, and if so, outputting the current optimal position, thereby obtaining the optimal smoothing factor σ parameter, and thereby establishing the optimal smart grid security situation assessment model; S6: Use the trained XGBOOST-ISSA-PNN attack detection model to perform attack detection on the sample test set, calculate the network security situation value based on the quantitative indicators of different attacks, and evaluate the security status of the network; S6 specifically includes the following steps: S61: Input the test sample data set obtained in step S4 into the XGBOOST-ISSA-PNN attack detection model with the optimal parameters to obtain its specific attack classification; S62: Security Situation Value Calculation: Comprehensively consider the vulnerability types, vulnerability hazard levels, related equipment asset information, number of attacks, number of security products, and other information of the smart grid system to extract security factors. The security situation value of the smart grid is evaluated by considering the three security factors of vulnerability index V, threat index T, and asset security risk index A. The security situation value formula is then introduced to quantitatively evaluate the security situation of the smart grid, as shown below: SA=g(ω1V+ω2T+ω3A) Where SA represents the security situation value of a cycle; ω1, ω2, and ω3 are weights; V represents the quantitative value of the vulnerability index of the system after the attack; T represents the quantitative value of the threat index of the system after the attack; A represents the quantitative value of the asset security risk index of the system after the attack; and g(·) is the normalization function. S63: Smart grid security situation assessment: Based on the attack detection situation and the situation value calculation formula, the security situation value of the smart grid is obtained. Combined with the security situation assessment level, the security situation assessment result of the smart grid is output.

2. The smart grid security situation assessment method based on an improved probabilistic neural network according to claim 1, characterized in that: The S2 specifically includes the following steps: S21: Eliminate invalid data from smart grid security data through data cleaning operations; S22: Normalization: Normalize the data to the range [0,1] according to the following formula: Where, X min and X max Respectively represent the minimum and maximum data in the experimental data set; X i is the i-th number in the experimental data, Y i For X i Normalized value.

3. The smart grid security situation assessment method based on an improved probabilistic neural network according to claim 1, characterized in that: In S3, the number of features of the samples in the training sample data set is used as the original input feature, XGBOOST is used to obtain the importance of each feature, and then a recursive method is used to screen and retain its key features.

4. The smart grid security situation assessment method based on an improved probabilistic neural network according to claim 1, characterized in that: In S4, 80% of the data of the smart grid security data set is determined as the training sample data set and 20% is determined as the test sample data set by the holdout method, and the training sample set D=(X i ,Y i ) and the test sample set T=(X i ,Y i ).

Images