Power SCADA system terminal dynamic trust evaluation method based on fuzzy logic

Abstract

The invention provides an electric power SCADA system terminal dynamic trust evaluation method based on fuzzy logic, and belongs to the technical field of electric power system information security. The method comprises the following steps: designing an overall architecture of a dynamic trust evaluation model, and defining key parameter setting and a definition range thereof; a parameter calculation module is constructed, terminal operation data are collected, and four input parameters including reliability parameters, safety parameters, response efficiency and historical satisfaction evaluation are calculated; constructing a fuzzy reasoning module, performing reasoning through a preset fuzzy rule base based on the input parameters, and outputting a direct trust value, a recommended trust value and a risk value; constructing a trust decision module, calculating a comprehensive trust value of the terminal in combination with a trust correlation value and a punishment mechanism, and outputting a trust level; defining an input and output mapping relation; and testing the effectiveness, the detection rate, the false detection rate and the fraud node identification capability of the model. According to the invention, real-time, accurate and dynamic evaluation of the trust state of the system terminal can be realized, and the security protection capability of the system is effectively improved.

Description

Technical Field

[0001] This invention belongs to the field of power system information security technology, specifically relating to a dynamic trust assessment method for terminals applicable to power data acquisition and monitoring control (SCADA) systems, and in particular, a method for constructing and analyzing a power SCADA system terminal trust assessment model based on fuzzy logic. Background Technology

[0002] With the continuous deepening of the intelligent and networked construction of power systems, the scale of power monitoring system terminal access continues to expand, making it a core hub connecting the information space and the physical power grid. During the data interaction and control command execution process between the terminal and the system, the reliability of its behavior directly determines the safety and stability of power production. However, existing security mechanisms generally focus on boundary protection and master station hardening, lacking effective dynamic assessment methods for the continuous trust status after terminal access. This makes it difficult to cope with the new security threat of "legitimate identity, malicious behavior" caused by terminal infiltration, which has become a weak link in the current power monitoring system security protection system.

[0003] Trust assessment of power monitoring system terminals faces significant challenges due to limited terminal resources, complex behavioral patterns, and high costs associated with manual monitoring. Therefore, a dynamic assessment mechanism capable of quantifying terminal trustworthiness in real time is urgently needed. Traditional static trust models or assessment methods based on single-dimensional indicators struggle to accurately depict the dynamic changes in terminal behavior, exhibiting lag in identifying malicious nodes and fraudulent activities. Furthermore, the assessment granularity is coarse, failing to provide precise evidence for fine-grained access control. Especially when dealing with the uncertainty and ambiguity of node behavior, existing methods often lack effective processing mechanisms, making it difficult to guarantee the accuracy and reliability of assessment results. Therefore, it is necessary to research a lightweight dynamic trust assessment method that integrates multi-dimensional behavioral characteristics, possesses uncertainty handling capabilities, and can rapidly respond to risk changes, thereby enhancing the proactive defense capabilities of power monitoring systems. Summary of the Invention

[0004] The purpose of this invention is to address the problems of existing trust assessment models in power SCADA system terminal applications, such as limited assessment dimensions, poor dynamic adaptability, and insufficient malicious behavior identification capabilities. Therefore, this invention provides a dynamic trust assessment method for power SCADA system terminals based on fuzzy logic. This method achieves real-time, accurate, and dynamic assessment of terminal trust values ​​by constructing multi-dimensional trust feature parameters, introducing a fuzzy inference mechanism, and a dynamic penalty strategy.

[0005] The objective of this invention is achieved through the following technical solution:

[0006] A dynamic trust assessment method for power SCADA system terminals based on fuzzy logic includes the following steps:

[0007] Step S1: Design the overall architecture of the dynamic trust assessment model, including: parameter calculation module, fuzzy reasoning module and trust decision module;

[0008] Step S2: Clarify the key parameter settings and their definition range for the dynamic trust assessment model in Step S1;

[0009] Step S3: Construct a parameter calculation module to collect the operating data of the power SCADA system terminal and calculate four input parameters: reliability parameters, safety parameters, response efficiency, and historical satisfaction evaluation.

[0010] Step S4: Construct a fuzzy reasoning module. Based on the key parameters of the dynamic trust evaluation model obtained in Step S2, perform fuzzy reasoning through a preset fuzzy rule base and output three types of trust-related values: direct trust value, recommended trust value, and risk value.

[0011] Step S5: Construct a trust decision module, combine the three types of trust-related values ​​and penalty mechanisms output in step S4, calculate the comprehensive trust value of the power SCADA system terminal, and output the trust level;

[0012] Step S6: Perform performance testing on the dynamic trust assessment model through a simulation platform to verify the effectiveness of the dynamic trust assessment model in terms of malicious node detection rate, false detection rate, and fraud node identification.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0014] 1. This invention comprehensively analyzes the multi-dimensional behavioral characteristics of power SCADA system terminals and establishes a four-dimensional evaluation index system including reliability parameters, security parameters, response efficiency, and historical satisfaction evaluation. The parameter calculation module enables precise quantification of each dimension. Specifically, reliability parameters are quantified through the access resource anomaly rate; security parameters are characterized by the illegal connection rate and malicious recommendation rate; response efficiency comprehensively considers the interaction success rate and transaction time; and historical satisfaction evaluation is calculated by the average evaluation of adjacent nodes, forming a comprehensive and objective foundation for trust assessment.

[0015] 2. This invention employs a trust reasoning method based on fuzzy logic, constructing a fuzzy rule base containing 81 rules. It addresses the uncertainty and fuzziness in the trust assessment process through triangular membership functions. A centroid method is used for defuzzification, mapping the four precise input parameters to three precise outputs: direct trust value, recommended trust value, and risk value. This effectively solves the shortcomings of traditional trust assessment models in handling subjective and fuzzy information.

[0016] 3. This invention introduces a dynamic penalty mechanism and a weight allocation strategy based on entropy weighting, comprehensively considering the combined effects of direct trust value, recommended trust value, and risk value in the trust decision-making module. It dynamically punishes malicious behavior through penalty factors and objectively determines the weights of each trust indicator using entropy weighting, avoiding biases caused by subjective weighting and ensuring the objectivity and fairness of trust assessment.

[0017] 4. This invention verifies the effectiveness of the dynamic trust assessment model through extensive simulation experiments. The results show that the model outperforms traditional trust assessment models in terms of malicious node detection rate, false detection rate, and fraudulent node identification. Particularly in identifying fraudulent nodes, due to the introduction of risk mechanisms and penalty factors, the model can quickly respond to malicious behavior of nodes, achieving rapid decay of trust values ​​and significantly improving the system's security protection capabilities.

[0018] 5. The dynamic trust assessment model proposed in this invention has good practicality and scalability. The model has a clear architecture and well-defined functions for each module. It can not only meet the high standards of power SCADA systems for secure terminal access, but also provide a reliable trust decision basis for subsequent dynamic access control, and has important engineering application value. Attached Figure Description

[0019] Figure 1 This is a flowchart of the dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to the present invention.

[0020] Figure 2 This is a flowchart illustrating the dynamic trust assessment method for power SCADA system terminals based on fuzzy logic, as described in this invention.

[0021] Figure 3 This is an architecture diagram of a dynamic trust assessment model based on fuzzy logic.

[0022] Figure 4 This is a schematic diagram of fuzzy reasoning and defuzzification.

[0023] Figure 5 This is a flowchart of the fuzzy inference module.

[0024] Figure 6 For fuzzy regular surface plots; where:

[0025] Figure 6 (a) is and right The influence of fuzzy rule surface plot;

[0026] Figure 6 (b) is and right The influence of fuzzy rule surface plot;

[0027] Figure 6 (c) is and right The influence of fuzzy rule surface plot;

[0028] Figure 6 (d) is and right The influence of fuzzy rule surface plot;

[0029] Figure 6 (e) is and right The influence of fuzzy rule surface plot;

[0030] Figure 6 (f) is and right The influence of fuzzy rule surface plot.

[0031] Figure 7 This is a graph showing how the overall trust value changes with the number of interactions.

[0032] Figure 8 This is a graph showing the change in detection rate as a function of malicious nodes.

[0033] Figure 9 This is a graph showing the change in false detection rate as a function of malicious nodes.

[0034] Figure 10 This is a graph showing how the overall trust value of a fraudulent node changes with the number of interactions. Detailed Implementation

[0035] The present invention will be further described in detail below with reference to the accompanying drawings: This embodiment is implemented under the premise of the technical solution of the present invention, and detailed implementation methods are given, but the protection scope of the present invention is not limited to the following embodiments.

[0036] like Figure 1 As shown, the dynamic trust assessment method for power SCADA system terminals based on fuzzy logic involved in this embodiment includes the following steps:

[0037] Step S1: Design the overall architecture of the dynamic trust assessment model;

[0038] Step S2: Construct a parameter calculation module to collect operating data from the power SCADA system terminal and calculate four input parameters: reliability parameters, safety parameters, response efficiency, and historical satisfaction evaluation.

[0039] Step S3: Construct a fuzzy reasoning module. Based on the key parameters of the dynamic trust assessment model obtained in Step S2, perform fuzzy reasoning through a preset fuzzy rule base and output three types of trust-related values: direct trust value, recommended trust value, and risk value.

[0040] Step S4: Construct a trust decision module. Combining the three types of trust-related values ​​and penalty mechanisms output in Step S4, calculate the comprehensive trust value of the power SCADA system terminal and output the trust level.

[0041] Step S5: Perform performance testing on the dynamic trust assessment model through a simulation platform to verify the effectiveness of the dynamic trust assessment model in terms of malicious node detection rate, false detection rate, and fraud node identification.

[0042] like Figure 2 As shown, the specific implementation steps are as follows:

[0043] Step S1: Design the overall architecture of the dynamic trust assessment model.

[0044] In this step, addressing the inherent subjectivity, ambiguity, quantifiability, and time decay characteristics of trust, a Dynamic Trust Evaluation Model Based on Fuzzy Logic (DTMFL) is constructed. This model includes a parameter calculation module, a fuzzy inference module, and a trust decision module. The parameter calculation module calculates input parameters such as reliability, security, response efficiency, and historical satisfaction ratings based on terminal operational data. The fuzzy inference module, based on the output of the parameter calculation module, performs inference through a pre-defined fuzzy rule base, outputting a direct trust value, a recommended trust value, and a risk value. The trust decision module combines the trust-related values ​​output by the fuzzy inference module with a built-in penalty mechanism to calculate the terminal's comprehensive trust value and output a trust level, providing reliable data support for subsequent dynamic access control. This model, combined with a dynamic access control mechanism, enables refined and dynamic permission management of terminal devices. The model architecture is as follows: Figure 3 As shown.

[0045] Step S2: Clarify the key parameter settings and their definition range for the model.

[0046] In this step, the key parameter settings for the constructed dynamic trust assessment model include: the definitions and value ranges of input and output parameters, as shown in Table 1, where: input parameters include reliability parameters. Safety parameters Response efficiency Historical satisfaction evaluation The input to the parameter calculation module is the direct trust value, the recommended trust value, and the risk value, which are the outputs of the fuzzy inference module; the output is the comprehensive trust value. This serves as the final output of the trust decision-making module; the overall trust value. The formation of trust is not only influenced by direct trust and recommended trust, but also by the combined effect of risk factors. In turn, direct trust, recommended trust and risk are further influenced by reliability parameters, security parameters, response efficiency and historical satisfaction evaluation. These factors are coupled with each other and jointly determine the overall level of trust in a certain node in the system.

[0047] Table 1 Input / output parameters and their definitions for the trust assessment model

[0048]

[0049] Step S3: Construct a trust assessment parameter calculation module, collect terminal operation data, and calculate four input parameters: reliability parameter, security parameter, response efficiency, and historical satisfaction evaluation.

[0050] In this step, the parameter calculation module is defined to include four input parameters: reliability parameter, safety parameter, response efficiency, and historical satisfaction evaluation. The corresponding quantitative calculation formulas are given, and the triangular membership function is used. The membership functions of each parameter are shown in Table 2.

[0051] Table 2 Membership Functions of Input Parameters

[0052]

[0053] Among them, Table 2 This indicates the value of the indicator. These represent the membership functions for low, medium, and high reliability, respectively. These represent the membership functions for low, medium, and high security, respectively. These represent the membership functions for low, medium, and high response efficiency, respectively. These represent the membership functions for low, medium, and high historical satisfaction ratings, respectively.

[0054] 1. Reliability parameters

[0055] Reliability parameters The abnormal rate of resource access by terminal device nodes is used for quantification, as shown in equation (1):

[0056] (1)

[0057] In the formula, —Number of times resource access errors occurred;

[0058] —Total number of resource accesses.

[0059] 2. Safety

[0060] Security The illegal connection rate and malicious recommendation rate of nodes are used to characterize this, as shown in equation (2):

[0061] (2)

[0062] In the formula, —The number of illegal connections to the node;

[0063] —Number of malicious recommendations by the node;

[0064] —The total number of times a node connects to and recommends another node.

[0065] 3. Response efficiency

[0066] Response efficiency The success rate of interactions between nodes and the transaction time are used to characterize this, as shown in equation (3):

[0067] (3)

[0068] In the formula, —Number of successful interactions;

[0069] —Total number of interactions;

[0070] —The time it takes for one interaction between nodes.

[0071] 4. Historical satisfaction evaluation

[0072] Historical satisfaction rating Used for nodes Directly connected node-to-node The satisfaction evaluation is represented by the expression shown in equation (4):

[0073] (4)

[0074] In the formula, — Other node-to-node The satisfaction rating score;

[0075] —Total number of interactions.

[0076] Step S4: Construct a fuzzy inference module. Based on the parameters obtained in step S3, perform fuzzy inference through a preset fuzzy rule base and output direct trust value, recommended trust value and risk value.

[0077] In this step, the fuzzy inference module constructs a rule base of 81 fuzzy rules, i.e., the preset fuzzy rule base, and outputs direct trust values, recommended trust values, and risk values ​​through the fuzzy inference mechanism. The membership functions of each output parameter are shown in Table 3. The definitions and mathematical expressions of each parameter are as follows:

[0078] 1. Direct Trust Value

[0079] The direct trust value refers to the degree of trust a node has in another node based on its existing interaction experience, and its expression is shown in equation (5):

[0080] (5)

[0081] In the formula, —Attenuation factor;

[0082] —Adjustment coefficient of the attenuation factor;

[0083] --node The average time interval between interactions with directly connected nodes;

[0084] —Direct trust value of fuzzy inference output.

[0085] 2. Recommended Trust Value

[0086] Recommendation trust is the degree of trust derived from the suggestions, evaluations, or recommendations of other nodes. The formula for calculating the recommendation trust value P2 is shown in equation (6):

[0087] (6)

[0088] (7)

[0089] (8)

[0090] (9)

[0091] (10)

[0092] In the formula, --node Recommended trust value;

[0093] ——No. The weight values ​​of each recommended path;

[0094] ——No. The recommendation trust value of each recommended path;

[0095] --path The endpoint;

[0096] --node and nodes Direct trust value between them;

[0097] --node and nodes The similarity between them;

[0098] —The recommended trust value output by fuzzy inference;

[0099] —Adjust the parameter growth rate control factor.

[0100] 3. Risk Value

[0101] Risk Value The risk level and risk value output by fuzzy inference. The expression is shown in equation (11).

[0102] (11)

[0103] in, The risk value output by fuzzy inference.

[0104] Depend on Figure 4 It can be seen that the fuzzy logic reasoning module can interpret the input through fuzzy reasoning rules. , , , Mapped to , and The fuzzy output is obtained by using the centroid method. , and Deblurring yields the precise value.

[0105] like Figure 5 As shown, the fuzzy inference module process is as follows:

[0106] Step S41: Variable fuzzification. This involves fuzzifying the input variables. , , , By converting the membership function set in Table 3 into a fuzzy set, the precise variables are mapped to the fuzzy variables.

[0107] Step S42: Define the fuzzy rule set. The system establishes a fuzzy rule set describing the input and output variables, transforming the membership distribution of the input fuzzy set into the output fuzzy set through conditional statement rules, thereby characterizing the logical relationships between variables.

[0108] Step S43: Fuzzy Inference. This involves implementing the inference process of mapping the input to the output fuzzy set based on a fuzzy rule base.

[0109] Step S44: Defuzzification. In the fuzzy inference system, in order to transform the output fuzzy set into a specific numerical value, this invention uses the centroid method for defuzzification. The basic idea of ​​the centroid method is to determine the centroid position based on the weighted average of the membership degrees of the fuzzy set. The centroid position can be calculated using the following formula:

[0110] (12)

[0111] In the formula, —The position of the center of gravity is blurred;

[0112] ——Fuzzy set exist The membership function at a given location.

[0113] Table 3 Membership Function of Output Parameters

[0114]

[0115] Among them, Table 3 This indicates the value of the indicator. These represent the membership functions for low, medium, and high direct trust values, respectively. These represent the membership functions for low, medium, and high recommendation trust values, respectively. These represent the membership functions for low, medium, and high risk, respectively.

[0116] The 81 fuzzy rules in the preset fuzzy rule base are as follows:

[0117] Rule 1. If (reliability is low) and (security is low) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0118] Rule 2. If (reliability is low) and (security is low) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0119] Rule 3. If (reliability is low) and (security is low) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0120] Rule 4. If (reliability is low) and (security is low) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0121] Rule 5. If (reliability is low) and (security is low) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is low) (recommended trust rating is low) (risk is high);

[0122] Rule 6. If (reliability is low) and (security is low) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0123] Rule 7. If (reliability is low) and (security is low) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0124] Rule 8. If (reliability is low) and (security is low) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is low) (risk is high).

[0125] Rule 9. If (reliability is low) and (security is low) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0126] Rule 10. If (reliability is low) and (security is medium) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0127] Rule 11. If (reliability is low) and (security is medium) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0128] Rule 12. If (reliability is low) and (security is medium) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0129] Rule 13. If (reliability is low) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is medium) (risk is medium).

[0130] Rule 14. If (reliability is low) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0131] Rule 15. If (reliability is low) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0132] Rule 16. If (reliability is low) and (security is medium) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is medium) (risk is medium).

[0133] Rule 17. If (reliability is low) and (security is medium) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is high) (recommended trust rating is medium) (risk is medium).

[0134] Rule 18. If (reliability is low) and (security is medium) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is medium).

[0135] Rule 19. If (reliability is low) and (security is high) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0136] Rule 20. If (reliability is low) and (security is high) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0137] Rule 21. If (reliability is low) and (security is high) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0138] Rule 22. If (reliability is low) and (security is high) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0139] Rule 23. If (reliability is low) and (security is high) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0140] Rule 24. If (reliability is low) and (security is high) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0141] Rule 25. If (reliability is low) and (security is high) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0142] Rule 26. If (reliability is low) and (security is high) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0143] Rule 27. If (reliability is low) and (security is high) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is low);

[0144] Rule 28. If (Reliability is Medium) and (Security is Low) and (Efficiency is Low) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Low) (Recommended Trust Rating is Low) (Risk is High);

[0145] Rule 29. If (Reliability is Medium) and (Security is Low) and (Efficiency is Low) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Low) (Recommended Trust Rating is Low) (Risk is High);

[0146] Rule 30. If (Reliability is Medium) and (Security is Low) and (Efficiency is Low) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Low) (Risk is High);

[0147] Rule 31. If (reliability is medium) and (security is low) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high);

[0148] Rule 32. If (reliability is medium) and (security is low) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0149] Rule 33. If (Reliability is Medium) and (Security is Low) and (Efficiency is Medium) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0150] Rule 34. If (Reliability is Medium) and (Security is Low) and (Efficiency is High) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Low) (Recommended Trust Rating is Low) (Risk is High);

[0151] Rule 35. If (Reliability is Medium) and (Security is Low) and (Efficiency is High) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0152] Rule 36. If (Reliability is Medium) and (Security is Low) and (Efficiency is High) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0153] Rule 37. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Low) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Low) (Recommended Trust Rating is Low) (Risk is Medium);

[0154] Rule 38. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Low) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0155] Rule 39. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Low) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0156] Rule 40. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Medium) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0157] Rule 41. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Medium) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0158] Rule 42. If (Reliability is Medium) and (Security is Medium) and (Efficiency is Medium) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is High) (Recommended Trust Rating is Medium) (Risk is Medium);

[0159] Rule 43. If (Reliability is Medium) and (Security is Medium) and (Efficiency is High) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Medium);

[0160] Rule 44. If (Reliability is Medium) and (Security is Medium) and (Efficiency is High) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is High) (Recommended Trust Rating is Medium) (Risk is Medium);

[0161] Rule 45. If (Reliability is Medium) and (Security is Medium) and (Efficiency is High) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is High) (Recommended Trust Rating is High) (Risk is Low);

[0162] Rule 46. If (Reliability is Medium) and (Security is High) and (Efficiency is Low) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Low) (Recommended Trust Rating is Low) (Risk is Medium);

[0163] Rule 47. If (Reliability is Medium) and (Security is High) and (Efficiency is Low) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Low);

[0164] Rule 48. If (Reliability is Medium) and (Security is High) and (Efficiency is Low) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Low);

[0165] Rule 49. If (Reliability is Medium) and (Security is High) and (Efficiency is Medium) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Low) (Recommended Trust Rating is Medium) (Risk is Low);

[0166] Rule 50. If (Reliability is Medium) and (Security is High) and (Efficiency is Medium) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Low);

[0167] Rule 51. If (Reliability is Medium) and (Security is High) and (Efficiency is Medium) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is High) (Recommended Trust Rating is High) (Risk is Low);

[0168] Rule 52. If (Reliability is Medium) and (Security is High) and (Efficiency is High) and (Historical Satisfaction Rating is Low), then (Direct Trust Rating is Medium) (Recommended Trust Rating is Medium) (Risk is Low);

[0169] Rule 53. If (Reliability is Medium) and (Security is High) and (Efficiency is High) and (Historical Satisfaction Rating is Medium), then (Direct Trust Rating is Medium) (Recommended Trust Rating is High) (Risk is Low);

[0170] Rule 54. If (Reliability is Medium) and (Security is High) and (Efficiency is High) and (Historical Satisfaction Rating is High), then (Direct Trust Rating is High) (Recommended Trust Rating is High) (Risk is Low);

[0171] Rule 55. If (reliability is high) and (security is low) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0172] Rule 56. If (reliability is high) and (security is low) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0173] Rule 57. If (reliability is high) and (security is low) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0174] Rule 58. If (reliability is high) and (security is low) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is high).

[0175] Rule 59. If (reliability is high) and (security is low) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0176] Rule 60. If (reliability is high) and (security is low) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is medium).

[0177] Rule 61. If (reliability is high) and (security is low) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is medium) (risk is medium).

[0178] Rule 62. If (reliability is high) and (security is low) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium);

[0179] Rule 63. If (reliability is high) and (security is low) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is medium).

[0180] Rule 64. If (reliability is high) and (security is medium) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0181] Rule 65. If (reliability is high) and (security is medium) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0182] Rule 66. If (reliability is high) and (security is medium) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0183] Rule 67. If (reliability is high) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0184] Rule 68. If (reliability is high) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is low);

[0185] Rule 69. If (reliability is high) and (security is medium) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is low);

[0186] Rule 70. If (reliability is high) and (security is medium) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is medium) (recommended trust rating is medium) (risk is medium).

[0187] Rule 71. If (reliability is high) and (security is medium) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is high) (recommended trust rating is high) (risk is low);

[0188] Rule 72. If (reliability is high) and (security is medium) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is high) (risk is low).

[0189] Rule 73. If (reliability is high) and (security is high) and (efficiency is low) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is low) (risk is medium).

[0190] Rule 74. If (reliability is high) and (security is high) and (efficiency is low) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is low) (risk is medium).

[0191] Rule 75. If (reliability is high) and (security is high) and (efficiency is low) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is medium) (risk is low);

[0192] Rule 76. If (reliability is high) and (security is high) and (efficiency is medium) and (historical satisfaction rating is low), then (direct trust rating is low) (recommended trust rating is medium) (risk is medium).

[0193] Rule 77. If (reliability is high) and (security is high) and (efficiency is medium) and (historical satisfaction rating is medium), then (direct trust rating is medium) (recommended trust rating is medium) (risk is low);

[0194] Rule 78. If (reliability is high) and (security is high) and (efficiency is medium) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is high) (risk is low);

[0195] Rule 79. If (reliability is high) and (security is high) and (efficiency is high) and (historical satisfaction rating is low), then (direct trust rating is medium) (recommended trust rating is medium) (risk is low);

[0196] Rule 80. If (reliability is high) and (security is high) and (efficiency is high) and (historical satisfaction rating is medium), then (direct trust rating is high) (recommended trust rating is high) (risk is low);

[0197] Rule 81. If (reliability is high) and (security is high) and (efficiency is high) and (historical satisfaction rating is high), then (direct trust rating is high) (recommended trust rating is high) (risk is low).

[0198] Step S5: Construct a trust decision module, combine the three types of trust-related values ​​and penalty mechanisms output in step S3, calculate the terminal's comprehensive trust value, and output the trust level.

[0199] In this step, the trust decision module obtains the information from the fuzzy inference module. , , Calculate the overall trust value , where: the comprehensive trust value is a comprehensive assessment of the trust level of a node in a specific context; the model introduces a penalty mechanism in the trust decision module, and a node will be punished when it performs malicious operations. The penalty function is shown in equation (13).

[0200] (13)

[0201] In the formula, —Punishment factor;

[0202] —Number of malicious interactions;

[0203] —Total number of interactions.

[0204] Overall Trust Value As shown in equation (14):

[0205] (14)

[0206] In the formula, —Direct Trust Value The weights;

[0207] —Recommended Trust Value The weights;

[0208] —Risk Value The weight.

[0209] In the above expression This model uses the entropy weight method to determine... , and The value, assuming the first The first sample The indicators are First, regarding the indicators Normalization is performed. , , The weights are calculated as shown in equation (15):

[0210] (15)

[0211] (16)

[0212] (17)

[0213] (18)

[0214] In the formula, ——No. Information entropy redundancy of the indicator;

[0215] ——No. The entropy value of the indicator;

[0216] ——No. The first item under the indicator The weight of each sample .

[0217] Step S6: Analyze the fuzzy regular surface to clarify the input-output mapping relationship.

[0218] In this step, DTMFL model code is written in simulation software and fuzzy regular surface plots are generated to evaluate the nonlinear mapping relationship between input and output variables, where: by Figure 6 From (a) and (b), we can see that the direct trust value Response efficiency and historical satisfaction evaluation The sensitivity is higher, especially when both inputs are greater than 0.8. To obtain the maximum value; by Figure 6 From (c) and (d), we can see that when security and response efficiency When the value is greater than 0.5, the trust value for the recommendation is... The effect stabilizes at 0.6, therefore right and The sensitivity is higher, and when and When all are greater than 0.8 To obtain the maximum value; by Figure 6 From (e) and (f), we can see that the risk value For security Most sensitive, even in reliability parameters At higher levels, if A level below 0.5 still presents a high risk.

[0219] Step S7: Conduct a model validity test to verify the rationality of the dynamic changes in the trust value.

[0220] In this step, the effectiveness of the DTMFL model is verified by analyzing the dynamic characteristics of the overall trust value of network nodes changing with the frequency of interaction. Specifically, three types of typical nodes—reliable nodes, unstable nodes, and malicious nodes—are randomly selected from the simulated network, with an initial trust value of 0.5. Ten nodes are randomly selected and each of the three types of nodes interacts with 20 times. The test results are as follows: Figure 7 As shown, after 20 interactions, the trust value of reliable nodes showed a stable upward trend, while the trust value of malicious nodes continued to decline. The trust value of unstable nodes fluctuated due to the uncertainty of their behavior. The test results show that the DTMFL model, combined with risk and penalty mechanisms, has a rapid response characteristic to abnormal behavior. The rate of increase in node trust value is lower than the rate of decrease, which achieves accurate differentiation between reliable and malicious nodes and verifies the rationality of trust value calculation.

[0221] Step S8: Conduct model detection rate and false detection rate tests to evaluate security performance.

[0222] In this step, the performance of three different access control models in terms of detection rate and false positive rate is evaluated. The comparison models include: the DTMFL model, the TAMFIS model without risk assessment, and the BBRMF model with risk assessment; Figure 8 As shown, the detection rate varies with the proportion of malicious nodes. Due to the introduction of risk mechanisms and penalty factors, the DTMFL model has a significantly higher detection rate for malicious nodes than the TAMFIS model, and is superior to the BBRMF model. Figure 9 As shown, the false detection rate varies with the proportion of malicious nodes. The DTMFL model maintains the lowest false detection rate, effectively reducing the misjudgment of malicious nodes. The combined results of detection rate and false detection rate indicate that the DTMFL model has the advantages of high detection rate and low false detection rate, effectively improving the security protection capability of the power system.

[0223] Step S9: Conduct a fraud node identification capability test to verify the model's resistance to spoofing attacks.

[0224] This step verifies the DTMFL model's ability to identify fraudulent nodes in the network. Fraudulent nodes are set to begin malicious behavior after accumulating a certain trust value; the detection performance of the three models—TMFIS, BBRMF, and DTMFL—is compared. Figure 10As shown, the change in the overall trust value of fraudulent nodes with the number of interactions is illustrated. The test results show that the growth rate of the overall trust value of nodes in the DTMFL model is lower than that of the TMFIS and BBRMF models. When fraudulent nodes exhibit malicious behavior, the trust value decay rate of the DTMFL model is the fastest, followed by the BBRMF model, and the TMFIS model is the slowest. This is because the DTMFL model has both a risk mechanism and a penalty factor, thus exhibiting the best detection performance for fraudulent nodes.

[0225] The above description is merely a preferred embodiment of the present invention. These specific embodiments are different implementations based on the overall concept of the present invention, and the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A dynamic trust assessment method for power SCADA system terminals based on fuzzy logic, characterized in that, Includes the following steps: Step S1: Design the overall architecture of the dynamic trust assessment model, including: parameter calculation module, fuzzy inference module and trust decision module; Step S2: Clarify the key parameter settings and their definition range for the dynamic trust assessment model in Step S1; Step S3: Construct a parameter calculation module to collect operating data from the power SCADA system terminal and calculate four input parameters: reliability parameters, safety parameters, response efficiency, and historical satisfaction evaluation. Step S4: Construct a fuzzy reasoning module. Based on the key parameters of the dynamic trust evaluation model obtained in Step S2, perform fuzzy reasoning through a preset fuzzy rule base and output three types of trust-related values: direct trust value, recommended trust value, and risk value. Step S5: Construct a trust decision module, combine the three types of trust-related values ​​and penalty mechanisms output in step S4, calculate the comprehensive trust value of the power SCADA system terminal, and output the trust level; Step S6: Perform performance testing on the dynamic trust assessment model through a simulation platform to verify the effectiveness of the dynamic trust assessment model in terms of malicious node detection rate, false detection rate, and fraud node identification.

2. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 1, characterized in that, The key parameter settings for the dynamic trust assessment model in step S2 include: the definition and value range of input and output parameters; input parameters include: reliability parameters. Safety parameters Response efficiency Historical satisfaction evaluation The input parameters are: direct trust value, recommended trust value, and risk value, which are used as inputs to the parameter calculation module. The output parameters include: direct trust value, recommended trust value, and risk value, which are used as outputs to the fuzzy reasoning module. The comprehensive trust value is the final output of the trust decision module. The comprehensive trust value is affected by the combined effects of direct trust, recommended trust, and risk factors, and is determined by the combined influence of each input parameter.

3. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 1, characterized in that, The reliability parameters mentioned in step S3 The security parameter is quantified by the abnormal access rate of terminal equipment nodes in the power SCADA system to access resources. The response efficiency is characterized by the illegal connection rate and malicious recommendation rate of nodes. Historical satisfaction evaluation is characterized by the success rate of interactions between nodes and transaction time. Used for nodes Directly connected node-to-node Satisfaction evaluation characteristics.

4. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 3, characterized in that, The specific steps for performing fuzzy inference using a preset fuzzy rule base as described in step S4 are as follows: Step S41: Variable fuzzification, fuzzing the input variables , , , By converting the variables into fuzzy sets using a defined membership function, precise variables are mapped to fuzzy variables. Step S42: Define a fuzzy rule set, establish a fuzzy rule set describing the input and output variables, and transform the membership distribution of the input fuzzy set into the output fuzzy set through conditional statement rules, thereby characterizing the logical relationship between variables; Step S43: Fuzzy reasoning, which is the reasoning process of mapping input to output fuzzy sets based on the fuzzy rule base; Step S44: Defuzzification. In the fuzzy inference module, in order to convert the output fuzzy set into specific numerical values, the centroid method is used for defuzzification. The basic idea of ​​the centroid method is to determine the centroid position based on the weighted average of the membership degrees of the fuzzy set. The centroid position is calculated using the following formula: In the formula, To blur the position of the center of gravity; For fuzzy sets exist The membership function at a given location.

5. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 4, characterized in that, The definitions and mathematical expressions of the direct trust value, recommended trust value, and risk value in step S4 are as follows: (1) Direct Trust Value Direct Trust Value This refers to the degree of trust a node has in another node based on its existing interaction experience, and its expression is as follows: In the formula, It is the attenuation factor; This is the adjustment coefficient for the attenuation factor; For nodes The average time interval between interactions with directly connected nodes; The direct trust value output by fuzzy inference; (2) Recommendation Trust Value Recommended Trust Value The level of trust is derived from the suggestions, evaluations, or recommendations of other nodes; the recommended trust value is... The calculation formula is as follows: In the formula, For nodes Recommended trust value; For the first The weight values ​​of each recommended path; For the first The recommendation trust value of each recommended path; For path The endpoint; For nodes and nodes Direct trust value between them; For nodes and nodes The similarity between them; The recommended trust value output by fuzzy inference; To correct the parameter growth rate control factor; (3) Risk Value Risk Value The expression for the risk level output by fuzzy inference is as follows: In the formula, The risk value output by fuzzy inference.

6. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 5, characterized in that, In step S5, the trust decision module obtains the information from the fuzzy inference module. , , Calculate the overall trust value Among them, the overall trust value It is a comprehensive assessment of the trust level of a node in a specific context; the dynamic trust assessment model introduces a penalty mechanism in the trust decision module, which will punish nodes when they perform malicious operations.

7. The method for dynamic trust assessment of power SCADA system terminals based on fuzzy logic according to claim 1, characterized in that, The specific steps for performing performance testing on the dynamic trust evaluation model using a simulation platform in step S6 are as follows: Step S61: Generate a fuzzy rule surface based on the dynamic trust assessment model, and analyze the fuzzy rule surface to clarify the mapping relationship between input variables and output variables; Step S62: Conduct an effectiveness test on the dynamic trust assessment model to verify the rationality of the dynamic change of the comprehensive trust value in step S5; Step S63: Conduct tests on the detection rate and false detection rate of the dynamic trust assessment model to evaluate its security performance; Step S64: Conduct a fraud node identification capability test to verify the resistance of the dynamic trust assessment model to spoofing attacks.

8. The dynamic trust assessment method for power SCADA system terminals based on fuzzy logic according to claim 7, characterized in that, The method for testing the effectiveness of the dynamic trust assessment model described in step S62 is as follows: the effectiveness of the dynamic trust assessment model is verified by analyzing the dynamic characteristics of the changes in the comprehensive trust value of network nodes with the frequency of interaction.

Images