Optimal hidden attack detection method and device in power grid system based on double adaptive observers and medium

Abstract

The invention relates to an optimal hidden attack detection method and device in a power grid system based on double adaptive observers and a medium, which are used for detecting an optimal hidden attack based on the vulnerability of an adaptive unknown input observer in a direct-current micro-grid system. Establishing a dynamic model of the distributed power generation unit; deploying a self-adaptive unknown input observer and a self-adaptive distributed observer for each distributed power generation unit, and monitoring the state of the distributed power generation unit when the distributed power generation unit is decoupled from a neighbor distributed power generation unit; the self-adaptive distributed observer receives the output of the neighbor self-adaptive unknown input observer and predicts the output of the distributed power generation unit; and judging whether the distributed power generation unit is attacked or not based on a detection residual error between the actual output of the system and the output of the system estimated by the self-adaptive unknown input observer and the self-adaptive distributed observer. Compared with the prior art, the reliability and the safety of the direct-current micro-grid system are effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of network security of a DC microgrid system, and in particular to an expandable detection method for optimal covert attacks on a DC microgrid system with a ZIP load. Background Art

[0002] The increasing proportion of renewable energy in the power grid has put forward new requirements for the architecture of the power system. DC microgrids can integrate flexible power loads, energy storage devices and distributed generation units to meet future energy needs. They have been widely used in electric vehicle charging stations, military bases and smart buildings. However, due to the extensive interconnectivity of DC microgrids and the openness of communication networks, hackers may exploit inherent vulnerabilities to launch malicious cyber attacks on distributed generation units, thereby triggering a cascading effect. If malicious adversaries continue to attack DC microgrids, it may cause large-scale power outages and cause irreversible damage to critical infrastructure. Therefore, it is crucial to detect attacks in a timely manner and implement appropriate protection measures.

[0003] Current detection techniques are mainly aimed at classic false data injection attacks, which can be summarized as model-based and learning-based methods. The former requires system model knowledge to design a residual generator that can monitor system anomalies in real time; while the main idea of ​​the latter is to use machine learning algorithms to learn patterns and rules from the collected data, so as to distinguish between normal and damaged data. However, the new optimal covert attack not only maximizes the attack effect under limited energy conditions, but also makes the detection residual or system data not have significant abnormal characteristics, effectively evading existing detection mechanisms. In addition, after the damaged distributed generation unit is unplugged from the DC microgrid system, the detection mechanism in the remaining distributed generation units needs to be globally redesigned, which not only brings a large computational load, but also may lead to omissions in detection. Summary of the invention

[0004] The purpose of the present invention is to provide a plug-and-play detection scheme for optimal covert attacks in a DC microgrid system in order to address the problem that existing detection methods lack scalability and are difficult to detect optimal covert attacks.

[0005] The purpose of the present invention can be achieved by the following technical solutions:

[0006] As a first aspect of the present invention, a method for detecting an optimal covert attack in a power grid system based on a dual adaptive observer is provided, which is used to detect an optimal covert attack based on the vulnerability of an adaptive unknown input observer in a DC microgrid system, comprising the following steps:

[0007] For the DC microgrid system with ZIP load, the dynamic model of distributed generation units is established;

[0008] Deploy an adaptive unknown input observer and an adaptive distributed observer for each distributed generation unit to monitor its state when it is decoupled from neighboring distributed generation units; wherein the adaptive distributed observer receives the output of the neighboring adaptive unknown input observer and predicts the output of the distributed generation unit;

[0009] Based on the detection residual between the actual system output and the system output estimated by the adaptive unknown input observer and the adaptive distributed observer, it is determined whether the distributed generation unit is under attack.

[0010] As a preferred technical solution, the dynamic model of the distributed generation unit considers the control problem of voltage balancing and current sharing in a DC microgrid system with a ZIP load, and the state equation of the distributed generation unit i is expressed as:

[0011]

[0012] In the formula, is the state variable, V i Indicates the voltage at the common connection point, I ti represents the RLC filter i current, v i Represents the integral of the voltage tracking error; and are the primary and secondary control inputs, V ti represents the output voltage of step-down inverter i; is the common connection point reference voltage; is the total load current, Y li , and P li is the load constant; I i , A ii , A ij , C i is the corresponding power grid system structure parameter matrix; i (t) is the output of distributed generation unit i.

[0013] As a preferred technical solution, the adaptive unknown input observer dynamics of the distributed generation unit i is expressed as:

[0014]

[0015] In the formula, and are the state of the adaptive unknown input observer i and the estimated system state, respectively; and is the gain of the adaptive unknown input observer i; is external input; coefficient matrix M i =[-1 / C ti ,0;0,0;0,1]; is an unknown input and satisfies

[0016] are the adaptive parameters of the unknown input observer.

[0017] As a preferred technical solution, the optimal covert attack injects false data into the channel to destroy the integrity of the data and thus reduce system performance;

[0018] The false data injection attack into the forward channel of DG unit i is described as:

[0019]

[0020] in, represents a damaged input signal, represents the normal input signal, ζ(·) represents the step function, Indicates the attack start time. Indicates an attack signal;

[0021] Optimal covert attack signal Described as:

[0022]

[0023] in:

[0024]

[0025] Satisfy the following dynamics:

[0026]

[0027] In the formula, are the coefficient matrices of the augmented systems of the attacked and normal systems, is the state of the system after the attack and normal system augmentation, It is the input of the system after the attack and normal system are augmented.

[0028] As a preferred technical solution, the adaptive distributed observer of the state of the adaptive distributed power generation unit is:

[0029]

[0030] In the formula, and is the state and output of DG unit i estimated by the observer, is the adaptive distributed observer gain, is the adaptive parameter of the adaptive distributed observer;

[0031] The global error dynamics of a DC microgrid containing n distributed generation units is expressed as:

[0032]

[0033] The sufficient condition for the error system to be asymptotically stable is:

[0034] For a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix Q i So that:

[0035]

[0036] in:

[0037]

[0038] The distributed observer gain is:

[0039]

[0040] As a preferred technical solution, the adaptive parameters of the adaptive distributed observer are The following adaptive rules are met:

[0041]

[0042] In the formula, is a positive adaptive parameter.

[0043] As a preferred technical solution, the adaptive distributed observer is designed in a decentralized manner, and the coupling between distributed power generation units is eliminated through a fixed structure relaxation matrix, giving the adaptive distributed observer a plug-and-play capability, and realizing scalable optimal covert attack detection, as follows:

[0044] The sufficient condition for the asymptotic stability of the adaptive distributed observer error system is transformed into:

[0045] For a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix So that:

[0046]

[0047] In the formula Distributed Observer Gain

[0048] As a preferred technical solution, the determination of whether the distributed power generation unit is under attack is specifically as follows:

[0049] Based on the designed adaptive unknown input observer and adaptive distributed observer, the system output is estimated to obtain the detection residual:

[0050]

[0051] Where: y i (t) is the actual output of the system, is the system output estimated by the observer;

[0052] like If the communication in distributed generation unit i is normal, it is considered that the distributed generation unit i is under attack. Represents the threshold for detecting residuals.

[0053] As a second aspect of the present invention, an electronic device is provided, comprising:

[0054] one or more processors;

[0055] A memory for storing one or more programs;

[0056] When the one or more programs are executed by the one or more processors, the one or more processors implement the optimal covert attack detection method in a power grid system based on dual adaptive observers as described above.

[0057] As a third aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the optimal covert attack detection method in a power grid system based on a dual adaptive observer are implemented as described above.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1) This paper proposes a scalable detection method for optimal covert attacks in DC microgrid systems based on dual adaptive observers, which fully considers the volatility of ZIP loads in DC microgrid systems. An attack detection scheme based on dual adaptive observers is designed, which monitors the status of distributed power generation units through distributed observers. The adaptive distributed observer not only uses local information, but also needs to accept the output of neighbor adaptive unknown input observers. It is not affected by neighbor power generation units and local optimal covert attacks, effectively improving the reliability and security of DC microgrid systems.

[0060] 2) The present invention also provides an adaptive distributed observer designed in a decentralized manner to eliminate the coupling between distributed power generation units through a relaxation matrix, thereby realizing a scalable design of an optimal covert attack detection mechanism in a DC microgrid with ZIP loads.

[0061] 3) The present invention also designs an optimal covert attack, which fully exploits the vulnerability of the DC microgrid system based on the unknown input observer, can effectively reduce the system performance and avoid the monitoring of traditional technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 A flow chart of a scalable detection method for optimal covert attacks in a DC microgrid system with a ZIP load according to an embodiment of the present invention;

[0063] Figure 2 It is a structural diagram of a DC microgrid system with a ZIP load in an embodiment of the present invention;

[0064] Figure 3 A topological diagram of a DC microgrid including six distributed power generation units in an embodiment of the present invention;

[0065] Figure 4 It is an attack effect diagram of the optimal concealed attack in the distributed power generation unit 2 in an embodiment of the present invention;

[0066] Figure 5 Schematic diagram of concealment of the optimal concealed attack in the distributed power generation unit 2 in an embodiment of the present invention;

[0067] Figure 6 Schematic diagram of detection results of a scalable detection mechanism based on dual adaptive observers in an embodiment of the present invention. DETAILED DESCRIPTION

[0068] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0069] Example 1

[0070] The present invention provides a scalable detection method for optimal covert attack in a DC microgrid system based on dual adaptive observers. Considering the influence of ZIP load disturbance in the DC microgrid system, a dynamic model of the system is established. The vulnerability of the DC microgrid system based on unknown input observer is analyzed, and the optimal covert attack is designed. An adaptive distributed observer is designed to monitor the state of distributed power generation units, and a fixed structure relaxation matrix is ​​designed to eliminate the coupling between distributed power generation units, thereby realizing the scalable design of the optimal covert attack detection mechanism in the DC microgrid system.

[0071] like Figure 1 As shown, in one embodiment, a scalable detection method for optimal covert attack in a DC microgrid system based on a dual adaptive observer comprises the following steps:

[0072] Step S1, analyzing the control problem of voltage balancing and current sharing in a DC microgrid system with ZIP loads, and establishing a dynamic model of distributed generation units.

[0073] Based on the quasi-static linearization assumption, Kirchhoff's voltage and current laws, and according to Figure 2 The DC microgrid structure diagram shown in the figure can be deduced:

[0074]

[0075] Where: R ti , L ti and C ti represents the capacitance, inductance and resistance of the RLC filter i; V i Indicates the voltage at the common connection point, I ti represents the RLC filter i current, Y li , and P li is the load constant.

[0076] Define the state vector v i represents the integral of the voltage tracking error, and the primary and secondary control inputs are and The common connection point reference voltage is The state equation of distributed generation unit i is:

[0077]

[0078] In the formula, is the total load current, I i , A ii , Aij , C i is a matrix of appropriate dimension.

[0079]

[0080] Step S2, analyze the vulnerability of the DC microgrid system based on the unknown input observer and design the optimal covert attack.

[0081] A DC microgrid is a typical system consisting of multiple interconnected subsystems (distributed generation units). When a distributed generation unit encounters an attack, it triggers a cascade effect. Therefore, it is necessary to deploy a diagnostic for each distributed generation unit to monitor its state when it is decoupled from its neighboring distributed generation units. An unknown input observer is an effective method that treats the interconnection information from neighbors as unknown inputs. Specifically, the unknown input observer dynamics of distributed generation unit i is formulated as:

[0082]

[0083] In the formula, is the state of observer i, is the system state estimated by observer i, As external input, Unknown input is recorded as And meet Coefficient matrix M i =[-1 / C ti , 0; 0, 0; 0, 1], the gain of observer i satisfies It's Hurwitz. is the adaptive parameter of the unknown input observer and satisfies:

[0084]

[0085] In the formula, is a positive adaptive parameter.

[0086] Malicious adversaries usually inject false data into the channel to destroy the integrity of the data and thus reduce system performance. In order to quantify the impact of the attack on system performance, the false data injection attack in the forward channel of distributed generation unit i can be described as:

[0087]

[0088] In the formula, ζ(·) represents the step function, Indicates the attack start time. Indicates normal input signal, Indicates an attack signal. Indicates a corrupted input signal.

[0089] Optimal covert attack signal It can be described as:

[0090]

[0091] Satisfy the following dynamics:

[0092]

[0093] in:

[0094]

[0095] In the formula, is the system coefficient matrix, state and input after attack and normal system augmentation.

[0096] The estimated error of the unknown input observer i is recorded as Combining equations (2), (3), (4) and (5), the error system of unknown input observer i of distributed generation unit i when it is subjected to optimal covert attack can be described as:

[0097]

[0098] The above formula does not contain the optimal hidden attack signal, which is consistent with the error system of the unknown input observer i when there is no attack.

[0099] Step S3, designing an adaptive distributed observer to monitor the status of distributed generation units, and establishing an observer gain design criterion under ZIP load.

[0100] The adaptive distributed observer i is designed for distributed generation unit i as follows:

[0101]

[0102] In the formula, and is the state and output of DG unit i estimated by the observer, is the designed distributed observer gain, is an adaptive parameter and satisfies the following adaptive rule:

[0103]

[0104] In the formula, is a positive adaptive parameter. is the estimation error of the adaptive distributed observer i, and the error system is described as:

[0105]

[0106] in:

[0107]

[0108] Considering that the DC microgrid contains n distributed generation units, the global error dynamics can be expressed as:

[0109]

[0110] A sufficient condition for the error system to be asymptotically stable is: for a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix Q i So that:

[0111]

[0112] in:

[0113]

[0114] Distributed observer gain:

[0115]

[0116] Prove the above sufficient conditions and construct the Lyapunov function:

[0117]

[0118] in:

[0119]

[0120] So:

[0121]

[0122] because, Pick Combining (8) we can get:

[0123]

[0124] in:

[0125]

[0126] From formula (11), we can see that J d (t)≤0, under zero initial condition, we can get:

[0127]

[0128] Since V d (∞∞)≥0, and J d (t)≤0, we can get:

[0129]

[0130] Then the global error system (10) is asymptotically stable and satisfies the H-infinity performance constraint.

[0131] Based on the above analysis, it can be seen that the adaptive distributed observer designed by the present invention can estimate the state of distributed generation units in the DC microgrid in real time. From (7), it can be seen that the adaptive distributed observer i not only uses local information, but also needs to accept the output of the neighbor adaptive unknown input observer and It is not affected by neighboring power generation units and is not affected by local optimal hidden attacks, that is, the signal received by the adaptive distributed observer i Therefore, the safety status of distributed generation unit i can be monitored in real time based on the residual calculated from the output of the adaptive distributed observer i.

[0132] Step S4, design a fixed-structure relaxation matrix to eliminate the coupling between distributed generation units and develop a scalable optimal covert attack detection algorithm.

[0133] It is worth noting that the information of global distributed generation units is included when designing the adaptive distributed observer i in step S3. This means that after removing the damaged distributed generation unit from the DC microgrid or inserting the repaired distributed generation unit, all distributed observers need to be redesigned. The following provides a decentralized way to design distributed observers, thereby giving the detection solution plug-and-play capabilities.

[0134] Furthermore, the sufficient condition for the asymptotic stability of the global error system (10) can be transformed into:

[0135] For a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix So that:

[0136]

[0137] Where: Distributed observer gain:

[0138]

[0139] To prove the above sufficient condition, firstly transform the above J d (t) is decomposed into:

[0140]

[0141] in:

[0142]

[0143] By the free matrix Q i From the structure, we can see that:

[0144]

[0145] So The positive or negative definiteness of is equivalent to the positive or negative definiteness of the following matrix:

[0146]

[0147] because is the Laplace matrix, that is Then there is also (12) makes In summary, J d (t)≤0. Similarly, the global error system (10) is asymptotically stable and satisfies the H-infinity performance constraint.

[0148] From the above analysis, it can be seen that when designing the adaptive distributed observer, this patent uses the free matrix Q i The structure of is restricted, and the Lyapunov function is divided into two categories, including global information and local information. The new matrix formed by taking the inverse of the matrix elements containing global information is exactly the Laplace matrix, which can determine the negative definiteness of the original matrix. Thus, we can judge J d (t) The positivity depends only on the information of the local generation unit, that is, the coupling between distributed generation units is eliminated.

[0149] Based on the adaptive unknown input observer and the adaptive distributed observer designed by the present invention, the diagnostic device in the distributed generation unit i can obtain the detection residual like If the communication in distributed generation unit i is normal, it is considered that the distributed generation unit i is under attack. Represents the threshold for detecting residuals.

[0150] A specific simulation experiment is provided below to verify the method of the present invention:

[0151] Consider a DC microgrid system consisting of 6 distributed generation units, whose topology is as follows: Figure 3As shown. Each distributed generation unit has a ZIP load, in which the constant impedance load parameters are 0.0333S, 0.0370S, 0.0303S, 0.0303S, 0.0333S, 0.0317S, the constant current load parameters are 1.6667A, 2A, 2.3333A, 2.1667A, 2.6667A, 2.5000A, and the constant power load parameters are 33.3333W, 40W, 35W, 31.6667W, 36.6667W, 35W. The common coupling point reference voltages are 47V, 48V, 45V, 50V, 46V, 49V. The power supply voltage of the distributed generation units is 80V, and the switching frequency of the insulated gate bipolar transistors is 15kHz.

[0152] Take the adaptive parameters as λ i =300, and the adaptive observer gain is obtained by solving LMI (12). The simulation time of this experiment is set to 5 seconds. First, each distributed generation unit operates independently. At the second second, each distributed generation unit operates according to Figure 3 The secondary controller is activated. Figure 4 As shown in Figure 2, proportional current sharing is achieved under the regulation of secondary control, namely At the same time, voltage balance is achieved, that is, Then, at the 4th second, DG unit 2 is attacked by the optimal covert attack (4). Figure 4 It can be seen that the malicious attack effectively destroyed the system performance, and the DC microgrid could not maintain the balance between load demand and power generation, that is, and The diagnostic results of the detector based on the adaptive unknown input observer are as follows: Figure 5 As shown, before and after the attack, the detection residuals in each distributed power generation unit did not change significantly, and the detection residuals did not exceed the threshold, that is, the optimal covert attack designed by the present invention can effectively reduce system performance and avoid monitoring by traditional technologies.

[0153] The present invention further conducts simulation experiments to verify the effectiveness of the designed detection scheme based on dual adaptive observers. Similarly, considering that the distributed generation unit is operating independently in the first 2 seconds, at the 3rd second, the distributed generation unit 2 is subjected to the optimal covert attack. The detection results of the scheme designed in this embodiment are as follows: Figure 6 As shown in the figure, the detection residual in DG unit 2 is obviously abnormal and quickly exceeds the detection threshold. At the 4th second, the operator actively unplugs DG unit 2 to avoid the spread of the attack and cause a cascade effect. At the same time, the neighbors of DG unit 2 autonomously update the detector structure, and the detectors in the other power generation units remain unchanged. At 4.5 seconds, DG unit 6 suffers the optimal covert attack. Figure 6 It can be seen that the detection residual in distributed power generation unit 6 is obviously abnormal and quickly exceeds the detection threshold, while the detection residuals in other power generation units are not abnormal.

[0154] The above simulation experiments show that, considering the DC microgrid with ZIP load, the optimal covert attack in the present invention can still destroy the system performance under limited energy conditions, while using system vulnerabilities to evade detection. The detection scheme based on dual adaptive observers in this embodiment can quickly and accurately detect and locate the optimal covert attack, while having good scalability.

[0155] Example 2

[0156] As a second aspect of the present invention, the present application also provides an electronic device, comprising: one or more processors; a memory for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement the optimal covert attack detection method in the power grid system based on the dual adaptive observer as described above. In addition to the above-mentioned processor, memory and interface, any device with data processing capability in which the device in the embodiment is located may also include other hardware according to the actual function of the device with data processing capability, which will not be described in detail.

[0157] Example 3

[0158] As a third aspect of the present invention, the present application also provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implements the optimal hidden attack detection method in a power grid system based on a dual adaptive observer as described above. The computer-readable storage medium may be an internal storage unit of any device with data processing capability described in any of the foregoing embodiments, such as a hard disk or a memory. The computer-readable storage medium may also be an external storage device, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), an SD card, a flash card (Flash Card), etc. equipped on the device. Further, the computer-readable storage medium may also include both an internal storage unit and an external storage device of any device with data processing capability. The computer-readable storage medium is used to store the computer program and other programs and data required by any device with data processing capability, and may also be used to temporarily store data that has been output or is to be output.

[0159] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. An optimal covert attack detection method in a power grid system based on dual adaptive observers, characterized in that: The optimal covert attack for detecting the vulnerability of an adaptive unknown input observer in a DC microgrid system includes the following steps: For the DC microgrid system with ZIP load, the dynamic model of distributed generation units is established; Deploy an adaptive unknown input observer and an adaptive distributed observer for each distributed generation unit to monitor its state when it is decoupled from neighboring distributed generation units; wherein the adaptive distributed observer receives the output of the neighboring adaptive unknown input observer and predicts the output of the distributed generation unit; Based on the detection residual between the actual system output and the system output estimated by the adaptive unknown input observer and the adaptive distributed observer, it is determined whether the distributed generation unit is under attack.

2. The optimal covert attack detection method in a power grid system based on a dual adaptive observer according to claim 1, characterized in that: The dynamic model of the distributed generation unit considers the control problem of voltage balancing and current sharing in a DC microgrid system with a ZIP load. The state equation of the distributed generation unit i is expressed as: In the formula, is the state variable, V i Indicates the voltage at the common connection point, I ti represents the RLC filter i current, υ i Represents the integral of the voltage tracking error; and are the primary and secondary control inputs, V ti represents the output voltage of step-down inverter i; is the common connection point reference voltage; is the total load current, Y li , and P li is the load constant; I i , A ii , A ij , C i is the corresponding power grid system structure parameter matrix; i (t) is the output of distributed generation unit i.

3. The optimal covert attack detection method in a power grid system based on a dual adaptive observer according to claim 2, characterized in that: The adaptive unknown input observer dynamics of the distributed generation unit i is expressed as: In the formula, and are the state of the adaptive unknown input observer i and the estimated system state, respectively; and is the gain of the adaptive unknown input observer i; is external input; coefficient matrix M i =[-1 / C ti ,0;0,0; 0, 1]; is an unknown input and satisfies are the adaptive parameters of the unknown input observer.

4. The optimal covert attack detection method in a power grid system based on a dual adaptive observer according to claim 3, characterized in that: The optimal covert attack injects false data into the channel to destroy the integrity of the data and thus reduce system performance; The false data injection attack into the forward channel of DG unit i is described as: in, represents a damaged input signal, represents the normal input signal, ζ(·) represents the step function, Indicates the attack start time. Indicates an attack signal; Optimal covert attack signal Described as: in: Satisfy the following dynamics: In the formula, are the coefficient matrices of the augmented systems of the attacked and normal systems, is the state of the system after the attack and normal system augmentation, It is the input of the system after the attack and normal system are augmented.

5. The optimal covert attack detection method in a power grid system based on a dual adaptive observer according to claim 3, characterized in that: The adaptive distributed observer of the state of the adaptive distributed generation unit is: In the formula, and is the state and output of DG unit i estimated by the observer, is the adaptive distributed observer gain, is the adaptive parameter of the adaptive distributed observer; The global error dynamics of a DC microgrid containing n distributed generation units is expressed as: The sufficient condition for the error system to be asymptotically stable is: For a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix Q i So that: in: The distributed observer gain is:

6. The optimal covert attack detection method in a power grid system based on dual adaptive observers according to claim 5, characterized in that: The adaptive parameters of the adaptive distributed observer are The following adaptive rules are met: In the formula, is a positive adaptive parameter.

7. The optimal covert attack detection method in a power grid system based on a dual adaptive observer according to any one of claims 5-6, characterized in that: The adaptive distributed observer is designed in a decentralized manner, and the coupling between distributed power generation units is eliminated through a relaxation matrix with a fixed structure, giving the adaptive distributed observer a plug-and-play capability, and realizing scalable optimal covert attack detection, as follows: The sufficient condition for the asymptotic stability of the adaptive distributed observer error system is transformed into: For a given adaptive parameter H Infinity Performance Index γ i > 0, there exists a matrix U of appropriate dimension i , P i >0 and free matrix So that: In the formula Distributed Observer Gain 8. The optimal covert attack detection method in a power grid system based on dual adaptive observers according to claim 1, characterized in that: The determination of whether the distributed generation unit is attacked is as follows: Based on the designed adaptive unknown input observer and adaptive distributed observer, the system output is estimated to obtain the detection residual: Where: y i (t) is the actual output of the system, is the system output estimated by the observer; like If the communication in distributed generation unit i is normal, it is considered that the distributed generation unit i is under attack. Represents the threshold for detecting residuals.

9. An electronic device, characterized in that: include: one or more processors; A memory for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the optimal covert attack detection method in a power grid system based on a dual adaptive observer as described in any one of claims 1 to 8.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the optimal covert attack detection method in a power grid system based on a dual adaptive observer are implemented as described in any one of claims 1 to 8.

Images