Failure performance compensation method for wind power plant grid-connected interconnection system actuator

Abstract

The invention belongs to the technical field of artificial intelligence, and discloses an actuator fault performance compensation method for a wind power plant grid-connected interconnection system, and the method comprises the steps: building a dynamic model of a nonlinear interconnection system containing a wind power plant; then, in combination with an actuator fault, constructing a state-space equation of a nonlinear system introducing a neural network; a composite observer with a dynamic event triggering mechanism is constructed, and estimation of the system state, external interference and actuator faults is achieved; according to the output of the composite observer, constructing a sliding mode safety compensation controller for realizing disturbance and fault compensation, and further obtaining an equivalent controller and a switching controller; and constructing a system master controller according to the equivalent controller and the switching controller. According to the invention, comprehensive estimation of the system state, external disturbance and actuator faults can be realized, active compensation of disturbance and faults can also be realized, and the system control stability is improved.

Description

Technical Field

[0001] The present invention belongs to the field of artificial intelligence technology and relates to a method for compensating for fault performance of an actuator in a wind farm grid-connected interconnection system. Background Art

[0002] As a highly complex networked dynamic system, modern power systems face numerous technical challenges in ensuring operational safety. From a system architecture perspective, these systems consist of multiple subsystems coupled through transmission networks and digital communication systems. These subsystems primarily include speed regulators, prime movers (such as steam turbines), generators, loads and interconnecting lines, and wind turbines, forming a typical wide-area cyber-physical system [see Shen Zengbo, Research on Secure Load Frequency Control of Interconnected Power Systems under Cyber ​​Attacks, Jiangnan University, 2024]. Within this distributed architecture, load frequency control plays a key role in maintaining system frequency stability and regional power balance. However, with the continuous increase in renewable energy penetration and the widespread use of power electronics, the system's dynamic characteristics are exhibiting significant nonlinearity, strong coupling, and time-varying characteristics.

[0003] At the operational level, power systems primarily face two typical abnormal operating conditions: first, random load disturbances, including natural fluctuations in user electricity demand and intermittent changes in distributed generation output; second, control channel failures, particularly hardware failures such as actuator jamming and failure. These abnormal conditions can directly disrupt the real-time balance between power generation and consumption, potentially causing frequency deviations beyond safe limits or, in severe cases, triggering cascading failures or even widespread power outages. Traditional control methods, including state feedback control, passive compensation strategies, and linear control methods, have significant limitations in addressing these issues: state feedback control relies on expensive sensor networks to achieve full state measurability, passive compensation strategies require precise boundary information for disturbances and faults, and conventional linear control methods struggle to cope with the inherent nonlinear characteristics of the system.

[0004] Of particular concern is the fact that three defining characteristics of modern power systems further complicate control design: first, system state variables (such as generator phase angle and speed) are often difficult to measure directly; second, load disturbances and actuator failures often occur simultaneously and are coupled; and third, bandwidth limitations of communication networks impose stringent requirements on the computational complexity of control algorithms. These challenges make it difficult for traditional model-based control methods to achieve the desired control performance in practical applications.

[0005] Sliding mode control (SMC) is a type of variable structure control and a nonlinear control method. It is primarily based on the dynamic adjustment of the state of the controlled object, so that the system state trajectory follows a designed sliding mode motion. Due to its inherent robustness advantages, it has attracted widespread attention and can be used in robotic control, power systems, mechanical systems, etc. However, traditional SMC methods have two key flaws: first, the design of its sliding surface strictly relies on real-time measurement of the system state; second, when dealing with complex disturbances, conservative boundary estimation methods are usually used, resulting in excessive control gains and prone to chattering problems. Although the introduction of observer technology in recent years has alleviated the problem of unmeasurable states, existing solutions still lack a systematic solution for dealing with simultaneous load disturbances and actuator failures. Summary of the Invention

[0006] In response to the above-mentioned problems existing in the prior art, the present invention provides a method for compensating for actuator fault performance in a wind farm grid-connected interconnected system, which can achieve comprehensive estimation of system status, external disturbances and actuator faults, and can also realize active compensation for disturbances and faults, thereby improving system control stability.

[0007] Based on the above analysis, the present invention provides a method for compensating for actuator fault performance in a wind farm grid-connected interconnection system, which comprises the following steps: S1: Establish a dynamic model of a nonlinear interconnected system containing a wind farm. Then, based on actuator failures, construct the state space equations of the nonlinear interconnected system with the introduction of a neural network. S2, constructing a composite observer with a dynamic event triggering mechanism to estimate the system state, external disturbances, and actuator faults; step S2 includes the following sub-steps: S21, build event triggering mechanism of sensor channel; S22, when the triggering condition of the event triggering mechanism is met, a composite observer is constructed based on the event triggering output of the sensor channel to estimate the system state; S23, constructing an error system based on the system state estimation of the composite observer; S24, construct the external disturbance of the system based on the error system; S25, constructs the system actuator failure factor based on the error system; S3, based on the output of the composite observer, a sliding mode safety compensation controller is constructed to achieve disturbance and fault compensation, and then an equivalent controller and a switching controller are obtained; S4, construct the system master controller based on the equivalent controller and switching controller.

[0008] In the above step S1, the nonlinear interconnected system including the wind farm is established to include more than two subsystems, each subsystem is provided with a corresponding wind turbine (the wind turbine is used to realize wind power generation), and the first subsystem is constructed. i The subsystem control dynamic equations and i The fan dynamic equation of the subsystem; i The control dynamic equation of each subsystem is shown in Equation (1a). i The dynamic equation of the fan of each subsystem is shown in formula (1b): (1a); in, 、 、 、 and For the i The state variables of each subsystem represent the frequency change, active power output change, governor valve opening change, integral control change and rotor angle deviation change respectively. represents the power system constant, represents the gain of the power system, represents the time constant of the steam turbine, Indicates the speed adjustment coefficient in the speed regulator, represents the speed regulator constant, and Represent the control gain and frequency offset coefficient respectively, represents the load disturbance, Indicates the i subsystems and j The interconnection system matrix between the subsystems, Indicates the system input voltage; (1b); in, and Respectively represent i The speed change and rotor current of the fan in each subsystem, , , and Represent the resistance of the fan stator and rotor respectively, and Represent the leakage inductive reactance of the fan stator and rotor respectively, represents the mutual inductance of the wind turbine rotor, represents the slip angular frequency, represents the inertia coefficient, , represents the friction coefficient, , represents the pitch angle of the wind turbine, represents the air density, represents the tip speed ratio, , Indicates the radius of the fan blade.

[0009] Will As a description of i For the convenience of description, the key state variables of each subsystem are defined as follows: , define the system output variable as , then (1a) and (1b) above can be formed into the following state space expression, where, represents the state quantity of the state space equation, and its components are defined as above: (2); in, , , , , , , N represents the number of subsystems, T Indicates transpose.

[0010] During the long-term operation of the system, the control channel is prone to failure. In the present invention, the control channel failure is considered to be an actuator failure (i.e., a control input failure), and its mathematical form is considered as follows: (3); in, represents the control input signal under fault conditions, represents the actual control input, Indicates a biased actuator fault, represents the actuator failure factor, 、 Represent the minimum and maximum thresholds respectively. For the power system, the actuator generally does not experience drastic changes during operation due to its working characteristics, so it is assumed that ,in is a positive scalar. For power systems, bias faults may be caused by changes in system parameters, so we further assume that In addition, it should be noted that not all actuators will fail at the same time, so a corresponding compensation strategy can be designed for some faulty actuators.

[0011] Using neural network (NNs) technology to Design. According to neural networks (NNs), any smooth nonlinear function It can be approximated by the following form: (4); in, represents the weight vector, Express Estimates, represents the radial basis function vector, Indicates the approach error and there is a relationship , here, express The upper bound of .

[0012] By introducing neural networks (NNs), combined with actuator faults, i The state space equation (2) of each subsystem can be reformulated as: (5); in, represents an external disturbance.

[0013] In the above step S2, the purpose is to design a composite observer with a dynamic event triggering mechanism to realize the state , external disturbance , weight vector and actuator failure factor Composite estimate of .

[0014] In the above step S21, i The event trigger mechanism of each subsystem sensor channel is designed as follows: (6); in, The sensor data transmission trigger time sequence, is the weight matrix of the trigger mechanism, h is the sampling period, , m represents an integer greater than zero, and Respectively expressed in and Output at any moment. is the adaptive gain, and its expression is: (7); and are two thresholds, satisfying , , whose value is chosen by the designer; and are two positive parameters, satisfying , whose value is chosen by the designer.

[0015] In the above step S22, when the trigger moment When the trigger condition is met, iSubsystem moments Output As event-triggered output of the sensor channel sent to the composite observer; Then, the composite observer is designed as follows: (8); in, 、 、 、 、 Respectively 、 、 、 and The estimated value of represents the estimated error compensation vector of the observer, represents the observer gain matrix, Event trigger output for the sensor channel. The time series of is determined by (6).

[0016] In the above step S23, define ,in, Representing the state observation error, the following error system can be obtained: (9); in, , , .

[0017] In the above step S24, it is noted that the error system (9) contains the overall disturbance , which shows that the disturbance information can be estimated by the state error of the system. Since the disturbance of the power system has boundary constraints in practical applications, it is assumed here that the overall disturbance has norm boundedness, that is, ,in is the unknown gain coefficient, is an auxiliary vector defined as ,in represents an intermediate variable whose update rate is as follows: (10); in, It represents a scalar greater than zero, and can be taken as a positive number. Indicates unknown gain The estimated value of The update law is , represents a constant, is a positive definite symmetric matrix, and Respectively and The Euler norm of .

[0018] according to The definition of (10) can be further calculated to obtain: (11); To estimate the external disturbance in the system, the following external disturbance estimation is designed: (12); in, express estimated value.

[0019] In the above step S25, the update rate of the actuator failure factor is as follows: (13); in , Represents a positive number greater than zero.

[0020] For the composite observer, the estimation error compensation vector is designed as: (14); in, , Represent the state estimation errors of the 1st, 2nd to Nth subsystems respectively, is a symmetric matrix, It is a vector that needs to be selected, and its value is determined by the designer.

[0021] The purpose of the above step S3 is to design a sliding mode safety compensation controller based on the output of the composite observer to enhance the system's robust compensation capability against disturbances and faults.

[0022] According to the output of the composite observer, a sliding mode safety compensation controller is constructed, and its second-order sliding mode function is expressed as follows: (15); in, Designed for , is a positive constant, is a vector to be selected, and the basis for selection is to satisfy The constant to be designed; Indicates the current i Control parameters of each subsystem.

[0023] Taking the derivative of the sliding mode function, we can get: (16); in, , .

[0024] Therefore, when the system state reaches the sliding surface, the equivalent controller Designed to: (17); Substituting the equivalent controller into the original system equation (5) yields: (18); in, , .

[0025] Furthermore, if all subsystems are considered globally (i.e., all systems i=1, 2,…,N are considered together), the entire nonlinear interconnected system can be expressed as: (19); in, , , , , , , , .

[0026] Control parameters It can be obtained in the following ways: , .

[0027] use performance indicators to suppress the influence of estimation errors and lumped disturbances, , Indicates the preset interference attenuation constant, For a given symmetric positive definite matrix and a positive constant , if the following conditions are met, the entire power system can be guaranteed to be stable and meet Performance indicators: (20); Among them, the entire matrix is ​​symmetrical, is a positive definite matrix, yes No. i elements and is itself a positive definite symmetric matrix, Indicates is a diagonal matrix with diagonal elements; , yes No. ielements and is itself in matrix form, whose values ​​are chosen by the designer; * indicates a symmetric matrix at the corresponding position. , , .

[0028] In order to satisfy the formula (5) i The trajectory of each subsystem converges to the designed sliding surface, and the switching controller is designed as follows : (twenty one); in, Indicated by Determine the sign function, , represents a constant, , yes The estimated value of represents the unknown gain constant, The update law is .

[0029] In the above step S4, the constructed i The overall controller of each subsystem is: (twenty two).

[0030] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention designs a composite observer with a dynamic event triggering mechanism based on a neural network, which can realize the composite estimation of the interconnected power grid system state, external disturbances and actuator failures; and reduces the computational burden through the dynamic event triggering mechanism; (2) The present invention uses the output of the composite observer to directly construct a sliding mode safety compensation controller without relying on state measurability or prior information of disturbances / obstacles, thus reducing conservatism; (3) This invention constructs an observer-based second-order sliding surface in a nonlinear interconnected power system for the first time, and combines it with precise estimation results to achieve active compensation for disturbances and faults, thereby enhancing the robust compensation capability of the interconnected power grid system for disturbances and faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] The following is a detailed explanation of the drawings in the specification.

[0032] Figure 1 Schematic diagram of the principle of the method for compensating for actuator fault performance in a wind farm grid-connected interconnected system provided by the present invention; wherein S represents the integral form of the transfer function; Figure 2 A schematic flow chart of a method for compensating for actuator fault performance in a wind farm grid-connected interconnection system provided by the present invention; Figure 3 Schematic diagram of the process flow for building a composite observer with dynamic event triggering mechanism to estimate system state, external disturbance and actuator fault; Figure 4 The state dynamic response of subsystem 1 in the embodiment of the present invention; Figure 5 The state dynamic response of subsystem 2 in the embodiment of the present invention; Figure 6 This is the state estimation effect of subsystem 1 in the embodiment of the present invention; Figure 7 This is the state estimation effect of subsystem 2 in the embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following will clearly and completely describe the technical solutions of various embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention. Example

[0034] This embodiment is simulated and verified on a dual-area nonlinear interconnected system of a typical power system to evaluate the control performance and stability of the actuator fault performance compensation method for the wind farm grid-connected interconnected system under fault and disturbance conditions.

[0035] This embodiment provides a method for compensating for actuator failure performance of a wind farm grid-connected interconnection system, such as Figure 1-Figure 3 As shown, it includes the following steps: S1, establish a dynamic model of the nonlinear interconnected system including the wind farm; then, combined with actuator faults, construct the state space equation of the nonlinear system with the introduction of neural networks.

[0036] The nonlinear interconnected system with wind farms constructed in this step i The dynamic model of the subsystem is shown in formula (1). Then the control input signal after considering the actuator failure is established, as shown in formula (3). Then the neural network is introduced to construct the nonlinear function shown in formula (4). Substituting formula (3) and formula (4) into formula (2) yields the nonlinear system. i The state space equation of each subsystem is shown in Equation (5).

[0037] S2, constructing a composite observer with a dynamic event triggering mechanism to estimate the system state, external disturbances, and actuator faults; step S2 includes the following sub-steps: S21, build the event triggering mechanism of the sensor channel.

[0038] Construct the event trigger mechanism of the sensor channel as shown in formula (6).

[0039] S22, when the triggering condition of the event triggering mechanism is met, a composite observer is constructed according to the event triggering output of the sensor channel to estimate the system state.

[0040] When the trigger When the trigger condition is met, i Subsystem moments Output The event-triggered output of the sensor channel is sent to the composite observer; the equation of the composite observer is shown in Equation (8).

[0041] S23, constructing an error system based on the system state estimation of the composite observer.

[0042] According to the system state obtained by the composite observer, the error system is obtained through equation (9).

[0043] S24, construct the system external disturbance based on the error system.

[0044] According to the formula (12) given above, the external disturbance can be estimated.

[0045] S25, construct the system actuator failure factor based on the error system.

[0046] According to the formula (13) given above, the actuator fault factor can be updated.

[0047] S3, based on the output of the composite observer, constructs a sliding mode safety compensation controller to achieve disturbance and fault compensation, and then obtains the equivalent controller and switching controller.

[0048] In this step, based on the output of the composite observer, the sliding mode safety compensation controller is used to determine the equivalent controller through Equation (17), and the switching controller is determined through Equation (21).

[0049] S4, construct the system master controller based on the equivalent controller and switching controller.

[0050] According to formula (22), the equivalent controller and the switching controller are added together to obtain i The overall controller of each subsystem.

[0051] Set the parameters of the power system as: , , , , , , , , , , , , ; In subsystem 1 , in subsystem 2 , The pitch angle is set to , the tip speed ratios are and , the radius of the wind wheel is , , and the remaining parameters are , , , , and , , The load disturbance is , . The actuator fault is set to , The control parameters are: , , , , . , , , .

[0052] Based on the above parameters, the dynamic response results of each state parameter of the two subsystems obtained by simulation through the actuator fault performance compensation method for wind farm grid interconnection system provided by this embodiment are as follows: Figure 4-Figure 5 As shown in , the state estimation effect of each state (reflected by observation error) is as follows Figure 6-Figure 7 As shown. Figure 4-Figure 7 It can be clearly seen that in the presence of actuator failures and external disturbances, the method of the present invention can still maintain a stable system state and has excellent convergence performance.

[0053] Those skilled in the art will appreciate that the embodiments herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art may make various other specific variations and combinations based on the technical teachings disclosed herein without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A method for compensating for actuator fault performance in a wind farm grid-connected interconnection system, characterized in that: The following steps are involved: S1: Establish a dynamic model of a nonlinear interconnected system containing a wind farm. Then, based on actuator failures, construct the state space equations of the nonlinear interconnected system with the introduction of a neural network. S2, constructing a composite observer with a dynamic event triggering mechanism to estimate the system state, external disturbances, and actuator faults; step S2 includes the following sub-steps: S21, build event triggering mechanism of sensor channel; S22, when the triggering condition of the event triggering mechanism is met, a composite observer is constructed based on the event triggering output of the sensor channel to estimate the system state; S23, constructing an error system based on the system state estimation of the composite observer; S24, construct the external disturbance of the system based on the error system; S25, constructs the system actuator failure factor based on the error system; S3, based on the output of the composite observer, a sliding mode safety compensation controller is constructed to achieve disturbance and fault compensation, and then an equivalent controller and a switching controller are obtained; S4, construct the system master controller based on the equivalent controller and switching controller.

2. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 1, characterized in that: In step S1, the nonlinear interconnected system including the wind farm is established to include more than two subsystems, each subsystem is provided with a corresponding wind turbine, and the first i The subsystem control dynamic equations and i The fan dynamic equations of each subsystem; Will As a description of i The state variables of each subsystem are defined as follows: , 、 、 、 and They represent frequency change, active power output change, governor valve opening change, integral control change, and rotor angle deviation change, respectively. and Respectively represent i The speed change and rotor current of the fan of each subsystem; the system output variable is defined as ; By introducing neural networks and combining actuator faults, i The state space equation of each subsystem is expressed as: ； in, , , , , , N represents the number of subsystems, T represents transpose; represents the actuator failure factor, 、 Represent the minimum and maximum thresholds respectively. represents the power system constant, represents the gain of the power system, represents the time constant of the steam turbine, Indicates the speed adjustment coefficient in the speed regulator, represents the speed regulator constant, and Represent the control gain and frequency offset coefficient respectively, represents the load disturbance, Indicates the i subsystems and j The interconnection system matrix between the subsystems, represents the actual control input, , , and Represent the resistance of the fan stator and rotor respectively, and Represent the leakage inductive reactance of the fan stator and rotor respectively, represents the mutual inductance of the wind turbine rotor, represents the slip angular frequency, represents the inertia coefficient, , represents the friction coefficient, represents the weight vector, Express Estimates, represents the radial basis function vector; represents external disturbance; in, Indicates the approach error and there is a relationship , express The upper bound of Indicates a biased actuator fault.

3. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 2, characterized in that: In step S21, i The event trigger mechanism of each subsystem sensor channel is designed as follows: ； in, The sensor data transmission trigger time sequence, is the weight matrix of the trigger mechanism, h is the sampling period, , m represents an integer greater than zero, and Respectively expressed in and Output at each moment; is the adaptive gain, and its expression is: ； and are two thresholds, satisfying , ; and are two positive parameters, satisfying .

4. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 3, characterized in that: In step S22, when the trigger moment When the trigger condition is met, i Subsystem moments Output As event-triggered output of the sensor channel sent to the composite observer; ； in, 、 、 、 、 Respectively 、 、 、 and The estimated value of represents the estimated error compensation vector of the observer, represents the observer gain matrix, Event trigger output for the sensor channel.

5. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 4, characterized in that: In step S23, define ,in, Representing the state observation error, the following error system can be obtained: ； in, , , .

6. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 5, characterized in that: In step S24, the external disturbance is estimated as: ； in, express The estimated value of represents a scalar greater than zero, is the unknown gain coefficient, Indicates unknown gain The estimated value of The update law is ;in, represents a constant, is a positive definite symmetric matrix, and Respectively and The Euler norm of ; is an auxiliary vector defined as ,in represents an intermediate variable whose update rate is as follows: 。 7. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 6, characterized in that: In step S25, the update rate of the actuator failure factor is as follows: ； in, ; Represents a positive number greater than zero; For the composite observer, the estimation error compensation vector is designed as: ； in, , Represent the state estimation errors of the 1st, 2nd to Nth subsystems respectively, is a symmetric matrix, is the vector to be determined; Represents a positive scalar.

8. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 7, characterized in that: According to the output of the composite observer, a sliding mode safety compensation controller is constructed. The second-order sliding mode function of the sliding mode safety compensation controller is expressed as follows: ； in, Designed for , is a positive constant, is a vector to be selected, and the basis for selection is to satisfy The constant to be designed; Indicates the i Control parameters of each subsystem; When the system state reaches the sliding surface, the equivalent controller Designed to: ； in, ; In order to make the i The trajectory of each subsystem converges to the designed sliding surface, and the switching controller is designed as follows : ； in, Indicated by Determine the sign function, , represents a constant, , yes The estimated value of represents the unknown gain constant, The update law is .

9. The method for compensating for actuator fault performance in a wind farm grid-connected interconnection system according to claim 8, characterized in that: In step S4, the constructed i The overall controller of each subsystem is: 。

Images