Large-scale power system load frequency integral sliding mode control method based on disturbance observer

Abstract

The invention provides a disturbance observer-based integral sliding mode control method for a large-scale power system, which is applied to the field of load frequency sliding mode control of the large-scale power system and comprises the following steps of: 1, establishing a dynamic model of the large-scale power system; 2, designing a disturbance observer; 3, designing a memory-free integral sliding mode control law; 4, designing an integral sliding mode control law based on memory; and 5, by taking the three-region interconnected power system as an example, verifying the effectiveness and optimization effect of the method. According to the method, a composite anti-interference control strategy for processing system load disturbance is comprehensively analyzed, a disturbance observer is designed to estimate the load disturbance, and meanwhile, a sliding mode control scheme is designed by using a disturbance estimation value to eliminate the negative influence of the load disturbance on the system.

Description

Technical Field

[0001] The present invention relates to the field of large-scale interconnected power system control, and in particular to a large-scale power system load frequency integral sliding mode control method based on a disturbance observer. Background Art

[0002] As the scale and complexity of modern power systems continue to increase, large-scale interconnected power systems have gradually become a research focus due to their multi-dimensional, highly coupled, and strongly interconnected characteristics. Load frequency control is widely used to ensure the power quality and stable operation of large-scale interconnected power systems because of its advantages in helping to restore system frequency and eliminate power deviations between interconnection lines. Scholars have proposed a variety of load frequency control schemes for power system safety control, optimization problems, and communication performance. However, due to the complex structure of large-scale interconnected power systems and the physical limitations of information transmission in various regions, it is difficult to study them. It is imperative to propose a load frequency scheme with better performance such as precise control and fast response.

[0003] Sliding mode control is widely used in smart grids, vehicle control, underwater robots and other fields due to its fast dynamic response, simple implementation, robustness to external disturbances, and insensitivity to model errors and parameter changes. In particular, (Journal: IEEE Transactions on Systems, Man, and Cybernetics: Systems; Author: H. Shen, D. Wang, J. H Park, et al.; Publication time: 2024; Article title: Switching-like event-triggered sliding mode load frequency control for networked power systems under energy-limited DoS attacks; Page number: 1589-1598) discussed the security issue of event-based power system load frequency control through a sliding mode control strategy; a fractional-order sliding mode control scheme for interconnected power systems with time-varying delay was proposed in (Journal: IEEE Transactions on Power Systems; Author: F. Yang, Y. Shen, D. Li, et al.; Publication time: 2024; Article title: Fractional-order sliding mode load frequency control and stability analysis for interconnected power systems with time-varying delay; Page number: 1006-1018). It is worth noting that the above control schemes are insensitive and fully robust to uncertainties or disturbances in the sliding mode, and may be sensitive to uncertainties or disturbances in the arrival mode. In order to ensure that the entire response of the system is insensitive or fully robust from the initial time, it is of research significance to use integral sliding mode control to study the load frequency control scheme of large-scale interconnected power systems.

[0004] Since the load changes of large-scale interconnected power systems are susceptible to external interference, the system often experiences frequency deviations beyond the allowable range, and even causes system instability or paralysis. In order to reduce or even eliminate the flutter effect caused by load disturbances in the power system, one of the common methods is to assume that the system load disturbance is a constant disturbance over a long period of time when designing an effective control scheme. For example, (Journal: IEEE Transactions on Smart Grid; Author: S.Hu, X.Ge, X.Chen, D.Yue; Publication Date: 2023; Article Title: Resilient load frequency control of islanded AC microgrids under concurrent false data injection and denial-of-service attacks; Pages: 690-700) Based on this assumption, a resilient and secure control scheme under false data injection attacks was studied. (Journal: IEEE Transactions on Automation Science and Engineering; Author: Y.Wang, Y.Liu, X.Yu, H.Zhang; Publication Date: 2024; Article Title: Jointed observer-based sliding mode predictive control for interconnected power systems with input delays; DOI: 10.1109 / TASE.2024.3353166) proposed an effective sliding mode predictive control scheme based on this assumption. Different from this, (Journal: IEEE Transactions on Industrial Electronics; Author: J.Yang, S.Li and X.Yu; Publication Date: 2013; Article Title: Sliding-mode control for systems with mismatched uncertainties via a disturbance observer; Pages: 160-169) introduced a disturbance observer in the sliding mode control to alleviate the disturbance effect. Its advantage is that the sliding mode control law is constructed by combining the disturbance estimation to directly compensate for the disturbance, which can better guarantee the control performance of the system. Therefore, it is necessary to design a disturbance observer to study the load frequency integral sliding mode control scheme of large-scale interconnected power systems, so as to improve the anti-interference ability and power quality of large-scale interconnected power systems.

[0005] Based on the above discussion, this paper focuses on the anti-disturbance control of large-scale interconnected power systems. The complex control structure of the system and the physical limitations of information exchange between subsystems bring difficulties to the controller design. First, a disturbance observer is designed to estimate the system load disturbance, and then the sliding mode control law is constructed by combining the disturbance estimation to directly compensate for the disturbance. Not only that, in order to ensure that the entire response of the system is insensitive or fully robust from the initial time, an integral sliding mode controller is designed. In addition, a memory-based integral sliding mode controller is designed, and the current and past state information of the system is utilized at the same time to improve the system performance. Finally, for the two sliding mode control laws, the control gain and disturbance suppression design schemes are proposed using the H-infinity control strategy. Summary of the invention

[0006] The purpose of the present invention is to solve the load disturbance problem of large-scale interconnected power systems and to propose a large-scale power system load frequency integral sliding mode control method based on a disturbance observer.

[0007] A large-scale power system load frequency integral sliding mode control method based on a disturbance observer comprises the following steps:

[0008] Step 1: Establish a dynamic model of a large-scale interconnected power system and analyze the load frequency control scheme of a large-scale interconnected power system containing any number of control areas;

[0009] Step 2: Design a disturbance observer to handle load disturbances in large-scale interconnected power systems;

[0010] Step 3: Design a memoryless integral sliding mode control law to improve the robustness of large-scale interconnected power systems;

[0011] Step 4: Design a memory-based integral sliding mode control law to improve the transient performance of large-scale interconnected power systems.

[0012] Furthermore, the specific process of establishing the large-scale interconnected power system dynamics model in step 1 is:

[0013] During the normal operation of a large-scale interconnected power system, a linearized model is used when the load changes slightly. The load frequency control problem is described by the following dynamic equation

[0014]

[0015] Where i = 1, 2, ..., N, N is the number of regions, the symbol Δ represents the degree of deviation from the steady state, Δf i (t) and Δf j (t) is the frequency change between region i and region j, ΔP mi (t) is the change in the speed regulator output command, ΔX gi(t) is the governor valve position of each zone, Δδ i (t) and Δδ j (t) is the change of rotor angle deviation between region i and region j, K Bi is the frequency deviation factor, K μ is the interconnection adjustment factor, β i is the frequency deviation factor, T s,ij is the power coefficient of the connecting line between area i and area j, T gi is the time constant of the speed regulator; T ti is the turbine time constant, T pi is the grid time constant; K pi is the grid gain, K Ei represents the adjustment parameter factor, R i Represents the droop coefficient of individual regions, ΔP di (t) is the incremental change of local load in each area, and ACE represents the linear sum of frequency deviation and tie line power deviation;

[0016] By defining the state variables, the system is represented in state space notation as follows

[0017]

[0018] Furthermore, the specific process of designing the disturbance observer in step 2 is as follows:

[0019] Consider a large-scale interconnected power system consisting of N interrelated subsystems; for the i-th subsystem, represents the state vector, Indicates the output, represents the local control input; in addition, j≠i represents the coupling term with other subsystems; is a constant matrix; suppose (A i ,B i ) is stable, (A i ,C i ) is considerable;

[0020] In order to ensure the stability and anti-interference of the system, the following definitions and assumptions are required:

[0021] Definition 1: If for all non-zero ΔP di (t)∈L2[0,∞) and zero initial condition, the following inequality holds and the system (2) satisfies H ∞ Performance index

[0022]

[0023] Assumption 1: Assume that the mismatch disturbance ΔP di (t) Satisfy the conditions Among them, α>0 and β>0 are two constants;

[0024] In order to deal with load disturbances, the following disturbance observer is designed

[0025]

[0026] in, is the perturbation estimate, Γ i is the Hurwitz matrix chosen by the designer, w i (t) is the internal variable vector of the observer; to simplify the expression, Under the assumption of , define the disturbance estimation error. According to assumption 1, prove It is bounded;

[0027] To advance the theoretical derivation, the following two lemmas are introduced;

[0028] Lemma 1: Assume A i is Hurwitzian, then there exists a scalar c>0 such that Established;

[0029] Lemma 2: For a given disturbance observer (3), the disturbance estimation error satisfy in is a scalar;

[0030] Prove; easy to obtain

[0031]

[0032] This means

[0033]

[0034] So, get

[0035]

[0036] Further

[0037]

[0038] Finally get

[0039]

[0040] By definition get satisfy According to the above derivation process, we can get the parameter matrix Γ in the disturbance observer (3): i must be Hurwitzian, and λ max(Γ i ) should be large enough.

[0041] Furthermore, the specific process of designing the memoryless integral sliding mode control law in step 3 is as follows:

[0042] Based on the designed disturbance observer, the following sliding surface is designed

[0043]

[0044] in And K i is the controller gain, K di is the disturbance suppression gain to be determined; unlike the traditional integral sliding surface, the disturbance observer (3) gives the disturbance estimate It is used to actively eliminate the unknown disturbance ΔP di (t) impact;

[0045] Design the following integral sliding mode controller

[0046]

[0047] in

[0048] The following theorem proves the reachability of the sliding surface;

[0049] Theorem 1: Under the action of controller (5), the state trajectory of the closed-loop system will be globally driven to the sliding surface in a finite time;

[0050] Proof: Design the following Lyapunov function

[0051]

[0052] Taking the derivative along time, we get

[0053] Therefore, we get

[0054]

[0055] Therefore, we get Therefore, the accessibility of the sliding surface can be guaranteed;

[0056] By solving θ i (t)=0, the equivalent control law can be obtained as follows: Will Substituting (2) into (2) we get

[0057]

[0058] Notice in The columns span The null space of Pick get

[0059]

[0060] definition The sliding mode dynamics equation can be expressed as

[0061]

[0062] The stability analysis of sliding mode dynamics (9) is as follows;

[0063] Theorem 2: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (9) with controller (5) is exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality

[0064]

[0065]

[0066] By solving inequality (11), we can obtain the control gain

[0067] Proof: Construct the following Lyapunov function

[0068]

[0069] If the above Lyapunov function satisfies

[0070]

[0071] get

[0072]

[0073] This means that the sliding mode dynamic equation (9) is exponentially stable;

[0074] Choose a free weight matrix L with appropriate dimensions i and Q i ,get

[0075]

[0076] Combining the above equations, we can calculate the derivative of the Lyapunov function (12):

[0077]

[0078] Next, we analyze the zero initial condition and non-zero ΔP di H under (t) condition ∞ performance

[0079]

[0080] In summary

[0081]

[0082] Assume J i =P i -1 ,X i =diag{J i ,J i ,J i ,J i}, Then use and X i Multiply inequality (15) on the left and right; therefore, inequality (11) is obtained by contract transformation;

[0083] Therefore, we get The results show that the sliding mode dynamics (9) can achieve exponential stability and effective control gains can be obtained by solving inequality (11).

[0084] Furthermore, the specific process of designing the memory-based integral sliding mode control law in step 4 is as follows:

[0085] Design the following memory-based integral sliding surface

[0086]

[0087] where τ i is the memory parameter, Obviously, the design of the above sliding surface contains information about the current state and the historical state;

[0088] Design the following memory-based integral sliding mode controller

[0089]

[0090] The following theorem proves the reachability of the sliding surface;

[0091] Theorem 3: Under the action of controller (16), the state trajectory of the closed-loop system will be globally driven to the sliding surface in a finite time;

[0092] Proof: Design the Lyapunov function as in (6), and we get

[0093]

[0094] So, get

[0095]

[0096] Therefore, we get Therefore, the accessibility of the sliding surface can be guaranteed;

[0097] Similarly, the sliding mode dynamics equation is obtained as follows:

[0098]

[0099] Pick and The sliding mode dynamics can be expressed as

[0100]

[0101] The stability analysis of sliding mode dynamics (9) is as follows;

[0102] Theorem 4: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (19) with controller (17) is exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality

[0103]

[0104]

[0105] By solving inequality (11), we can get the control gain Proof: Construct the following Lyapunov function

[0106]

[0107] Choose a free weight matrix E with appropriate dimensions i and F i ,get

[0108]

[0109] Combining the above equations, we can calculate the derivative of the Lyapunov function (22):

[0110]

[0111] Next, we analyze the zero initial condition and non-zero ΔP di H under (t) condition ∞ performance

[0112]

[0113] In summary

[0114]

[0115] set up X i =diag{M i ,M i ,M i ,M i ,M i},L i =ι i M i , choose Make inequality (20) hold, and then use and X i Left- and right-multiply inequality (26); therefore, inequality (21) is obtained by contract transformation.

[0116] Therefore, we can get The results show that the sliding mode dynamics (26) can achieve exponential stability, and the effective control gain can be obtained by solving the inequality (21)

[0117] The beneficial effects of the present invention are:

[0118] 1. Design a disturbance observer to estimate the system load disturbance for large-scale interconnected power systems with complex coupling structures. Then use the disturbance estimate to actively offset the disturbance in the integral sliding mode control design;

[0119] 2. In order to overcome the problem that the arrival mode in the sliding mode control of large-scale interconnected power systems is sensitive to load disturbances, an integral sliding mode controller based on coupling terms, integral sliding mode surface and disturbance estimation is designed. At the same time, the H-infinity control strategy is used to design effective control gain and disturbance suppression gain to improve the system robustness, fully considering the high coupling characteristics of the large-scale interconnected power system structure, and effectively improving the system robustness;

[0120] 3. Furthermore, in order to improve the transient performance of large-scale interconnected power systems, a memory-based integral sliding surface and integral sliding mode control strategy are proposed. At the same time, the current and past state information of the system is used to effectively improve the system performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0121] Figure 1 The present invention is a flow chart of the method.

[0122] Figure 2 is the control structure diagram of the i-th area.

[0123] Figure 3 The disturbance change and estimated value change of the load disturbance, where the upper figure is the estimated value of the disturbance and the lower figure is the disturbance value.

[0124] Figure 4 The exponential stability of the system state response under the integral sliding mode control and the integral sliding mode control strategy based on disturbance observer, where (a) is the first area of ​​the power system, (b) is the second area of ​​the power system, and (c) is the third area of ​​the power system.

[0125] Figure 5 To control the input changes under the integral sliding mode control and the integral sliding mode control strategy based on disturbance observer, the upper figure is the first area of ​​the power system, the middle figure is the second area of ​​the power system, and the lower figure is the third area of ​​the power system.

[0126] Figure 6 The system state response exponential stability under the integral sliding mode control and memory-based integral sliding mode control strategies with the disturbance observer function, where (a) is the first area of ​​the power system, (b) is the second area of ​​the power system, and (c) is the third area of ​​the power system. DETAILED DESCRIPTION

[0127] The specific implementation of the present invention is described below in conjunction with the above drawings for better understanding by the reader. It should be noted that in the following description, when the detailed description of known functions and designs may dilute the main content of the present invention, these descriptions will be omitted here.

[0128] like Figure 1 As shown, the present invention proposes a large-scale power system load frequency integral sliding mode control method based on a disturbance observer, which specifically includes the following steps:

[0129] (I) Establish a large-scale interconnected power system model; Consider a large-scale interconnected power system with load disturbances, such as Figure 2 During the normal operation of a large-scale interconnected power system, when the load changes slightly, a linearized model is used, and the load frequency control problem is described by the following dynamic equation, where i = 1, 2, ..., N, N is the number of regions, the symbol Δ represents the degree of deviation from the steady state, Δf i (t) and Δf j (t) is the frequency change between region i and region j, ΔP mi (t) is the change in the speed regulator output command, ΔX gi (t) is the governor valve position of each zone, Δδ i (t) and Δδ j (t) is the change of rotor angle deviation between region i and region j, K Bi is the frequency deviation factor, K μ is the interconnection adjustment factor, β i is the frequency deviation factor, T s,ij is the power coefficient of the connecting line between area i and area j, T gi is the time constant of the speed regulator; T ti is the turbine time constant, T pi is the grid time constant; K pi is the grid gain, K Ei represents the adjustment parameter factor, R i Represents the droop coefficient of individual regions, ΔP di (t) is the incremental change of local load in each area, and ACE represents the linear sum of frequency deviation and tie line power deviation;

[0130]

[0131] By defining the following state variables:

[0132] x i (t) = [Δf i (t)ΔP mi (t)ΔX gi (t)ΔE i (t)Δδi (t)].

[0133] A system with input delay can be represented in state space notation as follows

[0134]

[0135] (ii) Design a disturbance observer;

[0136] In order to deal with load disturbances, the following disturbance observer is designed

[0137]

[0138] in, is the perturbation estimate, Γ i is the Hurwitz matrix chosen by the designer, w i (t) is the internal variable vector of the observer. To simplify the expression, Under the assumption that According to hypothesis 1, it can be proved is bounded.

[0139] (iii) Design a memoryless integral sliding mode control law;

[0140] Based on the designed disturbance observer, the following sliding surface is designed

[0141]

[0142] in And K i is the controller gain, K di is the disturbance suppression gain to be determined. Different from the traditional integral sliding surface, the disturbance observer (3) gives the disturbance estimate It is used to actively eliminate the unknown disturbance ΔP di (t) impact.

[0143] Design the following integral sliding mode controller

[0144]

[0145] in

[0146] The following theorem proves the reachability of the sliding surface.

[0147] Theorem 1: Under the action of controller (5), the state trajectory of the closed-loop system will be globally driven to the sliding surface within a finite time.

[0148] By solving θ i (t)=0, the equivalent control law can be obtained as follows: Will Substituting (2) into (2) yields

[0149]

[0150] Notice in The columns span The null space of Pick get

[0151]

[0152] definition The sliding mode dynamics equation can be expressed as

[0153]

[0154] The stability analysis of sliding mode dynamics (9) is as follows.

[0155] Theorem 2: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (9) with controller (5) can be exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality

[0156]

[0157]

[0158] By solving inequality (11), we can get the control gain

[0159] In order to solve the controller gain matrix, taking Theorem 2 as an example, the following iterative algorithm is given.

[0160]

[0161]

[0162] (IV) Design of memory-based integral sliding mode control law

[0163] Design the following memory-based integral sliding surface

[0164]

[0165] where τ i is the memory parameter, Obviously, the design of the above sliding surface contains the information of current state and historical state.

[0166] Design the following memory-based integral sliding mode controller

[0167]

[0168] The following theorem proves the reachability of the sliding surface.

[0169] Theorem 3: Under the action of controller (16), the state trajectory of the closed-loop system will be globally driven to the sliding surface within a finite time.

[0170] Similarly, the sliding mode dynamics equation can be obtained as follows:

[0171]

[0172] Pick and The sliding mode dynamics can be expressed as

[0173]

[0174] The stability analysis of sliding mode dynamics (9) is as follows.

[0175] Theorem 4: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (19) with controller (17) can be exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality

[0176]

[0177]

[0178] By solving inequality (11), we can get the control gain

[0179] (V) Verification of the three-region interconnected power system

[0180] In this embodiment, the three-region interconnected power system is verified and set to the system parameters shown in Table 1 and the design parameters given during the simulation.

[0181] Table 1: System parameters and design parameters

[0182]

[0183]

[0184] Each subsystem has its own decentralized controller, and the control gains are chosen as

[0185] K1=[0.0579 0.7433-0.0153 0.0654 0.0638],

[0186] K2=[0.0590 0.7753-0.0127 0.0623 0.0612],

[0187] K3=[0.0562 0.7685-0.0176 0.0647 0.0656].

[0188] Assuming the load disturbance ΔP di (t) = 0.1sin(10t), then the observer parameters are selected as

[0189]

[0190] In order to verify the effectiveness of the designed disturbance observer, the changes in system response with and without the disturbance observer are compared. Select parameters σ = 0.60, exponential decay rate and x i (0) = [1,0,0,0,0] T .Disturbance suppression gain K di Designed for Based on Algorithm 1, H ∞ Performance Indicators Get the control gain K i As shown below.

[0191] K1=[0.0012-0.0128 0.1012-0.0012 0.0006],

[0192] K2=[0.0008-0.0219 0.2186-0.0058 0.0005],

[0193] K3=[0.0042-0.0322 0.3423-0.0081-0.0012].

[0194] According to the obtained control gain, Figure 3 The variation of actual and estimated load disturbances is shown. Figure 4 It shows that the system state response is exponentially stable and the disturbance observer is effective. In addition, the integral sliding mode control law and the change of the control input based on the disturbance observer law are as follows Figure 5 shown.

[0195] In order to verify whether the memory-based integral sliding mode control law improves the transient performance of the system, it is compared with the memoryless integral sliding mode control law. The parameters σ = 0.52 and the exponential decay rate and memory parameter τ i = 0.01s, Designed for

[0196]

[0197] Disturbance suppression gain K di Can be designed as According to Theorem 4, using H ∞ Performance Indicators The control gain K can be obtained as follows: i .

[0198] K1=[-0.0001 0.0002 0.0012-0.0007-0.0007],

[0199] K2=[0.0006 0.0003 0.0013 0.0001-0.0022],

[0200] K3=[0.0012-0.0005 0.0017-0.0097-0.0028].

[0201] According to the obtained control gain, Figure 6 It shows that the state response of the system based on memory and memoryless integral sliding mode control law is exponentially stable. Comparison shows that the memory-based integral sliding mode control law has better transient performance.

Claims

1. A large-scale power system load frequency integral sliding mode control method based on disturbance observer, characterized in that: The following steps are involved: Step 1: Establish a dynamic model of a large-scale interconnected power system and analyze the load frequency control scheme of a large-scale interconnected power system containing any number of control areas; Step 2: Design a disturbance observer to handle load disturbances in large-scale interconnected power systems; Step 3: Design a memoryless integral sliding mode control law to improve the robustness of large-scale interconnected power systems; Step 4: Design a memory-based integral sliding mode control law to improve the transient performance of large-scale interconnected power systems.

2. The method for load frequency integral sliding mode control of a large-scale power system based on a disturbance observer according to claim 1, characterized in that: The specific process of establishing the large-scale interconnected power system dynamics model in step 1 is as follows: During the normal operation of a large-scale interconnected power system, a linearized model is used when the load changes slightly. The load frequency control problem is described by the following dynamic equation Where i = 1, 2, ..., N, N is the number of regions, the symbol Δ represents the degree of deviation from the steady state, Δf i (t) and Δf j (t) is the frequency change between region i and region j, ΔP mi (t) is the change in the speed regulator output command, ΔX gi (t) is the governor valve position of each zone, Δδ i (t) and Δδ j (t) is the change of rotor angle deviation between region i and region j, K Bi is the frequency deviation factor, K μ is the interconnection adjustment factor, β i is the frequency deviation factor, T s,ij is the power coefficient of the connecting line between area i and area j, T gi is the time constant of the speed regulator; T ti is the turbine time constant, T pi is the grid time constant; K pi is the grid gain, K Ei represents the adjustment parameter factor, R i Represents the droop coefficient of individual regions, ΔP di (t) is the incremental change of local load in each area, and ACE represents the linear sum of frequency deviation and tie line power deviation; By defining the state variables, the system is represented in state space notation as follows in, 3. The method for load frequency integral sliding mode control of a large-scale power system based on a disturbance observer according to claim 2, characterized in that: The specific process of designing the disturbance observer in step 2 is as follows: Consider a large-scale interconnected power system consisting of N interrelated subsystems; for the i-th subsystem, represents the state vector, Indicates the output, represents the local control input; in addition, j≠i represents the coupling term with other subsystems; is a constant matrix; suppose (A i ,B i ) is stable, (A i ,C i ) is considerable; In order to ensure the stability and anti-interference of the system, the following definitions and assumptions are required: Definition 1: If for all non-zero ΔP di (t)∈L2[0,∞) and zero initial condition, the following inequality holds and the system (2) satisfies H ∞ Performance Indicators Assumption 1: Assume the mismatch disturbance ΔP di (t) satisfies the condition ||ΔP di (t)||≤α and Among them, α>0 and β>0 are two constants; In order to deal with load disturbances, the following disturbance observer is designed in, is the perturbation estimate, Γ i is the Hurwitz matrix chosen by the designer, w i (t) is the internal variable vector of the observer; to simplify the expression, Under the assumption of , define the disturbance estimation error. According to assumption 1, prove It is bounded; To advance the theoretical derivation, the following two lemmas are introduced; Lemma 1: Assume A i is Hurwitzian, then there exists a scalar c>0 such that Established; Lemma 2: For a given disturbance observer (3), the disturbance estimation error satisfy in is a scalar; Prove; easy to obtain This means So, get Further Finally get By definition get satisfy According to the above derivation process, we can get the parameter matrix Γ in the disturbance observer (3): i must be Hurwitzian, and λ max (Γ i ) should be large enough.

4. The method for load frequency integral sliding mode control of a large-scale power system based on a disturbance observer according to claim 3 is characterized in that: The specific process of designing the memoryless integral sliding mode control law in step 3 is as follows: Based on the designed disturbance observer, the following sliding surface is designed in And K i is the controller gain, K di is the disturbance suppression gain to be determined; unlike the traditional integral sliding surface, the disturbance observer (3) gives the disturbance estimate It is used to actively eliminate the unknown disturbance ΔP di (t) impact; Design the following integral sliding mode controller in The following theorem proves the reachability of the sliding surface; Theorem 1: Under the action of controller (5), the state trajectory of the closed-loop system will be globally driven to the sliding surface in a finite time; Proof: Design the following Lyapunov function Taking the derivative along time, we get Therefore, we get Therefore, we get Therefore, the accessibility of the sliding surface can be guaranteed; By solving θ i (t)=0, the equivalent control law can be obtained as follows: Will Substituting (2) into (2) we get Notice in The columns span The null space of Pick get definition The sliding mode dynamics equation can be expressed as The stability analysis of sliding mode dynamics (9) is as follows; Theorem 2: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (9) with controller (5) is exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality By solving inequality (11), we can obtain the control gain prove; Construct the following Lyapunov function If the above Lyapunov function satisfies get This means that the sliding mode dynamic equation (9) is exponentially stable; Choose a free weight matrix L with appropriate dimensions i and Q i ,get Combining the above equations, we can calculate the derivative of the Lyapunov function (12): Next, we analyze the zero initial condition and non-zero ΔP di H under (t) condition ∞ performance In summary set up X i =diag{J i ,J i ,J i ,J i }, Then use and X i Multiply inequality (15) on the left and right; therefore, inequality (11) is obtained by contract transformation; Therefore, we get The results show that the sliding mode dynamics (9) can achieve exponential stability and effective control gains can be obtained by solving inequality (11).

5. The method for load frequency integral sliding mode control of a large-scale power system based on a disturbance observer according to claim 4 is characterized in that: The specific process of designing the memory-based integral sliding mode control law in step 4 is as follows: Design the following memory-based integral sliding surface where τ i is the memory parameter, Obviously, the design of the above sliding surface contains information about the current state and the historical state; Design the following memory-based integral sliding mode controller The following theorem proves the reachability of the sliding surface; Theorem 3; Under the action of controller (16), the state trajectory of the closed-loop system will be globally driven to the sliding surface within a finite time; Proof: Design the Lyapunov function as in (6), and we get So, get Therefore, we get Therefore, the accessibility of the sliding surface can be guaranteed; Similarly, the sliding mode dynamics equation is obtained as follows: Pick and The sliding mode dynamics can be expressed as The stability analysis of sliding mode dynamics (9) is as follows; Theorem 4: For a given tuning parameter μ>∈>0, if there exists n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j >0,n i ×n i Symmetric Matrix P i >0,L i ,Q i ,n j ×n j Symmetric Matrix P j > 0, then the sliding mode dynamics (19) with controller (17) is exponentially stable when the decay rate is ρ = μ-∈, satisfying the following inequality By solving inequality (11), we can get the control gain prove; Construct the following Lyapunov function Choose a free weight matrix E with appropriate dimensions i and F i ,get Combining the above equations, we can calculate the derivative of the Lyapunov function (22): Next, we analyze the zero initial condition and non-zero ΔP di H under (t) condition ∞ performance In summary set up X i =diag{M i ,M i ,M i ,M i ,M i },L i =ι i M i , choose Make inequality (20) hold, and then use and X i Left- and right-multiply inequality (26); therefore, inequality (21) is obtained by contract transformation. Therefore, we can get The results show that the sliding mode dynamics (26) can achieve exponential stability, and the effective control gain can be obtained by solving the inequality (21)

Images