Frequency regulation method and device for power system under multiple attacks of deterministic network

Abstract

The application provides a power system frequency anti-interference regulation method and device under certainty network multiple attacks, and the method comprises the following steps: obtaining multiple groups of frequency regulation related parameters of a controlled area; inputting regional frequency deviation data and regional power deviation data in each group of frequency regulation related parameters into an adaptive network attack detection system to obtain estimated regional control error data; the adaptive network attack detection system utilizes sample data under multiple network attacks collected in advance to train a neural fuzzy system physical model; according to actual regional control error data and estimated regional control error data of the multiple groups of frequency regulation related parameters, root mean square error is calculated; according to the root mean square error and a preset range, a relieved regional control error is determined from the actual regional control error and the estimated regional error. The application comprehensively considers multiple attacks, can quickly detect new attacks, and improves power adjustment efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power systems, in particular to a deterministic network multi-attack power system frequency anti-disturbance regulation method and device. BACKGROUND

[0002] The progress of computer technology and communication technology promotes the development of cyber-physical systems, and the traditional power system is changing towards the trend of smart power system. Most smart power systems are connected to the control center through deterministic network channels for the regulation of important information. However, the transmission of important data in the communication link is vulnerable to various network attacks, including denial of service attacks, false data injection attacks and time delay attacks, which brings challenges to the frequency regulation of the power system. Power system frequency regulation is crucial for maintaining grid frequency and minimizing tie-line power deviation, and by obtaining data from the grid to adjust the generator output power. When the smart power system is subjected to network attacks, not only will the system frequency be severely affected, but also the stability and economic operation of the power system. This is because frequency deviation can trigger various protection measures, resulting in service interruption, infrastructure damage, and even large-scale power outages. Therefore, power system frequency control is crucial for enhancing the network security of the power system.

[0003] At present, only single time delay attack or single denial of service attack is considered for power system frequency control, and there is no solution to solve multiple attacks at the same time. In addition, existing control technologies, such as event-triggered mechanism or load prediction-based detection methods, although provide effective detection technology, require accurate system model and large amount of calculation, and have significant detection delay. In addition, different attacks have different effects on the power system, and more specific evaluation of multiple network attacks is needed, as well as consideration of real-time load disturbance and system nonlinearity of the actual power system. SUMMARY

[0004] The existing technology does not comprehensively consider the frequency regulation of the power system under network attack, and in addition, the existing power regulation has large amount of calculation and detection delay, and cannot quickly detect network attacks.

[0005] In order to solve the above technical problems, the first aspect of the present application provides a deterministic network multi-attack power system frequency anti-disturbance regulation method, comprising:

[0006] Obtaining a plurality of sets of frequency regulation related parameters of the controlled area, wherein each set of frequency regulation related parameters comprises regional frequency deviation data, regional power deviation data and actual regional control error data;

[0007] The regional frequency deviation data and the regional power deviation data in each set of frequency regulation related parameters are input into an adaptive network attack detection system to obtain estimated regional control error data, wherein the adaptive network attack detection system is obtained by training a neural fuzzy system physical model using sample data under multiple network attacks collected in advance; the sample data of each region includes regional frequency deviation, regional power deviation and regional control error; and the neural fuzzy system physical model is a combination of fuzzy logic and neural network.

[0008] According to the actual regional control error data and the estimated regional control error data of multiple sets of frequency regulation related parameters, a root mean square error is calculated.

[0009] If the root mean square error is within a preset range, no attack is detected, and the actual regional control error is taken as a mitigation regional control error; if the root mean square error exceeds the preset range, an attack exists, and the estimated regional control error is taken as the mitigation regional control error.

[0010] The mitigation regional control error is used to regulate the frequency of the power system.

[0011] As a further embodiment of the present application, an adaptive network attack detection system is established according to sample data of each region under multiple network attacks collected in advance and a neural fuzzy system physical model, and includes:

[0012] The following fuzzy rules are established:

[0013]

[0014] wherein x1 and x2 are regional frequency deviation and regional power deviation respectively, and are antecedents of interval type-2 fuzzy sets, Z is the output of fuzzy logic, i.e., estimated regional control error, k j1 , k j2 and l j are fuzzy system adaptive parameters, j=1, 2,..., N is the number of fuzzy rules, and R (j) is the jth fuzzy rule.

[0015] According to the fuzzy rules, a neural fuzzy system physical model is constructed.

[0016] The sample data under multiple network attacks collected in advance are divided into a training sample set and a test sample set.

[0017] The neural fuzzy system physical model is trained using the training sample set to obtain the adaptive network attack detection system.

[0018] The adaptive network attack detection system is tested by using the test sample, if the test is passed, the training is ended, if the test is not passed, the neural fuzzy system physical model is adjusted, and the training and test process is restarted.

[0019] As a further embodiment of the present application, the sample data determination process under multiple network attacks comprises:

[0020] The regional control error models under various network attacks are constructed in advance, wherein the regional control error models are functions of regional frequency deviation, regional power deviation and attack parameters;

[0021] According to multiple sets of actual regional frequency deviation and regional power deviation, the regional control error models under various network attacks are used to obtain multiple sets of sample data under various network attacks.

[0022] As a further embodiment of the present application, the network attacks include false data injection attack, scaling attack, denial of service attack and time delay attack.

[0023] As a further embodiment of the present application, the regional control error model under false data injection attack comprises:

[0024] The control error model of frequency measurement data of region i under false data injection attack is:

[0025]

[0026] wherein, is the regional control error of region i at time t under attack, β i is the frequency bias factor of region i, (Δf i +ψ a1 ) is the frequency deviation of region i at time t under attack, ψ a1 is the false data injection attack signal, a ij is the capacity factor between region i and region j, ΔP tie,ij (t) is the power deviation of the tie line between region i and region j at time t, t is the instantaneous time, and m is the total number of control regions;

[0027] The control error model of power measurement data between region i and region k under virtual data injection attack is:

[0028]

[0029] wherein, Δf i (t) is the frequency deviation of region i at time t, a ik is the capacity factor between region i and region k, (ΔP tie,ik +ψ a2(t) is the power deviation between region i and region k at time t under attack, ψ a2 is the false data injection attack signal.

[0030] The control error model of region i under the virtual data attack is:

[0031]

[0032] where ACE i (t) is the region control error of region i at time t, ψ a3 is the false data injection attack signal.

[0033] As a further embodiment of the present application, the region control error model under the scaling attack includes:

[0034] The control error model of frequency measurement data of region i under the scaling attack is:

[0035]

[0036] where, (t) is the region control error of region i at time t under attack, β i is the frequency bias factor of region i, (Δf i × λ a1 (t) is the frequency deviation of region i at time t under attack, λ a1 is the scaling attack signal, a ij is the capacity factor between region i and region j, ΔP tie,ij (t) is the power deviation between region i and region j at time t under attack, t is the instantaneous time, and m is the total number of control regions;

[0037] The control error model of power measurement data between region i and region k under the scaling attack is:

[0038]

[0039] where, Δf i (t) is the frequency deviation of region i at time t, a ik is the capacity factor between region i and region k, (ΔP tie,ik × λ a2 (t) is the power deviation between region i and region k at time t under attack, λ a2 is the scaling attack signal.

[0040] The control error model of control error data of region i under the scaling attack is:

[0041]

[0042] where ACE i (t) is the area control error of area i at time t, λ a3 is the scaled attack signal.

[0043] As a further embodiment of the present application, the area control error model under denial-of-service attack includes:

[0044]

[0045] where n is the number of periods, T is one period of the disturbance, T off is the idle time of the disturbance, is the minimum idle time of the disturbance, which is in the range of

[0046] the interval [(n-1)T,(n-1)T+T off ] is the idle time of the disturbance signal, indicating that the denial-of-service attack is not successful, and the frequency measurement data of the area received by the controller and the power measurement data of the area are correct;

[0047] the interval [(n-1)T+T off ,nT] is the active interval of the disturbance signal, indicating that the denial-of-service attack is activated, and the controller does not get any information.

[0048] As a further embodiment of the present application, the area control error model under time delay attack includes:

[0049] The control error model of the frequency measurement data of area i under time delay attack is:

[0050]

[0051] where, is the area control error of area i at time t under attack, β i is the frequency bias factor of area i, Δf i (t-t d ) is the frequency deviation of area i at time t under attack, t d is the time delay attack signal, a ij is the capacity factor between area i and area j, ΔP tie,ij (t) is the power deviation of the tie line between area i and area j at time t, t is the instantaneous time, and m is the total number of control areas;

[0052] The control error model of the power measurement data between area i and area k under time delay attack is:

[0053]

[0054] wherein, Δf i (t) is the frequency deviation of area i at time t, a ik is the capacity factor between area i and area k, ΔP tie,ik (t-t d ) is the power deviation of tie-line between area i and area k at time t under attack, t d is the time-delay attack signal.

[0055] The control error model of area i under time-delay attack is:

[0056]

[0057] wherein, ACE i (t-t d ) is the area control error of area i at time t-t d , t d is the time-delay attack signal.

[0058] As a further embodiment of the present application, the power system frequency is regulated by using the alleviated area control error, comprising:

[0059] inputting the alleviated area control error into a pre-constructed double-loop enhanced active disturbance rejection controller to obtain a control amount of the governor, and sending the control amount to the governor of the controlled area to regulate the frequency of the power system of the controlled area;

[0060] wherein, the double-loop enhanced active disturbance rejection controller is constructed by using an area power system model with disturbance and nonlinearity.

[0061] As a further embodiment of the present application, the construction process of the double-loop enhanced active disturbance rejection controller comprises:

[0062] constructing an area power system model with disturbance and nonlinearity represented by the following formula:

[0063]

[0064] wherein, y is the area control error ACE of area i, is the first derivative of y, is the second derivative of y, u is the control output of the governor, is the governor output speed coefficient, d is the external disturbance, and f is the overall system nonlinearity.

[0065] According to the area power system model, a state equation of the area power system is constructed, wherein the state quantity in the state equation is y, f.

[0066] According to the state equation of the regional power system, a self-disturbance state observer of the regional power system represented by the following formula is constructed:

[0067]

[0068] Wherein,

[0069] is the estimated value of y, is the observation value of the system state, L0=[ξ1,ξ2,ξ3] T is the observer gain;

[0070] According to the self-disturbance state observer of the regional power system, the observer gain in the stable state is solved by using the bandwidth method;

[0071] According to the observer gain in the stable state and the model of the regional power system, a feedforward control loop L1 and a feedback control loop L2 represented by the following formula are determined:

[0072]

[0073] Wherein, u1 is the control quantity of the feedforward control loop L1, u2 is the control quantity of the feedback control loop L2, is the output speed coefficient of the speed regulator, is the observation value of the overall system nonlinearity, k ti , k i Indicates the integral gain, η1 and η2 are proportional control parameters and derivative control parameters respectively, η1=ω c 2 , η2=2ω c , ω c is the bandwidth of the state error feedback control law, and s is the Laplace operator in the complex frequency domain;

[0074] According to the feedforward control loop L1 and the feedback control loop L2, a double-loop enhanced self-disturbance controller represented by the following formula is determined:

[0075] U(s)=u2(u1+ε o -γ);

[0076] Wherein, ε o is a predetermined error limit, and γ is a frequency deviation feedback signal of the generator output.

[0077] The second aspect of the present application provides a deterministic network under multiple attacks of power system frequency disturbance regulation device, comprising:

[0078] A parameter acquisition unit is configured to acquire a plurality of groups of frequency regulation related parameters of a region to be controlled, wherein each group of frequency regulation related parameters comprises region frequency deviation data, region power deviation data and actual region control error data.

[0079] A detection unit is configured to input the region frequency deviation data and the region power deviation data in each group of frequency regulation related parameters into an adaptive network attack detection system to obtain estimated region control error data, wherein the adaptive network attack detection system is trained by using sample data under a plurality of network attacks to obtain a neural fuzzy system physical model; the sample data of each region comprises region frequency deviation Δf i (t), region power deviation ΔP tie,ij (t) and region control error ACE i (t); the neural fuzzy system physical model is a combination of fuzzy logic and neural network.

[0080] A calculation unit is configured to calculate a root mean square error according to the actual region control error data and the estimated region control error data of the plurality of groups of frequency regulation related parameters.

[0081] A judgment unit is configured to, if the root mean square error is within a preset range, determine that no attack is detected, and take the actual region control error as a mitigation region control error; if the root mean square error exceeds the preset range, determine that an attack exists, and take the estimated region control error as the mitigation region control error.

[0082] A control unit is configured to regulate the frequency of the power system by using the mitigation region control error.

[0083] The third aspect of the present application provides a computer device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the method of any one of the preceding embodiments when executing the computer program.

[0084] The fourth aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor of a computer device to implement the method of any one of the preceding embodiments.

[0085] The fifth aspect of the present application provides a computer program product, wherein the computer program product comprises a computer program, and the computer program is executed by a processor of a computer device to implement the method of any one of the preceding embodiments.

[0086] The application provides a power system frequency anti-interference regulation method and device under certainty network multiple attacks.

[0087] In order to make the above and other objects, features and advantages of the present application more apparent, preferred embodiments will be described in detail below with reference to the accompanying drawings, and will be specifically explained in the following manner. BRIEF DESCRIPTION OF DRAWINGS

[0088] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0089] Figure 1 A training process flowchart of the adaptive network attack detection system of the embodiment of the present application is shown;

[0090] Figure 2 A flowchart of the power system frequency anti-interference regulation method of the embodiment of the present application under certainty network multiple attacks is shown;

[0091] Figure 3 A flowchart of the construction process of the double-loop enhanced active disturbance rejection controller of the embodiment of the present application is shown;

[0092] Figure 4 A structure diagram of the power system frequency anti-interference regulation device of the embodiment of the present application under certainty network multiple attacks is shown;

[0093] Figure 5 A structure diagram of the computer device of the embodiment of the present application is shown.

[0094] BRIEF DESCRIPTION OF DRAWINGS

[0095] 401, parameter acquisition unit;

[0096] 402, detection unit;

[0097] 403. Calculation Unit;

[0098] 404. Judgment Unit;

[0099] 405. Control unit;

[0100] 502. Computer equipment;

[0101] 504, Processor;

[0102] 506. Memory;

[0103] 508. Drive mechanism;

[0104] 510. Input / output module;

[0105] 512. Input devices;

[0106] 514. Output devices;

[0107] 516. Presentation equipment;

[0108] 518. Graphical User Interface;

[0109] 520. Network interface;

[0110] 522. Communication link;

[0111] 524. Communication bus. Detailed Implementation

[0112] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0113] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0114] The specification provides method operation steps as described in the embodiments or flowcharts, but can include more or less operation steps based on conventional or non-inventive labor. The order of steps listed in the embodiments is only one of the many execution orders of the steps, and does not represent the only execution order. In actual system or device product execution, the method order shown in the embodiments or the drawings can be executed in sequence or in parallel.

[0115] It should be noted that the data involved in the present application (including but not limited to data for analysis, stored data, displayed data, etc.) are all information and data authorized by the user or fully authorized by all parties, and the acquisition, transmission, storage, use and processing of related data comply with relevant national and regional laws, regulations and standards.

[0116] It should be noted that in the embodiments of the present application, some software, components, models and other industry existing solutions may be mentioned, which should be considered as exemplary, and the purpose is only to illustrate the feasibility of the implementation of the technical solutions of the present application, but it does not mean that the applicant has or will necessarily use the solution.

[0117] The prior art does not comprehensively consider the frequency regulation of the power system under network attack, in addition, the existing power regulation has large amount of calculation and detection delay, and cannot quickly detect network attacks. Based on this, in an embodiment of the present application, an adaptive network attack detection system capable of identifying regional control error under multiple network attacks is established in advance, whether the regional power system is attacked is identified based on the adaptive network attack detection system, and then the frequency of the power system is controlled according to the identification result, which can avoid the frequency regulation of the power system from being attacked by various network attacks, and quickly adjust the power when attacked.

[0118] Specifically, in an embodiment of the present application, a method for training an adaptive network attack detection system is provided, as shown in Figure 1 , comprising:

[0119] Step 101, the following fuzzy rules are established:

[0120]

[0121] Wherein, x1 and x2 are the regional frequency deviation and the regional power deviation, and are the antecedents of the interval type-2 fuzzy set, Z is the output of the fuzzy logic, that is, the estimated regional control error, k j1 , k j2 and l j are fuzzy system adaptive parameters, j=1,2,…,N is the number of fuzzy rules, R (j)is the jth fuzzy rule. In practice, the number of fuzzy rules can be set according to actual conditions, and the present application does not limit the specific value thereof.

[0122] Since the interval type-2 fuzzy set retains the processing ability of type-2 fuzzy set for high uncertainty, and has the advantages of fewer parameters and simple calculation, the embodiment uses interval type-2 fuzzy logic to design an adaptive network attack detection system.

[0123] In practice, in order to simplify the calculation, a single-valued fuzzy operator and product inference are used to calculate the activation interval of the input vector of each fuzzy rule, specifically:

[0124]

[0125] g j (x) and represent the upper and lower bounds of the activation interval; and are the lower membership function and the upper membership function, respectively, and the upper and lower membership functions are represented by Gaussian functions as follows:

[0126]

[0127]

[0128] The defuzzification output of the interval type-2 fuzzy logic system is:

[0129]

[0130] where the interval value g l (x) and g r (x) are obtained by Karnik-Mendel (KM) algorithm, and the KM algorithm is obtained by solving g j (x) and the activation interval set to iteratively calculate the center point interval value g l (x) and g r (x).

[0131] Step 102, according to the fuzzy rule, a neural fuzzy system physical model is constructed.

[0132] where the neural fuzzy system physical model is a combination of fuzzy logic and neural network. The frequency deviation Δf i and the power deviation ΔP tie,ij are used as inputs of the adaptive network attack detection system, and the area control error ACE is used as the output.

[0133] Step 103, the pre-collected sample data under multiple network attacks is divided into a training sample set and a test sample set.

[0134] In implementation, the pre-collected sample data under network attacks can be actual collected data, including frequency deviation Δf i , power deviation ΔP tie,ij and corresponding ACE under various network attacks, and frequency deviation Δf i , power deviation ΔP tie,ij and corresponding ACE under no network attack.

[0135] Considering the problem of small amount of attack data samples, attack sample data can also be determined based on the attack model, so as to improve the accuracy of the adaptive network attack detection system in determining the real ACE (no attack / under attack).

[0136] Step 104, the neural fuzzy system physical model is trained by using the training sample set to obtain the adaptive network attack detection system.

[0137] Step 105, the adaptive network attack detection system is tested by using the test sample, if the test is passed, the training is ended; if the test is not passed, the neural fuzzy system physical model is adjusted, and the training and testing process is restarted.

[0138] In implementation, whether the test is passed can be determined by the detection accuracy of the adaptive network attack detection system.

[0139] The adaptive network attack detection system trained in this embodiment can predict the predicted ACE based on the historical network attack. In implementation, if the error between the ACE predicted by the adaptive network attack detection system and the actually measured ACE is within the preset range, it is indicated that there is no new attack, and if the error between the ACE predicted by the adaptive network attack detection system and the actually measured ACE exceeds the preset range, it is indicated that there is a new attack.

[0140] In a specific embodiment of the present application, the training sample data is loaded into the neural fuzzy system physical model in MATLAB to generate the adaptive network attack detection system in the training stage. Then, in the test stage, the test sample data is input into the adaptive network attack detection system to verify the detection performance. The following is the design steps of the adaptive network attack detection system:

[0141] (1) The frequency deviation Δf i , power deviation ΔP tie,ij and regional control error ACE are obtained by applying and testing on the actual power system. Then Δf i and ΔP tie,ijAs input, corresponding ACE as output, the data is divided into training sample data and test sample data in the ratio of 80:20, wherein the test sample data and the training sample data are respectively used for training and testing of the adaptive network attack detection system.

[0142] (2) For training, load the training sample data into the neural fuzzy system physical model in Matlab. In order to generate the adaptive network attack detection system, the following grid partition is selected: MF=10, MF type=triangle. Output: MFtype=constant. The following parameters are selected to train the adaptive network attack detection system: Optimization method=hybrid, error tolerance=0, epochs=10. After the parameter setting is completed, the adaptive network attack detection system training is carried out.

[0143] (3) When testing, load the test sample data into the neural fuzzy system physical model to obtain the average test error of the test data. Reducing the test error makes the training more accurate.

[0144] In an embodiment of the present application, the sample data determination process under multiple network attacks includes:

[0145] A regional control error model under various network attacks is constructed in advance, wherein the regional control error model is a function of regional frequency deviation, regional power deviation and attack parameters;

[0146] According to multiple sets of actual regional frequency deviation and regional power deviation, multiple sets of sample data under various network attacks are obtained by using the regional control error model under various network attacks.

[0147] Specifically, the network attacks include false data injection attack, scaling attack, denial of service attack and time delay attack.

[0148] In order to ensure that the power deviation ΔP tie,ij (t) and the frequency deviation Δf i (t) of different regions of the power system remain within a safe range, for the i-th control region, at time t, the regional control error (ACE) can be expressed as:

[0149]

[0150] Wherein, ACE i (t) is the control error of region i at time t, β i is the frequency bias factor of region i, Δf i (t) is the frequency deviation of region i at time t, the coefficient a ij is the capacity factor between region i and region j, and ΔPtie,ij (t) is the power deviation between region i and region j, t is the instantaneous time, and m is the total number of control regions.

[0151] Δf i (t) = f i (t) - f ref ;

[0152] ΔP tie,ij (t) = P tie,ij (t) - P ref ;

[0153] wherein f ref , P ref are the preset frequency and power of the power system; f i (t), P tie,ij (t) are the actual frequency of region i at time t and the power deviation between region i and region j, which are collected by the phasor measurement unit.

[0154] In the smart power system, f i , P tie,ij and ACE i are the most important parameters for frequency regulation. Due to the vulnerability of the communication path between the phasor measurement unit and the control center and the control region, attackers can tamper with, interfere with or even delay the data uploaded by f i and P tie,ij to the control center or the ACE i data transmitted by the control center to the control region. In all cases, ACE i will eventually be affected, causing the controller in the control region to issue incorrect setting instructions to the governor, thereby disrupting the normal operation of the entire power system. We analyze the impact of different network attacks on the power system and establish a mathematical model of the regional control error under different network attacks as follows:

[0155] (1) The regional control error model under false data injection attack includes:

[0156] The control error model of the frequency measurement data of region i under false data injection attack is:

[0157]

[0158] wherein is the regional control error of region i at time t under attack, β i is the frequency bias factor of region i, (Δf i + ψ a1 )(t) is the frequency deviation of region i at time t under attack, and ψ a1a ij is the capacity factor between region i and region j, ΔP tie,ij is the power deviation between region i and region j at time t, t is the instantaneous time, and m is the total number of control regions;

[0159] The control error model of the power measurement data between region i and region k under the virtual data injection attack is:

[0160]

[0161] where Δf i is the frequency deviation of region i at time t, a ik is the capacity factor between region i and region k, (ΔP tie,ik + ψ a2 ) (t) is the power deviation between region i and region k at time t under attack, ψ a2 is the false data injection attack signal.

[0162] The control error model of the control error data of region i under the virtual data attack is:

[0163]

[0164] where ACE i (t) is the regional control error of region i at time t, ψ a3 is the false data injection attack signal.

[0165] In specific implementation, ψ a1 , ψ a2 , and ψ a3 may be a ramp, a pulse, or a random signal.

[0166] (2) The regional control error model under the scaling attack includes:

[0167] The control error model of the frequency measurement data of region i under the scaling attack is:

[0168]

[0169] where, is the regional control error of region i at time t under attack, β i is the frequency bias factor of region i, (Δf i × λ a1 ) (t) is the frequency deviation of region i at time t under attack, λ a1 is the scaling attack signal, a ij is the capacity factor between region i and region j, ΔP tie,ij(t) is the power deviation of tie-line between region i and region j at time t, t is the instantaneous time, m is the total number of control regions;

[0170] The control error model of power measurement data between region i and region k under the scaling attack is:

[0171]

[0172] where Δf i (t) is the frequency deviation of region i at time t, a ik is the capacity factor between region i and region k, (ΔP tie,ik × λ a2 )(t) is the power deviation of tie-line between region i and region k at time t under attack, λ a2 is the scaling attack signal;

[0173] The control error model of control error data of region i under the scaling attack is:

[0174]

[0175] where ACE i (t) is the region control error of region i at time t, λ a3 is the scaling attack signal.

[0176] (3) The region control error model under denial-of-service attack includes:

[0177]

[0178] where n is the number of periods, T is one period of interference, T off is the idle time of interference, is the minimum idle time of interference, which is in the range of

[0179] The interval [(n-1)T,(n-1)T+T off ] is the idle time of the interference signal, indicating that the denial-of-service attack is not successful, and the frequency measurement data of the region received by the controller, the power measurement data of the region is correct;

[0180] The interval [(n-1)T+T off ,nT] is the active interval of the interference signal, indicating that the denial-of-service attack is activated, and the controller does not get any information.

[0181] (4) The region control error model under time delay attack includes:

[0182] The control error model of frequency measurement data of region i under time delay attack is:

[0183]

[0184] in, To determine the region control error of region i at time t under attack, β i Let Δf be the frequency bias factor for region i. i (tt d ) represents the frequency deviation of region i under attack at time t, where t d For delay attack signals, a ij Let ΔP be the capacity factor between region i and region j. tie,ij (t) represents the power deviation of the link between region i and region j at time t, where t is the instantaneous time and m is the total number of control regions;

[0185] The control error model for the power measurement data between region i and region k under time delay attack is as follows:

[0186]

[0187] Where, Δf i (t) represents the frequency deviation of region i at time t, a ik Let ΔP be the capacity factor between region i and region k. tie,ik (tt d Let t be the power deviation of the link between region i and region k under attack at time t. d This is a delay attack signal;

[0188] The control error model for region i under a time delay attack is as follows:

[0189]

[0190] Among them, ACE i (tt d ) represents region i in tt d The regional control error at time t d This is a delay attack signal.

[0191] When frequency control is subjected to a delay attack, the optimal controller built for normal settings will deviate from the optimal state due to system latency, and the system may become unstable.

[0192] Following the adaptive network attack detection system, frequency anti-interference control can be performed based on the adaptive network attack detection system. Specifically, for example... Figure 2 As shown, the power system frequency disturbance rejection and control methods under various deterministic network attacks include:

[0193] Step 201: Obtain multiple sets of frequency modulation-related parameters for the region to be controlled.

[0194] wherein each set of frequency regulation related parameters comprises regional frequency bias data, regional power bias data and actual regional control error data.

[0195] Step 202, inputting the regional frequency bias data and the regional power bias data in each set of frequency regulation related parameters into the adaptive network attack detection system to obtain estimated regional control error data.

[0196] wherein the adaptive network attack detection system is obtained by training a neural fuzzy system physical model using pre-collected sample data under multiple network attacks; the sample data of each region comprises regional frequency bias, regional power bias and regional control error; and the neural fuzzy system physical model is a combination of fuzzy logic and neural network.

[0197] Step 203, calculating root mean square error according to the actual regional control error data and the estimated regional control error data of the multiple sets of frequency regulation related parameters.

[0198] wherein the root mean square error calculation formula is as follows:

[0199]

[0200] Step 204, if the root mean square error is within the preset range, no attack is detected, and the actual regional control error is taken as the mitigation regional control error.

[0201] In a specific embodiment, the preset range is, for example, 2.28x10 -4 .

[0202] Step 205, if the root mean square error exceeds the preset range, an attack exists, and the estimated regional control error is taken as the mitigation regional control error.

[0203] In denial of service attack, the ACE signal from network communication is unavailable, in false data injection attack, the ACE signal is distorted, and in time delay attack, the ACE signal is delayed. In these three cases, the RMSE is greater than the fault tolerance diagnosis range, indicating that an attack has occurred.

[0204] Step 206, regulating the frequency of the power system by using the mitigation regional control error.

[0205] In an embodiment of the present application, in order to improve the accuracy of power system frequency regulation, the above step 206 regulates the frequency of the power system by using the mitigation regional control error, comprising:

[0206] inputting the mitigation regional control error into a pre-constructed double-loop enhanced active disturbance rejection controller to obtain a control amount of the governor, and sending the control amount to the governor of the region to be controlled to adjust the frequency of the power system in the region to be controlled;

[0207] The double-loop enhanced active disturbance rejection controller is constructed by using a regional power system model with disturbance and nonlinearity.

[0208] The embodiment can avoid the influence of inherent nonlinearity limitation and uncertain load fluctuation of the power system on power adjustment, effectively offset the frequency deviation of the power system, and optimally distribute power of the power system.

[0209] As shown in the embodiment of the application, Figure 3 The construction process of the double-loop enhanced active disturbance rejection controller includes:

[0210] In step 301, a regional power system model with disturbance and nonlinearity represented by the following formula is constructed:

[0211]

[0212] Where y is a regional control error ACE of the region i, is a first derivative of y, is a second derivative of y, u is a control output of a speed governor, is a speed governor output rate coefficient, d is an external disturbance, and f is a whole system nonlinearity.

[0213] In step 302, a state equation of the regional power system is constructed according to the regional power system model.

[0214] Where the state quantity in the state equation is y, f. Specifically, the state equation of the power system of the region i can be represented as:

[0215]

[0216] Where x1=y, x3=f is an extended state of the system; is a differential of the extended state of the system.

[0217] In step 303, an active disturbance rejection state observer of the regional power system represented by the following formula is constructed according to the state equation of the regional power system:

[0218]

[0219] Where,

[0220] is an estimated value of y, is an observed value of the system state, and L0=[ξ1,ξ2,ξ3] T is an observer gain.

[0221] The active disturbance rejection state observer can also be rewritten as:

[0222]

[0223] Step 304, according to the active disturbance rejection state observer of the regional power system, the bandwidth method is used to solve the observer gain in stable state.

[0224] In order to simplify the adjustment process of the active disturbance rejection state observer parameters, the bandwidth method is introduced to adjust the parameters in this paper, and the state variable error is Then the error derivative can be written as:

[0225]

[0226] By selecting the appropriate observation gain matrix L0, the matrix A0-L0C0 is stable, so e→0, which can make The condition for the matrix A0-L0C0 to be stable is that the roots of its characteristic polynomial have negative real parts, and further, the expression of the matrix A-LC can be written as:

[0227]

[0228] The characteristic polynomial of the matrix A-LC is:

[0229]

[0230] According to the core idea of the bandwidth method, the poles of the characteristic polynomial are configured at-ω0, that is:

[0231] (s+ω0) 3 =s 3 +3ω0s 2 +3ω0 2 s+ω0 3

[0232] In the formula, ω0 is the bandwidth of the active disturbance rejection state observer; s is the root of the characteristic polynomial of the system.

[0233] The gains of the active disturbance rejection state observer can be obtained as follows:

[0234]

[0235] From the above analysis, it can be concluded that the roots of the characteristic polynomial are-ω0, as long as ω0 is greater than 0, the roots of the characteristic polynomial have negative real parts. Thus the matrix A-LC is stable, and e→0, that is

[0236] Therefore, we can get:

[0237]

[0238] wherein, denotes The f(x1(t), x2(t), d(t)) can be estimated, which contains the cumulative effect of the uncertain model and external disturbance, that is, the real-time influence of system disturbance.

[0239] In step 305, according to the observer gain in the steady state and the regional power system model, the feedforward control loop L1 and the feedback control loop L2 represented by the following formula are determined.

[0240]

[0241] wherein, u1 is the control quantity of the feedforward control loop L1, u2 is the control quantity of the feedback control loop L2, is the speed governor output rate coefficient, is the observation value of the overall system nonlinearity, k ti , k i denotes the integral gain, η1 and η2 are the proportional control parameter and the derivative control parameter respectively, η1 = ω c 2 , η2 = 2ω c , ω c is the bandwidth of the state error feedback control law, and s is the Laplace operator in the complex frequency domain, which is used to represent the frequency domain characteristics of the dynamic system.

[0242] In step 306, according to the feedforward control loop L1 and the feedback control loop L2, the dual-loop enhanced active disturbance rejection controller represented by the following formula is determined.

[0243] U(s) = u2(u1 + ε o -γ);

[0244] wherein, ε o is a predetermined error limit, and γ is the frequency deviation feedback signal of the generator output.

[0245] In the designed dual-loop enhanced active disturbance rejection controller, only ω0 and ω c two parameters need to be adjusted, that is, the bandwidth of the active disturbance rejection state observer and the bandwidth of the state error feedback control law. In order to maximize the performance of the controller, the optimization method is used to adjust the parameters of the controller.

[0246] The basic particle swarm optimization algorithm has the advantages of few parameters to be set and simple solving process, but has the defects of slow convergence speed and easy to fall into local optimal solution. In view of the defects of the basic particle swarm optimization algorithm, the adaptive particle swarm optimization algorithm is used to optimize and solve the parameters of the dual-loop active disturbance rejection controller. The adaptive particle swarm optimization algorithm combines the adaptive adjustment of the inertia weight value in the algorithm iteration process with the fitness function state, and its main feature can be represented as:

[0247]

[0248]

[0249] where, is the average fitness function value of the population at the kth iteration; is the fitness function value of the ith particle at the kth iteration; is the inertia weight value of the ith particle at the kth iteration; ω min , ω max are the minimum and maximum values of the inertia weight, respectively; is the minimum fitness function value of the population at the kth iteration. Based on this, the particle position and velocity update expressions of the adaptive particle swarm optimization algorithm can be written as:

[0250]

[0251] where, is the position of the ith particle at the kth iteration; is the velocity of the ith particle at the kth iteration; c1 is the individual optimal learning factor, used to track the current optimal position of the ith particle; c2 is the global optimal learning factor, used to track the current global optimal position of the entire colony; r1, r2 are random numbers distributed in the interval [0, 1]; is the optimal solution of the particle individual; gbest k is the optimal solution of the entire population.

[0252] The main task of the power system governor is to quickly and accurately track the command, and the control system designed based on the time multiplication absolute error integral criterion has a small overshoot and a relatively rapid system response speed, so the error evaluation function of the active disturbance rejection controller is designed as the fitness function in the adaptive particle swarm optimization algorithm, and its expression is:

[0253]

[0254] After determining the controller parameters that need to be adjusted and the fitness function, the adaptive particle swarm optimization algorithm is used to optimize and solve the parameters of the active disturbance rejection controller.

[0255] The specific optimization process of the adaptive particle swarm optimization algorithm mainly includes five modules: initialization module, inertia weight self-adaption module, particle update module, optimal particle selection module, and result output module, and the specific steps are as follows:

[0256] Step 1) Initialization module: First, set the parameters of the adaptive particle swarm optimization algorithm, including population size, iteration number, search space dimension, optimal learning factor, and inertia weight, etc.; then initialize the population and select the global optimal particle;

[0257] Step 2) inertia weight self-adaptive module: adaptively adjusting the inertia weight according to the fitness function of the particle and the average fitness function of the population;

[0258] Step 3) particle updating module: updating the speed and position of the particle;

[0259] Step 4) optimal particle selection module: selecting the individual optimal particle and the global optimal particle according to the particle fitness function;

[0260] Step 5) result output module: responsible for judging the iteration termination condition and outputting the optimization result.

[0261] Based on the same inventive concept, the present application also provides a deterministic network multi-attack power system frequency anti-disturbance regulation device, as described in the following embodiments. Since the deterministic network multi-attack power system frequency anti-disturbance regulation device solves the problem by the same principle as the deterministic network multi-attack power system frequency anti-disturbance regulation method, the implementation of the deterministic network multi-attack power system frequency anti-disturbance regulation device can be referred to the deterministic network multi-attack power system frequency anti-disturbance regulation method, and the repeated parts will not be described here.

[0262] Specifically, as shown in the Figure 4 deterministic network multi-attack power system frequency anti-disturbance regulation device includes:

[0263] The parameter acquisition unit 401 is configured to acquire a plurality of sets of frequency regulation related parameters of the controlled region, wherein each set of frequency regulation related parameters includes regional frequency deviation data, regional power deviation data, and actual regional control error data.

[0264] The detection unit 402 is configured to input the regional frequency deviation data and the regional power deviation data in each set of frequency regulation related parameters into an adaptive network attack detection system to obtain estimated regional control error data; wherein the adaptive network attack detection system is obtained by training a neural fuzzy system physical model using pre-collected sample data under multiple network attacks; the sample data of each region includes regional frequency deviation Δf i (t), regional power deviation ΔP tie,ij (t), and regional control error ACE i (t); the neural fuzzy system physical model is a combination of fuzzy logic and neural network;

[0265] The calculation unit 403 is configured to calculate the root mean square error according to the actual regional control error data and the estimated regional control error data of the plurality of sets of frequency regulation related parameters.

[0266] The judgment unit 404 is configured to, if the root mean square error is within the preset range, determine that no attack is detected, take the actual area control error as the mitigation area control error; and if the root mean square error exceeds the preset range, determine that an attack exists, and take the estimated area control error as the mitigation area control error.

[0267] The control unit 405 is configured to regulate the frequency of the power system by using the mitigation area control error. The specific implementation process of regulating the frequency of the power system by using the mitigation area control error can refer to the foregoing embodiments, and details are not described herein.

[0268] The detection algorithm based on the neural fuzzy physical model provided in the application can protect the frequency regulation of the power system from various network attacks without limiting the duration and location of the attacks, and can quickly identify and control the attacks when the attacks occur. Meanwhile, the application also provides a double-loop enhanced active disturbance rejection method with feedforward compensation, which can effectively offset the frequency deviation of the power system and optimally allocate the power of the power system, so as to solve the inherent nonlinear limitation and uncertain load fluctuation of the system.

[0269] In an embodiment of the application, a computer device is provided, as shown in Figure 5 The computer device 502 can include one or more processors 504, such as one or more central processing units (CPUs), each of which can implement one or more hardware threads. The computer device 502 can also include any memory 506 for storing any kind of information, such as code, settings, data, etc. Without limitation, for example, the memory 506 can include any one or a combination of the following: any type of RAM, any type of ROM, a flash memory device, a hard disk, an optical disk, etc. More generally, any memory can use any technology for storing information. Further, any memory can provide volatile or non-volatile retention of information. Further, any memory can represent a fixed or removable component of the computer device 502. In one case, the computer device 502 can perform any operation of the associated instructions when the processor 504 executes the associated instructions stored in any memory or combination of memories. The computer device 502 also includes one or more drive mechanisms 508, such as a hard disk drive mechanism, an optical disk drive mechanism, etc., for interacting with any memory.

[0270] The computer device 502 can also include an input / output module 510 (I / O) for receiving input (via input device 512) and for providing output (via output device 514). One specific output mechanism can include a presentation device 516 and associated graphical user interface 518 (GUI). In other embodiments, the input / output module 510 (I / O), input device 512, and output device 514 can not be included, and the computer device 502 can be a stand-alone computer device in a networked environment. The computer device 502 can also include one or more network interfaces 520 for exchanging data with other devices via one or more communication links 522. One or more communication buses 524 couple the above-described components so that each component can communicate with each other component.

[0271] The communication links 522 can be implemented in any manner, such as through a local area network, a wide area network (e.g., the Internet), a point-to-point connection, etc., or any combination thereof. The communication links 522 can include any combination of hardwired links, wireless links, routers, gateway functionality, name servers, etc., governed by any protocol or combination of protocols.

[0272] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program, and the computer program is run by a processor to execute the steps of the above method.

[0273] The embodiments of the present application also provide a computer readable instruction, wherein when the processor executes the instruction, the program in the processor executes the method of any of the above embodiments.

[0274] It should be understood that the size of the sequence number of each process described above does not mean the order of execution in various embodiments of the present application, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0275] It should also be understood that in the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, which means that there can be three relationships. For example, A and / or B can represent three cases: A exists alone, A and B exist together, and B exists alone. In addition, the character " / " in the present application generally represents an "or" relationship between the associated objects before and after it.

[0276] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0277] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the system, device and unit described above can refer to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0278] In several embodiments provided by the present application, it should be understood that the disclosed system, device and method can be implemented by other ways. For example, the device embodiments described above are merely schematic, for example, the division of the units is merely a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can also be electrical, mechanical or other forms of connection.

[0279] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed on a plurality of network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment of the present application.

[0280] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0281] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application or the whole or part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0282] The principles and implementation manners of the present application are described in the specific embodiments in the present application. The above embodiment descriptions are only used to help understand the method of the present application and its core idea. Meanwhile, for those skilled in the art, according to the idea of the present application, the specific implementation manners and application ranges will be changed. In summary, the content of the present description should not be understood as a limitation of the present application.

Claims

1. A method for frequency interference suppression and control of a power system under multiple deterministic network attacks, characterized in that, The method comprises the following steps: acquiring a plurality of groups of frequency regulation related parameters of a region to be controlled, wherein each group of frequency regulation related parameters comprises region frequency deviation data, region power deviation data and actual region control error data; inputting the region frequency deviation data and the region power deviation data in each group of frequency regulation related parameters into an adaptive network attack detection system to obtain estimated region control error data; wherein the adaptive network attack detection system is obtained by training a neural fuzzy system physical model using pre-collected sample data under a plurality of network attacks; the sample data of each region comprises region frequency deviation, region power deviation and region control error; the neural fuzzy system physical model is a combination of fuzzy logic and neural network; the sample data under the plurality of network attacks is determined by actual collection and attack models of the plurality of network attacks, and the plurality of network attacks comprise false data injection attack, scaling attack, denial of service attack and time delay attack; calculating a root mean square error according to the actual region control error data and the estimated region control error data of the plurality of groups of frequency regulation related parameters; if the root mean square error is within a preset range, it is indicated that the region to be controlled is not attacked, and the actual region control error is taken as a mitigation region control error; if the root mean square error exceeds the preset range, it is indicated that the region to be controlled is attacked, and the estimated region control error is taken as the mitigation region control error; regulating the frequency of the power system by using the mitigation region control error; wherein regulating the frequency of the power system by using the mitigation region control error comprises: inputting the mitigation region control error into a pre-constructed double-loop enhanced active disturbance rejection controller to obtain a control amount of a governor, and sending the control amount to the governor of the region to be controlled to adjust the frequency of the power system in the region to be controlled; wherein the double-loop enhanced active disturbance rejection controller is constructed by using a region power system model with disturbance and nonlinearity.

2. The method of claim 1, wherein, According to the pre-collected sample data of each region under a plurality of network attacks and the neural fuzzy system physical model, an adaptive network attack detection system is established, comprising: establishing the following fuzzy rules: where x1 and x2 are the regional frequency deviation and regional power deviation, respectively, and are the antecedents of the interval-valued intuitionistic fuzzy sets, Z is the output of the fuzzy logic, i.e., the estimated regional control error, k j1 , k j2 , and l j are the adaptive parameters of the fuzzy system, j = 1, 2, …, N is the number of fuzzy rules, and R (j) is the jth fuzzy rule. constructing a neural fuzzy system physical model according to the fuzzy rules; dividing the pre-collected sample data under the plurality of network attacks into a training sample set and a test sample set; training the neural fuzzy system physical model by using the training sample set to obtain the adaptive network attack detection system; testing the adaptive network attack detection system by using the test sample, if the test is passed, the training is ended; if the test is not passed, the neural fuzzy system physical model is adjusted, and the process of retraining and testing is performed.

3. The method of claim 1, wherein, The sample data under the plurality of network attacks are determined by the following process: pre-constructing a region control error model under various network attacks, wherein the region control error model is a function of region frequency deviation, region power deviation and attack parameters; obtaining a plurality of groups of sample data under various network attacks by using the region control error model under various network attacks according to a plurality of groups of actual region frequency deviation and region power deviation.

4. The method of claim 1, wherein, The network attacks comprise false data injection attack, scaling attack, denial of service attack and time delay attack.

5. The method of claim 4, wherein, The region control error model under the false data injection attack comprises: The control error model of the frequency measurement data of the region i under the false data injection attack is: wherein, is the area control error of the attacked area i at time t, β i is the frequency bias factor of area i, (Δf i + ψ a1 ) (t) is the frequency deviation of the attacked area i at time t, ψ a1 is the false data injection attack signal, a ij is the capacity factor between area i and area j, ΔP tie,ij (t) is the power deviation of the tie-line between area i and area j at time t, t is the instantaneous time, and m is the total number of control areas. The control error model of the power measurement data between the region i and the region k under the false data injection attack is: where Δf i (t) is the frequency deviation of area i at time t, a ik is the capacity factor between area i and area k, (ΔP tie,ik + ψ a2 )(t) is the power deviation of the tie-line between area i and area k at time t under attack, ψ a2 is the false data injection attack signal; The control error model of the control error data of the region i under the false data injection attack is: where ACE i (t) is the zone control error of zone i at time t, ψ a3 is the false data injection attack signal.

6. The method of claim 4, wherein, The region control error model under the scaling attack includes: The control error model of the frequency measurement data of the region i under the scaling attack is: where, is the zone control error of the aggressor zone i at time t, β i is the frequency bias factor of zone i, (Δf i × λ a1 is the frequency deviation of the aggressor zone i at time t, λ a1 is the scaled attack signal, a ij is the capacity factor between zone i and zone j, ΔP tie,ij is the power deviation of the tie-line between zone i and zone j at time t, t is the instantaneous time, and m is the total number of control zones. The control error model of the power measurement data between the region i and the region k under the scaling attack is: where Δf i (t) is the frequency deviation of area i at time t, a ik is the capacity factor between area i and area k, (ΔP tie,ik × λ a2 (t) is the power deviation of the tie-line between area i and area k at time t under attack, λ a2 is the scaled attack signal; The control error model of the control error data of the region i under the scaling attack is: where ACE i (t) is the zone control error of zone i at time t, λ a3 is the scaled attack signal.

7. The method of claim 4, wherein, The region control error model under the denial of service attack includes: where n is the number of periods, T is one period of interference, T off is the idle time of the interference, is the minimum idle time of the interference, which is in the range interval [(n-1)T, (n-1)T+T off ] is the idle time of the interference signal, indicating that the denial of service attack is not successful, the frequency measurement data of the area received by the controller, the power measurement data of the area are correct; Interval [(n-1)T + T off nT] is the active interval of the jamming signal, indicating that the denial-of-service attack is active and the controller gets no information.

8. The method of claim 4, wherein, The region control error model under the time delay attack includes: The control error model of the frequency measurement data of the region i under the time delay attack is: in, To determine the region control error of region i at time t under attack, β i Let Δf be the frequency bias factor for region i. i (tt d ) represents the frequency deviation of region i under attack at time t, where t d For delay attack signals, a ij Let ΔP be the capacity factor between region i and region j. tie,ij (t) represents the power deviation of the link between region i and region j at time t, where t is the instantaneous time and m is the total number of control regions; The control error model of the power measurement data between the region i and the region k under the time delay attack is: where Δf i (t) is the frequency deviation of zone i at time t, a ik is the capacity factor between zone i and zone k, ΔP tie,ik (t-t d ) is the power deviation of the tie-line between zone i and zone k at time t under attack, t d is the time delay attack signal; The control error model of the control error data of the region i under the time delay attack is: where ACE i (t-t d ) is the zone control error of zone i at time t-t d , and t d is the time delay attack signal.

9. The method of claim 1, wherein, The construction process of the dual-loop enhanced active disturbance rejection controller includes: A region power system model with disturbance and nonlinearity is constructed as follows: where y is the area control error ACE of area i, is the first derivative of y, is the second derivative of y, u is the control output of the governor, is the governor output rate coefficient, d is the external disturbance, and f is the overall system nonlinearity. According to a regional power system model, a state equation of a regional power system is constructed, wherein a state quantity in the state equation is y, f; According to the state equation of the region power system, a region power system active disturbance state observer is constructed as follows: wherein C0 = [1 0 0]; is an estimate of y, is an observation of the system state, L0= [ξ1, ξ2, ξ3] T is an observer gain; According to the active disturbance state observer of the region power system, the observer gain in the stable state is solved by using the bandwidth method; According to the observer gain in the stable state and the region power system model, a feedforward control loop L1 and a feedback control loop L2 are determined as follows: wherein u1 is a control quantity of the feedforward control loop L1, and u2 is a control quantity of the feedback control loop L2, is a speed governor output rate coefficient, is an observation value of the overall system nonlinearity, k ti , k i represents an integral gain, and η1 and η2 are proportional control parameters and derivative control parameters, respectively, η1 = ω c 2 , η2 = 2ω c , ω c is a bandwidth of a state error feedback control law, and s is a Laplace operator in a complex frequency domain. According to the feedforward control loop L1 and the feedback control loop L2, a dual-loop enhanced active disturbance rejection controller is determined as follows: U(s) = u2(u1+ ε o -γ); where ε o is a predetermined error limit, and γ is a frequency deviation feedback signal of the generator output.

10. A deterministic network multi-attack power system frequency disturbance regulation device, characterized in that, It includes: A parameter acquisition unit is configured to acquire a plurality of sets of frequency regulation related parameters of a region to be controlled, wherein each set of frequency regulation related parameters includes region frequency deviation data, region power deviation data, and actual region control error data; The detection unit is used for inputting the regional frequency deviation data and the regional power deviation data in each group of frequency regulation related parameters into an adaptive network attack detection system to obtain estimated regional control error data; wherein the adaptive network attack detection system is obtained by training a neural fuzzy system physical model by using sample data under multiple network attacks collected in advance; the sample data of each region includes regional frequency deviation Δf i (t), regional power deviation ΔP tie,ij (t), and regional control error ACE i (t); the neural fuzzy system physical model is a combination of fuzzy logic and neural network; the sample data under multiple network attacks are determined by actual collection and attack models of multiple network attacks, and the multiple network attacks include false data injection attack, scaling attack, denial of service attack, and time delay attack. A calculation unit is configured to calculate a root mean square error based on the actual region control error data and estimated region control error data of the plurality of sets of frequency regulation related parameters; A judgment unit is configured to, if the root mean square error is within a preset range, determine that the region to be controlled is not attacked, and use the actual region control error as a mitigation region control error; if the root mean square error exceeds the preset range, determine that the region to be controlled is attacked, and use the estimated region control error as the mitigation region control error; A control unit is configured to regulate the frequency of the power system by using the mitigation region control error. The regulation of the frequency of the power system by using the mitigation region control error includes: The mitigation region control error is input into a pre-constructed dual-loop enhanced active disturbance rejection controller to obtain a control amount of a speed regulator, and the control amount is sent to the speed regulator of the region to be controlled to adjust the frequency of the power system of the region to be controlled. The dual-loop enhanced active disturbance rejection controller is constructed by using a region power system model with disturbance and nonlinearity.

11. A computer device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the computer program to implement the method of any one of claims 1 to 9.

12. A computer-readable storage medium, the computer-readable storage medium storing a computer program, characterized in that, The computer program is executed by the processor of the computer device to implement the method of any one of claims 1 to 9.

13. A computer program product comprising a computer program, characterized in that, The computer program, when executed by a processor of a computer device, implements the method of any one of claims 1 to 9.

Images