Dissipative sliding mode control method under Delta operator framework
By designing a fuzzy state observer and a sliding mode controller under the Delta operator framework, the problem of sliding mode control method under the influence of unavailable time delay and external disturbance in the system state is solved, and the dissipation of the T-S fuzzy Delta operator system is realized. Sliding mode control is suitable for continuous and discrete systems, ensuring the stability and control effect of the system.
Patent Information
- Application Number
- CN202411065017.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-05
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2044-08-05
AI Technical Summary
The existing sliding mode control method is difficult to deal with the influence of time delay and external disturbances when the system state is unavailable, resulting in unsatisfactory control effect.
The dissipative sliding mode control method under the Delta operator framework is adopted. By establishing a dynamic model of the T-S fuzzy Delta operator system, a fuzzy state observer is designed, a fuzzy sliding mode surface and a fuzzy sliding mode controller are constructed, and the Lyapunov stability method is used. and convex optimization methods ensure asymptotic stability and strict dissipation of the system.
When the system state is unavailable or unknown, the impact of time delay and external disturbance on the controlled system is effectively solved, and the dissipation slip mode control of the T-S fuzzy Delta operator system is realized. It is suitable for continuous and discrete systems, ensuring that System stability and control effect.
Smart Images

Figure CN118963137A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a sliding mode control method, in particular to a dissipative sliding mode control method for a TS fuzzy Delta operator system with time lag and external disturbance when system state information is unavailable. Background Art
[0002] Among the existing control strategies, sliding mode control is an effective method to deal with model uncertainty, parameter changes and external disturbances. It has the advantages of strong robustness, fast response speed and easy implementation. Therefore, sliding mode control has been widely used in vehicle control, robot control and other fields. With the rapid development of science and technology, actual control systems are becoming more and more complex, and people's demand for high-speed sampling is increasing. If the mathematical model of the complex control system cannot be accurately given under high-speed sampling conditions, the control effect will be unsatisfactory. In order to make the control system stable and achieve the ideal control effect, the sliding mode control method combining the TS fuzzy model and the Delta operator framework came into being. In addition, when the control system suffers from time lag and external disturbances, the system state information cannot be fully obtained. The traditional sliding mode control method is difficult to meet the performance requirements of the system in this situation. The dissipative sliding mode control method based on the state observer estimates the unknown state of the system by designing the state observer and ensures the good operation of the system dissipative performance. Therefore, it has received widespread attention from scholars. Summary of the invention
[0003] In order to solve the problem that the existing sliding mode control method cannot handle the time lag and external disturbance affecting the system performance when the system state is not available, the present invention provides a dissipative sliding mode control method under the Delta operator framework. The method can realize the dissipative sliding mode control of the TS fuzzy Delta operator system with time lag and external disturbance when the system state is not available.
[0004] The objective of the present invention is achieved through the following technical solutions:
[0005] A dissipative sliding mode control method under a Delta operator framework comprises the following steps:
[0006] Step 1: Establish a dynamic model of the TS fuzzy Delta operator system with time lag and external disturbance, where:
[0007] The state space form of the dynamic model is:
[0008] System fuzzy rule i: If μ1(t) is μ2(t) is ,…,μ B (t) Yes So
[0009]
[0010] In the formula, i represents the fuzzy rule index; t represents the sampling time; μ1(t) is the first antecedent variable at time t, μ2(t) is the second antecedent variable at time t, and μ B (t) is the Bth antecedent variable at time t; is the first fuzzy set under the i-th fuzzy rule, is the second fuzzy set under the i-th fuzzy rule, is the Bth fuzzy set under the i-th fuzzy rule; x(t) is the state of the system at time t, x(td) is the state of the system with time delay at time t, d is the steady-state time delay; u(t) is the control input of the system at time t; y(t) is the measured output of the system at time t; φ(t) is the initial value of the system state; ν(t) is the external disturbance; A i , A id , B i , B iν , C and D are all known matrices of appropriate dimensions; δ(x(t)) is the Delta operator of the system state x(t) at time t;
[0011] Defuzzify the state space form of the dynamic model of the TS fuzzy Delta operator system and obtain:
[0012]
[0013] In the formula, τ i (μ(t)) is the membership function under the i-th fuzzy rule; represents the sum of the membership functions of F fuzzy rules; F is the total number of fuzzy rules;
[0014] Step 2: Design a fuzzy state observer for the dynamic model of the TS fuzzy Delta operator system with time delay and external disturbance established in step 1, where:
[0015] The formula of the fuzzy state observer is:
[0016]
[0017] In the formula, is the estimated value of the system state x(t) at time t, representing the state of the fuzzy state observer; is the estimated value of the system state x(td) with time lag at time t; is the estimated value of the system output y(t) at time t, representing the output of the fuzzy state observer; L i is the unknown gain matrix of the fuzzy state observer under the fuzzy rule of the i-th observer; ψ(t) is The initial value of
[0018] Defuzzify the fuzzy state observer and get:
[0019]
[0020] Step 3: Based on the dynamic model established in step 1 and the fuzzy state observer designed in step 2, the corresponding observation error system is obtained:
[0021]
[0022] Where δ(e(t)) is the Delta operator of the state error of the fuzzy state observer; is the state error of the fuzzy state observer; e(td) is the state error of the fuzzy state observer with time lag; is the output error of the fuzzy state observer;
[0023] Step 4: Based on the fuzzy state observer designed in step 2, construct the fuzzy sliding surface and the fuzzy sliding controller, where:
[0024] The specific form of the fuzzy sliding surface is:
[0025]
[0026] Where s(t) is the sliding mode function at time t; is the matrix B i The transpose of ; Q is the unknown matrix to be determined;
[0027] The specific form of the fuzzy sliding mode controller is:
[0028]
[0029] In the formula, K i is the unknown gain matrix of the fuzzy sliding mode controller; ε is a positive constant; yes The norm of ; sign(s(t)) is the sign function of the sliding mode function s(t);
[0030] Step 5: Using the fuzzy sliding surface and fuzzy sliding mode controller constructed in step 4, through the Lyapunov stability method, Delta operator method and convex optimization method, obtain the judgment conditions that ensure the asymptotic stability and strict dissipation of the fuzzy state observer system and the observation error system and the reachability of the sliding surface. The specific steps are as follows:
[0031] (1) If the external disturbance ν(t) and the output error z(t) of the fuzzy state observer satisfy:
[0032]
[0033] Then the fuzzy state observer system and the observation error system satisfy strict dissipativeness; where ν T (t) is the transpose of the external disturbance ν(t); λ is the limit of the dissipative performance; h t is a number greater than 0; J(t) is the energy supply function;
[0034] (2) To ensure strict dissipativeness, the Lyapunov stability method is:
[0035]
[0036] in:
[0037] V(t)=V1(t)+V2(t)+V3(t)+V4(t)+V5(t)
[0038]
[0039]
[0040] Where V(t) is the Lyapunov function at time t; δ(V(t)) is the Delta operator of V(t); yes The transpose of T (t) is the transpose of e(t), yes The transpose of T (t-ih) is the transpose of e(t-ih), w(t-jh)=e(t-jh)-e(t-(j-1)h), w T (t-jh) is the transpose of w(t-jh); Q, R, S and P are the positive definite matrices to be determined;
[0041] According to the above Lyapunov stability method, the judgment conditions for the fuzzy state observer system and the observation error system to satisfy asymptotic stability and strict dissipativeness are obtained as follows:
[0042] Λ ii <0,i=j
[0043] Λ ij +Λ ji <0,1≤i<j≤F
[0044] in:
[0045]
[0046] In the formula, * represents the symmetric part of the block matrix; Yes ij The first row and first column of the block matrix, Yes ij The first row and second column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 1st row and 4th column of the block matrix; (A i -B i K j ) T is the matrix (A i -B i K j ), (A i -L i C) T is the matrix (A i -L i C), (B iν -L i D) T is the matrix (B iν -L i D) is the transpose of C T is the transpose of matrix C, D T is the transpose of matrix D, is the matrix L i The transpose of is a matrix of matrix M2; α and β are greater than A known constant of
[0047] (3) To ensure that the sliding surface meets the accessibility requirement, the Lyapunov stability method is:
[0048]
[0049] In the formula, W is the positive definite matrix to be determined;
[0050] According to the above Lyapunov stability method, the judgment condition that the sliding surface satisfies the accessibility is obtained as follows:
[0051]
[0052] Ξ ii <0,i=j
[0053]
[0054] Where:
[0055]
[0056] Among them, γ ij is the scalar quantity to be sought; Yes ij The first row and first column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, Yes ij The 3rd row and 3rd column of the block matrix, Yes ij The 4th row and 4th column of the block matrix, Yes ij The 5th row and 5th column of the block matrix; is the matrix B i The transpose of is the matrix B j The transpose of is the matrix K j The transpose of is the matrix A id The transpose of is the matrix B iν The transpose of is a matrix The transpose of
[0057] Step 6: According to the judgment conditions of step 5, the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix are obtained. The specific steps are as follows:
[0058] Step 61: Initialize parameters α>0, β>0, ε>0, λ>0, and the maximum number of iterations N, and set t=0;
[0059] Step 62: By solving the following convex optimization problem, we get
[0060] minκ
[0061]
[0062] in:
[0063]
[0064]
[0065] in, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 4th row and 7th column of the block matrix, yes The 7th row and 7th column of the block matrix, yes The first row and first column of the block matrix, yes The first row and second column of the block matrix, yes The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column block matrix of 11 yes The first row and first column of the block matrix, Γ 12 yes The first row and second column of the block matrix, Γ 13 yes The first row and third column of the block matrix, Σ 11 yes The first row and first column of the block matrix, Σ 22 yes The 2nd row and 2nd column block matrix, Σ 33 yes The 3rd row and 3rd column of the block matrix, is the matrix Y j The transpose of is the matrix X i The transpose of
[0066] Step 63: If t<N, set t=t+0.15 and execute step 62; otherwise, execute the next step;
[0067] Step 64: If t ≥ N, exit the loop, find a feasible solution to the convex optimization problem, and pass L i =Q -1 X i and K j =Δ -1 Y j Obtain the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix;
[0068] Step 7: Substitute the fuzzy state observer gain matrix and fuzzy sliding mode controller gain matrix obtained in step 6 back into the fuzzy state observer and fuzzy sliding mode controller constructed in step 4 to realize dissipative sliding mode control of the TS fuzzy Delta operator system.
[0069] Compared with the prior art, the present invention has the following advantages:
[0070] 1. The present invention solves the dissipative sliding mode control problem in which the system state cannot be obtained under the influence of time lag and external disturbance. Compared with the traditional sliding mode control problem, the present invention establishes a unified model of TS fuzzy continuous system and TS fuzzy discrete system, and the obtained results are applicable to both the dissipative sliding mode control problem of TS fuzzy continuous system and the dissipative sliding mode control problem of TS fuzzy discrete system; a fuzzy state observer is designed to estimate the unknown state of the controlled system, solving the problem that the system state cannot be obtained in practical problems; the present invention simultaneously considers the influence of time lag and external disturbance on the dissipative performance of the controlled system. The present invention is applicable to the dissipative sliding mode control of TS fuzzy Detla operator system.
[0071] 2. The present invention establishes a unified model of TS fuzzy continuous system and TS fuzzy discrete system under the framework of Delta operator; solves the unknown phenomenon of system state by constructing fuzzy state observer; designs observer-dependent fuzzy sliding surface and fuzzy sliding controller; uses Lyapunov method, Delta operator method and convex optimization method to obtain the judgment conditions that ensure the asymptotic stability and strict dissipation of fuzzy state observer system and observation error system and the observer state can still be driven to the sliding surface under the influence of time lag and external disturbance. In the experiment of the present invention, when the sampling period h=0, the obtained results are consistent with the TS fuzzy continuous system; when the sampling period h=1, the obtained results are consistent with the TS fuzzy discrete system.
[0072] 3. The present invention can solve the dissipative sliding mode control problem of the TS fuzzy Delta operator system with time delay and external disturbance when the system state is unavailable or unknown. BRIEF DESCRIPTION OF THE DRAWINGS
[0073] Figure 1 It is a flow chart of the dissipative sliding mode control method under the Delta operator framework of the present invention;
[0074] Figure 2 is the true state trajectory diagram of the TS fuzzy Delta operator system; where “——” represents the trajectory of the first component of the system state, “---” represents the trajectory of the second component of the system state, and “-·-·-” represents the trajectory of the third component of the system state;
[0075] Figure 3 is the state trajectory diagram of the fuzzy state observer (i.e., the state estimation trajectory diagram of the TS fuzzy Delta operator system); wherein, “——” represents the trajectory of the first component of the fuzzy state observer state, “---” represents the trajectory of the second component of the fuzzy state observer state, and “-·-·-” represents the trajectory of the third component of the fuzzy state observer state;
[0076] Figure 4is the trajectory diagram of the observer error; where “——” represents the trajectory of the first component of the observer error, “---” represents the trajectory of the second component of the observer error, and “-·-·-” represents the trajectory of the third component of the observer error;
[0077] Figure 5 is the trajectory diagram of the sliding mode function; where “-” represents the trajectory of the first component of the sliding mode function, and “---” represents the trajectory of the second component of the sliding mode function;
[0078] Figure 6 is the trajectory diagram of the sliding mode controller; where “-” represents the trajectory of the first component of the sliding mode controller, and “---” represents the trajectory of the second component of the sliding mode controller. DETAILED DESCRIPTION
[0079] The technical solution of the present invention is further described below in conjunction with the accompanying drawings, but is not limited thereto. Any modification or equivalent replacement of the technical solution of the present invention without departing from the spirit and scope of the technical solution of the present invention should be included in the protection scope of the present invention.
[0080] The present invention provides a dissipative sliding mode control method under the Delta operator framework, such as Figure 1 As shown, the method comprises the following steps:
[0081] Step 1: Establish a dynamic model of the TS fuzzy Delta operator system with time delay and external disturbance.
[0082] In this step, the state space form of the dynamic model is:
[0083] System fuzzy rule i: If μ1(t) is μ2(t) is …, μ B (t) Yes So
[0084]
[0085] In the formula, i represents the fuzzy rule index; t represents the sampling time, satisfying t = ch, where c is a positive integer, and h represents the sampling period; μ1(t) is the first antecedent variable at time t, μ2(t) is the second antecedent variable at time t, and μ B (t) is the Bth antecedent variable at time t; is the first fuzzy set under the i-th fuzzy rule, is the second fuzzy set under the i-th fuzzy rule, is the Bth fuzzy set under the i-th fuzzy rule; is the state of the system at time t, x(td) is the state of the system with time delay at time t, d is a constant time delay satisfying d = nh, and n is a positive integer; is the control input of the system at time t; is the measured output of the system at time t; φ(t) is the initial value of the system state; the external disturbance belongs to l2[0,+∞), and its upper bound is is a constant greater than 0; in addition, A i , A id , B i , B iν , C and D are all known matrices of appropriate dimensions; Yes x Violette space, Yes u Violette space, Yes y Violette space, Yes ν dimensional Euclidean space; δ(x(t)) is the Delta operator of the system state x(t) at time t, which is defined as:
[0086]
[0087] In the formula, is the derivative of the system state x(t) at time t, and x(t+h) is the state of the system at time t+h; when the sampling period h=0, the system (1) represents a TS fuzzy continuous system; when the sampling period h≠0, the system (1) represents a TS fuzzy discrete system; it can be seen that the system (1) is a unified framework of the TS fuzzy continuous system and the TS fuzzy discrete system.
[0088] Defuzzifying the state space form of the dynamic model of the TS fuzzy Delta operator system yields:
[0089]
[0090] In the formula, τ i (μ(t)) is the membership function under the i-th fuzzy rule, which satisfies and 0≤τ i (μ(t))≤1; represents the sum of the membership functions of F fuzzy rules; F is the total number of fuzzy rules; membership function τ i (μ(t)) is defined as follows:
[0091]
[0092] Among them, θ ig (μg (t)) represents the g-th antecedent variable μ under the i-th fuzzy rule g The degree of membership of (t); It is the product of b membership degrees under the i-th fuzzy rule.
[0093] Step 2: Design a fuzzy state observer for the dynamic model of the TS fuzzy Delta operator system with time delay and external disturbance established in step 1.
[0094] In this step, the formula of the fuzzy state observer is:
[0095] Observer fuzzy rule i: If μ1(t) is μ2(t) is …, μ B (t) Yes So
[0096]
[0097] In the formula, is the estimated value of the system state x(t) at time t, representing the state of the fuzzy state observer; is the estimated value of the system state x(td) with time lag at time t; is the estimated value of the system output y(t) at time t, representing the output of the fuzzy state observer; L i is the unknown gain matrix of the fuzzy state observer under the fuzzy rule of the i-th observer; ψ(t) is The initial value of
[0098] Defuzzifying the fuzzy state observer, we can get:
[0099]
[0100] Step 3: According to the dynamic model established in step 1 and the fuzzy state observer designed in step 2, the corresponding observation error system is obtained.
[0101] In this step, the state error of the fuzzy state observer is defined as The output error of the fuzzy state observer is defined as According to the defuzzified TS fuzzy Delta operator system (2) and the defuzzified fuzzy state observer (4), the corresponding observation error system can be obtained:
[0102]
[0103] Where δ(e(t)) is the Delta operator of the state error of the fuzzy state observer; e(td) is the state error of the fuzzy state observer with time lag.
[0104] Step 4: Based on the fuzzy state observer designed in step 2, construct the fuzzy sliding surface and the fuzzy sliding controller.
[0105] In this step, the specific form of the fuzzy sliding surface is:
[0106]
[0107] Where s(t) is the sliding mode function at time t; is the matrix B i is the transpose of ; Q is the unknown matrix to be determined.
[0108] In this step, the specific form of the fuzzy sliding mode controller is:
[0109]
[0110] In the formula, K i is the unknown gain matrix of the fuzzy sliding mode controller; ε is a positive constant; yes The norm of ; sign(s(t)) is the sign function of the sliding mode function s(t).
[0111] Step 5: Using the fuzzy sliding surface and fuzzy sliding mode controller constructed in step 4, through the Lyapunov stability method, Delta operator method and convex optimization method, obtain the judgment conditions that ensure the asymptotic stability and strict dissipation of the fuzzy state observer system and the observation error system and the reachability of the sliding surface. The specific steps are as follows:
[0112] (1) If the external disturbance ν(t) and the output error z(t) of the fuzzy state observer satisfy:
[0113]
[0114] Then the fuzzy state observer system and the observation error system satisfy strict dissipativeness; where ν T (t) is the transpose of the external disturbance ν(t); λ is the limit of dissipative performance, and λ>0; h t is a number greater than 0; J(t) is the energy supply function, which is defined as follows:
[0115] J(t)=z T (t)M1z(t)+2z T (t)M2ν(t)+ν T (t)M3ν(t)
[0116] Where M1 is a semi-negative definite real matrix; M2 is a real matrix; M3 is a symmetric real matrix; z T (t) is the transpose of z(t).
[0117] (2) To ensure strict dissipativeness, the Lyapunov stability method is:
[0118]
[0119] in:
[0120] V(t)=V1(t)+V2(t)+V3(t)+V4(t)+V5(t)
[0121]
[0122] Where V(t) is the Lyapunov function at time t; δ(V(t)) is the Delta operator of V(t); yes The transpose of T (t) is the transpose of e(t), yes The transpose of T (t-ih) is the transpose of e(t-ih), w(t-jh)=e(t-jh)-e(t-(j-1)h), w T (t-jh) is the transpose of w(t-jh); Q, R, S and P are the positive definite matrices to be determined;
[0123] According to the above Lyapunov stability method, the judgment conditions for the fuzzy state observer system and the observation error system to satisfy asymptotic stability and strict dissipativeness are obtained as follows:
[0124] Λ ii <0,i=j
[0125] Λ ij +Λ ji <0,1≤i<j≤F
[0126] in:
[0127]
[0128] In the formula, * represents the symmetric part of the block matrix; Yes ij The first row and first column of the block matrix, Yes ij The first row and second column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 1st row and 4th column of the block matrix; (A i -B i K j ) T is the matrix (A i -B i K j ), (A i -L i C) T is the matrix (A i -L i C), (B iν -L i D) T is the matrix (B iν -L i D) is the transpose of C T is the transpose of matrix C, D T is the transpose of matrix D, is the matrix L i The transpose of is a matrix of matrix M2; Q, R, S and P are positive definite matrices to be determined; α and β are greater than known constant.
[0129] (3) To ensure that the sliding surface meets the accessibility requirement, the Lyapunov stability method is:
[0130]
[0131] In the formula, W is the positive definite matrix to be determined;
[0132] According to the above Lyapunov stability method, the judgment condition that the sliding surface satisfies the accessibility is obtained as follows:
[0133]
[0134] Where:
[0135]
[0136] Among them, γ ij is the scalar quantity to be sought; Yes ij The first row and first column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, Yes ij The 3rd row and 3rd column of the block matrix, Yes ij The 4th row and 4th column of the block matrix, Yes ijThe 5th row and 5th column of the block matrix; is the matrix B i The transpose of is the matrix B j The transpose of is the matrix K j The transpose of is the matrix A id The transpose of is the matrix B iν The transpose of is a matrix The transpose of .
[0137] Step 6: According to the judgment conditions of step 5, the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix are obtained. The specific steps are as follows:
[0138] Step 61: Initialize parameters α>0, β>0, ε>0, λ>0, and the maximum number of iterations N, and set t=0;
[0139] Step 62: By solving the following convex optimization problem, we get
[0140] minκ
[0141]
[0142] in:
[0143]
[0144]
[0145] in, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 4th row and 7th column of the block matrix, yes The 7th row and 7th column of the block matrix, yes The first row and first column of the block matrix, yes The first row and second column of the block matrix, yes The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column block matrix of 11 yes The first row and first column of the block matrix, Γ 12 yes The first row and second column of the block matrix, Γ 13 yes The first row and third column of the block matrix, Σ 11 yes The first row and first column of the block matrix, Σ 22 yes The 2nd row and 2nd column block matrix, Σ 33 yes The 3rd row and 3rd column of the block matrix, is the matrix Y j The transpose of is the matrix X i The transpose of
[0146] Step 63: If t<N, set t=t+0.15 and execute step 62; otherwise, execute the next step;
[0147] Step 64: If t ≥ N, exit the loop, find a feasible solution to the convex optimization problem, and pass L i =Q -1 X i and K j =Δ -1 Y j Obtain the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix.
[0148] Step 7: Substitute the fuzzy state observer gain matrix and fuzzy sliding mode controller gain matrix obtained in step 6 back into the fuzzy state observer and fuzzy sliding mode controller constructed in step 4 to realize dissipative sliding mode control of the TS fuzzy Delta operator system.
[0149] The following examples are used to verify the beneficial effects of the present invention:
[0150] Consider a TS fuzzy Delta operator system with two fuzzy rules. The specific selection of all parameters involved in the system is as follows:
[0151] System parameters:
[0152]
[0153] d=1,α=0.2,β=0.3,ε=0.4
[0154] The external disturbance is expressed as ν(t) = 2sin(0.6t)exp(-0.8t); the strict dissipative performance limit is λ = 0.2, and other dissipative performance parameters are:
[0155] M3=0.5I
[0156] In addition, the sampling period is h = 0.15, and the initial value of the system state is x(0) = [0 0 0] T (t≤0), the initial value of the fuzzy state observer is ; The membership function is selected as:
[0157]
[0158] τ2(μ1(t))=1-τ1(μ1(t))
[0159] By solving the convex optimization problem, we can obtain:
[0160]
[0161] use The fuzzy state observer gain matrix can be calculated as:
[0162]
[0163] use The fuzzy sliding mode controller gain matrix can be calculated as:
[0164]
[0165] The effect of the constructed sliding mode controller is shown in Figure 2-6 .Depend on Figure 2-3 It can be seen that the fuzzy state observer designed by the present invention can effectively estimate the true state of the TS fuzzy Delta operator system. Figure 4 It can be seen that the difference between the state estimated by the fuzzy state observer and the state of the real system is very small; Figure 5 and Figure 6 The trajectory diagrams of the fuzzy sliding mode variable and the fuzzy sliding mode control variable are described respectively. It can be seen that the curves of these two variables gradually tend to 0.
Claims
1. A dissipative sliding mode control method under the Delta operator framework, characterized in that The method comprises the following steps: Step 1: Establish a dynamic model of the TS fuzzy Delta operator system with time lag and external disturbance, where: The state space form of the dynamic model is: System fuzzy rule i: If μ1(t) is μ2(t) is μ B (t) Yes So In the formula, i represents the fuzzy rule index; t represents the sampling time; μ1(t) is the first antecedent variable at time t, μ2(t) is the second antecedent variable at time t, and μ B (t) is the Bth antecedent variable at time t; is the first fuzzy set under the i-th fuzzy rule, is the second fuzzy set under the i-th fuzzy rule, is the Bth fuzzy set under the i-th fuzzy rule; x(t) is the state of the system at time t, x(td) is the state of the system with time delay at time t, d is the steady-state time delay; u(t) is the control input of the system at time t; y(t) is the measured output of the system at time t; φ(t) is the initial value of the system state; ν(t) is the external disturbance; A i , A id , B i , B iν , C and D are all known matrices of appropriate dimensions; δ(x(t)) is the Delta operator of the system state x(t) at time t; Defuzzify the state space form of the dynamic model of the TS fuzzy Delta operator system and obtain: In the formula, τ i (μ(t)) is the membership function under the i-th fuzzy rule; represents the sum of the membership functions of F fuzzy rules; F is the total number of fuzzy rules; Step 2: Design a fuzzy state observer for the dynamic model of the TS fuzzy Delta operator system with time delay and external disturbance established in step 1, where: The formula of the fuzzy state observer is: In the formula, is the estimated value of the system state x(t) at time t, representing the state of the fuzzy state observer; is the estimated value of the system state x(td) with time lag at time t; is the estimated value of the system output y(t) at time t, representing the output of the fuzzy state observer; L i is the unknown gain matrix of the fuzzy state observer under the fuzzy rule of the i-th observer; ψ(t) is The initial value of Defuzzify the fuzzy state observer and get: Step 3: Based on the dynamic model established in step 1 and the fuzzy state observer designed in step 2, the corresponding observation error system is obtained: Where δ(e(t)) is the Delta operator of the state error of the fuzzy state observer; is the state error of the fuzzy state observer; e(td) is the state error of the fuzzy state observer with time lag; is the output error of the fuzzy state observer; Step 4: Based on the fuzzy state observer designed in step 2, construct the fuzzy sliding surface and the fuzzy sliding controller, where: The specific form of the fuzzy sliding surface is: Where s(t) is the sliding mode function at time t; is the matrix B i The transpose of ; Q is the unknown matrix to be determined; The specific form of the fuzzy sliding mode controller is: In the formula, K i is the unknown gain matrix of the fuzzy sliding mode controller; ε is a positive constant; yes The norm of ; sign(s(t)) is the sign function of the sliding mode function s(t); Step 5: Using the fuzzy sliding surface and fuzzy sliding mode controller constructed in step 4, through the Lyapunov stability method, Delta operator method and convex optimization method, obtain the judgment conditions that ensure the asymptotic stability and strict dissipation of the fuzzy state observer system and the observation error system and the reachability of the sliding surface. The specific steps are as follows: (1) If the external disturbance ν(t) and the output error z(t) of the fuzzy state observer satisfy: Then the fuzzy state observer system and the observation error system satisfy strict dissipativeness; where ν T (t) is the transpose of the external disturbance ν(t); λ is the limit of the dissipative performance; h t is a number greater than 0; J(t) is the energy supply function; (2) To ensure strict dissipativeness, the Lyapunov stability method is: in: V(t)=V1(t)+V2(t)+V3(t)+V4(t)+V5(t) Where V(t) is the Lyapunov function at time t; δ(V(t)) is the Delta operator of V(t); yes The transpose of T (t) is the transpose of e(t), yes The transpose of T (t-ih) is the transpose of e(t-ih), w(t-jh)=e(t-jh)-e(t-(j-1)h), w T (t-jh) is the transpose of w(t-jh); Q, R, S and P are the positive definite matrices to be determined; According to the above Lyapunov stability method, the judgment conditions for the fuzzy state observer system and the observation error system to satisfy asymptotic stability and strict dissipativeness are obtained as follows: L ii <0,i=j L ij +L ji <0.1≤i<j≤F in: In the formula, * represents the symmetric part of the block matrix; Yes ij The first row and first column of the block matrix, Yes ij The first row and second column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 1st row and 4th column of the block matrix; (A i -B i K j ) T is the matrix (A i -B i K j ), (A i -L i C) T is the matrix (A i -L i C), (B iν -L i D) T is the matrix (B iν -L i D) is the transpose of C T is the transpose of matrix C, D T is the transpose of matrix D, is the matrix L i The transpose of is a matrix of matrix M2; α and β are greater than A known constant of (3) To ensure that the sliding surface meets the accessibility requirement, the Lyapunov stability method is: In the formula, W is the positive definite matrix to be determined; According to the above Lyapunov stability method, the judgment condition that the sliding surface satisfies the accessibility is obtained as follows: X ii <0,i=j Where: Among them, γ ij is the scalar quantity to be sought; Yes ij The first row and first column of the block matrix, Yes ij The 2nd row and 2nd column block matrix, Yes ij The 3rd row and 3rd column of the block matrix, Yes ij The 4th row and 4th column of the block matrix, Yes ij The 5th row and 5th column of the block matrix; is the matrix B i The transpose of is the matrix B j The transpose of is the matrix K j The transpose of is the matrix A id The transpose of is the matrix B iν The transpose of is a matrix The transpose of Step 6: According to the judgment conditions of step 5, the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix are obtained. The specific steps are as follows: Step 61: Initialize parameters α>0, β>0, ε>0, λ>0, and the maximum number of iterations N, and set t=0; Step 62: By solving the following convex optimization problem, we get minκ s.t. in: in, yes The first row and first column of the block matrix, yes The 4th row and 4th column of the block matrix, yes The 4th row and 7th column of the block matrix, yes The 7th row and 7th column of the block matrix, yes The first row and first column of the block matrix, yes The first row and second column of the block matrix, yes The 2nd row and 2nd column block matrix, yes The first row and first column of the block matrix, yes The 4th row and 4th column block matrix of 11 yes The first row and first column of the block matrix, Γ 12 yes The first row and second column of the block matrix, Γ 13 yes The first row and third column of the block matrix, Σ 11 yes The first row and first column of the block matrix, Σ 22 yes The 2nd row and 2nd column block matrix, Σ 33 yes The 3rd row and 3rd column of the block matrix, is the matrix Y j The transpose of is the matrix X i The transpose of Step 63: If t<N, set t=t+0.15 and execute step 62; otherwise, execute the next step; Step 64: If t ≥ N, exit the loop, find a feasible solution to the convex optimization problem, and pass L i =Q -1 X i and K j =Δ - 1 Y j Obtain the fuzzy state observer gain matrix and the fuzzy sliding mode controller gain matrix; Step 7: Substitute the fuzzy state observer gain matrix and fuzzy sliding mode controller gain matrix obtained in step 6 back into the fuzzy state observer and fuzzy sliding mode controller constructed in step 4 to realize dissipative sliding mode control of the TS fuzzy Delta operator system.
2. The dissipative sliding mode control method under the Delta operator framework according to claim 1 is characterized in that In step 1, δ(x(t)) is defined as: In the formula, is the derivative of the system state x(t) at time t, and x(t+h) is the state of the system at time t+h.
3. The dissipative sliding mode control method under the Delta operator framework according to claim 1 is characterized in that In the step 1, τ i (μ(t)) is defined as follows: Among them, θ ig (μ g (t)) represents the g-th antecedent variable μ under the i-th fuzzy rule g The degree of membership of (t); It is the product of b membership degrees under the i-th fuzzy rule.
4. The dissipative sliding mode control method under the Delta operator framework according to claim 1 is characterized in that In step 5, J(t) is defined as follows: J(t)=z T (t)M1z(t)+2z T (t)M2ν(t)+ν T (t)M3ν(t) Where M1 is a semi-negative definite real matrix; M2 is a real matrix; M3 is a symmetric real matrix; z T (t) is the transpose of z(t).
Citation Information
Patent Citations
Delta operator-based secondary chemical reactor fault estimation method
CN113159647A
