CN113949064B - Load frequency fault-tolerant control system and method based on reverse row reconstruction - Google Patents
Load frequency fault-tolerant control system and method based on reverse row reconstruction Download PDFInfo
- Publication number
- CN113949064B CN113949064B CN202111300633.0A CN202111300633A CN113949064B CN 113949064 B CN113949064 B CN 113949064B CN 202111300633 A CN202111300633 A CN 202111300633A CN 113949064 B CN113949064 B CN 113949064B
- Authority
- CN
- China
- Prior art keywords
- controller
- controlled system
- fault
- control
- mse
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 36
- 238000004364 calculation method Methods 0.000 claims abstract description 16
- 238000013461 design Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 36
- 230000006399 behavior Effects 0.000 claims description 22
- 238000005259 measurement Methods 0.000 claims description 7
- 230000000694 effects Effects 0.000 claims description 6
- 238000005516 engineering process Methods 0.000 claims description 5
- 230000002159 abnormal effect Effects 0.000 claims description 4
- 230000003416 augmentation Effects 0.000 claims description 4
- 230000001550 time effect Effects 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 238000001514 detection method Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 238000011426 transformation method Methods 0.000 claims description 2
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 101100481404 Danio rerio tie1 gene Proteins 0.000 description 3
- 101100481406 Mus musculus Tie1 gene Proteins 0.000 description 3
- 238000004088 simulation Methods 0.000 description 3
- 230000001360 synchronised effect Effects 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000011217 control strategy Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/001—Methods to deal with contingencies, e.g. abnormalities, faults or failures
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Feedback Control In General (AREA)
- Safety Devices In Control Systems (AREA)
Abstract
Load frequency fault-tolerant control system based on reverse row reconstruction, information acquisition module: and acquiring and sensing the operation parameters of the controlled system by using a sensor arranged on the controlled system, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters. The controller library module designs corresponding controllers aiming at the expected information/physical faults to ensure the dynamic performance of a fault system, and forms a controller library, wherein the number of controllers in the basic controller library is M+N; the reverse-direction-motion reconstruction module reversely reconstructs the dynamic behavior of the controlled system in a parallel mode according to the output quantity y '(t), the control instruction u' (t) and the controller library of the controlled system, and obtains the reconstructed reference inputThe performance index calculation module inputs the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE; the controller selection module selects the controller investment corresponding to the smallest performance indicator MSE. The application also provides a load frequency fault-tolerant control method based on reverse-direction reconstruction.
Description
Technical field:
the application relates to the technical field of active fault-tolerant control of load frequency of an interconnected power grid, in particular to a load frequency fault-tolerant control system and method based on reverse-direction reconstruction.
The background technology is as follows:
the new energy installation capacity of continuous grid connection in the novel power system is obvious, and the randomness and uncertainty of the new energy installation capacity increase various dominant and recessive faults in the system, so that the active output of the power transmission network is obviously reduced, and the frequency is obviously deviated from a rated operating point. In addition, as the informatization degree of the power grid is continuously improved, various uncertainty factors (such as delay, packet loss and hacking) in the network transmission process bring new potential safety hazards to various real-time control systems. The load frequency control (Load frequency control, LFC) exchanges power through the control area link to achieve stable control between the system frequency input bias to the controller output control quantity.
The comprehensive sensing and real-time stable control of the running state are basic requirements of various real-time control systems in the electric power Internet of things. However, the tight coupling of the information subsystem and the physical subsystem, resulting in malicious network attacks and sudden physical failures, will cause structural changes to the power internet of things, and such uncertainty has violated the assumption of small-range fluctuation of parameters in the robust control theory. The traditional fault-tolerant control strategy based on hardware redundancy or fault identification has the risk of mismatching between a controller and a controlled system due to the fact that the controlled system parameters and a large number of operation state parameters need to be relied on accurately and priori, and fault-tolerant failure is caused.
The application comprises the following steps:
in view of the foregoing, it is desirable to provide a load frequency fault tolerant control system based on reverse-direction reconstruction that does not rely on fault identification.
There is also a need to provide a method of load frequency fault tolerant control based on reverse-direction-based reconstruction that does not rely on fault identification.
The load frequency fault-tolerant control system based on reverse row reconstruction comprises an information acquisition module, a controller library module, a reverse row reconstruction module, a performance index calculation module and a controller selection module:
and the information acquisition module is used for: and acquiring and sensing the operation parameters of the controlled system by using a sensor arranged on the controlled system, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters.
And a controller library module: aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and the number of the controllers in the basic controller library is Sigma c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) Controller pool Σ c Representing a specific shipmentM refers to a controller set designed by a row scene, wherein the controlled system has M normal working scenes, and N refers to the controlled system has N typical fault scenes;
and (3) a reverse reconstruction module: based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converges to an ideal reference input r (t).
The performance index calculation module is used for: input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE;
the controller selection module: and selecting the controller input corresponding to the minimum performance index MSE, thereby realizing zero trial-and-error one-time accurate switching.
A load frequency fault tolerance control method based on reverse direction reconstruction comprises the following steps:
aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and the sigma is given by c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The method comprises the steps that a controller set designed for a specific operation scene is represented, M means that a controlled system has M normal operation scenes, and N means that the controlled system has N typical fault scenes;
receiving operation parameters of a controlled system acquired and sensed by a sensor, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters;
receiving a control instruction u' (t) generated by a control center according to the operation parameters of the sensor acquisition sensing controlled system;
based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converging to an ideal reference input r (t);
input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE;
and selecting the controller input corresponding to the minimum performance index MSE, thereby realizing zero trial-and-error one-time accurate switching.
In the load frequency fault tolerance control system and method based on reverse-direction reconstruction, a reverse-direction reconstruction strategy is provided, a corresponding controller is designed aiming at the expected information/physical faults to ensure the dynamic performance of the fault system, the prior fault information is not required to be acquired, and the fault tolerance control system only depends on the input and output measured values of the system, so that fault tolerance can be realized as much as possible, and better performance is provided for the stability control and the stable operation of the system of the information/physical tightly-coupled power Internet of things under the complex and diverse fault conditions.
Description of the drawings:
FIG. 1 is a functional block diagram of a load frequency fault tolerant control system based on reverse-direction-based reconstruction;
FIG. 2 is a schematic diagram of a load frequency fault tolerant control system architecture based on reverse-direction behavior reconstruction;
FIG. 3 is a schematic diagram of a controller library framework;
FIG. 4 is a block diagram of a three-area interconnected power system;
FIG. 5 is a normal communication latency for Area 1-Area 3;
FIG. 6 is a diagram illustrating the communication latency of Area 2 when it is subject to a latency attack;
FIG. 7 is an Area 1-Area 3 load surge;
FIG. 8 is Δf for different fault conditions 1 ΔP tie1 Is a mean square error of (c).
In the figure: the load frequency fault tolerance control system based on reverse row reconstruction 10, the information acquisition module 20, the controller library module 30, the reverse row reconstruction module 40, the performance index calculation module 50 and the controller selection module 60.
The specific embodiment is as follows:
the application aims at providing a load frequency fault-tolerant control system which does not depend on fault identification aiming at information/physical faults of an electric power Internet of things, and the core idea is to put a most matched controller into a controller loop in a zero trial-and-error mode. First, a corresponding controller is designed for the expected information/physical faults to ensure the dynamic performance of the fault system, and a basic controller library is formed. And then, designing a controller switching strategy, and putting the controller into an optimal controller according to a performance evaluation result to ensure safe and stable operation of the system. The following describes in further detail the embodiments of the present application with reference to the drawings and examples. The following examples are illustrative of the application and are not intended to limit the scope of the application.
Referring to fig. 1 to 3, a load frequency fault tolerance control system 10 based on reverse row reconstruction includes an information acquisition module 20, a controller library module 30, a reverse row reconstruction module 40, a performance index calculation module 50, and a controller selection module 60:
information acquisition module 20: and acquiring and sensing the operation parameters of the controlled system by using a sensor arranged on the controlled system, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters.
Controller library module 30: aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and the number of the controllers in the basic controller library is Sigma c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) Controller libraryΣ c Representing a controller set designed for a specific operation scene, wherein M refers to a controlled system having M normal operation scenes, and N refers to a controlled system having N typical fault scenes;
the reverse row reconstruction module 40: based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converges to an ideal reference input r (t). The reconstruction process can always reconstruct the dynamic behaviors of all preset controllers in the controller library only through the control parameters and the control instructions of the control center in any detection time period without sequentially switching the preset controllers into the control loop, thereby laying a foundation for the performance evaluation of the controllers.
Performance index calculation module 50: input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE; wherein the mean square error MSE is defined as the reconstructed reference input +.>A measure of the degree of difference between two variables of the controlled system output y' (t), i.e
The MSE index can be used for accurately expressing the control deviation of different controllers under the condition of uncertain information. Wherein, as the control target of the real-time control system is the output y' (t) of the controlled system, the ideal reference input r (t) is tracked in real time, such asFor the actual controlled system, the control target of the state quantities such as the rotation speed, the frequency f and the like of the generator is that the regional control error ACE is 0, namely r (t) -y' (t) =0, so that the reconstruction reference input is adoptedThe MSE index of the deviation between the output y' (t) of the controlled system is used as a measure of the dynamic performance of the controller.
Controller selection module 60: and selecting the controller input corresponding to the minimum performance index MSE, thereby realizing zero trial-and-error one-time accurate switching. For example, according to the calculated performance index { MSE } 1 ,MSE 2 ,…,MSE M+N Selecting the most matched controller to be put into the control loop in a zero trial and error mode, namely selecting the preset controller input with the minimum performance index MSE, namely
Wherein, the controller set in the basic controller library is obtained by the following way:
1) Establishing a real-time control system model of information/physical faults:
giving a controlled object model in a controlled system under normal conditions:
wherein x is R p 、u∈R q 、ω∈R r And y.epsilon.R v The state variable, the input variable, the external disturbance variable, and the output variable, respectively, and the matrix A, B, H, C, D is a dimension-adaptive matrix.
Modeling according to the expected fault action effect in the controlled system:
(1) Fault type modeling
A. Modeling data faults, wherein under the faults, the actual control quantity/quantity measurement and reality meet the following relations:
control amount: u (u) f (t)＝Γ P Γ C u(t)
Measuring the amount: y is f (t)＝Ψ C Ψ P y(t)
Where the subscript f indicates a fault condition,representing an actuator physical fault matrix; />Representing a network fault matrix between the controller and the actuator;representing a network information fault matrix between the sensor and the controller;representing a sensor physical fault matrix;
B. modeling time-effect type faults, and letting tau sc Representing the network delay of the sensor to the controller, τ ca In order to delay the network from the controller to the executor in the controlled system, the control instruction sent by the controller is u '(t), the received output quantity of the controlled system is y' (t), and the control instruction u (t) provided by the control center actually received by the executor in the controlled system and the actual output quantity y (t) of the controlled system satisfy the following conditions
u(t)＝u'(t+τ ca )
y'(t)＝y(t+τ sc )
Further, the pure time-lag links are approximated by adopting Pade approximation technology, and the approximated state space is expressed as
Wherein x is ca (t)、x sc (t) intermediate variable introduced for Pade approximation, A k ,B k ,C k ,D k Respectively is
Where k= { sc, ca }, l k To approximate the order, a i And b i (i＝1,2,…,l k ) To approximate coefficients, the following equation is given:
C. component fault modeling, under the fault, the original running structure of the controlled system is obviously changed, which is expressed as abnormal state space description, namely
In the formula, A ', B ', H ', C ', D ' represent the matrix of the controlled system after the fault occurs.
(2) Information/physical hybrid fault modeling
Based on mathematical modeling of three types of faults, the augmentation vector x= [ X, X ca ,x sc ] T The method comprises the steps of establishing a state equation of an electric power Internet of things real-time control system considering information/physical hybrid faults as follows
Wherein A is f ＝A'，B f ＝B'Γ P Γ C ，H f ＝H'，C f ＝Ψ C Ψ P C'，D f ＝Ψ C Ψ P D'Γ P Γ C ；
2) The basic controller library normative design method comprises the following steps:
the real-time control system state equation has model uncertainty, so the controller design must have excellent robustness to the model uncertainty, and the application refers to the mixed H 2 /H ∞ The method is used for designing a basic controller to realize robust control on known faults;
considering that the controlled system has M normal working scenes and N typical fault scenes, the number of controllers in the basic controller library is M+N, so that Σ c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The } represents a set of controllers designed for a particular operating scenario, having the following general form:
wherein j=1, 2, …, m+n, x uj (t) intermediate variables introduced for the controller, A uj ,B uj ,C uj ,D uj Is a matrix of controller systems.
Furthermore, the application also provides a load frequency fault-tolerant control method based on reverse-direction reconstruction, which comprises the following steps:
step S300, aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of controllers in the basic controller library is M+N, and the sigma is given by c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The method comprises the steps that a controller set designed for a specific operation scene is represented, M means that a controlled system has M normal operation scenes, and N means that the controlled system has N typical fault scenes;
step S303, receiving the operation parameters of the controlled system acquired and sensed by the sensor, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters;
step S305, a control command u' (t) generated by a control center according to the operation parameters of the sensor acquisition sensing controlled system is received;
step S307, based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converging to an ideal reference input r (t);
step S309, performing root mean square error calculation on the output quantity y' (t) of the controlled system and the reconstructed reference input to obtain a performance index MSE;
in step S311, the controller input corresponding to the minimum performance index MSE is selected, so as to realize zero trial-and-error one-time accurate switching.
Wherein, the corresponding controllers are designed to ensure the dynamic performance of the fault system aiming at the predicted information/physical faults to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and Sigma is formed c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The method comprises the steps of designing a controller set for a specific operation scene, wherein M refers to that a controlled system has M normal operation scenes, N refers to that the controlled system has N typical fault scenes, and the steps are as follows:
1) Establishing a real-time control system model of information/physical faults:
giving a controlled object model in a controlled system under normal conditions:
wherein x is R p 、u∈R q 、ω∈R r And y.epsilon.R v The state variable, the input variable, the external disturbance variable and the output variable are respectively, and the matrix A, B, H, C, D is a proper dimensionA matrix.
Modeling according to the expected fault action effect in the controlled system:
(1) Fault type modeling
A. Modeling data faults, wherein under the faults, the actual control quantity/quantity measurement and reality meet the following relations:
control amount: u (u) f (t)＝Γ P Γ C u(t)
Measuring the amount: y is f (t)＝Ψ C Ψ P y(t)
Where the subscript f indicates a fault condition,representing an actuator physical fault matrix; />Representing a network fault matrix between the controller and the actuator;representing a network information fault matrix between the sensor and the controller;representing a sensor physical fault matrix;
B. modeling time-effect type faults, and letting tau sc Representing the network delay of the sensor to the controller, τ ca In order to delay the network from the controller to the executor in the controlled system, the control instruction sent by the controller is u '(t), the received output quantity of the controlled system is y' (t), and the control instruction u (t) provided by the control center actually received by the executor in the controlled system and the actual output quantity y (t) of the controlled system satisfy the following conditions
u(t)＝u'(t+τ ca )
y'(t)＝y(t+τ sc )
Further, the pure time-lag links are approximated by adopting Pade approximation technology, and the approximated state space is expressed as
Wherein x is ca (t)、x sc (t) intermediate variable introduced for Pade approximation, A k ,B k ,C k ,D k Respectively is
Where k= { sc, ca }, l k To approximate the order, a i And b i (i＝1,2,…,l k ) To approximate coefficients, the following equation is given:
C. component fault modeling, under the fault, the original running structure of the controlled system is obviously changed, which is expressed as abnormal state space description, namely
Wherein A ', B ', H ', C ', D ' represent the matrix of the controlled system after the fault occurs;
(2) Information/physical hybrid fault modeling
Based on mathematical modeling of three types of faults, the augmentation vector x= [ X, X ca ,x sc ] T The method comprises the steps of establishing a state equation of an electric power Internet of things real-time control system considering information/physical hybrid faults as follows
Wherein A is f ＝A'，B f ＝B'Γ P Γ C ，H f ＝H'，C f ＝Ψ C Ψ P C'，D f ＝Ψ C Ψ P D'Γ P Γ C ；
2) The basic controller library normative design method comprises the following steps:
the real-time control system state equation has model uncertainty, so the controller design must have excellent robustness to the model uncertainty, and the application refers to the mixed H 2 /H ∞ The method is used for designing a basic controller to realize robust control on known faults;
considering that the controlled system has M normal working scenes and N typical fault scenes, the number of controllers in the basic controller library is M+N, so that Σ c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The } represents a set of controllers designed for a particular operating scenario, having the following general form:
wherein j=1, 2, …, m+n, x uj (t) intermediate variables introduced for the controller, A uj ,B uj ,C uj ,D uj A controller system matrix;
wherein, the control command u '(t) and the existing controller library Sigma are "according to the controlled system output y' (t) c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically receiptsThe step of converging on the ideal reference input r (t) "is specifically:
obtaining dynamic behavior of a real-time control system:
define a signal vector s (t) = [ ω (t), r (t), y '(t), u' (t)] T Based on Laplace transformation and polynomial matrix primary equal-line transformation method, the dynamic behavior of the controlled system and the controller can be described as
And the controlled system:
B P ＝{s＝(r,y′,u′) T |[-N p2 (ξ)0D p (ξ)-N p1 (ξ)]s(t)＝0}
and (3) a controller:
B cj ＝{s＝(r,y′,u′) T |[0N cj (ξ)-N cj (ξ)-D cj (ξ)]s(t)＝0}
in the formula, xi=d/dt, N p1 (ξ)、N p2 (ξ)、D p (ξ)、N cj (ξ)、D cj (ζ) is a matrix of a dimension-adaptive polynomial;
the controlled object and the controller can be known by the two formulas through common variables The dynamic behavior of the controlled object can be achieved by coupling variables [ r (t), y '(t), u' (t) with the controller] T Interconnection is realized;
from the dynamic behavior model of the controlled system, the controller Σ cj Dynamic behavior B in arbitrary detection period cj Complete reconstruction can be achieved only through y '(t) and u' (t), without the need for accurate prior system operating parameters and fault information; order theThe dynamic behavior of the reconstructed controller meets the following conditions
In the method, in the process of the application,for sigma cj Reconstructed reference input, if Σ cj For a controller that matches the current operating state of the controlled system, the reference input +.>Asymptotically converging to an ideal reference input r (t);
wherein, the output quantity y' (t) of the controlled system is input into the reconstructed referenceThe step of calculating the root mean square error to obtain the performance index MSE' is specifically as follows;
input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE; wherein the mean square error MSE is defined as the reconstructed reference input +.>A measure of the degree of difference between two variables of the controlled system output y' (t), i.e
The MSE index can be used for accurately expressing the control deviation of different controllers under the condition of uncertain information. Wherein, as the control target of the real-time control system is the controlled system output y '(t) to track the ideal reference input r (t), such as the state quantity of the generator rotation speed, the frequency f, etc., for the actual controlled system, the control target is the regional control error ACE of 0, i.e. r (t) -y' (t) =0, therefore, the reconstruction reference input is adoptedThe MSE index of the deviation between the output y' (t) of the controlled system is used as a measure of the dynamic performance of the controller;
the step of selecting the controller input corresponding to the minimum performance index MSE so as to realize zero trial-and-error one-time accurate switching specifically comprises the following steps:
according to the calculated performance index { MSE } 1 ,MSE 2 ,…,MSE M+N Selecting the most matched controller to be put into the control loop in a zero trial and error mode, namely selecting the preset controller input with the minimum performance index MSE, namely
In the load frequency fault tolerance control system and method based on reverse-direction reconstruction, a reverse-direction reconstruction strategy is provided, a corresponding controller is designed aiming at the expected information/physical faults to ensure the dynamic performance of the fault system, the prior fault information is not required to be acquired, and the fault tolerance control system only depends on the input and output measured values of the system, so that fault tolerance can be realized as much as possible, and better performance is provided for the stability control and the stable operation of the system of the information/physical tightly-coupled power Internet of things under the complex and diverse fault conditions.
The following illustrates the effect of the reconstruction-AFTC strategy of the present application by way of example:
the application takes the load frequency control of the three-region interconnection power system as shown in fig. 4 as a research object to carry out simulation experiments, the power grid frequency can be obtained by a synchronous phasor measurement device (Phasor Measurement Unit, PMU), and is uploaded to a control center through a communication network, meanwhile, a control instruction of the control center is sent to a synchronous generator set through the communication network, and the active power supply and demand balance of the power system under the rated frequency is maintained by adjusting the active power output of the synchronous generator set in real time, and the simulation parameters are shown in table 1. The simulation experiment parameters of table 1 are shown in the table 1
Consider that each region only has a single failure, for a duration of 50s. The normal communication delay of Area 1-Area 3, the attack delay of Area 2 and the load step fluctuation of Area 1-Area 3 are respectively shown in fig. 5, 6 and 7. Table 2 shows a possible operation scenario of the three-area interconnection system, 8 kinds in total. The adopted comparison method is as follows: consider a Passive Fault Tolerant Control (PFTC) strategy that progressively stabilizes for all fault conditions and an active fault tolerant control strategy based on fault identification (fault identification-AFTC). For comparison analysis of the superiority of the proposed fault tolerance strategy with the existing fault tolerance strategy, the root mean square error (Mean square error, MSE) in the case of a fault is calculated and quantified as compared to the normal case, as shown in fig. 8.
TABLE 2 interconnected Power System possible operation scenarios (Single failure per region)
As can be seen from FIG. 8, Δf is determined using the PFTC strategy 1 ΔP tie1 Is generally greater than AFTC, and is only slightly better than the AFTC strategy (Δf in the rest of the cases) in failure cases 2 and 6 1 ΔP tie1 The root mean square error is at least 18.40% and 220% higher). The performance index of the reconstruction-AFTC strategy is approximately similar to that of the fault identification-AFTC strategy in control performance, but the scheme only depends on the input/output of the controlled system when the best matching controller is selected, does not need to depend on accurate prior system parameter knowledge, can achieve similar fault-tolerant effect, and has stronger practicability.
Claims (6)
1. The utility model provides a load frequency fault-tolerant control system based on reverse row is reconstruction which characterized in that: the system comprises an information acquisition module, a controller library module, a reverse row reconstruction module, a performance index calculation module and a controller selection module:
and the information acquisition module is used for: acquiring and sensing the operation parameters of the controlled system by using a sensor arranged on the controlled system, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters;
and a controller library module: aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and the number of the controllers in the basic controller library is Sigma c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) Controller pool Σ c Representing a controller set designed for a specific operation scene, wherein M refers to a controlled system having M normal operation scenes, and N refers to a controlled system having N typical fault scenes;
and (3) a reverse reconstruction module: based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converging to an ideal reference input r (t);
the performance index calculation module is used for: input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE;
the controller selection module: selecting a controller input corresponding to the minimum performance index MSE, so as to realize zero trial-and-error disposable accurate switching;
the set of controllers in the base controller library is obtained by:
1) Establishing a real-time control system model of information/physical faults:
giving a controlled object model in a controlled system under normal conditions:
wherein x is R p 、u∈R q 、ω∈R r And y.epsilon.R v The state variable, the input variable, the external disturbance variable and the output variable are respectively, and the matrix A, B, H, C, D is a dimension-adaptive matrix;
modeling according to the expected fault action effect in the controlled system:
(1) Fault type modeling
A. Modeling data faults, wherein under the faults, the actual control quantity/quantity measurement and reality meet the following relations:
control amount: u (u) f (t)＝Γ P Γ C u(t)
Measuring the amount: y is f (t)＝Ψ C Ψ P y(t)
Wherein the subscript f represents a fault condition Γ P Diag (γp1, γp2, …, γpp), representing an actuator physical fault matrix; Γ -shaped structure C =diag (yc 1, yc 2, …, ycp) represents the network failure matrix between the controller and the actuator; psi C =diag (ψc1, ψc2, …, ψcq), representing a network information failure matrix between the sensor and the controller; psi P Diag (ψp1, ψp2, …, ψpq), representing the sensor physical fault matrix;
B. modeling time-effect type faults, and letting tau sc Representing the network delay of the sensor to the controller, τ ca In order to delay the network from the controller to the executor in the controlled system, the control instruction sent by the controller is u '(t), the received output quantity of the controlled system is y' (t), and the control instruction u (t) sent by the control center actually received by the executor in the controlled system and the actual output quantity y (t) of the controlled system satisfy the following conditions
u(t)＝u'(t+τ ca )
y'(t)＝y(t+τ sc )
Further, the pure time-lag links are approximated by adopting Pade approximation technology, and the approximated state space is expressed as
Wherein x is ca (t)、x sc (t) intermediate variable introduced for Pade approximation, A k ,B k ,C k ,D k Respectively is
Where k= { sc, ca }, l k To approximate the order, a i And b i (i＝1,2,…,l k ) To approximate coefficients, the following equation is given:
C. component fault modeling, under the fault, the original running structure of the controlled system is obviously changed, which is expressed as abnormal state space description, namely
Wherein A ', B ', H ', C ', D ' represent the matrix of the controlled system after the fault occurs;
(2) Information/physical hybrid fault modeling
Based on mathematical modeling of three types of faults, the augmentation vector x= [ X, X ca ,x sc ] T The method comprises the steps of establishing a state equation of an electric power Internet of things real-time control system considering information/physical hybrid faults as follows
Wherein A is f ＝A'，B f ＝B'Γ P Γ C ，H f ＝H'，C f ＝Ψ C Ψ P C'，D f ＝Ψ C Ψ P D'Γ P Γ C ；
2) The basic controller library normative design method comprises the following steps:
the real-time control system state equation has model uncertainty, so the controller design must have excellent robustness to the model uncertainty, and the application refers to the mixed H 2 /H ∞ The method is used for designing a basic controller to realize robust control on known faults;
considering that the controlled system has M normal working scenes and N typical fault scenes, the number of controllers in the basic controller library is M+N, so that Σ c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The } represents a set of controllers designed for a particular operating scenario, having the following general form:
wherein j=1, 2, …, m+n, x uj (t) intermediate variables introduced for the controller, A uj ,B uj ,C uj ,D uj Is a matrix of controller systems.
2. The reverse-behavior-based fault-tolerant control system of load frequencies for reconstruction as claimed in claim 1, wherein: mean square error MSE is defined as the reconstructed reference inputA measure of the degree of difference between two variables of the controlled system output y' (t), i.e
The MSE index can be used for accurately expressing the control deviation of different controllers under the condition of uncertain information.
3. A load frequency fault tolerance control method based on reverse direction reconstruction comprises the following steps:
aiming at the predicted information/physical faults, designing corresponding controllers to ensure the dynamic performance of a fault system to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and the sigma is given by c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The method comprises the steps that a controller set designed for a specific operation scene is represented, M means that a controlled system has M normal operation scenes, and N means that the controlled system has N typical fault scenes;
receiving operation parameters of a controlled system acquired and sensed by a sensor, and generating a corresponding output quantity y' (t) of the controlled system according to the received operation parameters;
receiving a control instruction u' (t) generated by a control center according to the operation parameters of the sensor acquisition sensing controlled system;
based on the controlled system output y '(t), the control command u' (t) and the existing controller library Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->Asymptotically converging to an ideal reference input r (t);
input of the output quantity y' (t) of the controlled system and the reconstructed referencePerforming root mean square error calculation to obtain a performance index MSE;
selecting a controller input corresponding to the minimum performance index MSE, so as to realize zero trial-and-error disposable accurate switching;
wherein, the corresponding controllers are designed to ensure the dynamic performance of the fault system aiming at the predicted information/physical faults to form a basic controller library, wherein the number of the controllers in the basic controller library is M+N, and Sigma is formed c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The method comprises the steps of designing a controller set for a specific operation scene, wherein M refers to that a controlled system has M normal operation scenes, N refers to that the controlled system has N typical fault scenes, and the steps are as follows:
1) Establishing a real-time control system model of information/physical faults:
giving a controlled object model in a controlled system under normal conditions:
wherein x is R p 、u∈R q 、ω∈R r And y.epsilon.R v The state variable, the input variable, the external disturbance variable and the output variable are respectively, and the matrix A, B, H, C, D is a dimension-adaptive matrix;
modeling according to the expected fault action effect in the controlled system:
(1) Fault type modeling
A. Modeling data faults, wherein under the faults, the actual control quantity/quantity measurement and reality meet the following relations:
control amount: u (u) f (t)＝Γ P Γ C u(t)
Measuring the amount: y is f (t)＝Ψ C Ψ P y(t)
In the middle, lowerThe label f indicates the fault condition Γ P Diag (γp1, γp2, …, γpp), representing an actuator physical fault matrix; Γ -shaped structure C =diag (yc 1, yc 2, …, ycp) represents the network failure matrix between the controller and the actuator; psi C =diag (ψc1, ψc2, …, ψcq), representing a network information failure matrix between the sensor and the controller; psi P Diag (ψp1, ψp2, …, ψpq), representing the sensor physical fault matrix;
B. modeling time-effect type faults, and letting tau sc Representing the network delay of the sensor to the controller, τ ca In order to delay the network from the controller to the executor in the controlled system, the control instruction sent by the controller is u '(t), the received output quantity of the controlled system is y' (t), and the control instruction u (t) sent by the control center actually received by the executor in the controlled system and the actual output quantity y (t) of the controlled system satisfy the following conditions
u(t)＝u'(t+τ ca )
y'(t)＝y(t+τ sc )
Further, the application adopts Pade approximation technology to approximate the pure time lag links, and the state space after approximation is expressed as
Wherein x is ca (t)、x sc (t) intermediate variable introduced for Pade approximation, A k ,B k ,C k ,D k Respectively is
Where k= { sc, ca }, l k To approximate the order, a i And b i (i＝1,2,…,l k ) To approximate coefficients, the following equation is given:
C. component fault modeling, under the fault, the original running structure of the controlled system is obviously changed, which is expressed as abnormal state space description, namely
Wherein A ', B ', H ', C ', D ' represent the matrix of the controlled system after the fault occurs;
(2) Information/physical hybrid fault modeling
Based on mathematical modeling of three types of faults, the augmentation vector x= [ X, X ca ,x sc ] T The method comprises the steps of establishing a state equation of an electric power Internet of things real-time control system considering information/physical hybrid faults as follows
Wherein A is f ＝A'，B f ＝B'Γ P Γ C ，H f ＝H'，C f ＝Ψ C Ψ P C'，D f ＝Ψ C Ψ P D'Γ P Γ C ；
2) The basic controller library normative design method comprises the following steps:
the real-time control system state equation has model uncertainty, so the controller design must have excellent robustness to the model uncertainty, and the application refers to the mixed H 2 /H ∞ The method is used for designing a basic controller to realize robust control on known faults;
considering that the controlled system has M normal working scenes and N typical fault scenes, the number of controllers in the basic controller library is M+N, so that Σ c ＝{Σ c1 ,Σ c2 ,…,Σ c(M+N) The } represents a set of controllers designed for a particular operating scenario, having the following general form:
wherein j=1, 2, …, m+n, x uj (t) intermediate variables introduced for the controller, A uj ,B uj ,C uj ,D uj Is a matrix of controller systems.
4. A load frequency fault tolerant control method based on reverse-run reconstruction as claimed in claim 3, characterized in that "based on the controlled system output y '(t), the control command u' (t) and the existing controller bank Σ c Reversely reconstructing dynamic behaviors of the controlled system in a parallel mode and obtaining a reconstructed reference inputReconstructed reference input->The step of asymptotically converging to an ideal reference input r (t) "is specifically:
obtaining dynamic behavior of a real-time control system:
define a signal vector s (t) = [ ω (t), r (t), y '(t), u' (t)] T Based on Laplace transformation and polynomial matrix primary equal-line transformation method, the dynamic behavior of the controlled system and the controller can be described as
And the controlled system:
B P ＝{s＝(r,y′,u′) T |[-N p2 (ξ) 0 D p (ξ) -N p1 (ξ)]s(t)＝0}
and (3) a controller:
B cj ＝{s＝(r,y′,u′) T |[0 N cj (ξ) -N cj (ξ) -D cj (ξ)]s(t)＝0}
in the formula, xi=d/dt, N p1 (ξ)、N p2 (ξ)、D p (ξ)、N cj (ξ)、D cj (ζ) is a matrix of a dimension-adaptive polynomial;
the controlled object and the controller can be known by the two formulas through common variables The dynamic behavior of the controlled object can be achieved by coupling variables [ r (t), y '(t), u' (t) with the controller] T Interconnection is realized;
from the dynamic behavior model of the controlled system, the controller Σ cj Dynamic behavior B in arbitrary detection period cj Complete reconstruction can be achieved only through y '(t) and u' (t), without the need for accurate prior system operating parameters and fault information; order theThe dynamic behavior of the reconstructed controller meets the following conditions
In the method, in the process of the application,for sigma cj Reconstructed reference input, if Σ cj For a controller that matches the current operating state of the controlled system, the reference input +.>Asymptotically converges to an ideal reference input r (t).
5. The load frequency fault tolerance control method based on reverse-direction reconstruction as claimed in claim 4, wherein: "input of the output y' (t) of the controlled system and the reconstructed referenceThe step of calculating the root mean square error to obtain the performance index MSE' is specifically as follows;
using a reconstructed reference inputThe mean square error MSE index between the output of the controlled system y' (t) and the MSE index is used as a measurement index, and the MSE is defined as the reconstructed reference input +.>A measure of the degree of difference between two variables of the controlled system output y' (t), i.e
The MSE index can be used for accurately expressing the control deviation of different controllers under the condition of uncertain information.
6. The load frequency fault tolerance control method based on reverse-direction reconstruction as claimed in claim 5, wherein: the step of selecting the controller input corresponding to the minimum performance index MSE so as to realize zero trial and error disposable accurate switching is specifically as follows:
according to the calculated performance index { MSE } 1 ,MSE 2 ,…,MSE M+N Selecting the most matched controller to be put into the control loop in a zero trial and error mode, namely selecting the preset controller input with the minimum performance index MSE, namely
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111300633.0A CN113949064B (en) | 2021-11-04 | 2021-11-04 | Load frequency fault-tolerant control system and method based on reverse row reconstruction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111300633.0A CN113949064B (en) | 2021-11-04 | 2021-11-04 | Load frequency fault-tolerant control system and method based on reverse row reconstruction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113949064A CN113949064A (en) | 2022-01-18 |
| CN113949064B true CN113949064B (en) | 2023-11-14 |
Family
ID=79337498
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111300633.0A Active CN113949064B (en) | 2021-11-04 | 2021-11-04 | Load frequency fault-tolerant control system and method based on reverse row reconstruction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113949064B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015078474A1 (en) * | 2013-11-28 | 2015-06-04 | Vestas Wind Systems A/S | A power plant controller for generating a power reference to wind turbine generators |
| CN107069771A (en) * | 2016-12-23 | 2017-08-18 | 重庆大学 | The method of LOAD FREQUENCY control of the multi-area Interconnected Power System based on failure tolerant |
| CN107861383A (en) * | 2017-10-23 | 2018-03-30 | 天津大学 | Satellite failure diagnosis and fault tolerant control method based on Adaptive Observer |
| CN110112940A (en) * | 2019-06-14 | 2019-08-09 | 安徽工业大学 | A kind of PWM rectifier adaptive sliding mode QPIR control method under the β coordinate system based on α |
| WO2020142984A1 (en) * | 2019-01-10 | 2020-07-16 | 大连理工大学 | Active fault tolerant control method of aero-engine based on error interval observer |
-
2021
- 2021-11-04 CN CN202111300633.0A patent/CN113949064B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015078474A1 (en) * | 2013-11-28 | 2015-06-04 | Vestas Wind Systems A/S | A power plant controller for generating a power reference to wind turbine generators |
| CN107069771A (en) * | 2016-12-23 | 2017-08-18 | 重庆大学 | The method of LOAD FREQUENCY control of the multi-area Interconnected Power System based on failure tolerant |
| CN107861383A (en) * | 2017-10-23 | 2018-03-30 | 天津大学 | Satellite failure diagnosis and fault tolerant control method based on Adaptive Observer |
| WO2020142984A1 (en) * | 2019-01-10 | 2020-07-16 | 大连理工大学 | Active fault tolerant control method of aero-engine based on error interval observer |
| CN110112940A (en) * | 2019-06-14 | 2019-08-09 | 安徽工业大学 | A kind of PWM rectifier adaptive sliding mode QPIR control method under the β coordinate system based on α |
Non-Patent Citations (3)
| Title |
|---|
| 参数变化对异步机解耦控制的影响;王秀芝, 王红, 许镇琳, 车延博, 天津大学, 马小亮;天津大学学报(第04期);全文 * |
| 基于混沌优化的多模型容错控制方法;李炜;马克;鲁保云;;控制工程(第03期);全文 * |
| 基于逆系统多模型内模主动容错控制;李炜;程应峰;许德智;;计算机仿真(第03期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113949064A (en) | 2022-01-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Amini et al. | Dynamic load altering attacks in smart grid | |
| Yin et al. | Data-driven design of robust fault detection system for wind turbines | |
| CN108234492B (en) | Power information physical cooperative attack analysis method considering load data false injection | |
| CN104850889B (en) | Airplane rotation actuator drive unit adaptive fault detection, isolation and confidences assessment method | |
| CN105976257A (en) | Power grid vulnerability evaluation method based on membership function fuzzy comprehensive evaluation method | |
| CN108075470A (en) | A kind of power grid transient frequency feature Forecasting Methodology based on SFR and ELM | |
| CN107968409A (en) | A kind of probability load flow calculation method and system for considering imbalance power distribution | |
| CN103262379B (en) | Power inverter, the control device of power inverter and the control method of power inverter | |
| JP5710303B2 (en) | Parameter estimation method and apparatus for power system contraction model | |
| Mondai et al. | Analysis of limiting bounds for stalling of natural gas genset in the CERTS microgrid test bed | |
| Djukanovic et al. | Neural net based determination of generator-shedding requirements in electric power systems | |
| CN113949064B (en) | Load frequency fault-tolerant control system and method based on reverse row reconstruction | |
| CN109066651A (en) | The calculation method of wind-powered electricity generation-load scenarios limit transmitted power | |
| CN106786602B (en) | A kind of distribution power system load flow calculation method | |
| CN109638871B (en) | Main network dividing method of large-scale alternating current-direct current hybrid system considering wind power access | |
| CN105334747A (en) | Ship dynamic positioning three-redundancy computer data voting synchronization method | |
| Hu et al. | Frequency prediction model combining ISFR model and LSTM network | |
| Wang et al. | A three-machine equivalent system frequency response model and its closed-form solution | |
| Rahimilarki et al. | Grey-box model identification and fault detection of wind turbines using artificial neural networks | |
| CN111860969A (en) | Power transmission network extension planning method based on reinforcement learning algorithm | |
| Negnevitsky et al. | Preventing large-scale emergencies in modern power systems: AI approach | |
| CN113221064B (en) | Model processing method, device, computer equipment and storage medium | |
| Callenes et al. | Dynamic reconfiguration for resilient state estimation against cyber attacks | |
| Pan et al. | Modelica-supported attack impact evaluation in cyber physical energy system | |
| Baros et al. | Towards resilient cyber-physical energy systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |