Safety control method for electric power system with double-layer hidden Markov jump

Abstract

The invention discloses a safety control method of a power system with double hidden Markov jumps, and belongs to the technical field of Markov system control. According to the method, an asynchronous behavior between a power system and a controller and the difficulty of acquiring system state information are considered, and a hidden Markov power system model is established; secondly, based on random processes (transmission success, transmission failure and DoS attack) of communication transmission and difficulty in acquiring mode information of the communication processes, establishing an HMM (Hidden Markov Model) for describing the communication processes; therefore, a novel double-layer hidden Markov jump power system is established.

Description

Technical Field

[0001] The invention relates to the technical field of Markov system control, and in particular to a safety control method for a power system with a double-layer hidden Markov jump. Background Art

[0002] As a complex nonlinear system, the power system often has low-frequency oscillation problems during operation. These continuous oscillations can seriously affect the stability of the system and cause significant economic losses. The power system stabilizer (PSS) can provide supplementary damping for the synchronous machine rotor oscillation, thereby suppressing the oscillation, and has therefore been widely studied and applied. Considering the changes in the operating conditions of the power system, there are currently two main methods for large-scale stabilization of the power system: robust PSS and adaptive PSS. Compared with the adaptive PSS that requires relatively strict continuous excitation conditions, the robust PSS has received more attention because it can still maintain good dynamic performance under large load changes and system nonlinearity. For example, experts discussed the control problem of nonlinear robust coordinated PSS automatic voltage regulators. However, these studies did not consider the impact of random topology mutations on system stability. It is of great significance and challenge to study the impact of random mutations in power system topology on system stability.

[0003] The faults and external disturbances experienced by the power system are unpredictable. Therefore, the Markov theory can be used to model the mutations of the system structure and parameters. For example, some scholars have used the Markov model to describe the random mutations of the power system. However, the existing Markov power system controller designs are mostly pattern-independent or pattern-dependent. Since data loss or delay may occur during the data transmission from the sensor to the controller, partial loss of system pattern information may occur, which may cause the controller to be out of sync with the system. Secondly, in actual work, the acquisition of power system state information is also a daunting challenge. Considering these factors, the hidden Markov model (HMM) is undoubtedly a better choice. HMM can not only characterize the phenomenon of out of sync, but also solve the problem of difficulty in obtaining system pattern information. Therefore, it is necessary to study the power system based on HMM.

[0004] On the other hand, data in the power system may be subject to malicious network attacks during transmission, the most common of which is a DoS attack that causes packet loss. In addition, due to the unreliability of the transmission channel, transmission failures are prone to occur. Faced with such challenges, some scholars use discrete hidden Markov chains to model the probability of transmission failures. However, most current communication transmission studies ignore the possibility of DoS attacks and transmission failures occurring at the same time. It is worth noting that network failures that cause packet loss and energy-limited DoS attacks are different. Transmission failures are usually unintentional and occur randomly, while DoS attacks are intentional and may last for a period of time. Based on the above discussion, how to deal with these random processes in hidden Markov jump power systems is the main motivation of this invention.

[0005] Similar to obtaining pattern information in the Markov power system model, obtaining model information for communication transmission is also challenging. Fortunately, studies have shown that HMM provides a feasible solution to this challenge. Therefore, the communication transmission process can be modeled as a hidden Markov jump communication model. In some studies, HMM is effectively used to characterize channel dynamics. In other studies, finite state HMM is used to study the state estimation problem in unreliable communication channels, but these studies are limited by the assumption that the transition probability matrix or the observation probability matrix is ​​fully known, and there are great limitations in theory. To address this problem, some scholars consider using partial information transfer probability matrices when using HMM to describe the random process of packet loss and time delay. In addition, some scholars have studied Markov jump systems based on sliding modes, considering partially known observation probability matrices. Based on the above research, the present invention explores a more general scenario, that is, the transition probability matrix and the observation probability matrix may both be partially known, or one of them may be partially unknown. Based on the above discussion, overcoming the asynchronous behavior of controllers and system modes and data in unreliable network transmission, especially under unknown probability conditions, is a very challenging problem. Summary of the invention

[0006] The technical problem to be solved by the present invention is how to overcome the asynchronous behavior of controller and system mode and data in unreliable network transmission, and provide a safety control method for power system with double-layer hidden Markov jump. ∞ The combination of performance indicator control can achieve security control of the system under DoS attacks.

[0007] The present invention solves the above technical problems through the following technical solutions, and the present invention comprises the following steps:

[0008] S1: Considering the circuit breaker switching action caused by line faults in the power system and the asynchronous behavior between the power system and the controller, a hidden Markov jump power system model is introduced;

[0009] S2: In order to deal with the packet loss and DoS attacks that may be encountered when data is transmitted in an unreliable network channel, another independent hidden Markov model is used to describe this situation. The transition probability matrix and observation probability matrix in the hidden Markov model may be unknown.

[0010] S3: Build a controller model to obtain a closed-loop system model of the system under DoS attack;

[0011] S4: Give the system obtained in step S3 that is randomly stable under disturbance and satisfies H ∞ Linear matrix inequality conditions for performance index σ;

[0012] S5: Use Lyapunov function and performance index function to prove that the inequality condition in step S4 is valid;

[0013] S6: Solve the gain matrix of the controller;

[0014] S7: Implementing safety control of the double-layer hidden Markov power system according to the gain matrix of the controller in step S6 and the given system parameters.

[0015] Furthermore, in step S1, the specific processing process is as follows:

[0016] S11: Establish the power system dynamic model as follows:

[0017]

[0018] Where δ is the rotor angle, x d is the synchronous reactance along the d-axis, x' d is the transient reactance along the d-axis, T e is the electrical torque, V is the infinite bus voltage, E fd is the generator field voltage, u is the stable signal, T do ' is the d-axis open-circuit transient time constant, V t is the terminal voltage, E q ' is the q-axis voltage after the transient reactance, x e is the external line reactance, M is the inertia coefficient, k E ,T E is the actuator gain and time constant, T m is the mechanical torque;

[0019] S12: The fourth-order state space model of the power system in step S11 is expressed as:

[0020]

[0021] in:

[0022] x T (υ)=[Δδ Δω ΔE′ q ΔE fd ];

[0023]

[0024] x(υ) and u(υ) represent the state variable and control input of the υth node respectively, y(υ) is the measured output, is the internal coupling matrix between nodes, w(υ) is the interval Additional interference on is the sensor’s measurement matrix, is the output matrix of the system, k1, k2...k6 are the linearization model constants of the synchronous motor, and Δω is the speed deviation;

[0025] S13: Introduce a hidden Markov process with unknown partial information and establish the following hidden Markov jump power system:

[0026]

[0027] in, is a discrete-time Markov chain;

[0028] The transition probability matrix R in the hidden Markov jump power system is R={η ψj}as follows:

[0029]

[0030] Among them, η ψj ∈[0,1],

[0031] Observation Probability Matrix in Hidden Markov Jump Power System as follows:

[0032]

[0033] in,

[0034] Furthermore, in step S2, the energy of the DoS attack is limited and may last until the Nth time point. The mode of the DoS attack is as follows:

[0035] Mode 1: indicates successful data transmission (λ(μ) = 1); in this mode, the probability that the system will continue to succeed in the next time step is τ 11, the probability of transmission channel failure is τ 12 , the probability of being attacked by DoS is τ 13 ;

[0036] Mode 2: indicates communication transmission failure (λ(μ) = 2); in this mode, the probability of successful transmission in the next time step is τ 21 , the probability of transmission channel failure is τ 22 , the probability of being attacked by DoS is τ 23 ;

[0037] Mode 3: represents transmission under DoS attack (λ(μ) = 3); the probability of normal transmission in the next time step is τ 31 , the probability of transmission channel failure is τ 32 , the probability of continuing to be attacked by DoS is τ 34 ;

[0038] According to the same principle, we extend it to mode λ(μ)=N+1. Since the attack energy is limited when the attack reaches λ(μ)=N+2, the system may attack with probability τ in the next time step. (N+2)1 and τ (N+2)2 Return to Mode 1 or Mode 2.

[0039] Furthermore, in step S2, the specific processing process is as follows:

[0040] S21: Consider using the Markov variable q(μ) to simulate the switching of DoS attacks and determine the transition probability matrix as follows:

[0041]

[0042] in,

[0043] S22: The variable λ(μ) is observed through q(μ) and and determine the observation probability matrix as follows:

[0044]

[0045] in,

[0046] S23: Analyze the situation where the transition probability matrix and the observation probability matrix are partially unknown in:

[0047]

[0048] There are three specific situations:

[0049]

[0050] Furthermore, in step S3, the specific processing process is as follows:

[0051] S31: Build the following Hidden Markov Model-based asynchronous state feedback controller:

[0052]

[0053] in, represents the controller gain matrix, Variable θ ζ belong Indicates the transmission status. When ζ=1, it indicates that the data transmission is successful; when ζ≠1, it indicates that the current status information cannot be obtained and the last successfully transmitted data is used.

[0054] S32: Then we get the following two-layer hidden Markov jump power system:

[0055]

[0056] Among them, x(t+1) represents the state variable of the t+1th node, y(t) is the measured output, is a given system parameter matrix, w(t) is a matrix in the interval Additional interference on is the sensor’s measurement matrix, is the output matrix of the system;

[0057] S33: Augment the double-layer hidden Markov jump power system in step S32 to obtain a closed-loop system model as follows:

[0058]

[0059] in,

[0060] Furthermore, in step S4, given a scalar σ>0, if there exists a matrix K i , ρ=a,b; and the symmetric matrix Then when the following inequality condition holds, the system in step S32 can ∞ Achieve random stability under performance index σ:

[0061]

[0062] in:

[0063]

[0064] Furthermore, in step S5, the specific processing process is as follows:

[0065] S51: For w(t)≡0 and any initial value, when the condition When it is established, the system in step S32 is stochastically stable; when all non-zero When the zero initial state satisfies the following inequality, the system in step S32 can reach the specified H ∞ Performance index σ:

[0066]

[0067] S52: Consider the following Lyapunov functionals:

[0068]

[0069] in,

[0070] make get:

[0071]

[0072] in,

[0073] Through the inequality below in step S51 and the above equation, we can obtain:

[0074]

[0075] in:

[0076] S53: Through the inequality above in the inequality condition in step S4, we get:

[0077]

[0078] in,

[0079] S54: Substitute the inequality in step S53 into the inequality below in the inequality condition in step S4 to obtain:

[0080]

[0081] S55: Then, based on the inequality Complementing with Schur, we can get the inequality below in the inequality condition in step S4;

[0082] S56: According to the inequality in step S54, we get:

[0083]

[0084] If w(t)≡0, then The system is stochastically stable;

[0085] Under zero initial conditions, w(t)≠0, we know

[0086] The system is stochastically stable and satisfies the specified H ∞ Performance index σ.

[0087] Furthermore, in step S6, the specific processing process is as follows:

[0088] S61: Order K i =T i -1 U i ;

[0089] S62: Perform an equivalent transformation on the inequality below in the inequality condition of step S4 as follows:

[0090]

[0091] in:

[0092]

[0093] S63: Then use the simulation software to calculate the value of the controller gain matrix by giving the matrix parameters.

[0094] Compared with the prior art, the present invention has the following advantages:

[0095] 1. Considering the asynchronous behavior between the power system and the controller, and the difficulty of obtaining system status information, a hidden Markov power system model is established; secondly, based on the random process of communication transmission (transmission success, transmission failure and DoS attack) and the difficulty of obtaining the pattern information of these communication processes, an HMM describing the communication process is established; then a new double-layer hidden Markov jump power system is established;

[0096] 2. For random processes in communication, HMM with limited information is used to describe them. Specifically, this limited information may exist in the transition probability matrix, the observation probability matrix, or both. BRIEF DESCRIPTION OF THE DRAWINGS

[0097] Figure 1 is a flow chart of a safety control method for a power system with a double-layer hidden Markov jump in an embodiment of the present invention;

[0098] Figure 2 is a system mode and controller mode diagram in an embodiment of the present invention;

[0099] Figure 3 is a sequence diagram of different attacks considering the impact of hybrid network attacks on power system stability in an embodiment of the present invention;

[0100] Figure 4 is a trajectory diagram of the controller when the gain matrix of the controller is solved in an embodiment of the present invention;

[0101] Figure 5 is a system state trajectory diagram when the controller is not included in the embodiment of the present invention;

[0102] Figure 6 is a system state trajectory diagram when a controller is included in an embodiment of the present invention;

[0103] Figure 7 is a state trajectory diagram of x1(t) in an embodiment of the present invention;

[0104] Figure 8 is a top view of x4(t) in an open-loop system according to an embodiment of the present invention;

[0105] Fig. 9 is a three-dimensional graph of x4(t) under an open-loop system in an embodiment of the present invention;

[0106] Fig.10 is a top view of x4(t) in a closed-loop system in an embodiment of the present invention;

[0107] Fig.11 is a top view of x4(t) in a closed-loop system in an embodiment of the present invention;

[0108] Fig.12 It is a schematic diagram of the transmission information construction structure in an embodiment of the present invention. DETAILED DESCRIPTION

[0109] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented on the premise of the technical solution of the present invention, and a detailed implementation method and a specific operation process are given, but the protection scope of the present invention is not limited to the following embodiment.

[0110] This embodiment provides a technical solution: a safety control method for a power system with a double-layer hidden Markov jump structure. First, considering the circuit breaker switching action caused by line faults in the power system and the asynchronous behavior between the power system and the controller, a hidden Markov jump power system model is introduced. In order to deal with the packet loss and denial of service attack (DoS attack) that may be encountered when data is transmitted in an unreliable network channel, another independent hidden Markov model (HMM) is used to describe this situation, and the transition probability matrix (TPM) and observation probability matrix (OPM) in the HMM may be unknown. Secondly, based on the power system with a double-layer hidden Markov jump structure, some sufficient conditions are established to ensure that the closed-loop system achieves random stability during mode asynchronous and unreliable data network transmission. Finally, the correctness of the double-layer hidden Markov jump structure power system theory is verified by simulation examples.

[0111] like Figure 1 As shown, the above-mentioned safety control method of the power system with double-layer hidden Markov jumps specifically includes the following steps:

[0112] Step S1: Considering the circuit breaker switching action caused by line faults in the power system and the asynchronous behavior between the power system and the controller, a hidden Markov jump power system model is introduced;

[0113] The dynamic model of the power system is established as follows:

[0114]

[0115] Where δ is the rotor angle, x d is the synchronous reactance along the d-axis, x' d is the transient reactance along the d-axis, T e is the electrical torque, V is the infinite bus voltage, E fd is the generator field voltage, u is the stable signal, T do ' is the d-axis open-circuit transient time constant, V t is the terminal voltage, E q ' is the q-axis voltage after the transient reactance, x e is the external line reactance, M is the inertia coefficient, k E ,T E is the actuator gain and time constant, T m is the mechanical torque;

[0116] The fourth-order state space model of the above power system (1) is expressed as:

[0117]

[0118] in:

[0119] xT (υ)=[Δδ Δω ΔE′ q ΔE fd ];

[0120]

[0121] x(υ) and u(υ) represent the state variable and control input of the υth node respectively, y(υ) is the measured output, is the internal coupling matrix between nodes, w(υ) is the interval Additional interference on is the sensor’s measurement matrix, is the output matrix of the system, k1, k2...k6 are the linearization model constants of the synchronous motor, and Δω is the speed deviation.

[0122] Introducing a partially unknown hidden Markov process, the following hidden Markov jump power system is established:

[0123]

[0124] in, is a discrete-time Markov chain;

[0125] The transition probability matrix R = {η ψj}:

[0126]

[0127] Among them, η ψj ∈[0,1],

[0128] Considering the asynchronous behavior and the difficulty of obtaining system status information, a hidden Markov jump power system model is established in this embodiment. The observation probability matrix s(t) associated with r(t) follows:

[0129]

[0130] in,

[0131] Step S2: In order to handle the packet loss and denial of service attacks that may be encountered when data is transmitted in an unreliable network channel, another independent hidden Markov model is used to describe this situation, and both the transition probability matrix and the observation probability matrix in the hidden Markov model may be unknown.

[0132] In the study of power systems, communication channels may encounter transmission failures and denial of service attacks (DoS attacks) during transmission. These situations can damage or even lose data packets, seriously affecting system stability. Among them, the energy of DoS attacks is limited and may last until the Nth time point. The mode of DoS attacks is as follows:

[0133] Mode 1: indicates successful data transmission (λ(μ) = 1); in this mode, the probability that the system will continue to succeed in the next time step is τ 11 , because the probability of transmission channel failure is τ 12 , the probability of being attacked by DoS is τ 13 ;

[0134] Mode 2: indicates communication transmission failure (λ(μ) = 2); in this mode, the probability of successful transmission in the next time step is τ 21 , the probability of transmission channel failure is τ 22 , the probability of being attacked by DoS is τ 23 ;

[0135] Mode 3: represents transmission under DoS attack (λ(μ) = 3); the probability of normal transmission in the next time step is τ 31 , the probability of transmission channel failure is τ 32 , the probability of continuing to be attacked by DoS is τ 34 ;

[0136] According to the same principle, it is extended to mode λ(μ)=N+1. Since the attack energy is limited when the attack reaches λ(μ)=N+2, the system may attack with probability τ in the next time step. (N+2)1 and τ (N+2)2 Return to Mode 1 or Mode 2.

[0137] Consider using Markov variables q(μ) to simulate the switching of DoS attacks, the transition probability matrix

[0138]

[0139] in,

[0140] The variable λ(μ) is observed through q(μ) and The observation probability matrix

[0141]

[0142] in,

[0143] Analyze the situation where the transition probability matrix and the observation probability matrix are partially unknown.

[0144]

[0145] in,

[0146] Step S3: construct a controller model to obtain a closed-loop system model of the system under DoS attack.

[0147] Build the following Hidden Markov Model-based asynchronous state feedback controller:

[0148]

[0149] in, represents the controller gain matrix, and the two subscripts of K represent the system mode and the observed mode respectively.

[0150] A structural diagram can be constructed based on the modeling information of transmission success, transmission failure, and DoS attack, such as Fig.12 The variable θ ζ belong Indicates the transfer status.

[0151]

[0152] When ζ=1, it indicates that the data transmission is successful; on the contrary, when ζ≠1, it means that the current state information cannot be obtained, so the data successfully transmitted last time is used. The specific expression is as follows:

[0153]

[0154] In summary, the following two-layer hidden Markov jump power system can be obtained:

[0155]

[0156] Among them, x(t+1) represents the state variable of the t+1th node, y(t) is the measured output, is a given system parameter matrix, w(t) is a matrix in the interval Additional interference on is the sensor’s measurement matrix, is the output matrix of the system.

[0157] By augmenting formula (6), the closed-loop system model can be obtained as follows:

[0158]

[0159] in,

[0160] Step S4: Give the system in step S3 that is stochastically stable under disturbance and satisfies H ∞ Linear matrix inequality conditions for the performance index σ.

[0161] Given a scalar σ>0, if there exists a matrix K i , ρ=a,b, and symmetric matrix Then, when the following inequality condition holds, system (7) can be ∞ Achieve random stability under performance index σ:

[0162]

[0163] in:

[0164]

[0165] Step S5: Use the Lyapunov function and the performance index function to prove that the inequality condition in step S4 is valid.

[0166] When system (7) holds under given conditions, it not only behaves stochastically stable but also satisfies H ∞ Performance index σ.

[0167] For w(t)≡0 and any initial value, the system (7) is stochastically stable when the condition When founded:

[0168] When all non-zero When the zero initial state satisfies the following inequality, the system (7) can reach the specified H ∞ Performance index σ.

[0169]

[0170] Consider the following Lyapunov functionals:

[0171]

[0172] in,

[0173] make We can get:

[0174]

[0175] in,

[0176] Through formula (10) and the above formula, we can get:

[0177]

[0178] in: Through formula (8), we can get:

[0179]

[0180] in, Substituting the above inequality into equation (9), we can obtain:

[0181]

[0182] Then, based on the inequality And Schur complement, we can get formula (9);

[0183] According to formula (12), we can get:

[0184]

[0185] If w(t)≡0, we know The system is stochastically stable.

[0186] Under zero initial conditions, w(t)≠0, we can get The system is stochastically stable and satisfies the specified H ∞ Performance index σ.

[0187] Step S6: Solve the gain matrix of the controller.

[0188] Let K i =T i -1 U i .

[0189] The equivalent transformation of some inequality conditions in step S4 is as follows:

[0190]

[0191] in:

[0192]

[0193] At this point, simulation software can be used to calculate the controller gain matrix K by giving matrix parameters. i The value of .

[0194] Step S7: Implementing safety control of the double-layer hidden Markov power system according to the gain matrix of the controller in step S6 and the given system parameters.

[0195] When a line fails, the hidden Markov jump power system of formula (3) has two modes of change according to the on-off state of the circuit breaker. The specific parameters are as follows:

[0196] Mode 1:

[0197]

[0198] Mode 2:

[0199]

[0200] The remaining parameters in this embodiment are assumed as follows:

[0201] σ=20, w(t)=0.3*cos(0.3t)e -0.56t , S1 = [0.6 0.6 0.45 0.3] T , S2 = [1.2 0.45 0.45 0.45] T ;

[0202] Then, the controller gain matrix is ​​as follows:

[0203] K1=[-0.0076 -0.048 0.0393 0.0043];

[0204] K2=[-0.0239 -0.2067 0.0107 -0.0021].

[0205] According to the above parameters, the following simulation diagram can be obtained. Figure 4 The output responses u(t) on 10 sampling paths are shown. Figure 2 shows the evolution of the system and controller modes over time. Figure 3 Three transmission modes under HMM are described. Figure 5 It shows that without a controller, the system state trajectory is unstable on the 10 sampling paths. Figure 7 In Figure 2, after applying the controller, the state trajectory x1(t) clearly shows that the system eventually stabilizes. Figure 6 This indicates that all state trajectories have reached stability. In addition, Figure 8-Figure 11 A 3D plot of x4(t) is shown for a single sampling path. Figure 8 and Fig. 9 The system state in is divergent from a three-dimensional perspective (without a controller), and then under the intervention of the controller, the system state tends to be stable, such as Fig.10 and Fig.11 shown.

[0206] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A safety control method for a power system with double-layer hidden Markov jumps, characterized in that: The following steps are involved: S1: Considering the circuit breaker switching action caused by line faults in the power system and the asynchronous behavior between the power system and the controller, a hidden Markov jump power system model is introduced; S2: In order to deal with the packet loss and DoS attacks that may be encountered when data is transmitted in an unreliable network channel, another independent hidden Markov model is used to describe this situation. The transition probability matrix and observation probability matrix in the hidden Markov model may be unknown. S3: Build a controller model to obtain a closed-loop system model of the system under DoS attack; S4: Give the system obtained in step S3 that is randomly stable under perturbations and satisfies H ∞ Linear matrix inequality conditions for performance index σ; S5: Use Lyapunov function and performance index function to prove that the inequality condition in step S4 is valid; S6: Solve the gain matrix of the controller; S7: Implementing safety control of the double-layer hidden Markov power system according to the gain matrix of the controller in step S6 and the given system parameters.

2. A safety control method for a power system with double-layer hidden Markov jumps according to claim 1, characterized in that: In step S1, the specific processing process is as follows: S11: Establish the power system dynamic model as follows: Where δ is the rotor angle, x d is the synchronous reactance along the d-axis, x' d is the transient reactance along the d-axis, T e is the electrical torque, V is the infinite bus voltage, E fd is the generator field voltage, u is the stable signal, T do ' is the d-axis open-circuit transient time constant, V t is the terminal voltage, E q ' is the q-axis voltage after the transient reactance, x e is the external line reactance, M is the inertia coefficient, k E ,T E is the actuator gain and time constant, T m is the mechanical torque; S12: The fourth-order state space model of the power system in step S11 is expressed as: in: x T (υ)=[Δδ Δω ΔE′ q D.E. fd ]; C = diag{0,1,0,0}; x(υ) and u(υ) represent the state variable and control input of the υth node respectively, y(υ) is the measured output, is the internal coupling matrix between nodes, w(υ) is the interval Additional interference on is the sensor's measurement matrix, is the output matrix of the system, k1, k2k6 are the linearized model constants of the synchronous motor, and Δω is the speed deviation; S13: Introduce a hidden Markov process with unknown partial information and establish the following hidden Markov jump power system: in, is a discrete-time Markov chain; The transition probability matrix R in the hidden Markov jump power system is R={η ψj }as follows: Among them, η ψj ∈[0,1], Observation Probability Matrix in Hidden Markov Jump Power System as follows: in, ξ ψi ∈[0,1], 3. A safety control method for a power system with double-layer hidden Markov jumps according to claim 2, characterized in that: In step S2, the energy of the DoS attack is limited and may last until the Nth time point. The mode of the DoS attack is as follows: Mode 1: indicates successful data transmission (λ(μ) = 1); in this mode, the probability that the system will continue to succeed in the next time step is τ 11 , the probability of transmission channel failure is τ 12 , the probability of being attacked by DoS is τ 13 ; Mode 2: indicates communication transmission failure (λ(μ) = 2); in this mode, the probability of successful transmission in the next time step is τ 21 , the probability of transmission channel failure is τ 22 , the probability of being attacked by DoS is τ 23 ; Mode 3: represents transmission under DoS attack (λ(μ) = 3); the probability of normal transmission in the next time step is τ 31 , the probability of transmission channel failure is τ 32 , the probability of continuing to be attacked by DoS is τ 34 ; According to the same principle, we extend it to mode λ(μ)=N+1. Since the attack energy is limited when the attack reaches λ(μ)=N+2, the system may attack with probability τ in the next time step. (N+2)1 and τ (N+2)2 Return to Mode 1 or Mode 2.

4. A safety control method for a power system with double-layer hidden Markov jumps according to claim 3, characterized in that: In step S2, the specific processing process is as follows: S21: Consider using the Markov variable q(μ) to simulate the switching of DoS attacks and determine the transition probability matrix as follows: in, S22: The variable λ(μ) is observed through q(μ) and and determine the observation probability matrix as follows: in, S23: Analyze the situation where the transition probability matrix and the observation probability matrix are partially unknown: in: There are three specific situations: Case1: Case2: Case3:

5. A safety control method for a power system with double-layer hidden Markov jumps according to claim 4, characterized in that: In step S3, the specific processing process is as follows: S31: Build the following Hidden Markov Model-based asynchronous state feedback controller: in, represents the controller gain matrix, Variable θ ζ belong Indicates the transmission status. When ζ=1, it indicates that the data transmission is successful; when ζ≠1, it indicates that the current status information cannot be obtained and the last successfully transmitted data is used. S32: Then we get the following two-layer hidden Markov jump power system: Among them, x(t+1) represents the state variable of the t+1th node, y(t) is the measured output, is a given system parameter matrix, w(t) is a matrix in the interval Additional interference on is the sensor's measurement matrix, is the output matrix of the system; S33: Augment the double-layer hidden Markov jump power system in step S32 to obtain a closed-loop system model as follows: in, 6. A safety control method for a power system with double-layer hidden Markov jumps according to claim 5, characterized in that: In step S4, given a scalar σ>0, if there exists a matrix K i , ρ=a,b; and the symmetric matrix Then when the following inequality condition holds, the system in step S32 can ∞ Achieve random stability under performance index σ: in:

7. A safety control method for a power system with double-layer hidden Markov jumps according to claim 6, characterized in that: In step S5, the specific processing process is as follows: S51: For w(t)≡0 and any initial value, when the condition When it is established, the system in step S32 is stochastically stable; when all non-zero When the zero initial state satisfies the following inequality, the system in step S32 can reach the specified H ∞ Performance index σ: S52: Consider the following Lyapunov functionals: in, make get: in, Through the inequality below in step S51 and the above equation, we can obtain: in: S53: Through the inequality above in the inequality condition in step S4, we get: in, S54: Substitute the inequality in step S53 into the inequality below in the inequality condition in step S4 to obtain: S55: Then, based on the inequality Complementing with Schur, we can get the inequality below in the inequality condition in step S4; S56: According to the inequality in step S54, we get: If w(t)≡0, then The system is stochastically stable; Under zero initial conditions, w(t)≠0, we know The system is stochastically stable and satisfies the specified H ∞ Performance index σ.

8. A safety control method for a power system with double-layer hidden Markov jumps according to claim 7, characterized in that: In step S6, the specific processing process is as follows: S61: Order K i =T i -1 U i ; S62: Perform an equivalent transformation on the inequality below in the inequality condition of step S4 as follows: in: S63: Then, the simulation software is used to calculate the value of the controller gain matrix by giving the matrix parameters.

Images