Nonlinear generalized Markov jump system finite time fault-tolerant control method based on observer
By designing a nonlinear generalized Markov jump system finite time fault-tolerant control method based on observer, the problem of insufficient model description and fault estimation capabilities in the prior art is solved, and accurate estimation and robust control of system status and faults are achieved , ensuring the transient performance and stability of the system for a limited time.
Patent Information
- Application Number
- CN202411031882.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
The prior art lacks flexible model description and fault estimation capabilities in nonlinear generalized Markov jump systems, especially when multiple faults occur simultaneously, and the robust performance is poor, so it cannot accurately estimate the system state and fault type. Maintain system stability and interference attenuation for a limited time.
A finite time fault-tolerant control method for nonlinear generalized Markov jump system based on observers was designed. By establishing a nonlinear model, designing a lower triangular matrix observer and a finite time fault-tolerant controller, combining the Lyapunov function and linear matrix inequality, Implement simultaneous estimation of system status, actuator failures and sensor failures, and ensure that the H∞ finite time boundary of error systems and closed-loop systems.
It realizes robustness estimation and control of nonlinear generalized Markov jump system, ensures the system's transient performance in a limited time, improves system stability and interference attenuation capabilities, and solves the flexibility of the model in the existing technology. Insufficient sexual and robust performance.
Smart Images

Figure CN118963129A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of nonlinear system control, and in particular to a finite-time fault-tolerant control method for a nonlinear generalized Markov jump system based on an observer. Background Art
[0002] A generalized Markov jump system is a random hybrid system composed of a finite number of subsystems or modes. The jumps between its modes follow the Markov process, which can accurately describe some real systems whose structures and parameters undergo random mutations due to environmental influences, such as financial economics, communication networks, mechanical engineering, etc. In recent years, with the wide application of Markov jump systems (MJS) in practice and the enrichment of generalized system theory, many scholars have begun to pay attention to the stability analysis and control problems of generalized Markov jump systems.
[0003] However, in the current research of MJS, the state of the real system is often impossible to observe directly or requires huge costs to measure, so it is extremely important to design a system controller based on the observer. On the other hand, fault-tolerant control is a popular research direction in control theory. In real life, various faults will occur during the operation of the system. These faults caused by the aging of actuators and sensors, the wear of internal components, and sudden changes in the working environment will have a bad impact on the performance of the system. It is often impossible to directly obtain the size and type of the fault. It is necessary to use fault estimation technology to diagnose and estimate the fault in a timely manner, and take measures to maintain the stable operation of the system in response to different types of faults. Therefore, fault-tolerant control is of great significance.
[0004] Nonlinear generalized Markov jump systems can currently describe real systems well, but there are few studies on fault estimation problems, especially the transient performance of real systems in finite time.
[0005] The existing technology considers MJS, but lacks certain modeling flexibility. MJS cannot accurately describe the characteristics of actual complex systems in real life; when multiple faults occur at the same time, general observers cannot accurately estimate the system state, actuator faults and sensor faults at the same time. The existing technology research on MJS fault-tolerant control mainly focuses on the Lyapunov asymptotic stability problem of MJS in infinite time domain, and does not pay attention to the transient performance of the system in a limited time in modern industrial systems; the general observer based on MJS has poor robust performance, cannot keep the system stable under abnormal conditions, and has a low interference attenuation level. Summary of the invention
[0006] Purpose of the invention: In view of the problems pointed out in the background technology, the present invention discloses a finite-time fault-tolerant control method for a nonlinear generalized Markov jump system based on an observer, designs a finite-time observer and a fault-tolerant controller, has good robustness to unknown inputs, can simultaneously estimate the state of the system, actuator faults and sensor faults, and ensures the H of the error system ∞ Finite time bounded, ensuring the transient performance of the system within a finite time.
[0007] Technical solution: The present invention discloses a finite-time fault-tolerant control method for a nonlinear generalized Markov jump system based on an observer, comprising the following steps:
[0008] Step 1: Establish a nonlinear generalized Markov jump system model and give the conditions that the nonlinear part must satisfy;
[0009] Step 2: Design a lower triangular matrix observer that can simultaneously estimate the system state, actuator faults, and sensor faults;
[0010] Step 3: Based on step 2, design a finite-time fault-tolerant controller and select the Lyapunov function. According to the linear matrix inequality LMI, find the error system and the closed-loop system H ∞ Finite time bounded conditions;
[0011] Step 4: Solve the fault-tolerant controller gain K based on the finite time bounded condition in step 3 i and the observed controller gain L i , and then control the nonlinear generalized Markov jump system.
[0012] Furthermore, the nonlinear generalized Markov jump system model constructed in step 1 is specifically as follows:
[0013]
[0014] in, is a singular matrix, assuming rank(E) = r < n, represents the state vector, represents the control input, Indicates the measurement output; and They represent actuator failure and sensor failure respectively; Indicates unknown input, represents nonlinear perturbation; is n-dimensional, m-dimensional, p-dimensional, q-dimensional, s-dimensional, k-dimensional, r-dimensional Euclidean space, is an n×n-dimensional Euclidean space, r(t) is a right-continuous Markov chain, taking values in the finite set S = {1,2,...,s}, and its state transition rate is:
[0015]
[0016] Among them, Δ>0, π ij is the transfer rate from mode i to j. When i≠j, π ij >0 and follows the formula:
[0017]
[0018] Among them, A(r(t)), B(r(t)), C(r(t)), D a (r(t)),D s (r(t)), H(r(t)) and G(r(t)) are concrete matrices of appropriate dimensions, B(r(t)) and D s (r(t)) is a column full rank matrix. For ease of writing, define r(t) = i and use the matrix A i , B i , C i , D si , D ai , H i and G i To express.
[0019] Furthermore, in step 1, the nonlinear part of the nonlinear generalized Markov jump system model satisfies the following conditions:
[0020] Condition 1) The nonlinear perturbation f(x,u) satisfies the one-sided Lipschitz condition as follows:
[0021]
[0022] Among them, δ i is a one-sided Lipschitz constant, f(x,u) is a function of the state vector x(t), It is an estimate of the nonlinear part f(x,u);
[0023] Condition 2) The nonlinear perturbation f(x,u) satisfies the one-sided Lipschitz condition as follows:
[0024]
[0025] Among them, ρ i and φ i are all known constants, and x is used to represent the state vector x(t). represents the estimate of the state vector x(t).
[0026] Furthermore, the specific process of designing the lower triangular matrix observer in step 2 is as follows:
[0027] Step 2.1: Perform a linear transformation of the lower triangular matrix for the nonlinear generalized Markov jump system:
[0028] Construct an augmented matrix, let:
[0029]
[0030] Formula (1) is equivalent to the following augmented system:
[0031]
[0032] in,
[0033] Defining new state variables Perform a lower triangular matrix linear transformation on equation (7):
[0034]
[0035] Formula (1) can be further expressed as:
[0036]
[0037] in, J is an arbitrary constant matrix, I m and I n They are all second-order identity matrices;
[0038] For equation (9), the following lower triangle observer is designed:
[0039]
[0040] in, is the correct answer to formula (9) The state estimate of Yes The derivation of is the observer output, L i is the observation controller gain matrix to be solved, i∈S;
[0041] Define the estimation error:
[0042]
[0043]
[0044] in, is the estimate of the state vector x(t), They are respectively a (t), f s (t) estimate;
[0045] Step 2.2: Subtract equation (9) from equation (10) to obtain the error system:
[0046]
[0047] in, Represents the new state variable The corresponding nonlinear perturbation is Represents the nonlinear part The estimate of η T (t) represents the transpose of the unknown input η(t), Represents the transpose of the first derivative of f(t).
[0048] Furthermore, the finite time fault-tolerant controller designed in step 3 is specifically:
[0049] Assume that for the matrix B i Any generalized inverse matrix of and any rank(B i )=rank([B i D ai ]), D ai , there is always Established;
[0050] The observer-based time-limited fault-tolerant controller has the following form:
[0051]
[0052] Among them, K i is the fault-tolerant controller gain, i∈S;
[0053] The closed-loop system composed of equation (12) and equation (1) has the following form:
[0054]
[0055] Among them, B i,μ =[-B i K i D ai 0],G1=[0G i ].
[0056] Furthermore, the Lyapunov function selected in step 3 is used to obtain the error system and the closed-loop system H according to the linear matrix inequality LMI.∞ The conditions for finite time are as follows:
[0057] Choose from the following Lyapunov functions:
[0058]
[0059] Where P i , W i is a positive definite symmetric matrix,
[0060] If for a given time constant T>0, there exists a constant c2>c1>0, given a constant φ i and constant d and the existence of constant θ 1i >0,θ 2i >0,θ 3i >0,θ 4i >0,β i >0,γ>0,ε<0,v0>0,v1>0,λ0>0,λ1>0,where i∈S,and there exists the following matrix W j >0,P i >0,P j >0,O i >0, R i >0 and the identity matrix I, so that the following inequality holds, then the closed-loop system is about (c1, c2, T, R i ,d)Satisfy H ∞ Finite time bounded conditions:
[0061]
[0062] v0R i ≤λ0R i ≤O i ≤λ1R i ≤v1R i (18)
[0063] in,
[0064]
[0065] Beneficial effects:
[0066] The present invention aims at a nonlinear generalized Markov jump system. Assuming that the generalized Markov jump system is regular and pulse-free, a fault estimation observer and a finite-time fault-tolerant controller with a lower triangular matrix are designed when actuator fault and sensor fault occur simultaneously. First, the original system is equivalently converted into a newly expanded system through the linear transformation of the lower triangular matrix. A more comprehensive lower triangular matrix observer is designed for the expanded system, which realizes state estimation, actuator fault estimation and sensor fault estimation at the same time. On the above basis, a finite-time fault-tolerant controller is designed to ensure the H of the error system and the closed-loop system. ∞ It is bounded in finite time, and by constructing the Lyapunov function, a sufficient condition for the finite time boundedness of the generalized Markov jump system is given.
[0067] In actual systems under complex conditions, the lower triangular matrix observer is used to solve the practical problem that the system state cannot be accurately measured in actual engineering applications, or it requires huge costs to obtain. With its good robustness to unknown inputs, the system can also accurately observe the system state and multiple faults at the same time under complex operating conditions. On the other hand, a finite-time fault-tolerant controller is designed for the system, which can greatly improve the transient performance of the system within a limited time, thereby solving the problem of resource and cost waste. BRIEF DESCRIPTION OF THE DRAWINGS
[0068] Figure 1 It is the principle diagram of nonlinear generalized Markov jump system;
[0069] Figure 2 The technical roadmap of the control method of the present invention DETAILED DESCRIPTION
[0070] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and cannot be used to limit the protection scope of the present invention.
[0071] The present invention aims at a nonlinear generalized Markov jump system. Assuming that the generalized Markov jump system is regular and pulse-free, a finite-time observer and a fault-tolerant controller are designed when actuator fault and sensor fault occur simultaneously. First, the original system is equivalently converted into a newly expanded system through the linear transformation of the lower triangular matrix. A more comprehensive lower triangular matrix observer is designed for the expanded system, which realizes state estimation, actuator fault estimation and sensor fault estimation at the same time. On the above basis, a finite-time fault-tolerant controller is designed to ensure the H of the error system and the closed-loop system. ∞ Finite time bounded. Then, by constructing the Lyapunov function, sufficient conditions for the finite time boundedness of the generalized Markov jump system are given.
[0072] Step 1: Establish a nonlinear generalized Markov jump system model and give the conditions that the nonlinear part must satisfy;
[0073] Step 2: Design a lower triangular matrix observer that can simultaneously estimate the system state, actuator faults, and sensor faults;
[0074] Step 3: Based on step 2, design a finite-time fault-tolerant controller and select the Lyapunov function. According to the linear matrix inequality (LMI), find the error system and the closed-loop system H ∞ Finite time bounded conditions;
[0075] Step 4: Demonstrate the effectiveness of the designed lower triangular matrix observer and finite-time fault-tolerant controller.
[0076] 2. Refine each of the above steps.
[0077] Step 1: Establish a nonlinear generalized Markov jump system model and give the conditions that the nonlinear part must satisfy.
[0078] Step 1.1: Nonlinear generalized Markov jump system model
[0079] The present invention targets a type of nonlinear generalized Markov jump system having the following form:
[0080]
[0081] Where, assuming rank(E) = r < n, represents the state vector, represents the control input, Indicates the measurement output; and They represent actuator failure and sensor failure respectively; Indicates unknown input, represents a nonlinear perturbation. yes is a singular matrix, n-dimensional, m-dimensional, p-dimensional, q-dimensional, s-dimensional, k-dimensional, r-dimensional Euclidean space, is an n×n-dimensional Euclidean space, r(t) is a right-continuous Markov chain, taking values in the finite set S = {1,2,...,s}, and its state transition rate is
[0082]
[0083] Where Δ>0, π ij is the transfer rate from mode i to j. When i≠j, π ij >0 and follows the formula:
[0084]
[0085] Among them, A(r(t)), B(r(t)), C(r(t)), D a (r(t)),D s (r(t)), H(r(t)) and G(r(t)) are concrete matrices of appropriate dimensions. B(r(t)) and D s (r(t)) is a column full rank matrix. For ease of writing, define r(t) = i, and use the matrix A i , B i , C i , D si , D ai , H i and G i To express.
[0086] Step 1.2 gives the conditions that the nonlinear part satisfies
[0087] Assumption 1 The nonlinear part of the system f(x,u) satisfies the unilateral Lipschitz condition as follows
[0088]
[0089] Among them, δ i is a one-sided Lipschitz constant, f(x,u) is a function of the state x(t), It is an estimate of the nonlinearity f(x,u);
[0090] Assumption 2 The nonlinear perturbation f(x,u) satisfies the unilateral Lipschitz condition as follows
[0091]
[0092] Among them, ρ i and φ i are all known constants, and x is used to represent the state vector x(t). represents the estimate of the state vector x(t).
[0093] Step 2: Lower triangular matrix observer design
[0094] Step 2.1: Perform a lower triangular matrix linear transformation on the nonlinear generalized Markov jump system (1):
[0095] First, construct an augmented matrix, let
[0096]
[0097] The original system (1) is equivalently transformed into the following augmented system:
[0098]
[0099] in,
[0100] Defining new state variables Perform a lower triangular matrix linear transformation on system (7).
[0101]
[0102] System (1) can be further expressed as:
[0103]
[0104] in, J is an arbitrary constant matrix, I m and I n are all second-order identity matrices.
[0105] For system (9), the following lower triangle observer is designed:
[0106]
[0107] in, is the state estimate of the system (9), Yes The derivation of is the observer output, L i (i∈S) is the observation controller gain matrix to be solved.
[0108] Defining the estimation error
[0109]
[0110] in, is the estimate of the state vector x(t), f a (t), f s (t) is estimated.
[0111] Step 2.2: Get the error system
[0112] Then subtract the system (9) from the observer (10) to obtain the error system
[0113]
[0114] in, Represents the new state variable The corresponding nonlinear perturbation is Represents the nonlinear part The estimate of η T (t) represents the transpose of the unknown input η(t), Represents the transpose of the first derivative of f(t).
[0115] Step 3: Based on step 2, design a finite-time fault-tolerant controller and select the Lyapunov function. According to the linear matrix inequality (LMI), find the error system and the closed-loop system H ∞ Finite time bounded conditions;
[0116] Step 3.1: Design a finite-time fault-tolerant controller
[0117] Assumption 3
[0118] Lemma 1 For the matrix B i Any generalized inverse matrix of and any D satisfying Assumption 3 ai , there is always Established.
[0119] The observer-based fault-tolerant controller has the following form:
[0120]
[0121] Among them, K i (i∈S) is the fault-tolerant controller gain.
[0122] The closed-loop system consisting of the controller (12) and the system (1) has the following form:
[0123]
[0124] Among them, B i,μ =[-B i K i D ai 0],G1=[0G i ].
[0125] Step 3.2: Select the Lyapunov function
[0126] Choose from the following Lyapunov functions:
[0127]
[0128] Among them, P i , W i are all positive definite symmetric matrices,
[0129] Step 3.3: Sufficient conditions for the existence of finite-time observers and fault-tolerant controllers
[0130] If for a given time constant T>0, there exists a constant c2>c1>0, and the constant φ i and d, and the existence of a constant θ 1i >0,θ 2i >0,θ 3i >0,θ 4i >0,β i >0,γ>0,ε<0,v0>0,v1>0,λ0>0,λ1>0,where i∈S,and there exists the following matrix W j >0,P i >0,P j >0,O i >0, R i >0 and the identity matrix I, so that the following inequality holds, then the closed-loop system is about (c1, c2, T, R i ,d)Satisfy H ∞ Finite time bounded conditions:
[0131]
[0132] v0R i ≤λ0R i ≤O i ≤λ1R i ≤v1R i (18)
[0133] in,
[0134]
[0135]
[0136] Step 4: Demonstrate the effectiveness of the designed lower triangular matrix observer and finite-time fault-tolerant controller.
[0137] Step 4.1: Demonstrate that the designed lower triangular matrix observer can effectively estimate the system state, actuator faults, and sensor faults;
[0138] Step 4.2: Demonstrate that the designed fault-tolerant controller can make the error system and the closed-loop system H ∞ Bounded in finite time.
[0139] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent transformation or modification made according to the spirit of the present invention should be included in the protection scope of the present invention.
Claims
1. An observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems, characterized in that: The steps include: Step 1: Establish a nonlinear generalized Markov jump system model and give the conditions that the nonlinear part must satisfy; Step 2: Design a lower triangular matrix observer that can simultaneously estimate the system state, actuator faults, and sensor faults; Step 3: Based on step 2, design a finite-time fault-tolerant controller and select the Lyapunov function. According to the linear matrix inequality LMI, find the error system and the closed-loop system H ∞ Finite time bounded conditions; Step 4: Solve the fault-tolerant controller gain K based on the finite time bounded condition in step 3 i and the observed controller gain L i , and then control the nonlinear generalized Markov jump system.
2. The observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems according to claim 1, characterized in that: The nonlinear generalized Markov jump system model constructed in step 1 is specifically as follows: in, is a singular matrix, assuming represents the state vector, represents the control input, Indicates the measurement output; and They represent actuator failure and sensor failure respectively; Indicates unknown input, represents nonlinear perturbation; is n-dimensional, m-dimensional, p-dimensional, q-dimensional, s-dimensional, k-dimensional, r-dimensional Euclidean space, is an n×n-dimensional Euclidean space, r(t) is a right-continuous Markov chain, taking values in the finite set S = {1,2,...,s}, and its state transition rate is: Among them, Δ>0, π ij is the transfer rate from mode i to j. When i≠j, π ij >0 and follows the formula: Among them, A(r(t)), B(r(t)), C(r(t)), D a (r(t)),D s (r(t)), H(r(t)) and G(r(t)) are concrete matrices of appropriate dimensions, B(r(t)) and D s (r(t)) is a column full rank matrix. For ease of writing, define r(t) = i and use the matrix A i , B i , C i , D si , D ai , H i and G i To express.
3. The observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems according to claim 2, characterized in that: In step 1, the nonlinear part of the nonlinear generalized Markov jump system model satisfies the following conditions: Condition 1) The nonlinear perturbation f(x,u) satisfies the one-sided Lipschitz condition as follows: Among them, δ i is a one-sided Lipschitz constant, f(x,u) is a function of the state vector x(t), It is an estimate of the nonlinear part f(x,u); Condition 2) The nonlinear perturbation f(x,u) satisfies the one-sided Lipschitz condition as follows: Among them, ρ i and φ i are all known constants, and x is used to represent the state vector x(t). represents the estimate of the state vector x(t).
4. The observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems according to claim 2, characterized in that: The specific process of designing the lower triangular matrix observer in step 2 is as follows: Step 2.1: Perform a linear transformation of the lower triangular matrix for the nonlinear generalized Markov jump system: Construct an augmented matrix, let: Formula (1) is equivalent to the following augmented system: in, Define a new state variable x(t) and perform a lower triangular matrix linear transformation on equation (7): Formula (1) can be further expressed as: in, J is an arbitrary constant matrix, I m and I n They are all second-order identity matrices; For equation (9), the following lower triangle observer is designed: in, is the correct answer to formula (9) The state estimate of Yes The derivation of is the observer output, L i is the observation controller gain matrix to be solved, i∈S; Define the estimation error: in, is the estimate of the state vector x(t), f a (t), f s (t) estimate; Step 2.2: Subtract equation (9) from equation (10) to obtain the error system: in, represents the nonlinear perturbation corresponding to the new state variable x(t), Represents the nonlinear part The estimate of η T (t) represents the transpose of the unknown input η(t), Represents the transpose of the first derivative of f(t).
5. The observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems according to claim 4, characterized in that: The finite time fault-tolerant controller designed in step 3 is specifically: Assume that for the matrix B i Any generalized inverse matrix of and any rank(B i )=rank([B i D ai ]), D ai , there is always Established; The observer-based time-limited fault-tolerant controller has the following form: Among them, K i is the fault-tolerant controller gain, i∈S; The closed-loop system composed of equation (12) and equation (1) has the following form: Among them, B i,μ =[-B i K i D ai 0],G1=[0G i ].
6. The observer-based finite-time fault-tolerant control method for nonlinear generalized Markov jump systems according to claim 5, characterized in that: The Lyapunov function selected in step 3 is used to solve the error system and the closed-loop system H according to the linear matrix inequality LMI. ∞ The conditions for finite time are as follows: Choose from the following Lyapunov functions: Where P i , W i is a positive definite symmetric matrix, P i >0,W i >0; If for a given time constant T>0, there exists a constant c2>c1>0, given a constant φ i and constant d and the existence of constant θ 1i >0,θ 2i >0,θ 3i >0,θ 4i >0,β i >0,γ>0,ε<0,v0>0,v1>0,λ0>0,λ1>0,where i∈S,and there exists the following matrix W j >0,P i >0,P j >0,O i >0, R i >0 and the identity matrix I, so that the following inequality holds, then the closed-loop system is about (c1, c2, T, R i ,d)Satisfy H ∞ Finite time bounded conditions: v0R i ≤λ0R i ≤O i ≤λ1R i ≤v1R i (18) in,
