CN110472839A - Thermal power plant's control system Information Security Evaluation system based on SA-PSO-AHP - Google Patents

Abstract

The present invention relates to a thermal power plant control system information security assessment system based on SA-PSO-AHP, comprising a data input module for obtaining scoring data of the control system to be evaluated; The threat assessment layered progressive model of the built control system, the model combines the scoring data input by the data input module to construct the assessment matrix; the matrix correction module is used to judge the consistency of the assessment matrix, and the assessment matrix The simulated annealing optimization particle swarm algorithm is used for consistency correction; the data output module is used to obtain the threat ranking of the power plant control system by using the hierarchical algorithm, and obtain the comprehensive threat score combined with the threat frequency. Compared with the prior art, the present invention can fully reflect the ambiguity of the evaluation elements and process, reduce the influence of personal subjective factors, and better obtain objective and sufficient evaluation results.

Description

SA-PSO-AHP-based information security assessment system for thermal power plant control system

technical field

The invention relates to the technical field of information security, in particular to an SA-PSO-AHP-based thermal power plant control system information security evaluation system.

Background technique

Thermal power plants are the center of gravity in the energy field, and the information security of their industrial control systems has become particularly important. With the continuous advancement of information construction by enterprises, the degree of informatization has gradually increased, the connection between control systems and enterprise management systems has become closer and closer, the openness of the system has been continuously improved, and various general information products and technologies have been introduced into the system, such as Ethernet The scope of application of network technology and commercial operating systems in control systems continues to expand, which is manifested in the increasing dependence of power systems on information technology. On the other hand, the interconnection between the power plant control system and the power dispatching system has caused the power plant information security risk to continue to expand, and even seriously endanger the safe operation of the power plant. As a result, the information system security problems of electric power enterprises are increasing day by day, so it is very important to evaluate the information security of the power plant control system.

Commonly used control systems in power plants are mainly distributed control system (DCS), programmable control system (PLC) and field bus system (FCS). However, the control system of thermal power plants generally adopts special software and hardware, operating system and communication protocol, and exists in a closed network, so it is often neglected and there are many security risks. Therefore, power plants should establish a multi-technical protection system to achieve multi-angle and all-round protection of physical, network, terminal, and data. Through the assessment of security threats, the security threats existing in the power plant control system can be found, and suggestions for safety rectification of the power plant control system can be proposed and implemented to ensure the life cycle security of the power plant control system.

Existing evaluation methods and systems generally use Markov decision theory, fuzzy mathematics theory, etc. to analyze only quantifiable indicators related to decision-making issues, so it is difficult to judge and comprehensively evaluate qualitative factors such as assumptions and implementation suggestions. There are problems of strong subjectivity and uncertainty, and it is impossible to obtain objective and accurate security assessments.

Contents of the invention

The purpose of the present invention is to provide a SA-PSO-AHP-based thermal power plant control system information security assessment method in order to overcome the above-mentioned defects in the prior art.

The purpose of the present invention can be achieved through the following technical solutions:

A thermal power plant control system information security assessment system based on SA-PSO-AHP, including:

The data input module is used to obtain the scoring data of the control system to be evaluated;

A matrix building module, which pre-stores a layered and progressive model of threat assessment of the control system built using the AHP, which combines the scoring data input by the data input module to construct an assessment matrix;

The matrix correction module is used to judge the consistency of the evaluation matrix, and use the simulated annealing optimization particle swarm optimization algorithm to perform consistency correction on the evaluation matrix that does not meet the consistency standard;

The data output module is used to obtain the threat ranking of the power plant control system by using the hierarchical classification algorithm, and obtain the comprehensive threat score in combination with the threat frequency.

Further, the following steps are specifically performed in the matrix correction module:

S1. Initialize the particles, set the particle number n, initial position and speed, set the annealing coefficient α and initial temperature T;

S2. Obtain the new position of the particle from the current position and velocity of the particle;

S3. Determine whether the number of iterations is less than the specified number of iterations, if so, execute step S4; if not, transmit the corrected judgment matrix data to the data output module;

S4. Calculate the particle fitness value. If the particle fitness value is better than the original individual extreme value pbest, then update the current particle fitness value pbest, then update the current particle fitness value pbest, and the global extreme value gbest is based on the particle individual extreme value pbest Find and update the position and velocity of the particle; if not, execute step S2;

S5. Calculate the variation ΔE of the fitness value of the two positions of the current particle and the updated particle; if ΔE≤0 or satisfy exp(-ΔE/T)>rand(0,1), then accept the updated value to perform cooling, and change the T value Update to αT; otherwise, refuse to accept the updated value and go to step S2.

Further, in the step S2, the expression of speed update is:

v _k+1 ＝c ₀ v _k +c ₁ (pbest _k -x _k )+c ₂ (gbest _k -x _k )

Among them, v _k is the velocity vector of the particle; v _k+1 is the velocity vector of the updated particle; x _k is the position of the current particle; pbest _k is the position where the particle itself finds the optimal solution; gbest _k is the current finding of the entire population The position of the optimal solution; c ₀ , c ₁ , and c ₂ represent the group cognition coefficient, c ₀ generally takes a random number between (0,1), and c ₁ and c ₂ take a random number between (0,2) random number between.

Further, in the step S2, the expression of location update is:

x _k+1 ＝x _k +v _k+1

Among them, x _k is the current particle position; x _k+1 is the updated particle position; v _k+1 is the updated particle velocity vector.

Further, the threat assessment hierarchical structure includes a target result layer, a judging criterion layer and a threat element layer, wherein the judging criterion layer includes:

The impact of threats on the confidentiality of the system, including control or destruction of equipment or software, unauthorized use of remote maintenance ports, password brute force attacks, software backdoors or backdoor software, malware, and unauthorized connection of equipment to the network;

The impact of threats on system availability, including password brute force attacks, improper configuration and operation, denial of service attacks, remote overflow attacks, and remote file access;

The impact of threats on the integrity of the system, including the control or destruction of equipment or software, unauthorized use of remote maintenance ports, malware, denial of service attacks, remote overflow attacks and traffic overload.

Further, in the matrix correction module, when CR<0.1 or λ _max =n, CI=0, the judgment matrix consistency meets the requirements, wherein CR is the consistency test coefficient; CI is the consistency index, when When CI=0, there is complete consistency; when CI is close to 0, there is satisfactory consistency; the larger the CI, the more serious the inconsistency; λ _max is the maximum eigenvalue of the judgment matrix, and n is the order of the judgment matrix.

Further, the acceleration factor of the simulated annealing optimized particle swarm optimization algorithm is 1.49445, the number of evolutions is 200, the population size is 20, the weight coefficient range is [0.4,0.9], the speed range is [-1,1], and the position range is [1/9,9], the starting temperature is 1000, and the annealing coefficient is 0.9.

Further, the calculation expression of the comprehensive threat score Ta is:

Among them, ω _Pi is the global weight of the i-th threat, T _Pi is the occurrence frequency of the i-th threat.

Compared with the prior art, the present invention has the following advantages:

The present invention is based on SA-PSO-AHP (Simulate Anneal-Particle Swarm Optimization-Analytic Hierarchy Process, based on the Analytic Hierarchy Process of Simulated Annealing Improved Particle Swarm Algorithm) to evaluate the information security threat of the power plant control system.

Firstly, the comprehensive decision-making conditions for participating in the evaluation after quantifying the qualitative factors through the analytic hierarchy process can fully reflect the ambiguity of the evaluation elements and process, reduce the influence of personal subjective factors, and obtain objective and sufficient evaluation results; and then Correcting matrix consistency by particle swarm optimization algorithm overcomes the problem that if the evaluation matrix does not have a satisfactory consistency index, the weight calculated from this evaluation matrix is not reliable enough to be used as judgment evidence; finally, optimizing particle swarm algorithm by simulated annealing algorithm can In the particle swarm algorithm, the new position is constrained, and the new solution is accepted with a certain probability, which can obtain better consistency results with fewer iterations, and overcomes the slow convergence caused by the poor new position of the particle swarm algorithm. question.

Therefore, the present invention can not only fully reflect the ambiguity of the evaluation elements and process, reduce the influence of personal subjective factors, obtain more accurate, more objective and sufficient evaluation results, but also improve the efficiency of thermal power plants in the risk evaluation process , by analyzing the main threats in the control system, it is convenient for the timely maintenance and improvement of the power plant.

Description of drawings

Fig. 1 is the structural representation of present embodiment;

Fig. 2 is the schematic flow chart of this embodiment;

Fig. 3 is the network topology structure diagram of corresponding unit DCS system in the embodiment;

Fig. 4 is a hierarchical structure and judgment matrix diagram in the embodiment.

Detailed ways

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is carried out on the premise of the technical solution of the present invention, and detailed implementation and specific operation process are given, but the protection scope of the present invention is not limited to the following embodiments.

As shown in Figure 1, this embodiment provides a thermal power plant control system information security assessment system based on SA-PSO-AHP, including:

The data input module is used to obtain the scoring data of the control system to be evaluated;

A matrix building module, which pre-stores a layered and progressive model of threat assessment of the control system built using the AHP, which combines the scoring data input by the data input module to construct an assessment matrix;

The matrix correction module is used to judge the consistency of the evaluation matrix, and use the simulated annealing optimization particle swarm optimization algorithm to perform consistency correction on the evaluation matrix that does not meet the consistency standard;

The data output module is used to obtain the threat ranking of the power plant control system by using the hierarchical classification algorithm, and obtain the comprehensive threat score in combination with the threat frequency.

The working principle of this system is:

Before using this system, first analyze the threats affecting information security according to the relevant standards of industrial control system information security, and establish the corresponding hierarchical structure and pre-store it in the matrix building module; organize an expert group to judge the evaluated factors, and input the scoring data into the data input Module, and the evaluation matrix of each factor is obtained by the matrix construction module; then the consistency check of the evaluation matrix is carried out through the matrix correction module, and if the matrix does not meet the consistency requirements, the consistency correction is carried out; finally, the data output module passes the level Analytics to get systemic threats.

As shown in Figure 2, the specific operation process of this embodiment is as follows:

Step S1 , analyzing the network architecture of the DCS system to be evaluated, identifying and acquiring threats to the control system, and inputting scoring data of the control system to be evaluated in the data input module.

Step S2, the matrix construction module obtains the scoring data, and then generates an assessment matrix through the preset threat assessment layered progressive model.

Step S3 , in the matrix correction module, evaluate the consistency of the evaluation matrix. When CR<0.1 or λ _max =n, CI=0, it is considered that the consistency of the matrix meets the requirement.

Among them: the formula for calculating the consistency index CI is:

When CI=0, there is complete consistency; when CI is close to 0, there is satisfactory consistency; the larger the CI, the more serious the inconsistency;

The formula for calculating the consistency test coefficient CR is:

λ _max is the maximum eigenvalue of the judgment matrix, and n is the order of the judgment matrix.

Consistency correction is performed on the evaluation matrix that does not meet the consistency standard, and the particle swarm optimization algorithm is optimized by simulated annealing to make the evaluation matrix have a satisfactory consistency index. The simulated annealing algorithm is based on the similarity between the annealing process of solid matter in physics and the general combinatorial optimization problem. With each cooling, the temperature gradually decreases. For the particle swarm algorithm, new particle states are continuously generated, and new particle states are accepted or discarded with a certain probability. state, and finally the particles gradually become orderly, thus reaching an equilibrium state. The specific expansion is as follows:

Step S31, use simulated annealing improved particle swarm optimization algorithm (SA-PSO) to find the optimal value of the consistency check coefficient CR and optimize the matrix parameters. First, initialize the particles, set the number of particles n, the initial position and speed, and set Annealing coefficient α and starting temperature T.

Step S32, obtain the new position of the particle from the current position and velocity of the particle.

The expression for speed update is:

v _k+1 ＝c ₀ v _k +c ₁ (pbest _k -x _k )+c ₂ (gbest _k -x _k )

Among them: v _k is the velocity vector of the particle; x _k is the position of the current particle; pbest _k is the position where the particle itself finds the optimal solution; gbest _k is the position of the optimal solution found by the entire population; c ₀ , c ₁ , c ₂ represents the group cognition coefficient, c ₀ generally takes a random number between (0,1), and c ₁ and c ₂ take a random number between (0,2).

The speed update limits it to V _max , and limits the particle speed, so that the new speed reduces the possibility of deterioration, thereby alleviating the problem of slowing down the convergence speed to a certain extent.

The expression for location update is:

x _k+1 ＝x _k +v _k+1

Among them: x _k+1 is the updated particle position; v _k+1 is the updated particle velocity vector.

Step S33 , judging whether the number of iterations is less than the specified number of iterations, if yes, execute step S34 ; if not, execute step S4 .

Step S34, calculate the particle fitness value, if the particle fitness value is better than the original individual extremum pbest, then update the current particle fitness value pbest, the global extremum gbest is searched according to the particle individual extremum pbest, and update the position and velocity of the particle; If not, execute step S32;

Step S35, calculate the variation ΔE of the fitness value of the current particle and the particle after the update; if ΔE≤0 or satisfy exp(-ΔE/T)>rand(0,1), accept the updated value, perform cooling, and The T value is updated to αT; otherwise, the updated value is rejected, x _k+1 is still x _k , and step S32 is executed.

Step S4, using the AHP calculation method to calculate the threat ranking of the power plant control system, and calculating the comprehensive threat score in combination with the threat frequency, the calculation expression of the comprehensive threat score Ta is:

Among them, ω _Pi is the global weight of the i-th threat, T _Pi is the occurrence frequency of the i-th threat.

In order to verify the effectiveness of the method, in this embodiment, the assessment system is used to assess the threat of a DCS system of a unit, construct a hierarchical analysis structure, and perform consistency correction on the matrix that does not meet the consistency standard.

As shown in Figure 3, it is the network architecture of the system evaluated in the example. This control system consists of 6 operator stations, engineer stations, history stations, link stations, and SIS workstations (interface machines) connected to the switch through the A/B network . The switch is connected with BMS furnace safety monitoring system, CCS coordinated control, SCS sequence control system, DAS thermal power plant computer monitoring system, and ECS thermal power plant electrical control system to control the control points and process systems in the production process, and complete the collection of production data, Sending tasks such as operation instructions.

As shown in Figure 4, it is the hierarchical analysis model and the corresponding judgment matrix. Through threat identification, the impact of each underlying threat (P1~P11) on the system's confidentiality (Confidentiality), integrity (Integrity) and availability (Availability) is obtained. , matrix C represents the impact of threats P1-P5 on system confidentiality, matrix I represents the impact of threats P3, P7-P10 on system integrity, matrix A represents the impact of threats P1, P2, P5, P8, P9, and P11 on system availability Impact, matrix S represents the impact of system confidentiality, integrity, and availability on the overall threat score of the system.

Table 1 to Table 3 are the correction results to the evaluation matrix that does not meet the consistency standard, and compare the method (SAPSO) proposed by the present invention with the traditional particle swarm optimization algorithm (PSO), and Table 4 is obtained respectively by the method of hierarchical analysis The global weight of the underlying threat.

The acceleration factor of the particle swarm optimization algorithm is 1.49445, the number of evolutions is 200, the population size is 20, the weight coefficient range is [0.4,0.9], the speed range is [-1,1], and the position range is [1/9,9] ], the optimized parameters of simulated annealing were added with an initial temperature of 1000 and an annealing coefficient of 0.9.

Table 1 Comparison of consistency correction results of integrity assessment matrix

I12 I13 I14 I15 I23 I24 I25 I34 I35 I45 initial 1 2 3 3 3 1/2 4 2 5 4 PSO 1.7745 1.841 2.999 6.313 0.985 1.575 3.9911 1.719 4.638 2.7785 SAPSO 1.7714 1.879 3.000 6.855 1.053 1.693 3.979 1.618 3.909 2.4112

Table 2 Comparison of consistency correction results of usability evaluation matrix

A12 A13 A14 A15 A16 A23 A24 A25 A26 A34 A35 A36 A45 A46 A56 initial 2 3 1/3 1/2 1/3 2 1/3 1/2 1/2 1 1/3 1 3 2 1 PSO 1.539 1.684 1.022 2.013 1.462 1.118 0.666 1.329 0.959 0.604 1.197 0.866 2.010 1.431 0.725 SAPSO 2.854 2.944 1.041 1.821 2.372 0.908 0.334 0.727 0.778 0.351 0.684 0.844 2.336 2.706 1.249

Table 3 Evaluation matrix consistency correction parameter comparison

Table 4 Global weight of threats

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 PSO 0.1506 0.1202 0.1237 0.0225 0.1034 0.0125 0.0579 0.1962 0.1046 0.0142 0.0943 SAPSO 0.173 0.0873 0.1259 0.0225 0.0802 0.0125 0.0596 0.1945 0.1611 0.0149 0.0685

Table 5 Threat occurrence frequency

P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 Frequency T 1 1 2 1 2 1 3 1 1 4 1

The calculation results show that the simulated annealing optimization particle swarm optimization algorithm adopted in this embodiment can obtain better consistent results with fewer iterations, and can avoid the problem of slow convergence caused by the poor new position of the particle swarm optimization algorithm. Among the 11 types of system threats obtained by the two methods, the highest threat weights are P8 denial of service attack and P1 equipment or software being controlled or destroyed. The threat scores calculated according to Table 5 are 2.7712 and 2.74, respectively. The results are divided into high, high, medium, low and low, and the corresponding ranges are 10-8, 8-6, 6-4, 4-2, 2-0, all of which indicate that the threat level of the DCS network is low, The overall loss that may be borne by threats is relatively low.

Comprehensively consider the information security threat quantitative assessment results of the DCS system, the existing control measures and possible risk consequences, and improve the system security measures to enhance the level of security protection. It can be concluded that the threats that have a greater impact on the system are denial of service attacks and equipment or software being controlled or destroyed. If too many connections are caused by denial of service supply, it is necessary to change the security policy in time, take preventive measures, and cooperate with multiple parties. Address threats. The system should update the vulnerability database in a timely manner, add vulnerability inspection equipment or software, and add network security tools such as industrial firewalls, intrusion detection products (IDS), one-way isolation devices, etc., and pay attention to encryption and authentication in the process of interconnecting with dispatching data networks and other external networks The access of the device ensures the encrypted transmission of data; shut down the redundant services of the host system, configure the password complexity policy, upgrade the ransomware patch in time or configure the corresponding security response strategy. The application of threat assessment proposed by the invention can not only improve the efficiency of the thermal power plant in the process of risk assessment, but also analyze the main threats in the system to facilitate timely maintenance and improvement of the power plant.

The preferred specific embodiments of the present invention have been described in detail above. It should be understood that those skilled in the art can make many modifications and changes according to the concept of the present invention without creative efforts. Therefore, all technical solutions that can be obtained by those skilled in the art based on the concept of the present invention through logical analysis, reasoning or limited experiments on the basis of the prior art shall be within the scope of protection defined by the claims.

Claims (6)

1. a kind of thermal power plant's control system Information Security Evaluation system based on SA-PSO-AHP characterized by comprising

Data input module, for obtaining the marking data to control system to be assessed；

Matrix builds module, and the threat assessment layering that the control system built using analytic hierarchy process (AHP) is prestored in the module is progressive The marking data of model, models coupling data input module input construct evaluating matrix；

Matrix correction module, for carrying out consistency judge to evaluating matrix, to the evaluating matrix benefit for being unsatisfactory for consistency criterion Consistent correction is carried out with simulated annealing optimization particle swarm algorithm；

Data outputting module, for obtaining this control system of power plant threat sequercing using the algorithm of level point, in conjunction with threat frequency Obtain synthetic threat scoring.

2. thermal power plant's control system Information Security Evaluation system according to claim 1 based on SA-PSO-AHP, special Sign is, specifically executes following steps in the matrix correction module:

S1, initializing set, setting population n, initial position and speed, setting annealing coefficient α and starting temperature are carried out to particle Spend T；

S2, the new position of this particle is obtained by the current Position And Velocity of particle；

S3, judge whether the number of iterations is less than regulation the number of iterations, if so, thening follow the steps S4；If it is not, then sentencing revised Disconnected matrix data is transmitted to data outputting module；

S4, particle adaptive value is calculated, if it is suitable to update current particle better than original individual extreme value pbest for the adaptive value of particle It should be worth for pbest, then updating current particle adaptive value is pbest, and global extremum gbest is according to particle individual extreme value pbest It finds, the position and speed of more new particle；If it is not, thening follow the steps S2；

S5, the adaptive value variation delta E for calculating two positions of particle after current particle and update；If Δ E≤0 meets exp (- Δ E/T) > rand (0,1), then receive updated value and execute cooling, T value is updated to α T；Otherwise updated value is rejected, is held Row step S2.

3. thermal power plant's control system Information Security Evaluation system according to claim 1 based on SA-PSO-AHP, special Sign is that the threat assessment hierarchical structure includes objective result layer, judges rule layer and threaten element layer, wherein is judged Rule layer includes:

The influence to the confidentiality of system is threatened, including equipment or software are controlled or destroyed, remote maintenance port is unauthorized Ground use, the brute force attack of password, software back door or Backdoor Software, Malware and network is connected devices to without permission；

The influence to the availability of system is threatened, brute force attack including password, irrelevantly configuration and operation, Denial of Service attack It hits, long-range flooding and remote document access；

The influence to the integrality of system is threatened, including equipment or software are controlled or destroyed, remote maintenance port is unauthorized Ground use, Malware, Denial of Service attack, long-range flooding and overload.

4. thermal power plant's control system Information Security Evaluation system according to claim 1 based on SA-PSO-AHP, special Sign is, in the matrix correction module, as CR < 0.1 or λ_maxWhen=n, CI=0, then judgment matrix approach, which reaches, wants It asks, wherein CR is consistency check coefficient；CI is coincident indicator, as CI=0, there is complete consistency；CI close to 0, There is satisfied consistency；CI is bigger, inconsistent more serious；λ_maxFor judgment matrix maximum eigenvalue, n is judgment matrix order.

5. thermal power plant's control system Information Security Evaluation system according to claim 1 based on SA-PSO-AHP, special Sign is that the acceleration factor of the simulated annealing optimization particle swarm algorithm is 1.49445, and evolution number is 200, population scale It is 20, weight coefficient section is [0.4,0.9], and speed interval is [- 1,1], and position section is [1/9,9], and initial temperature is 1000, annealing coefficient 0.9.

6. thermal power plant's control system Information Security Evaluation system according to claim 1 based on SA-PSO-AHP, special Sign is, synthetic threat scores the calculation expression of Ta are as follows:

Wherein, ω_PiThe global weight threatened for i-th, T_PiThe occurrence frequency threatened for i-th.